-~.Vl.‘

. .u...
E .5
:. ......,.. .. .
.vls re “film... . x.” .. . ..
....x.... . .: 2..
r ‘ V ‘6. 1
2

 

 

 

 

 

 

 

. :3:
3.1...

.-

{z .7. 3.5....
i: . x171.
. .. nu! .i x

":5:

3.1,

325!

l
b.

n:
in”. ..
Q If» I. . I

,n... w:
knapsqsdwni

.7 :66 an
5.: ..'9o-.sl

‘fixm
\. t... ’35.} )1
.2925 3.2:...

c s ‘ 3.:

 

k

s;
{$5
4-

 

Illllll’illlllfillllllllfilllllllll

1293 01420 1663
THEBJS 3

This is to certify that the

dissertation entitled

GENETIC VARIATION IN IRON BIOAVAILABILITY
FROM AMARANTHUS SPECIES

presented by

Anusuya Rangarajan

has been accepted towards fulfillment
of the requirements for

Ph. D. degree in Horticulture

~9an

Date October 23,1995

MS U i: an Affirmative Action/Equal Opportunity Institution 0- 12771

 

 

LIBRARY
Mlchigan State
Unlversity

 

 

 

PLACE II RETURN BOXto romovo this ohookoutfrom your rooord.
TO AVOID FINES rotum on or baton duo duo.

    

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

‘7

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Oppommlty lnotttmlon
mm:

ABSTRACT

GENETIC VARIATION IN IRON BIOAVAILABILIT Y
FROM MRANYH US SPECIES

By

Anusuya Rangarajan

If green leafy vegetables, such as Amarmzthus species, are to be promoted as sources of iron
(Fe) in the human diet, the bioavailability of Fe in these plant foods needs to be improved.
The following research investigated the potential for use of plant breeding to manipulate the
amomut of bioavailable Fe in this green leafy vegetable. Genetic variation in Fe bioavailability
was identified within the genus Amamnthus using an in vitro assay which simulates
gastrointestinal digestion and confirmed using the hemoglobin regeneration assay in anemic
rats. Initial experiments involved screening of 46 lines of amaranth, from 12 species, using
an in vitro assay. Significant difl‘erences in the percent of the total bioavailable Fe were
detected among species of amaranth, with a range of 6 to 12% of total Fe as bioavailable.

Although total Fe concentrations of leaves did not correlate with Fe measured as bioavailable,
those species which accumulated higher total Fe, such as A. tricolor and A. livicbts, tended
to provide more bioavailable Fe to the diet. Much of the Fe in lines with high total Fe
concentrations was sequestered in insoluble, unavailable forms. Large fluctuations in total
Fe of leaves over the growing season did not lead to similar changes in bioavailable Fe.

When fed to rats, a selected line of A. tricolor had lower percent of total bioavailable Fe than
lines of A. lopochoncb'iacus or A. cruentus. However, when the same quantity of each

amaranth was added to diets, the A. tricolor line, which contained the highest total Fe

concentration, supported the largest gain in hemoglobin. The hemoglobin repletion
efficiencies of the lines were equivalent, suggesting that any differences in other plant
compounds had minimal cfi‘ect on the Fe bioavailability. These results confirmed the relative
relationships observed in the in vitro assay and supported the use of the in vitro assay in
breeding programs focused on improving the Fe nutritional quality of amaranth and other

green leafy vegetables.

To Rob and My Favorite Celestial Orbs

iv

ACKNOWLEDGEMENTS

I would like to convey a special thank you to my major professor, Dr. John F. Kelly,
of the Horticulture Department. I arrived in Lansing five years ago, starry-eyed and
idealistic. The challenges of this PhD. program could have destroyed that, but I think that
Jack's timely support and unending confidence in me helped minimize the damage. I do not
think I am so starry-eyed anymore, but I am still a dreamer.

It was not easy. I thought that everyone exaggerated that part. I do know that I
could not have survived without the love and support of my family and fiiends. Rob hung
on for the roller-coaster ride of the process. Banu and Babu continued to believe that I could
feed the world. Mom and Dad always wondered how I would make a living in gardening!
Regardless of all of the illusions, it was their faith in me that made it happen.

My three dearest fiiends, Cindy, Scott and Michael, were there through the tough
times. They continually offered diversions which they knew would inspire me to get my
work done. I do not think that Lansing is a paradise, but "Paradise" is in Lansing. Many
an evening was spent killing brain cells in that Garden of Eden.

Thank you to my special mentors- all strong women who inspired me along the way:
Sue and Anna Kann, Dr. Wanda Chenoweth and Dr. Gretchen Hill. It was a long haul, but
I cannot complain. I live by the words of MK. Gandhi:

There is more to life than increasing its speed.

TABLE OF CONTENTS

LIST OF TABLES .......................................................................................................... viii
LIST OF FIGURES .......................................................................................................... x
INTRODUCTION ............................................................................................................ 1
LITERATURE REVIEW .................................................................................................. 5
Human Iron Absorption ............................................................................. 5
Iron Nutritional Quality of Green Leafy Vegetables .................................. 17
CHAPTER PAGE
1. Genetic Variation of Iron Bioavailability from Amaranthus Species ....................... 40
Abstract .................................................................................................... 40
Introduction ............................................................................................. 41
Materials and Methods .............................................................................. 43
Results ...................................................................................................... 46
Discussion ................................................................................................. 48
Literature Cited ......................................................................................... 6O

2. Changes in Iron Concentration and Bioavailability with Amaranth

LeafDevelopment ................................................................................................. 63
Abstract .................................................................................................... 63
Introduction .............................................................................................. 64
Materials and Methods ............................................................................... 65
Results ...................................................................................................... 68
Discussion ................................................................................................ 70
Literature Cited ......................................................................................... 80

3. Evaluation of Iron Bioavailability from Amaranthus Species

Determined by Hemoglobin Repletion in Anemic Rats ........................................... 82
Abstract ................ 82

Introduction .............................................................................................. 83
MaterialsandMethods ............................................................................... 85

Results ...................................................................................................... 88

Discussion ................................................................................................ 90

Literature Cited ......................................................................................... 10]

SUMMARY .................................................................................................................... 103
APPENDIX A Recommended Daily Allowances of Iron .............................................. 109

APPENDIX B. Total, dialyzable (pg-g") and percent dialyzable Fe in leaves
of A. tricolor (tri)(Ames 5113), A. hwochondriacus (hyp)
(Ames2171) and A. cruentus (cru)(PI 433228) grown in
the field in two different years ................................................................. 110

LIST OF REFERENCES ............................................................................................... 11 1

vii

LIST OF TABLES

TABLE PAGE
1. Iron-containing enzymes and storage proteins and their cellular locations ................. 35
2. Past research on iron bioavailability of green leafy vegetables ................................ 36
3. Amaranthus accessions evaluated for total and bioavailable iron under

greenhouse and/or field conditions ........................................................................ 54

4. Total, dialyzable (pg-g" dry wt.) and percent dialyzable Fe in
leaves of 35 lines of Amaranthus grown in the field ............................................. 56

5. Total, dialyzable (pg-g“l dry wt.) and percent dialyzable Fe in
leaves of 12 species Amaranthus grown in the field ............................................. 57

6. Total, dialyzable (Mg-g'1 dry wt.) and percent dialyzable Fe in
leaves of 24 lines of Amaranthus grown in the greenhouse .................................. 58

7. Total, dialyzable (lug-gl dry wt.) and percent dialyzable Fe in

leaves of four species of Amaranthus grown in the greenhouse ............................. 59
8. Amaranthus accessions evaluated for total and bioavailable
Fe under field conditions on three harvest dates .................................................... 75

9. Total, dialyzable (Dia‘l.)(;.¢g-g’l dry weight) and percent dialyzable
(%Dial) Fe in leaves of 18 lines of Amaranthus grown in the field
and harvested on three dates ................................................................................ 76

10. Dialyzable, soluble and total Fe (pg-g" dry weight) and percent dry matter

(%DM) in leaf 4 of two amaranth species grown in the greenhouse and
harvested on five difierent dates ........................................................................... 77

viii

ll.

12.

13.

14.

15.

16.

17.

18.

Total, dialyzable (rig-g“l dry wt.) and percent dialyzable Fe and nutritional
composition of amaranth lines evaluated by hemoglobin regeneration in

anemic rats ............................................................................................................ 95

Modified AIN93-G diets utilized for hemoglobin repletion in anemic rats .............. 96

Analysis of diet compositions. Added and actual Fe values are represented

as [.tg-g'l dry weight .............................................................................................. 97

Average weight gain, diet consumed and Fe intake for rats fed control
and treatment diets during the 14-day repletion period, and calculated
hemoglobin repletion efficiency (HRE) and relative biological value (RBV)

of dietary Fe sources ............................................................................................. 98

Analysis of variance of hemoglobin gain in rats to test for satisfaction
of criteria for slope ratio analysis. Actual diet Fe levels were used in

the analysis ............................................................................................................ 99

Linear regression equations and calculations of relative Fe bioavailability,
using slope ratio analysis, of two amaranth species compared to the FeSO4

control Fe source ................................................................................................... 100

Recommended Daily Allowances of Iron ............................................................... 109

Total, dialyzable (,ug-g" dry wt.) and percent dialyzable Fe in leaves of
A. tricolor (tri)(Ames 5113), A. lopocondriacus (hyp)(Ames 2171) and

A. mentus (cru)(PI 433228) grown in the field in two different years ................ 110

ix

LIST OF FIGURES

FIGURE PAGE

1. Change in total number of leaves (> 2 cm) and length of leaf 4
of two amaranth lines grown in the greenhouse and harvested
on five dates ....................................................................................................... 78

2. Change in percent of soluble Fe or percent available Fe of leaf 4
of two amaranth lines grown in the greenhouse and harvested
on five dates ....................................................................................................... 79

Introduction

Iron deficiency is the most prevalent mineral nutrient deficiency afi‘ecting humans,
especially among women of child-bearing age and young children (Scrimshaw, 1991). An
estinmted 0.75 to 1 billion people around the world (US of the population) (Chidambaram et
al., 1989) including 1-6% of the population of the United States (National Research Council,
1989) sufi‘er from some degree of iron deficiency. However, the actual number of persons
afl‘ected by iron deficiency is dificult to determine, due to variation in methods used to
evaluate the deficiency, the broad range of normal hemoglobin concentrations, and the
subtlety of iron deficiency symptoms.

Iron nutritional anemia, the most common deficiency disease, results from the
decrease in functional capacity of hemoglobin in the red blood cells to transport oxygen and
carbon dioxide in the body. Severe iron deficiency anemia results in decreased hemoglobin
levels accompanied by decreased red blood cell size (packed cell volume less than 3.5)
(National Research Council, 1989). In males and females over 14 years of age, hemoglobin
levels below 13 and 12 g/dL, respectively, indicate anemia. Some symptoms ascribed to
anemia include: lethargy, irritability, poor attention span, susceptibility to secondary
infections, reduced learning capacity and decreased tolerance to cold temperatures. Many
of these symptoms occur to different degrees prior to observable changes in hematological

status. Although most of these symptoms are difiicult to measure, generally these can

2

contribute to poor physical and mental capacity, afi‘ecting performance in work and school
(National Research Council, 1989; Prasad and Prasad, 1991; Scrimshaw, 1991).

The incidence of iron deficiency is particularly high in developing regions of the world.
Tropical populations with predominately vegetable food diets, and also having high incidence
of intestinal infections, show higher degrees of iron deficiency (Monsen, 1988a; Thompson,
1988). The widespread poverty observed in many of these areas is the most significant factor
contributing to this deficiency. Resource-poor individuals in either rural or urban areas may
lack money to buy food or land to produce adequate amounts of food (Oyejola and Bassir,
1975). In regions with poor crop production conditions or unstable political situations,
limited supplies and distribution of food contribute to nutritional deficiencies. The incidence
of secondary disease infections increases in those with anemia (Fritz et al., 1970), and poor
health and sanitation conditions can increase amounts of iron lost as blood in stool fi'om
ulcers, or infection by intestinal parasites (Monsen, 1988).

In addition to the limited quantities of food, poor access to highly bioavailable iron
sources increases the incidence of iron deficiency. Bioavailability of a mineral is a measure
of the potential absorption and use of that mineral by an organism (Smith, 1983). Although
there is between 6-7 mg iron/ 1000 kcal in most food supplies, the bioavailability of this iron
can vary, depending on the source (Monsen, 1988a; Chidambaram et al., 1989). As incomes
increase, a greater proportion of animal products generally can be found in the diet (Walker,
1982). Meat has a higher iron bioavailability than plant foods, such as cereal and beans,
which often replace meat as incomes decrease (Oyejola and Bassir, 1975; Reinhold et al.,
1981; Omueti, 1982). Because iron supplements often are utilized better than many food

iron sources, these are popular interventions to combat severe iron deficiency (Fritz et al.,

3

1970). However, the effectiveness of iron supplementation programs may be limited by the
socioeconomic conditions and isolation of the populations in developing countries (Monsen,
1988a). In these cases, well selected diets with high iron content and improvements in the
bioavailability of iron fi'om the foods regularly consumed may be more eficient and effective
methods to combat iron deficiency than provision of externally supplied iron supplements
(Zhang et al., 1985; Monsen 1988a).

Green leafy vegetables (GLVs) are one food source to focus efi‘orts for improvement
of iron bioavailability. GLVs have a high concentration of iron per unit calorie (Miller, 1987)
and have wide use and acceptability, providing flavor, texture and variety to meals. These
vegetables often have very high productivity, being grown intensively and cheaply for local
sale on small farms or in home gardens or collected fi'om the wild in many developing regions
(Pirie, 1985). GLV consumption for adults can average up to 100 g/day and for children 30
g/day (Pirie, 1985). Amaranth (Amwanthus spp.) is a GLV that is cultivated and consumed
in many developing regions of the world and is described as perhaps the most extensively
cultivated leafy vegetable in West Afiica (Schmidt, 1971). The low cost, high productivity,
and stress tolerance of amaranths makes these GLVs an ideal addition to the home garden.
Amaranth has a nutritional value that can be compared to spinach, but has higher dry matter,
protein, calcium and iron (Grubben, 1976), and it is native to many tropical regions where it
is consumed and thus better suited for production in warm climates. Improvement in the iron
nutritional quality of this vegetable could provide a significant source of iron for many
resource-poor individuals around the world, especially women and children.

Many previous studies have examined the iron nutritional quality of amaranth (Oyejola

and Bassir, 1975; Grubben, 1976; Reddy and Kulkami, 1976; Stafford et al., 1976; Deutsch,

4

1978; Ifon and Bassir, 1978; Devadas and Saroja, 1980; Smith, 1982; Makus, 1984;
Chweya, 1985; Chawla et al., 1988). However, many of these studies have been conducted
by food scientists using a ”market basket” approach, in which foods are collected fi'om local
markets and analyzed for iron content (Welch and Gabelrnan, 1984). Although this method
does provide information on the amount of iron people may consume, it provides no
indication of the difi‘erences in iron bioavailability due to genetic variation, cultural practices,
or postharvest handling methods. The goal of this research was to expand the understanding
of genetic and physiological factors which influence the bioavailability of iron fiom amaranth.
The objectives included identification of genetic lines of high iron bioavailability which may
be used in plant breeding programs, and examination of cultural practices for effects on the
amount of bioavailable iron. Prior to describing the results of this research efl‘ort, a review
of relevant literature in the areas of human iron nutrition and iron bioavailability from GLVs

is provided.

Literature Review

Human Iron Absorption Interactions

Iron (Fe) is an essential trace element for humans (as well as all other animals and
plants). Fe is part of hemoglobin and myoglobin, and in the heme configuration of these
proteins, Fe facilitates transport of oxygen and carbon dioxide in blood and muscle. The Fe-
adequate adult human body contains approximately four to five grams of Fe, of which 74%
is in hemoglobin, 8% in myoglobin, 0.2% bound to transferrin (an Fe-transport protein in the
plasma), and 13% in the Fe-storage compounds ferritin and hemosiderin (Hunt and Grofi;
1990). The remaining Fe is found in minute quantities, but is essential to the function of many
enzymes which transfer electrons during metabolism, including oxidases, catalases,
reductases, and peroxidases (Scrimshaw, 1991). The nutritional requirement for Fe for men
is about 1 mg/day, and for women, 2 mg/day. However, the U.S.-Recommended Daily
Allowance for men and women is 10 and 15 mg/day, due to the complex interactions that
afl'ect Fe bioavailability from foods (National Research Council, 1989).

Although Fe is abundant in the food supply, averaging between 6-7 mg Fe/ 1000 kcal
(Monsen, 1988a), only 10-15% of this Fe is absorbed from the diet (Hunt and Grofi‘, 1990).
Fe absorption is governed by complex homeostatic controls, only now being elucidated.
Intrinsic human factors, such as age, health and Fe status, as well as extrinsicfactors such as

the physical and chemical forms of Fe in food can influence Fe absorption in the human gut

6

(Bothwell et al., 1979; Hallberg, 1981; Latunde-Dada and Neale, 1986). This review
provides a general description of these factors which afi‘ect Fe absorption in humans. More
detailed information may be found in some excellent review articles in human nutrition
literature (Bothwell et al., 1979; Hallberg 1981; Latunde-Dada and Neale, 1986; Clydesdale,

1988; Thompson, 1988).

Intrinsic Human Fe Absorption Factors

Many intrinsic human factors influence the efficiency of Fe absorption, including sex,
age, Fe status and health In order to maintain Fe levels in the body, men must absorb about
1.3 mg/day and women 1.8 mg/day during premenopausal years, due to the regular monthly
blood loss (Monsen, 1988a). During pregnancy, greater quantities of Fe must be absorbed
to maintain the expanded blood supply and needs of the developing embryo. Infants
generally have Fe stores sufficient to cover needs for the first three months of life. Children
must absorb different amounts depending on age, and need to support rapid growth
processes. With age (>50 years) and decreasing activity, Fe requirements generally decrease,
to 10 mg/day for both men and women (Appendix A) (National Research Council, 1989).

The Fe status or Fe stores of the individual can influence the amount of Fe absorbed
in the gut (Bothwell et al., 1979; Hallberg, 1981). An Fe-deficient individual will absorb up
to 25% of dietary Fe. The mechanisms of this increased efficiency are not well understood
(Monsen, 1988a).

General health and activity also affect Fe absorption efficiency. Infection by intestinal
parasites, ulcers or other diseases which cause chronic blood loss, will increase the Fe

requirement (Fritz et al., 1970). In addition, abnormalities in digestive processes, such as

7

insufficient hydrochloric acid production in the stomach (achlorhydria), decrease the Fe
release from foods and consequently the amount of soluble Fe available for absorption

(Bothwell et al., 1979).

Fe Forms and Interactions in the Diet

To be bioavailable, Fe must be soluble in the upper small intestine. It is assumed that
this soluble Fe must be in the form of low molecular weight conjugates to be absorbed, based
on research with low molecular weight Fe compounds (Bemer and Miller, 1985). Two
difi‘erent forms of Fe can be found in foods- heme and nonheme Fe. These two forms of Fe
behave very difi‘erently in the digestive milieux (Cook, 1983). Home Fe, as found in
hemoglobin, myoglobin and the cytochrome enzymes, is conjugated within a protoporphyrin
ring. After side chains are lysed from the central porphyrin ring, this Fe-containing ring is
absorbed intact into the mucosal cell (Hunt and Grofi‘, 1990). Inside the mucosal cell, the
hemeoxygenaseenzymelysesthering and liberatestheFe fortransport into the body. Fe in
this home structure remains soluble in the small intestine and is isolated fi'om other food
components which might otherwise react with the Fe and affect it's solubility. Because of this
isolation, the efficiency of absorption for heme Fe is high, between 15 and 35%, depending
on the Fe status of the individual (Hunt and Grofl‘, 1990). Home Fe is found in animal and
fish meats at high levels, averaging about 40% of total tissue Fe (National Research Council,
1989). Red meat has 50% of its Fe in home structures (Monsen, 1988b).

Nonheme Fe is any other form of Fe in food and it represents the most significant
source ofFe in the diet (Hallberg, 1981; Cook, 1983; Monsen, 1988b). Almost 100% ofFe

intake in developing countries and 90% of intake in developed countries (despite the high

8
dietary meat content) is firm nonheme Fe forms (Clydesdale et al., 1991). Greater than 90%

of the Fe in noncellular animal food products (eggs, milk, cheese), vegetables, grains and
fiuits is nonheme (Monsen, 1988b).

The efliciency of nonheme Fe absorption is much lower than heme, ranging from 2-
20%, depending on diet composition and Fe stores of the individual (Bothwell et al., 1979;
Hallberg, 1981; Cook, 1983; Monsen, 1988a). The mechanism, chemical forms and
receptors for nonheme Fe absorption are not well understood. Ferrous Fe is transferred
across the mucosa via a carrier, and once inside the mucosal cell, is oxidized to ferric Fe and
bound to a transport protein (transferrin) for movement in the blood (Hunt and Grofi‘, 1990).

Solubility ofthe nonheme Fe is the critical factor afi‘ecting the availability of nonheme
Fe for absorption. Reduction potentials and pH of foods, as well as the type of complexes
formed between the Fe and other compounds, all will change nonheme Fe solubility and
therefore its availability (Smith, 1983; Clydesdale, 1988). Nonheme Fe becomes part of a
luminal pool of Fe which may interact with other meal components that alter the availability
of nonheme Fe up to ten-fold, depending on the diet composition (Hallberg, 1974; Hallberg
and Rossander, 1982). Meal components which move the equilibrium of Fe” and Fe3+
towards reduced Fe2+ increase the Fe availability and absorption. Those meal components
which react with Fe in the gut may be grouped as enhancers or inhibitors based on their efl‘ect
on Fe solubility and availability. All enhancer of mineral bioavailability is a molecular species
which forms soluble compounds with a mineral that 1) can be absorbed as such, 2) can be
cleaved to release the mineral, or 3) has stability constants that allow transfer of the mineral
to a mucosal acceptor. An inhibitor forms insoluble compounds which cannot be cleaved

and cannot deliver the mineral to the mucosal acceptor (Clydesdale, 1988). Some of the

9

meal components associated with changes in Fe availability include organic acids, amino
acids, protein digestion intermediates, carbon structural compounds (fiber, lignin and pectin),
sugars and polyphenols.

Enhancers of nonheme Fe availability include reducing or chelating agents such as
organic acids, ascorbic and citric acids, meat protein digestion intermediates and certain
amino acids. Ascorbic acid is more efi‘ective than other organic acids- citrate, malate or
lactate, in increasing Fe absorption (Gillooly, 1983), due to its ability to decrease meal pH,
change the meal reduction potential and form soluble chelates with Fe (Hallberg, 1981;
Clydesdale et al., 1991). Ascorbic acid most effectively reduces ferric to ferrous Fe between
pH 1.5 and 5 and probably contributes to the liberation of Fe from ligands which may be
insoluble during acid digestion (Kojima et al., 1981). Even if partially destroyed by heat,
ascorbic acid in orange juice may enhance Fe availability compared to controls with no added
ascorbic acid (Hazell and Johnson, 1987b). A molar ratio of greater than 7.5 ascorbate to
Fe (100 mg ascorbatez3 mg Fe) will give only moderate increases in Fe availability (Monsen,
1988a). Citrate may be more effective than ascorbate at enhancing nonheme Fe availability,
at levels found in fi'uits and vegetables, and has maximum enhancing activity at pH 7.0.
When the two acids are combined, such as in orange juice, Fe absorption increases to a
greater extent than with the acids supplemented individually (Hazell and Johnson, 1987).
Theadditiveefl‘octofthesetwoacidsalsohasbeenobserved inFe solubility fi'om pinto beans
(Chidambaram et al., 1989). Citric acid also decreases Fe-binding by dietary fiber,
particularly neutral detergent fiber (Reinhold et al., 1981; Suzuki et al., 1994).

Oxalic acid has been cited in the literature as inhibiting Fe absorption, yet there have

been no detailed studies documenting this effect (Darrell Van Campen, personal

10

commmieation; Welch and House, 1984). Oxalate forms insoluble precipitates with calcium
and can decrease calcium availability, but its effect on Fe has not been confirmed. When
added to a meaL oxalate has been found to enhance Fe availability both in vitro (Kojima et
al, 1981) and in animals (Van Campen and Welch, 1980; Gordon and Chao, 1984) or have
no cfl‘ect (Hallberg, 1981). Both ferric and ferrous oxalates are water-soluble and would be
present as low-molecular—weigllt complexes which are available for absorption (Van Campen
and Welch, 1980). Between 50 and 70% of an Fe dose was absorbed by Fe suficient rats
in the presence of 0.75% oxalate (Van Campen and Welch, 1980). When oxalate was added
to a pinto bean digest, soluble Fe concentrations increased (Kojima et al., 1981).

The intermediate breakdown products of some cellular proteins fiom meats have an
enhancing effect on nonheme Fe availability. Addition of meat (beef, fish or poultry) to a
vegetable diet enhanced Fe absorption 1.5 to 4 times (Cook and Monsen, 1976). Not all
meats have the same effect, beef being more enhancing, followed by chicken and fish.
Breakdown products from noncellular animal products, such as eggs and milk, do not enhance
availability and may even decrease it (Cook and Monsen, 1976; Berner and Miller, 1985).
Protein digestion products may either solubilize Fe by complexing food Fe to prevent
precipitation, or decrease Fe solubility through formation of insoluble complexes or soluble
complexes which do not release Fe to the mucosal receptors (Berner and Miller, 198 5). The
quality of the protein digestion intermediates, particularly high-cysteinecontaining
polypeptides, appears to be more important than the amino acid end products (Slatkavitz and
Clydesdale, 1988). However, this protein efi‘ect may be very dificult to interpret or

distinguish in high-Fe-containing materials, such as soybeans (Berner and Miller, 1985). The

l 1
rate of protein digestion also may be very important to enhancing ability, by influencing the
size and number of polypeptides able to chelate the ferric Fe to maintain solubility.

A few amino acids have been found to act as enhancers and may act to chelate Fe to
prevent precipitation (Berner and Miller, 1985). I-Iistidine and lysine significantly enhanced
Fe absorption in ligated duodenal segments of rats, and modification of the amino-acid-
ionizable groups eliminated the activity (Van Campen and Gross, 1969). Cysteine also was
reported to enhance Fe availability, but fiee sulfliydryl groups were essential for activity
(Monsen, 1988a).

Certain vitamins enhance Fe bioavailability and assimilation. The efi‘ects of ascorbic
acid (vitamin C), have been described. Vitamin C also is required for hemoglobin synthesis
and Fe reduction for binding to fenitin (Lonnerdal, 1988). Addition of both ascorbic acid and
vitamin E to a diet increased bioavailability of ferrous sulphate by 33%, but not of ferric
orthophosphate, a poorly available Fe compound (Fritz et al., 1970). Vitamin A deficiency
impairs Fe incorporation into the erythrocyte in hematopoiesis. Massive doses of vitamin A
can improve Fe status within two weeks of administration, increasing hemoglobin, serum Fe
and hernatopoiesis (Bloem et al., 1990). Under conditions of riboflavin deficiency, mucosal
ferritin-Fe release may be decreased, thus contributing to greater Fe loss by mucosal
sloughing (Lonnerdal, 1988).

Inhibitors of nonheme Fe absorption include noncellular animal proteins, tannins,
phytate, calcium phosphate, and high levels of other divalent cations, such as Zn and Cu.
The inhibitory efi‘ect of fiber is debated (Cook, 1983; Gordon and Chao, 1984; Reinhold et
al., 1981; Clydesdale, 1988). Lignin associated with fiber may be responsible for the

inhibitory efi‘ect, because it has been shown consistently to bind Fe with high affinity

12
(Clydesdale, 1988). Noncellular animal proteins, such as eggs, milk and cheese, have been

shown to decrease Fe absorption in meals (Monsen, 1988b). Polyphenols, such as tannins
in tea, which have a high content of Fe—birlding galloyl groups can reduce Fe absorption from
a meal byup to 80% (Disler et al., 1975; Brune et al., 1989, Monsen, 1988a). Phytate, the
primary phosphorous storage form in seeds (grains and beans) also decreases Fe availability
(Simpson et al., 1981). Variability in phytate content of seeds, such as soybean, may afi‘ect
Fe availability from these foods (Berner and Miller, 1985). Inclusion of high-phytate-
containing oat products in abreakfast meal reduced the nonheme Fe absorption 40-50% when
compared to the standard meal (Rossander et al., 1990). These oat products had been heat-
processed, destroying endogenous phytase activity. Other research has not found any change
in Fe availability due to naturally occurring phytate, but other minerals, such as zinc, may be
effected (Welch and Van Campen, 1975; Welch and House, 1982; Hallberg and Rossander,
1982)

Excessive intake of other divalent ions, such as zinc, cobalt, copper and manganese,
may decrease Fe availability, suggesting common or competing pathways for absorption
(Monsen, 1988a). Zinc, which is in the same transition series as Fe, may inhibit Fe
absorption competitively, especially under conditions of Fe deficiency or Fe excess in a zinc-
deficient individual. Excessive intakes of either mineral may influence absorption of the other
(Solomons, 1986). Calcium and sodium phosphate, when individually added to a semi-
synthetic meal, have a small effect on nonheme Fe absorption, yet when added as calcium
phosphate, they decreased Fe absorption. This compound may form aggregates with Fe

compounds to decrease the release of Fe (Monsen and Cook, 1976).

13

Inorganic Fe from soils has a very low bioavailability. In some regions, geophagia
(eating of soil) has been observed. However, the inorganic molearles containing Fe are highly
resistant to the moderate acid conditions of the stomach and are not attacked by the

hydrolytic enzymes of the gastrointestinal tract (W apnir, 1990).

Methods for Estimation of Fe Bioavailability from Foods

Cunent methods used for estimating Fe bioavailability include human, animal and in
vitro protocols. Accurate measurement of Fe availability to humans is expensive, dificult
and time-consuming (Miller and Shricker, 1982). Comparing estimates from different
laboratories may be problematic due to subtle difi‘erences in protocol (Forbes et al., 1989)
and lack of agreement among researchers on the definition of bioavailability (Van Campen,
1983). Some of the problems associated with trying to estimate the availability of Fe fiom
plant foods include: a) the major chemical forms of Fe absorbed fiom plants is unknown, b)
meals have inhibitors and enhancers of Fe absorption as well as antinutritional factors, c)
human diets are complex and diverse, (1) human Fe absorption is not well understood, and e)
it is difficult to use human subjects for research (Welch and House, 1984). In addition,
contamination of plant material with soil may lead to error in estimations of total Fe and
relevance of percent bioavailability data (Van Campen, 1983).

Human F e—feeding studies have been conducted utilizing balance studies (McMillan
and Johnson, 1951), bernoglobin repletion in anemic individuals (Devadas and Saroja, 1980),
or radiolabelled Fe feeding (Layrisse et al., 1969; Hallberg, 1980; Martinez-Torres et al.,
1986). However, relatively few researchers are licensed to administer radioisotopes to

human subjects (Miller and Schricker, 1982). Balance studies involve measurement of Fe

14

intake and output (in wastes), and hemoglobin repletion efliciency is calculated based on
estimated Fe absorption In studies using radiolabelled Fe, retention of the label from a single
meal is measured, with the assumption that the absorption of Fe from the test meal reflects
the average Fe absorption over time. However, some researchers have questioned the
appropriateness of these tests, especially with vegetable Fe (Smith, 1983). Researchers
conducting long-term feeding studies with humans (greater than four weeks), have found the
Fe absorption fi'om GLVs was greater than absorption M a meat-plus-GLV diet (McMillan
and Johnson, 1951). Short-term studies, especially those involving the administration of a
single radiolabelled Fe test dose, may not reflect the actual Fe absorption (Zhang et al., 1989;
Thompson, 1988).

Of the radiolabelling tests, two predominate: the extrinsic tag and the intrinsic tag.
In the extrinsic tag technique, subjects are fed diets spiked with SS'Fe, and percent absorption
is determined by the amount of labelled Fe assimilated into hemoglobin after two weeks
(Consul and Lee, 1983). Intrinsic tag studies involve incorporation of the radiolabelled Fe
into the test material, by growing or feeding the test materials (plants or animals) with
radiolabelled Fe, and then measuring the amount of radiolabelled Fe absorbed from these test
materials after feeding to the subject (human or animal) (Layrisse et al., 1969). The extrinsic
tag method involves adding labelled Fe to the diet and measuring absorption or assimilation
at a later date. It is assumed that the tag is completely in exchange with the nonheme Fe pool
in the gut and behaves as intrinsically bound Fe, although it is now recognized that not all
food sources of nonheme Fe undergo complete exchange with this pool (Amine and Hegsted,
1971; Bothwell et al., 1979; Smith, 1983). Exchangeability of Fe from soy protein extracts

tagged both extrinsically and intrinsically were not equal (Smith, 1983). When incubated

1 5
with added Fe, the extrinsic tag was observed to exchange with the added Fe, whereas the

intrinsic tag did not exchange. In turnip greens, extrinsic tagging overestimated the amount
of bioavailable Fe, when compared to intrinsic tagging (VVlen et al., 1975).

Animal studies allow bioassay of Fe availability with an intact biological system. They
ofi‘er greater flexibility for testing Fe bioavailability, depending on animal size and
expendability, and are relatively simple and widely available to many researchers. Rat or
chick hemoglobin repletion tests are common methods for Fe availability estimation and have
been proposed as standardized bioavailability tests (Van Campen, 1983; Forbes, et al., 1989).
The piglet also has been described as a good model for Fe absorption studies (Howard et al.,
1993). Labelled Fe can be fed to animals, whole body counts taken, and hemoglobin and
liver Fe stores determined to calculate fairly accurately the levels of Fe absorbed (Smith,
1983). However, results obtained with these assays are relative measures of Fe
bioavailability and may be difliarlt to extrapolate to humans. Use of Fe-deficient animals will
contribute to greater absorption levels, and some animals (i.e. rat) have been found to have
difi‘erent absorption efiiciencies than humans (Reddy and Cook, 1991).

In vitro methods for assessing Fe bioavailability have the advantage of low cost, high
repeatability and rapid processing, especially where there are many samples to be compared
(Miller and Schricker, 1982; Latunde-Dada, 1991). These methods are useful for making
relative comparisons of Fe availability, but do not provide absolute numbers for absorption.
Variation in in vitro bioavailability estimates of the same food material occurred when
analyzed by different laboratories (Forbes et al., 1989). Subtle difi‘erences in analytical
methods may explain these different results. In vitro methods cannot provide information on

variation caused by human intrinsic factors, such as Fe status or transit times (Hazell and

16
Johnson, 1987b). However, substantial agreement (r2=.96) between in vivo and in vitro Fe
availability measurements has been reported (Schricker et al., 1981).

Most current methods of in vitro Fe availability estimation are based on simulation of
gastrointestinal digestion of a food, using commercially available enzymes, and measurement
of the soluble Fe released fi'om that food. Measurement of only soluble Fe, without
digestion, is inadequate for estimating bioavailability of Fe, because it may be chelated in
soluble forms resistant to the digestive enzymes, or are too large for adsorption to mucosal
receptors (Reddy et al., 1986). The Rao and Prabuvathi (1978) method measures 'ionizable'
Fe by colorimetric reaction with a,a' dipyridyl. This method exposes the test food to
digestion by hydrochloric acid and pepsin An aliquot is centrifirged and filtered to determine
acid-soluble Fe. Another aliquot of the supernatant is adjusted to pH 7.5 and incubated. This
subsequent digest is filtered again and analyzed for ionizable Fe. No pancreatic enzymes are
used, and the insoluble food contents are separated fi'om the soluble Fe components in the
second step. Precipitation or polymerization of Fe may be changed if pH modification of the
sample includes the complete insoluble digest fraction (Miller and Schricker, 1982).

The Miller methods (Miller and Schricker, 1981; Kapsokefalou and Miller, 1991) of
in vitro Fe availability estimation assume that Fe will react with ligands in both the acid and
pancreatic digestion steps (Miller and Schricka, 1982). In addition, the Miller methods make
twomajorchangesintheapproach. First, abufferisusedtoraisethe pH slowly after the acid
digestion. Rapid changes in pH can precipitate Fe. Chelates may be unstable in the presence
of concentrated base, and thus may precipitate (Miller and Schricker, 1982). Second, a piece
of dialysis tubing is used to screen for low-molecular-weight Fe compounds. Bufi‘er inside

the dialysis tubing difl‘uses slowly and allows gradual change of digest pH. Previous research

1?
concluded that the low molecular weight, soluble Fe compounds are the most important for

absorptionbecauseoftbeireasytransporttothenalcosal receptors (BernerandMiller, 1985).
Using dialysis tubing, one can select for small size, 'dialyzable' Fe and equate this with

bioavailable Fe (Schricker et al., 1981).

Iron Nutritional Quality of Green Leafy Vegetaqu

Iron content of green leaves is higher than in other plant organs, not including seeds
(Quarterman, 1973; Gupta, 1992). Based on total Fe concentration, GLVs would appear to
be excellent sources of this mineral to the diet (Quarterman, 1973). Spinach Fe content has
been reported to be two thirds the level found in meat on a fresh weight basis and two to
three times the level on a dry weight basis (Duhaiman, 1988; Zhang et al., 1989). However,
the bioavailability of Fe in GLVs is generally low compared to other sources of dietary Fe.
Explanations for the low availability of Fe fi'om GLVs are based on the prevalence of
nonheme forms of Fe in plant tissue. The specific chemical forms of nonheme Fe that may
be more available to humans have not been identified (Quarterrnan, 1973; Welch and House,
1984). In addition, several factors can influence Fe accumulation and bioavailability in
GLVs, including the site and methods of production, genetic variation among species, the
physiological age and maturity of harvested plant parts, and postharvest handling and
processing A general description of the localization and function of Fe compounds in plants
is provided prior to reviewing estimates of Fe bioavailability from GLVs and factors which

may influence these values.

18
Iron Levels, Forms, Functions and Localization in Green Leaves

Iron is an essential nutrient required for normal plant growth and development. Due
to the potent ability of Fe to generate hydroxyl radicals (mutagenic), Fe concentration in most
organisms, including plants, is highly regulated (Guerinot and Yi, 1994). Because of these
elemental properties of Fe, plants (and all aerobic organisms) have developed sophisticated
absorption, transport, andstoragemechanismstomaintaincontrolofFe concentrationswithin
the cell (Neilands, 1987; Romheld, 1987; Bienfait, 1989; Williams, 1990).

The interaction of soil, plant and climatic factors affect Fe accumulation in plants.
Although Fe is abundant in the soil environment, it has limited solubility at neutral pH under
aerobic conditions (Romheld, 1987; Williams, 1990; Kabata-Pendias and Pendias, 1992).
The rate of Fe uptake in plants is under metabolic control and can be affected by plant
physiological state, genetic background, and plant Fe status. In addition, the Fe uptake rate
is dependent on the concentration of the soluble, chelated Fe forms which predominate in the
soil solution (Marschner, 1987). This soluble soil Fe fiaction can be afi‘ected by the soil pH,
organic matter content, bicarbonate and water content, and divalent cation concentrations
(Quarterman, 1973; Tinker, 1981; Romheld and Marschner, 1986; Marschner, 1987; Bienfait,
1989; Romheld, 1990; Kabata-Pendias and Pendias, 1992; Guerinot and Yi, 1994). Soil pH
may be considered the most important factor afi‘ecting Fe content of plants, and an increase
in pH from 5 to 6 decreases the availability of soil cation trace elements to plants by one half
(West, 1981). Plants have evolved several mechanisms, both specific and nonspecific, to
improve uptake of these Fe compounds by roots (Bienfait, 1987 ; Romheld, 1990; Guerinot
and Yr, 1994). These systems primarily involve the release of protons or organic acids into

the rhizosphere to reduce or chelate Fe for transport to the root surface, or to stimulate

l9

microbial activity indirectly for the same efi‘ect. Plant breeding to select for high Fe uptake
eficiencies is considered the most promising approach to improve Fe uptake in crop species,
especially for production in high pH soils (Romeld, 1987; Kabata-Pendias and Pendias, 1992).

The interaction of these plant and soil factors with climatic variables can contribute
to large variation in the total Fe concentration of a plant. Up to a 500% difference in total
Fe concentration has been observed with crop production at difl‘erent locations (Bear et al.,
1948). Greenhouse-produced crops consistently have lower Fe concentration compared to
field-grown cr0ps, despite the greater availability of Fe fi'om nutrient solutions (Romheld,
1987; Abadia, 1992). The difl‘erences in light intensities between these two enviromnents has
been considered the most significant factor contributing to the variation in total Fe
accumulation (Abadia, 1992). However, the potential interactions of field grown plants with
rhizosphere microorganisms which enhance soil Fe solubility or transport to roots also may
be important (Romheld, 1987).

Within plants, the Fe3+ ion dominates in Fe transport and storage and is the
predominate form bound to proteins (F e2+ does not bind as strongly to proteins as Fe”)
(Williams, 1990). Because of the small size of Fe” , stearic hinderances prevent bulky protein
ligands fiom binding directly with Fe”. Therefore, Fe” in proteins generally is bound as a
mixed valence poly-metal center (Fe,s,) (nonheme Fe proteins) or in a specialized ligand-the
porphyrin (heme Fe proteins). Nonheme Fe proteins include ferredoxin and several
roductases (Table l). The F e,S, centers have more negative redox potentials than the heme
centers and function in single electron transfers (i.e. ferredoxin) (Williams, 1990) and are

important in nitrite and sulfite reduction reactions (Kabata-Pendias and Pendias, 1992).

20

Iron, a transition metal, tends to form coordination compounds with six ligands,
especially those containing oxygen, nitrogen and sulfur, such as organic acids, porphyrin rings
or proteins (Guerinot and Yi, 1994). These ligands contribute to the large redox range
observed for these coordinate groups and allow for multiple firnctions of heme Fe enzymes,
including: 1) the formation of coordination complexes, such as a bridge function between
substrate and enzyme and 2) reversible redox reactions in the energy transfer pathways of
respiration, photosynthesis and nitrogen reduction (Marschner, 1987; Williams, 1990;
Guerinot and Yi, 1994). These proteins containing Fe often are bound to membranes,
limiting Fe exchangability fiom these compounds.

Estimates of the amount of heme Fe in plant materials range fi'om 0.1% (DeKock,
1960; Mengel and Kirkby, 1979) to 9%, most of which (90%) is found as cytochromes
(Hewitt, 1983). The remaining fiactions of total plant Fe include 19% as nonheme Fe
(predominately as ferredoxin) and an additional 35% bound to ferritin, totaling 63% of plant
Fe as protein bound (Hewitt, 1983). A significant fi'action of Fe may be found in the cell wall
(Tinker, 1981). However, the function of most of the Fe in plants remains unknown (Hewitt,
1983; Smith, 1984).

Within the plant cell, most of the Fe is found within the strongly reducing
enviromnents of the mitochondria and chloroplasts (Neilands et al., 1987). The chloroplast
stroma contains 80% of the total plant Fe (Price, 1968; Tirnperley, 1982; Smith, 1984). In
rapidly growing sugar beet leaves, 3/5 of plant Fe was found in the thylakoids, 1/5 in the
stroma and 1/5 external to the chloroplast (Terry and Low, 1982). Even under conditions
of Fe deficiency, the chloroplast contains most of the plant Fe; however, the subcellular

localization changes fi'om the stroma to the lamellar chloroplast fiactions, probably due to

21
decrease in the stromal storage Fe form, phytoferritin (Seckback, 1982). The mitochondrial

Fe concentrations are lower than the chloroplast concentrations, and Fe concentrations
decrease fiirther within the cytosol (Neilands et al., 1987). The role of the vacuole in cell Fe
storage or sequestering has not been well investigated (Laulhere and Briat, 1993).

Most of the chloroplast stromal Fe is in the storage form, phytoferritin, which can be
observed in the stroma as a crystalline array (Smith, 1984). The structure of phytoferritin
is very similar to animal ferritin, suggesting a common evolutionary pathway (Andrews et al.,
1992). Phytoferritin consists of a protein shell, comprised of 20-24 identical subunits of a
globular polypeptide chain, surrounding an Fe phosphate core with a molecular formula of
(F eOOH),(FeO:OPO,Hz). An estimated 6200 Fe3+ ions can be bound within the core
(Smith, 1984). Phytoferritin can contain between 10 to 35% Fe by dry weight (T erry and
Low, 1982; Marschner, 1987) and has a molecular weight (including Fe) of 900 kDa (Smith,
1984).

The proportion of F e present as phytoferritin in the leaf depends on the physiological
state of the leaf and availability of Fe fi'om the soil environment (Seckback, 1982). Rapidly
growing leaves assimilate high proportions of Fe into chloroplast membranes and less into
phytoferritin storage (Terry and Abadia, 1986). Fe-deficient plants grown in -Fe nutrient
solutions deposited Fe into fenitin upon transfer to +F e nutrient solutions (Seckback, 1968).
Dark-grown plants stored 50% of Fe as phytoferritin, but upon light exposure, the ferritin
levels decreased (Smith, 1984). Mature bean leaves, grown in Fe-adequate nutrient
solutions, contained 7% of total Fe as phytoferritin, but the number of atoms of Fe per ferritin
molecule were lower than found in leaves grown under Fe-deficient conditions (van der Mark

et al., 1982). Under continuous supply of Fe, no ferritin was detectable in roots or leaves of

22

pea (Piston sutivrmr L.) (Lobreaux and Briat, 1991). However, as leaves senesce, greater
portions of leaf Fe may be found in phytoferritin (Laulhere and Briat, 1993).

Outside the cell, plant Fe may be found complexed to cell wall celluloses, pectins and
ligrrins (Tinker, 1981). Due to the oxidizing environment outside the cell, strong complexing
agents would be required to maintain Fe solubility (Williams, 1990). The soluble Fe
transport form, Fe” citrate, may be found in the xylem and in the phloem sap in small

quantifies (Tinker, 1931, Bienfait, 1989).

Fe Bioavailability from GLVs
Amounts and Forms of Available Fe. Although many GLVs contain high concentrations of
foliar Fe, this Fe is poorly absorbed in the human gastrointestinal tract (Layrisse et al., 1969).
Hmnan and animal feeding and in vitro methods have been used to develop estimates of Fe
bioavailability fiom GLVs (Table 2). Most studies involving GLVs have examined spinach
(Soinacea oleracea L.), a temperate GLV. A few researchers have screened popular local
tropical greens. It is diEcult to make comparisons among different studies due to the
difi‘erent protocols and lengths of studies, and methods for collecting the GLV. In human
feeding studies evaluating spinach, the percent of Fe bioavailable ranged fi'om 1.3 to 13%
(Moore and Dubach, 1951; McMillan and Johnson, 1951; Moore and Dubach, 1956;
Layrisse et al., 1969). In feeding studies using anemic rats, the bioavailability of spinach
ranged fi'om 41 to 70% (Table 2). Using Fe-replete rats, spinach Fe bioavailability ranged
fiom 26 to 59% of total Fe. In a comparison of ten tropical GLVs using anemic rats, Ifon
and Bassir (1978) found a range of bioavailabilities from 7.7 to 36.2% and identified a line

of Amarwltlms hybridus L. which Ind the highest percent bioavailable Fe of all species tested.

23
In vitro studies have identified similar levels of total Fe as available for absorption

(1‘ able 2). Bioavailable Fe from spinach, estimated using the Rao and Prabavathi (197 8) in
vitro method, ranged fiom 0.96 to 4.43%, depending on the cultivar and treatment (Reddy
and Malewar, 1992). In another study using a different in vitro method, 7.5% of total Fe
from spinach was bioavailable (Duhaiman, 1988). In a comparison of six GLVs for total and
dialyzable (estimate for bioavailable) Fe, dialyzable Fe was not correlated with total Fe
(Chawla et al., 1988). Total Fe ranged from 51 to 160 mg-kg", but available Fe only ranged
fi'om 3 to 10 mg-kg". Analysis of 14 tropical GLVs, collected fi'om noncultivated areas, for
total and available Fe also indicated availability of Fe not proportional to total Fe (Reddy and
Kulkarni, 1986). Bioavailability ranged from 3.1 to 53.6 %, depending on the species.
Amantlnrs hybridus had the highest percent bioavailable Fe (10.1%) of 12 tropical GLVs
(Latunde-DaDa, 1990). The estimated in vitro bioavailabilities ranged from 1.8 to 10.1%.
The forms of F e in plant material may effect the amount of Fe solubilized. However,
specific bioavailable forms have not been identified (Welch and House, 1984). Upon
fiactionating the total Fe into soluble and insoluble forms, 93% of spinach Fe was found in
the insoluble fraction (Lee and Clydesdale, 1980). After extracting homogenates of turnip
leaftissue with a series of acids, 36% of the total Fe was found in the insoluble fi'action (Wien
et al., 1975). Chromatographic separation of Fe-binding proteins from soybeans grown in
soil treated with ”F e indicated the majority of Fe in a large protein fi'action (>600 kDa), fi'om
which Fe was not exchangeable (Smith, 1983). In soybean flour, most of the Fe is bound to
phytoferritin The low bioavailability of Fe from soybean may be due to this ferritin form of
storage Fe (Lynch and Covell, 1987). Layrisse et al. (1975) suggested that absorption of

animal ferritin Fe also is afi‘ected by enhancers and inhibitors in the diet, but is absorbed to a

24

lesser extem than other nonheme Fe forms. Plant ferritins, which have similar structures to
animal ferritins (Smith, 1984; Andrews et al., 1992), also may be absorbed to a lesser extent
than other nonheme Fe forms. However, the significance of ferritin in GLV Fe bioavailability
is questionable, due to the low levels of ferritin detected in the leaves of plants grown with

an adequate Fe supply (van der Mark et al., 1982; Lobreaux and Briat, 1991).

Interactions of Fe Bioavailability and Secondary Plant Compounds. Many compounds
which have been correlated with changes in the bioavailability of Fe in meal studies are
present in leaf tissue. These compounds may interact with the nonheme Fe pool in the
gastrointestinal tract either to enhance or inhibit Fe absorption. Organic acids, tannins,
phosphorous, calcium, fiber and protein are compounds which are present in leaf tissue and
may influence Fe bioavailability. The concentration of many of these compounds may vary
withthecultivar, growingerlvironmentandpostharvesthandling. This reviewwillfocus only
on the changes in Fe bioavailability of GLVs associated with the presence of these
compounds.

Organic acids predominately interact with Fe to enhance bioavailability. Ascorbic,
citric and oxalic acids have been studied in relation to GLV Fe bioavailability. Ascorbic acid
concentration correlated positively to the percent of bioavailable Fe in spinach (Reddy and
Malewar, 1992). Addition of ascorbic acid to 12 tr0pical GLVs prior to in vitro digestion
increased the percent Fe bioavailable fi'om a range of 1.8-10.1% to 6.1-16.9% (Latunde-
Dada, 1990). However, another researcher found no relationship between endogenous
ascorbic acid content and in vitro estimates of Fe bioavailability from spinach and amaranth

(Chawla et al., 1988). The amaranth species contained three times the ascorbic acid as the

25
spinach, but had the same fraction of Fe available, 2.8% of total Fe. Other plant compound

interactions may have decreased amaranth Fe bioavailability. Analysis of ascorbic acid,
oxalic acid and phosphorous in spinach showed a combined influence (RL—.99) in multiple
regression analysis among these three variables and the percent of total Fe bioavailable
(Reddy and Malewar, 1992). Strong positive correlations were detected between the percent
of Fe bioavailable and ascorbic acid (r=0. 82) and oxalic acid (r=0.85) and a negative
correlation with phosphorous (r=0.53). However, in a second study by the same researcher
(Reddy et al., 1993), no correlation was detected between percent Fe bioavailability and
ascorbic acid content. The negative correlation between percent Fe bioavailability and
phosphorous content was confirmed (r=-.82).

Both ascorbic and citric acid showed positive correlations with Fe bioavailability fiom
16 different fruits and vegetables, although no GLVs were analyzed (Hazell and Johnson,
1987b). When combined, citric and ascorbic acid showed greater activity in enhancing Fe
availability than if supplemented separately. However, Hazel] and Johnson argued that citric
acid was more important than ascorbic acid in Fe interactions, because citrate enhanced Fe
availability fiom meals when present at levels commonly found in plants, whereas ascorbate
was required at levels greater than are commonly present. Citric acid was inefi‘ective at
solubilizing Fe fi'om pinto beans at low pHs, but maximum solubilization occurred at pH 6
(Kojima et al., 1981). The sz of citric acid are at 3.5, 4.5 and 5.8. Low pH protonation
may decrease the ability of citric acid to bind and liberate Fe from foods. Citric acid also
strongly inhibits binding of Fe by dietary fiber (Reinhold et al., 1981), although the role of
dietary fiber in Fe availability is questioned. No studies have examined citric acid in relation

to Fe bioavailability fi'om GLVs.

26

Oxalic acid has been cited in the literature as an inhibitor of Fe absorption, yet there
have been no detailed studies documenting this efi‘ect (Darrell Van Campen, personal
communication; Welch and House, 1984). Solubility of simple oxalates of divalent cations
depends upon pH, concentration of competing cations and concentration of oxalate (W elch
et al., 197 7). Oxalate forms insoluble precipitates with calcium and decreased calcium
availability to rats, but increased Zn availability (Welch et al., 1977). Both ferric and ferrous
oxalates are water-soluble, and should yield low molecular weight complexes which are
available for absorption (Van Campen and Welch, 1980). When added to a meal, oxalate has
been found to enhance Fe availability, both in vitro and in rats. Between 50 and 70% of Fe
from spinach was absorbed by non-anemic rats when amended with 0.75% oxalate (Van
Campen and Welch, 1980). Other research on Fe availability from GLVs (including spinach)
found no correlation with oxalate Concentration (Gillooly et al., 1983; Gordon and Chao,
1984; Oyejola and Bassir, 1975). Amaranthus spinosus and spinach, which had the highest
average levels of oxalate of six GLVs examined, 7. 83 and 6.62 g-kg'l respectively, tended
to exhibit low Fe availability, however, no statistical analysis was possible due to the
experimental design (Chawla et al., 1988).

Tannins (polyphenols) have been associated with poor availability of Fe from plant
foods (Gillooly et al., 1983). Tea, having a high concentration of tannins, has been shown to
reduce Fe bioavailability in humans (Disler et al., 1975) and rats (Fairweather-Tait et al.,
1991). In a survey of Fe bioavailability fiom several vegetables, including spinach, beet
greens and cabbage, condensed polyphenol content was correlated negatively with Fe

bioavailability (Gillooly et al., 1983).

27

Increasing phosphorous content in spinach leaves has been associated with decreased
Fe bioavailability (Reddy and Malewar, 1992; Reddy et al., 1993). Phytate, the phosphorous
storage form found in seeds, such as cereals, legumes and nuts, decreased Fe availability fiom
meals (Gillooly et al., 1983; Hazel] and Johnson, 1987). Phytate forms insoluble precipitates
with many cations at pHs similar to those found in the small intestine. However, phytate is
not important in Fe interactions of GLVs.

Fiber exhibited no correlation with available Fe in one study surveying several
different vegetables (no GLVs) (Hazell and Johnson, 1987). However, in early research on
Fe bioavailability from spinach, it was suggested that the high 'roughage' content of GLVs
decreased Fe absorption due to mechanical obstruction (Smith and Otis, 193 7). Gordon and
Chao (1984) examined the fiber components of spinach and wheat bran, and then tested the
effect on Fe availability to rats of different combinations of cellulose, phytic acid, lignin, and
oxalate at levels found in these crops. Spinach had some ”factors” other than those tested
that decreased the availability of Fe. In cereals, the neutral detergent fiber accounts for
nearly all ofthe Fe binding capacity (Reinhold et al., 1981). The amount ofFe bound to this
fiber depends on the Fe concentration, the pH (higher pH= higher binding), the fiber quantity,
and the presence of inhibitors of Fe binding, such as ascorbic, citric and phytic acids, cysteine
and phosphorous and calcium Due to the high ascorbic and citric acid levels in GLVs, little

Fe binding by the endogenous dietary fiber is expected (Reinhold et al., 1981).

Variation in Bioavailable Fe due to Production Site/Soil Fe Concentrations. Few studies
have examined the changes in Fe bioavailability by site of production. Plant Fe concentration

can vary up to 500% when grown in different environments (Vlflen et al., 1975). Total Fe

28

content of lettuce varied fi'om 9 to 16 ug-g" dry weight by location (Bear et al., 1948).
Certain intensive agricultlu'al systems may contribute to trace element depletion in poorer soil
types and decreased nutrient content in food (Welch and House, 1984 ). Soil management
practices which impact Fe solubility (i.e. organic matter management, liming, drainage) can
afi‘ect plant Fe levels (Kabata-Pendias and Pendias, 1992). Micronutrients important to
animals, such as Se, V, Sn, Cr or As, are especially prone to depletion under intensive
production systems (Grunes and Allaway, 1985). The soil pH and moisture content
indirectly will affect dietary Fe intake through their influence on total Fe accumulated in edible
plant parts (Wapnir, 1990). In a comparison of difi‘erent sites with soil pH between 5.1 and
7.0, little difi‘erence was detected in plant Fe content of several vegetables and grasses
(Gupta, 1992). Other soil or climatic factors could have been interacting as well.

The use of Fe fertilization is questionable in relation to increasing Fe content and
bioavailability from plants (Welch and House, 1984; Grunes and Allaway, 1985; Reddy and
Malewar, 1992). Soil media Fe concentrations can have a variable efl‘ect on plant Fe content.
No difi‘erence in Fe absorption or Fe bioavailability was noted in spinach grown in nutrient
solutions with two Fe levels in a growth chamber (Van Campen and Welch, 1980).
Increasing solution Fe concentration led to small increases in turnip leaf tissue Fe (Wien et
al., 1975) and spinach leaf Fe (El-Sherifet al., 1984).

Two studies estimated changes in Fe bioavailability from spinach grown under
difl‘erent soil Fe concentrations in the greenhouse (Reddy and Malewar, 1992; Reddy et al.
1993). In both studies, soils were collected from the field and modified with added levels of
Fe, fi'om 0 up to 50 ppm. Plant Fe concentrations did not increase with increasing soil Fe

concentrations in the initial study, but more consistent changes in plant Fe were observed in

29

the second study. Bioavailable Fe levels did not increase with increasing soil and plant Fe.
An inverse relationship was observed between total Fe and the percent bioavailable Fe. Plants
accumulating higher Fe levels had lower percent of bioavailable Fe. Plants with no added Fe
had the highest percent of dialynble Fe but not the highest absolute amount of dialyzable Fe.

Nitrogen (N) fertilization increased the concentration of beta carotenes, oxalates,
nitrates, and protein (Grubben, 1976; Stafford, 1976; Walters et al., 1988 ) and had a variable
effect on fiber content (Walters et al., 1988), but had little effect on Fe concentrations or
ascorbic acid levels in amaranth greens. Increasing N fertility in soil decreased total Fe
content of eight tropical vegetables (Schmidt, 1971). The increased grth rate associated
with higher N fertility rates may contribute to a 'dilution effect in which the concentrations
of minerals in leaves is decreased, provided mineral concentrations are not growth-limiting
(Tinker, 1981). When turnips were grown in solutions deficient in macronutrients, leaf
tissue Fe concentrations were elevated, but plant grth significantly curtailed (“lien et al.,
1975). When fed to anemic rats, the Fe fi'om plants grown under nutrient-deficient

conditions had a lower percent bioavailability than Fe from nutrient-adequate plants.

Genetic Vtmotion in Fe Bioavailability. Carefirl study of difi‘erences in Fe bioavailability due
to genetic variation within GLVs has not been reported. Genetic difi‘erences in Fe uptake
and bioavailability may be more important than fertilizer utilization for mineral content, but
environmental effects cannot be ignored (Quarterman, 1973; Kelly and Rhodes, 197 5).

Difi‘erent plant species and cultivars within a species may differ in Fe accumulation, but the
availability of the Fe is rarely tested. Investigators may state that because a plant has high

Fe concentrations, it is a good source of Fe for human consumption (Deutsch, 1979). The

30
validity of this statement is questionable, because little is known of the forms of Fe available

fi'om plant foods, or the portion of the total Fe that may be available (W elch and House,
1984). Included in the analysis of 18 wild type GLVs for total and bioavailable Fe (in vitro

estimated) were four amaranth species, which had Fe concentrations as follows:

Amaranth species Total Fe Avail. Fe %Avail.
(ms-1003") (ms-1003")

Arnaranthus spinosrrs 17.3 2.6 15.2

Amaranthus virdis 13.2 3.3 25.0

Amarantlms species 4.8 1.5 31.7

Amaranthus polygamous 8.2 4.4 53.6

A. polygamous had a very high percentage of Fe available (Reddy and Kulkami, 1986).
From these results, it appears that there may be genetic difi‘erences in Fe availability which
might be exploited through plant breeding. However, these materials were gathered in the

wild, so no control of environmental variation was possible.

Variation in Fe Bioavailability due to Physiological Age or Maturity. Although GLVs are
utilized as mineral nutrient supplements to diets, few research efforts have investigated the
changes in nutrient availability that occur with change in physiological state of the plant.

Some studies have examined Fe content and sometimes bioavailability, over different harvest

dates. Comparisons among these studies are difiicult, due to the widely varied growing

31

conditions of the experiments. Variability in Fe availability among leaves of different
physiological states within the plant has not been explored extensively.

Sweet potato tips were harvested on three difl‘erent days after planting and examined
for total Fe content and percent availability (Pace and Bonsi, 1987). Total Fe content
decreased with age, but the percent of Fe available increased. The actual amount of Fe
available for absorption did not vary significantly among harvest dates. The cause for the
increase in Fe bioavailability as the plants matured was not investigated, but authors proposed
that the increase may have been due to some undetermined enhancing factor in older geens,
or to a change in the forms of Fe available with increasing age.

Average Fe accumulation of four lines of Celosia argentea, another tropical GLV,
increased fiom week 5 to week 15 after planting (Omueti, 1982). However, the total Fe
values between adjacent harvests within a line were highly variable. From week 7 to week
9, total Fe accunurlation of three lines increased 50%, but on subsequent harvests, each line
behaved difi‘erently. This increase may have been due to environmental conditions, but no
explanation was given. The optimum time for harvest to achieve highest Fe accumulation
was determined to be fiom five to seven weeks alter planting.

A doubling of total Fe content was detected in Amaranthus hybridus when harvest
was delayed fiom day 49 to day 63 afier planting (Stafi‘ord et al., 1976). Deutsch (1979)
harvested lines from several amaranth species every five days fi'om day 21 to day 46 after
planting Foralllines, widevariationinFecontentwasdetected between day 26 and day 32,
Fe levels being the highest on day 32. After day 32, Fe levels decreased, but not at the same

rate in all of the lines. This trend was observed in three separate experiments at two

32

locations. However, these changes in Fe content were associated with environmental
conditions and not physiological changes in the plant.

Few of the studies discussed thus far provided detailed information on the leaves
sampled and none have compared the Fe availability from different leaves fi'om the same
plant. As leaves expand and mature, changes in Fe bioavailability may occur. In mineral
analysis of ‘younger‘ (first ten centimeters fi'om apical meristem) and 'older' (second ten
centimeters) leaves of sweet potato, 'oldel’ leaves accumulated higher amounts of Fe
compared to 'younger leaves (Pace et al., 1985). However, the more immature plant parts
generally are preferred for consumption, even if the older sections may have higher nutrient
content. Young turnip leaves contained higher levels of Fe than older leaves but availability
of Fe was not tested (\Vlen et al., 1975).

Other nutritional characteristics of leaves may vary by location on the plant.
Nutritional and anti-nutritional qualities (not including Fe) were examined in different leaves
on the same plant for an Amwanthus tricolor line four and eight weeks after planting
(Prakash and Pal, 1991). Within plants, leaves seven and eight accumulated the highest
protein and carotenoid content, but nitrate concentrations were similar at all leaf positions.
Oxalate levels of leaves increased progessively as leaf position was closer to the apical
meristern Up to 300-fold difference may be detected in vitamin A content between inner to
outer leaves of cabbage (Pirie, 1985).

Fe accumulation and Fe forms also may change as plants make the transition from
vegetative to reproductive states. In Fe-depleted Xanthiwn plants, floral induction decreased
total Fe uptake and ferritin deposition as compared to non-induced plants (Seckback, 1982).

Changes in Fe bioavailability of GLVs in relation to changing physiological state have not

33

been explored. Due to the reported decrease in leaf palatiblity of GLVs which occurs with
flowering (Reddy and Kulkami, 1986), such changes may not have relevance for nutritional

quality research.

Variation in Fe Bioavailability due to Postharvest Handing and Processing. Few studies
have investigated the changes in Fe bioavailability that occur due to postharvest handling and
processing practices. Changes in ascorbic acid levels of GLVs during postharvest handling
may alter Fe bioavailability indirectly. Because ascorbic acid is highly reactive, and can be
deactivated by light, oxygen, heat, enzymes and metals and leached by water during cooking,
prediction of the concentrations in plant material is difiicult without detailed postharvest
storage information (Albrect et al., 1991; Clydesdale et al., 1991). In studies on postharvest
preservation of tropical GLVs, ascorbic acid content decreased regardless of method
employed (refiigeration, blanching and fi'eezing, air drying or sun drying) (Akpapunam, 1984;
Onayemi and Badifu, 1987). Blanching followed by freezing was found to be the best
preservation method for ascorbic acid retention in tropical GLVs, but is not practical in most
developing countries (Akpapunan, 1984; Onayemi and Badifu, 1987). Cooking methods also
contributed to siglificant decreases in oxalate, beta carotene and ascorbic acid, but steam
blanching preserved the most of these nutrients in amaranth and other tropical GLVs
(Stafi‘ord, 1976; Ajayi et al., 1980; Akpapunan, 1984; Onayemi and Badifir, 1987)
Postharvest washing and cooking has been investigated for effects on the
bioavailability of Fe. Between 28 and 47% of the total Fe concentration of 12 tropical
GLVs was discarded in the water used for leaf washing and blanching. A decrease in the

percent of bioavailable Fe was detected after blanching, but addition of ascorbic acid to the

34
blanched material increased the percent of bioavailable Fe siglificantly (Latunde-DaDa,

1990).

The effect of heat processing (canning) on Fe solubility and bioavailability also has
been investigated. Canning of several fiuits and vegetables increased Fe bioavailability (in
vitro estimated); however, no GLVs were examined (Hazell and Johnson, 1987a). Canning
of spinach increased the amount of soluble Fe from 7 to 15% compared to the unprocessed
control (Lee and Clydesdale, 1980) and from 20 to 40% in another study (Miller, 1987).
However, canning of spinach did not siglificantly increase the Fe bioavailability to rats
(Miller, 1987). The rats received 18 mg-kg'l Fe from either raw or blanched spinach, and
the relative biological value (compared to FeSO,) was 44 or 52%, respectively. The
increased soluble Fe detected alter processing may contribute to different levels of

bioavailable Fe but further study is needed.

35

Table 1. Iron-containing mes and stage proteins and their cellular locations.‘

 

Enzyme Fe form Location
cytochromes heme mitochondria, chloroplast
catalase heme microbodies
peroxidase heme microbodies, cell wall
nitrate reductase heme cytoplasm

ferredoxin nonheme chloroplast

succinic dehydrogenase nonheme mitochondria

NADH dehydrogenase nonheme mitochondria
xanthine dehydrogenase nonheme cytoplasm

pyruvate dehydrogenase nonheme mitochondria
aconitase nonheme mitochondria

nitrite reductase nonheme chloroplast

sulfite reductase nonheme chloroplast
phytoferritin nonheme chloroplast

 

 

lTable adapted from B. Smith, 1984, Iron in higher plants: Storage and metabolic role.
Journal ofPlant Nutrition, 7(1-5):759-766.

36

fib— .._o .0 =83

Sansone-21:5

Smo— 1.538535

8.2 tease-nae.

22 91.1.8.8

Sagas?

82 €ng

e2. £35 2382
:2 6853 138.3%:
32 £352.28:

 

.iséeaassaefieasiaaeséssaiaeeg.i§3 .33:-
Saiacflsosa grasses: easoauaaasiueaaaaustiesnzasu service‘s-use aa>§§L>m§

N

Baggage?

.3.— 01:55
E8303. o..—

.585 a .e .53.... .85. a seam"

33:32..

.3. 38.83.51.333“.
Sega 628388830".
8838383.."—

 

fine

$0 8
38 cm
38 an
38 «N

88

8.2

hum
v."
v. —

8o:
9

h.—
2
m—

99:8 A63
.33. .58 a £13

£33....

 

 

 

.8

58::

5:5:

5:5:
5:5:
5.5:

in
382

3:50.. mo ape—ago... co: no 53 35 .N 033.

37

  

5.393.892 2383.955: .gésaéscguasieefiasu 336325.988... 81> insane-9mm.

2.9 .3 a:- co..—

aa. 33:36

 

 

.isés?88ana.§a§8§£a.§§.39§§£§la§§ seas.

 

5883329...."

88353-

5033.33.38
anemogseaaasesaa
€393. .282..st

 

8.5 an:
6.8. 4.3.
8.2. E
6.3 3.3
8.5 2.8
ads 82
8...... Gen
2.8. 8.3
$3.. 8.2
GB 8..."
8% X
a... 3.
82. no
88 a
.8: S
a... a.
88 .n
as .v
E. m.
A .8 s.

 

o;

Orv

20—

a;

 

are... state...
assures eases
3% 3.93.
2.8. 3.33..»
23.5.32 5.80
8.52....» ham
33%.... 3.95.1
gang 535
3.5.3 seesaw
3535 size.

.Eu. 253.33..ng

a; 3310
.52». 385 gob

.5265 553....
its. $83...
g 83.3
3&3: .53
at? 38.8.
33a 58.3.

 

 

”5.33.8430

.92 8 AER—v in
.?< 30H: 9"— ah.

as a 2......

 

we

38

.85.853§§§§8§£§.§§8§E§Ia8§ 28......
8.32.3.8... 88.828.888.288: .§§§§§.§§.2§22§. 858388823988... .3=>_88222»§A>me.
J

 

 

 

 

 

 

2... 2». 8.86

2% 3.328188%

«.9. 883233.:

3.. 8.88328...

a... 832338....

S" 8......» 2&5

3.. 38...... 38.38....

28 5983.385.

8.8 .382... 3225....

can 888.3853...

5. .m 8...»... 28.385.

22 S...» .3385...

65.35»... 3:38.283. 1.. %355:
88132235»... 3:332:22... 82. 8828.28 2: 28...». 3.8.2.2... a... s

a... .25.. as on
82 .1835 882...»... as s. as 8.5... .8
$2 .852 a... 2...»... 3. 8.5... .8
«81.18.55 . 822...»... 5. =85... .8
3818.62.55 822...»... as." 8...... .8
82 88802.86 282...»... an. 8. =85... .8
a... .282. a. an :

823832-8933. 83335.3»... 282...»... 8a.... .8
in
.282

 

'5

39

.sfiésfiguagisaziaggissgggg «3......
3393.52 £383.32 §§§a§.§.89§§.§ 35633855588... .8_.>_833m§_.3u$m6_

 

a... £3.63.

82 533.2 v5 >33—

.8185;

82 ducal—=9

83:18:38.0

  

 

83$ 3506 2.33.5 v:- and

28:33:35

afiwaiasgsa
3833355335233.

 

«.m—
«v.9

end
mad
avd
he.“
86
3.—
2:
g."
«Eh
vnd
8.“

v.3
as
ad
hm
of
fin
9n

 

um—

N»—
Na—
finm
fine
v.9
”.2.
03.
how
ofin
mdm
93

N5
an _

 

goo—fir

=85...

sages... 5.26
£35398 3%:
:5? SEE:
§§o£ £3.33.
858969.. £3.33“.
gang .35.:

33.3 3.88:1
3.5.8. 33...:

.38 :85. 3.2.... 38.6
.38 3.. 3.2.... 3.8.6
in}... 53.58820

:85...

88.33 83......
at»: gawk
igufigth
a»: Bug:
£29535. 33830
39...... 3.88...‘

.956 8 955
.ma>< 18.—3 on .83.

 

2.3...

.88 .N as:

Chapter 1

Genetic Variation of Iron Bioavailability from Amaranthus Species

Abstract
Grew leafy vegetables have the potential to contribute significant amounts of iron to the diet,
if the bioavailability from these foods is improved. In order to determine the potential for
genetic manipulation of this trait, 46 lines from 12 species of Amwanthus were evaluated for
total and bioavailable iron in greenhouse and field experiments. The bioavailable iron was
estimated using an in vitro assay for dialyzable, low-molecular-weight iron. Significant
differences (p<.01) were detected among lines for total iron and bioavailable iron. Total iron
ranged from 358 to 880 ppm in field-grown plants and 55 to 123 ppm in greenhouse-grown
plants. Bioavailable iron ranged fi'om 41 to 63 ppm in field-grown plants and fi'om 24 to 51
ppm in greenhouse-grown plants. A. tricolor and A. Iividus had the highest total and
bioavailable iron, while AJgpochoncb'iacus had the lowest levels of the species tested. Field-
grown Amaranthus accumulated higher levels of total and bioavailable iron, but a greater
proportion of the total iron was sequestered in insoluble, unavailable forms. Generally,
species with higher total iron had higher levels of bioavailable iron. These analyses indicated
potential for genetic improvement of iron nutritional quality from Amaranthus, especially

within the species A. tricolor.

4O

41

Introduction

Iron (Fe) deficiency anemia is one of the most prevalent nutritional deficiencies around
the world, particularly afi‘ecting poor women and children in developing countries
(Scrimslmw, 1991). Most intervention programs utilize Fe supplementation or fortification
to combat this deficiency. However, constraints to this type of intervention include the need
for access to health care providers or nutritionists, adequate and continuous supply of the
supplement or fortified food, compliance by recipients, simple administration and minimal cost
to the recipient (Monsen, 1988; MacPhail and Bothwell, 1989). An alternative approach is
to attempt to increase the Fe content of the diet. Inclusion of a meat source in the diet would
increase Fe intake, but social, economic and cultural factors may limit access to and/or
acceptability of meat as a component of the diet (Monsen, 1988). Green leafy vegetables
(GLVs) provide a viable option throughout the world as a source of Fe in the diet. Access
to GLVs rarely is limited by economic constraints, due to local or home production and
consumption, or even collection of wild types. Amwanthus species are examples of such
commonly cultivated and foraged GLVs.

Investigators may assume that GLVs with high Fe concentrations are good sources
of Fe for human consumption (Gueimot and Yi, 1994; Chweya, 1985; Deutsch, 1978).
However, little is known about the forms of bioavailable Fe in food plants (W elch and House,
1984). Although many GLVs contain high concentrations of foliar Fe, this Fe is absorbed
poorly in the human gastrointestinal tract. Estimates of Fe bioavailability from GLVs range
fi'om 1 to 70% of the total Fe, depending on the method utilized (Oyejola and Bassir, 1975;
Ifon and Bassir, 1978; Van Campen and Welch, 1980; Chawla et al., 1988; Layrisse et al.,

1969; Reddy and Kulkami, 1986). Previous studies on Fe bioavailability fi'om GLVs often

’ 42
have utilized a ”market basket" approach. Samples are collected fiom local markets and

analyzed for Fe. Although this type of sampling can provide information on consumer
nutritional intake, there is no characterization of the efi‘ect of cultural practices or genetic
variation on Fe bioavailability of the GLVs.

Examination of differences in Fe bioavailability due to genetic variation among
Amanthus species (or other GLVs) has not been reported. Differences in total Fe
accumulation by amaranth have been detected (Makus, 1984; Deutsch, 1978; Elias, 1977),
and the bioavailability of the leaf Fe has been tested (Oyejola and Bassir,]975; Ifon and
Bassir, 1978; Reddy and Kulkami, 1986; Chawla et al., 1988, Latunde-Dada, 1991).
Analysis of four amaranth species (gathered in the wild but utilized as greens) for total and
bioavailable Fe indicated that the proportion of bioavailable Fe was not fixed across species
(Reddy and Kulkami, 1986). Species with the highest total Fe concentration did not have
the highest available Fe concentrations. These results suggest that genetic variability for Fe
bioavailability may exist among species in the genus Amaranthus. This variability among
species might be used to enhance the Fe bioavailability fi'om amaranths, which are known for
their ability to form interspecific hybrids (Kulakow and Jain, 1990).

To explore the genetic variation in bioavailable Fe fi'om amaranth species, field and
greenhouse studies were conducted. An in vitro assay for estimation of bioavailable Fe was
adapted for analysis of multiple samples simultaneously (Kapsokefalou and Miller, 1991).
This method for estimating bioavailable Fe has demonstrated good correlation with in vivo
measurements (Schricker et al., 1981). This assay is also useful to make relative comparisons
among many different plant materials (Latunde-Dada, 1991). The Fe content of four of the

lines grown in both the greenhouse and the field was fiactionated firrther to compare soluble

43

Fe fiactions. Genetic screening of amaranth lines, under known environmental conditions,

may indicate suficient variability to support breeding for improved Fe nutritional quality.

Materials and Methods
Field Embration Thirty-five lines of Amanthus, fi'om 12 species, were selected from the
Plant Introduction Station collection at Ames, Iowa (Table 3). The lines included both
cultivated and wild types and grain and vegetable types to represent the broad geographic and
genotypic range of the genus. Local cultivar names have been provided, when known, for
popular types from certain countries.

The field evaluation was conducted at the Michigan State University Horticulture
Teaching and Research Center. A site with a sandy loam soil (Mariette fine sandy loam,
mixed mesic Glassoboric Hapludalt) was fertilized with 60 lb nitrogen per acre as urea. Lines
were direct-seeded in a randomized complete block design, with four blocks, on July 12,
1990. A tar-foot row was seeded per line in each block and after one week, thinned to a 6-
inch in-row spacing. Irrigation supplemented rainfall to provide a minimum of one inch of
water per week. No pesticides were applied, and all weeding was done by hand.

Plots were harvested on day 35 afier seeding, at 7:30 am, to minimize field heat.
No flowering was detected in any of the lines on this date. Five to seven whole plants were
selected randomly fiom each line, cut at the soil line, bagged and placed on ice in a cooler for

transport. Samples then were stored at 4° C for up to five hours prior to processing.

Greenhouse evaluation. Twenty-four lines fiom four species of amaranth (A. Iivr’dus, A.

tricolor, A. chlbius, and A. lopochondriacus) were grown in the greenhouse to determine if

44
species and line differences in Fe nutritional quality could be detected in plants grown under
controlled environmental conditions (Table 3). Lines were seeded into plastic cell flats, and
afler 14 days, uniform seedlings were selected and transplanted into 20-cm diameter plastic
pots, with two seedlings per pot. Three pots per line were placed under supplemental
lighting (high pressure sodium lamps providing 420 urnol-m'z-sec“) in a randomized complete
block design, with three blocks. Diurnal temperatures were maintained at 30° C day and
18° C night (1 5° C). Plants were fertilized with 200 ppm of nitrogen with fertilizer
formulation of 20-20-20 N-P205-K20, three times per week. Irrigation water was acidified
to pH 6.0.

Pots were harvested on day 28 after transplanting. The plants from each block were
cut ofi‘ at the soil line and pooled, and the samples were stored at 4° C for up to two hours

prior to processing.

LequroceMrg. The upper whorl of leaves, up to the most fully expanded leaf, was removed
from each stem and bulked for each line. For the lines of A. spinosus, all leaves were
harvested due to the small leaf size and prostrate growth habit. To remove surface soil
contamination, leaves were washed as follows: three rinses with distilled water, lS-second
rinse in 0.1N I-INO3 and three more distilled water rinses. Excess water was removed by
spinning in a salad spinner device, and samples were held at -20° C prior to fieezedrying.

The fleece-dried samples were ground in a stainless steel erey Mill, with a 40-mesh screen.

Fe qrawmficatiom. Total and bioavailable Fe were determined in duplicate, for each sample.

Total Fe was quantified using atomic absorption spectrophotometry (AAS) (Video 12,

45

Instrumentation Laboratory AA Spectrophotometer, Andover, MA). Samples were wet-
ashed with perchloric acid and hydrogen peroxide, diluted to 10 ml, and Fe concentration was
determined by comparison of absorption at 248.3 nm to a stande curve (Alder and Wilcox,
1985).

Bioavailable Fe was estimated using the in vitro assay described by Kapsokefalou and
Miller (1991). This assay estimates bioavailable Fe fi'om the amount of low-molecular-
weight Fe (dialyzable) which remains in solution after a simulation of gastrointestinal
digestion. The assay was modified to maximize the number of samples per run (28) and
minimize the amount of dried leaf material required (0.5 g). Included in each run of the assay
was two replicates of a spinach (Spinacea oIeracea) sample developed in the lab, as a
standard to compare runs. In a 30-ml sample bottle, 0.5 g of ground leaf sample was mixed
with 10 ml distilled deionized water and the solution was adjusted to pH 2 with 0.5N HCl.
The final weight of this solution was increased to 14 g. This prepared sample was frozen
overnight and thawed for the enzymatic digestion the following day. One ml of pepsin
solution (4 g pepsin in 100 ml 0.1 N HCl, Sigma P-7000, Sigma Chemical Company, St.
Louis, Missouri) was added to each sample bottle and these bottles were placed in a shaking
water bath, at 37° C, for 2 h. After this initial peptic digestion, dialysis tubing (Spectrapore
I, 6000-8000 MW cutofl‘, 32 mm flat width) containing 10 ml of 0. 15 M PIPES buffer, pH
6. 87, was added to each of the bottles. Solutions were allowed to pH-equilibrate for l h
prior to addition of 5 ml pancreatin and bile enzyme solution. This solution contained 0.2
g pancreatin (Sigma P-1750) and 1.2 g bile salts (Sigma B-8631) in 100 ml of 0.1 M
NaHCO;,. After the two-hour pancreatin digestion, dialysis tubes were removed from the

bottles, rinsed with distilled water, and the contents were emptied into plastic vials.

46

Colorrnetric determination of reduced Fe in the dialyzed samples was conducted using a
ferrozine chromogen (Sigma P9762) as described by Kapsokefalou and Miller (1991), due
to the greater sensitivity and lower volume requirement as compared to AAS.

The percent dialyzable Fe was calculated fiom the values obtained in the above
amlyses, by dividing the amount of Fe measured as bioavailable by the total Fe concentration
and multipling the result by 100. Analysis of variance was performed for total, dialyzable,
and percent dialyzable Fe for each experiment.

Four lines (liv 1, hyp 2, tri 4 and tri 7)(Table 2), representing high and low total Fe
concentrations from each experiment, were analyzed for soluble Fe. After the removal of the
dialysis tubing in the bioavailable Fe assay, the remaining digest was centrifirged (1500 rpm,
15 min), and Fe concentration of the supernatant was determined using AAS. This Fe
fraction included those soluble Fe compounds having molecular weight greater than 8000, the

cutofl‘ for the dialysis tubing.

Results
Field evaluation. Significant differences (p<.01) were detected among the lines (Table 4)
and species (Table 5) for total, dialyzable, and percent dialyzable Fe. Total Fe content of the
leaves ranged from 358 to 880 ppm. The 10 lines accumulating the highest concentration
of total Fe were fi'om A. Iividus, A. tricolor, and A. spinosus species (Table 4). Lines liv-l
and [iv-3 firm A. lividus, accumulated the highest concentration of total Fe. The two lines
with lowest total Fe, A. lopoehondriacw lines hyp-1 and hyp-2, contained approximately half

the total Fe concentration of the two highest lines.

47

In vitro digestion provided estimates of bioavailable Fe (dialyzable) from 41 to 63
ppm, significant differences (p<.01) being detected among lines (Table 4) and species (Table
5). No significant correlation was detected between total and dialyzable Fe in the field-
grown material (r = 0.21). The speciesA. dubius accumulated a medium level of total Fe but
ranked second highest in dialyzable Fe. Mthin-specics differences in dialyzable Fe were
significant for A. tricolor and A. lividus.

The percentages of dialyuble Fe ranged fiom 6 to 12 (Table 4) and was negatively
correlated to total Fe concentration (r = -0.80). Lines fiom A. tricolor and A. lividus had the

lowest percents dialyzable Fe, whereas A. lopochondriacus had the highest (Table 4).

Greenhouse evaluation. Significant differences (p<.01) also were detected for both total and
dialyzable Fe among lines (Table 6) and species (Table 7) grown in the greenhouse. Total
Fe ranged fi'om 55 to 123 ppm, and dialyzable Fe ranged fi'om 24 to 51 ppm. Total Fe was
correlated with dialyzable Fe (r = 0.84). The species A. lividus accumulated the highest
levels of total and dialyzable Fe, while A. lopochondriacus accumulated the lowest levels
(Table 7), similar to results fi'om the field evaluation (Table 5). A. tricolor was the only
species with significant differences in total Fe among individual lines (LSD.m for total Fe=l4).
The percent dialyzable Fe did not vary significantly among greenhouse-grown plants. An
average of 43% of the total Fe was measured as bioavailable.
When comparing the Fe levels for the 13 lines grown in both the greenhouse and the
field (Table 3), weak correlations were observed for both the total Fe (r = 0.64) and

dialyzable Fe (r = 0.31).

48
SolubleFe comparisons. The four lines analyzed for soluble Fe (liv 1, hyp 2, tri 4 and tri 7)

represented the high and low total Fe concentrations observed among the species tested.
Soluble Fe firm the four lines analyzed ranged fi'orn 115 to 176 ppm for the field-grown lines
and fi'om 60 to 92 ppm for the same lines grown in the greenhouse, or about one half the
levels observed in the field-grown material. This fiaction represented 20-29% of the total
Fe in the field-grown plants and 78-84% of the total Fe in greenhouse-grown material. A
slightly ln'gher correlation was observed between the soluble and dialyzable Fe levels (r-=. 78)

than the total and dialyzable Fe levels (r=.72) in both the growing environments.

Discussion

This study represented the first comprehensive attempt to determine if genetic
variation exists for Fe bioavailability from a GLV. The significant difl‘erences observed for
total and bioavailable (dialyzable) Fe among the species of Amaranthus indicated potential
genetic variability that may be utilized to improve the Fe nutritional quality of this GLV.
This genus is known for its ability to form interspecific hybrids (Kulakow and Jain, 1990),
allowing for such variability to be utilized to enhance Fe bioavailability. A. tricolor, A.
lividus, and A. spinosus, which accumulated high levels of total Fe, are used predominately
as GLVs, whereas A. hypochondriacus, which accumulated the lowest levels of total Fe, is
predominately cultivated for grain (Martin and Ruberte, 1977; Huang, 1979). Although A.
spinosus is consumed as a GLV in some regions of Asia (Martin and Ruberte, 1977), this
species also has been referred to as a noxious weed (Kegel and Rubin, 1985) and has a
prostrate growth habit and axillary spines (Huang, 1979). A. lividus (also referred to as A.

blitum) is related to A. tricolor, and both species have historical use as cultivated GLVs

49

(Huang, 1979). Within the species A. tricolor, significant differences were observed for
total and dialyzable Fe among lines tested both production environments. A. tricolor has
been reported previously to accumulate higher levels of total Fe than other Ammanthus
species (Makus 1984; Deutsch, 1979; Elias, 1977; Grubben, 1976). The high Fe
concentrations observed among A. lividus and A. tricolor and their popularity as cultivated
vegetable types suggests these two species as good candidates for breeding efi‘orts to improve
the Fe nutritional quality of this GLV.

Although significant difi‘erences in dialyzable Fe were detected in both environments,
the narrow range of values observed in both environments suggested a fairly constant fraction
of bioavailable Fe for this leaf material (Tables 4 and 6). The dialyzable Fe of the tested
amaranth species was between 41 and 63 ppm in the field-grown plants and 23 and 51 ppm
inthe greenhouse-grown plants. Using aconversion ofdryweight equal to 10% fresh weight
for field-produced plants (average value from data not reported), the low and high dialyzable
lines produced in the field contributed 0.41 and 0.63 mg bioavailable Fe per 100 g fresh
weight. Chawla et al. (1988) detected a similar value for Fe bioavailability from amaranth,
0.45 mg-100 g" fiesh weight. Although the dialyzable Fe assay has been reported to be
highly correlated with in vivo bioavailability measurements (Schricker et al., 1981), it is useful
only to make relative comparisons among tested materials. It does not indicate absolute Fe
bioavailability. Determination of actual bioavailability would require animal or human
feeding studies and an examination of meal interactions (Forbes et al., 1989). Ifthe values
for bioavailable Fe estimated in this study were actual levels, a 100 3 portion of amaranth
couldprovide40to60% ofthedailyFerequirernentformenandchildren (RDA ofl myday)

and 27 to 41% for women (RDA of 1.5 mg/day) (National Research Council, 1989).

50

Another common method of presenting the Fe contribution of a food source is by
calculating the percent bioavailable Fe (Thompson, 1988; Reddy and Malewar, 1992;
Latunde-Dada, 1991; Hazell and Johnson, 1987). Foods which contain a high proportion
of bioavailable Fe or which contain other compounds that enhance or inhibit Fe bioavailability
may be identified by examining the percent bioavailable Fe. The field-grown amaranth lines
accumulated between six and eleven times the amount of total Fe when compared to
greenhouse-grown plants, but the percent bioavailable (dialyzable) Fe was low (642%, Table
4) compared to greenhouse-grown plants (38-47%, Table 6). However, these percent
bioavailability data can mask the actual quantity of Fe contributed to the diet. A food with
a high Fe concentration but low percent bioavailable Fe may be a significant source of Fe to
the Fe—deficient individual (Thompson, 1988). The absolute amount of Fe measured as
bioavailable from field-grown plants was greater than the amount measured for the same lines
when grown in the greenhouse (Table 4 and 6). The significant negative correlation
detected between the percent bioavailable Fe and total Fe concentration in the field study does
not imply that increasing total Fe concentration contributes to decreasing percent bioavailable
Fe. The low percentages of bioavailable Fe may indicate the presence of interactions among
Fe and other compounds within the field-grown plant material.

To examine firrther the difi‘ercnces in the total Fe fractions of plants grown in the
greenhouse and field, four lines grown in both locations were analyzed for soluble Fe. The
actual levels for the soluble Fe from the field-grown plants were about twice the values
observed in the greenhouse-grown plants. However, in field-grown material, an average of
only 23% of the total Fe could be solubilized during the in vitro digestion, compared to 82%

of the total Fe that was soluble in greenhouse-grown plants. The insoluble Fe fiaction of the

51

in vitro digest included insoluble Fe precipitates and those Fe compounds resistant to the
digestion conditions of the assay (Miller and Schricker, 1982). These data provided further
evidence that a large portion of Fe accumulated by field-grown plants was resistant to
digestion and would not be included in bioavailable Fe fi'actions. An examination of the
fiactions of Fe in raw spinach revealed 93% of the plant Fe was bound chemically in the
insoluble fi‘action (Lee and Clydesdale, 1981). Although this type of soluble Fe profile is not
equivalent to available Fe, it may indicate the potential for improving Fe bioavailability if
soluble Fe levels can be increased (Thompson, 1988).

Two potentially interrelated hypotheses exist to explain these observations. The
dialymble Fe concentrations were similar for both environments, but total Fe concentration
varied. The dialyzable Fe fraction may consist of specific Fe compounds within plants that
are physiologically active, may not vary greatly in actual concentrations in difi‘erent
production environments (metabolically essential enzymes), and can be solubilized readily
under the conditions of the in vitro assay. Such 'active' Fe compounds have been extracted
fi'om leaf material using weak acids or Fe chelators (DeKock et al., 1979; Abadia et al., 1984;
Mehotra and Chandra, 1990). The concentration of these 'active' Fe compounds has
exhibited higher correlation with plant Fe status (or deficiency) than total Fe concentration.

The relationship between the 'active' Fe fiaction of leaves and the bioavailable Fe fiaction has
not been investigated.

Alternatively, field-grown plants may accumulate higher levels of other plant
compounds (ascorbic acid, tannins, phosphorous) which may change the Fe bioavailability
(Gillooly, et al., 1983; Reddy and Malewar, 1992). In one study, bioavailable Fe in spinach

(grown in the greenhouse) was correlated positively to ascorbate and oxalate, and negatively

52

with phosphorous (Reddy and Malewar, 1992), but results of a second study indicated
correlation only between Fe bioavailability and phosphorous content (Reddy et al., 1993).
The levels of these other compounds in relation to bioavailable Fe have not been examined
in field-grown material. More information is needed on the relative contributions of the
forms of bioavailable Fe and the presence of certain secondary plant compounds to Fe
bioavailability in GLVs. Changes in the forms of Fe or levels of secondary compounds may
indicate alternative approaches for genetic improvement of Fe nutritional quality.

In this study, the greenhouse-grown lines contained an average of 43% bioavailable
Fe, compared to 6-12% for field-grown lines. A high percent bioavailable Fe has been
observed in other GLVs grown in greenhouse environments Men, et al., 1975; Van Campen
and Welch, 1980). Researchers have concluded that the Fe from GLVs is highly available,
based on calculations of percent bioavailable Fe fiom a greenhouse evaluation. The results
fi'om this study indicated that conclusions regarding Fe bioavailability based solely on
greenhouse evaluations may be incorrect. General differences among the amaranth species
tested were maintained between the two production environments, but the efi‘ects of
secondary compounds or the change in the Fe form of field-grown material had a great efi‘ect
on the percent of bioavailable Fe. Although differences were detected among species in the
greenhouse evaluation, field analyses must be utilized for final determinations of Fe nutritional
quality.

Attempts to increase the total Fe concentrations in plants grown under controlled
environments generally have been unsuccessfirl (Abadia, 1992). Some investigators have
had some success increasing the amount of Fe in spinach leaves grown controlled

environments by increasing Fe concentrations in the growing media (either nutrient solutions

53
or soil) (El-Sherif et al., 1984; Reddy and Malewar, 1992; Reddy et al., 1993) while others

reported no increase in Fe levels in plant tissue grown in high-Fe media (\Men et al., 1975,
Van Campen and Welch, 1980). The differences in the soil pH, the light intensity and light
quality of controlled environments may be more important than soil Fe concentrations in
explaining the difi‘erences observed in total Fe accumulation from plants grown in field
environments (Abadia, 1992). In studies focusing on Fe biochemical activity in plants, high
correlation has been found between results fiom greenhouse and field environments (Abadia,
1992). Identification of other controlled environment production methods which result in
plants with levels of Fe similar to field-grown plants may increase the correlation between Fe
bioavailability levels fi'om both environments.

The modification of the in vitro method of estimating Fe bioavailability developed by
Kapsokefalou and Miller (1991) used in this study is a low-cost and simple method for
evaluation of plant material. The small amount of plant material needed and the higher
number of samples analyzed per run favored its applicability to the evaluation of plant
material. This method also may be adaptable for the study of nutritional contribution of other
plant minerals (Reykdal and Lee, 1991) or materials (Latunde-Dada, 1991). The use oftotal
Fe analysis for broad screening of total Fe content, combined with in vitro dialyzable and
soluble Fe analyses for more specific screening of select high and low Fe lines, is a very
eficient method for evaluation of genetic variability of Fe nutritional quality and identification

of interesting species and lines for further study.

54

Table 3. Amaranthus accessions evaluated for total and bioavailable iron

under greenlmusgnd/or field conditions.

 

 

 

 

Lme‘ Loc.2 P.I. Number Species and cultivar Origin

hi 1 f,g Ames 5379 A. tricolor 'Tiger leaf Taiwan

hi 2 f,g PI 419057 A tricolor 'Red leaf Taiwan

hi 3 f Ames 5147 A tricolor 'Red leaf Taiwan

hi 4 f,g Ames 2154 A tricolor 'Red leaf Taiwan

hi 5 f Ames 2205 A tricolor 'White leaf Hong Kong
hi 6 f Ames 2209 A. tricolor 'White leaf Hong Kong
tri 7 f,g PI 173837 A tricolor 'Chulai' India

hi 8 f Ames 1980 A. tricolor Zaire

hi 9 f Ames 511 1 A. tricolor 'Aupamalip' Papua New Guinea
tri 10 f Ames 5113 A tricolor 'Duradera' Taiwan

tri 11 g Ames 2208 A tricolor 'Tiger leaf Hong Kong
iii 12 g Ames 5166 A tricolor India

dub l f,g Ames 2098 A dubius India

dub 2 f,g Ames 1967 A dubius India

dub 3 f,g Ames 5114 A dubius 'Stubby‘ Taiwan

dub 4 f,g Ames 5674 A dubius 'Imbondwe' Zambia

dub 5 g Ames 1997 A. dubius 'Noudom' Ghana

dub 6 g Ames 5105 A dubius Seychelles
hyb l f P1482049 A hybridus Zimbabwe
hyb 2 f P1494768 A hybridus Zambia

hyb 3 f PI 210995 A. hybridus Afghanistan
can 1 f PI 166045 A. caudatus 'Chua' India

cau 2 f PI 490609 A. caudatus Equador

vir l f P1540445 A. virdis Indonesia
cm 1 f Ames 5598 A cruentus Benin

cm 2 f PI 433228 A cruentus Guatemala
cm 3 f Ames 8269 A cruentus United States
gm 4 f PI 511715 A cruentus Mal;

 

 

 

‘ Initial three letters are an abbreviation of the species.
2 Location of evaluation; f= field, g= greenhouse.

 

Table 3. gears-9t.)

 

55

 

 

 

 

 

 

Linel Loc.2 P.I. Number S ies andcultivar Ori '
spn l f Ames 2043 A. spinosus Indonesia
spn 2 f Pl 482058 A. spinosus Zimbabwe
liv 1 f,g P1288277 A. lividus India
liv 2 f,g Ames 2206 A lividus Hong Kong
liv 3 f,g Ames 5146 A lividus India
liv 4 g Ames 2035 A. lividus India
liv 5 g Ames 5103 A. lividus Hong Kong
liv 6 g Ames 5387 A Iividus India
hyp l f,g PI 477917 A. hwochondriacus Mexico
hyp 2 f,g Ames 2171 A hwochondriacus Nepal
hyp 3 g Ames 5474 A. hwochondriacus Mexico
hyp 4 g Ames 5691 A. hypochondriacus 'Julma' Nepal
hyp 5 g Ames 2156 A. hypochondriacus Nepal
hyp 6 g PI 477915 A. hfirochondriacus India
nrd l f Ames 10827 A. rudis United States
pal 1 f Ames 5306 A palmerii Senegal
spe 1 f PI 511752 A. species Peru

2 f PI 337611 A species Uganda

 

 

‘ Initial three letters are an abbreviation of the species.
2 Location of evaluation; f= field, g= greenhouse.

56

Table 4. Total, dialyzable (ug-g‘ dry wt.) and percent dialyzable
Fe in leaves of 35 lines of Amaranthus grown in the field

 

 

 

 

Line Total Fe Dialyzable Fe % Dialyzable Fe
hi 1 764 60 8
hi 2 579 43 9
hi 3 641 52 8
hi 4 575 49 9
hi 5 702 52 8
hi 6 748 57 8
hi 7 805 62 8
hi 8 723 44 6
hi 9 754 57 8
hi 10 756 45 6
dub l 527 57 1 1
dub 2 572 58 1 1
dub 3 560 54 10
dub 4 525 55 l l
hyb l 596 53 9
hyb 2 661 51 8
hyb 3 450 52 12
can 1 615 55 10
can 2 492 48 10
vir 1 581 43 8
cm 1 595 51 9
cm 2 495 48 1 1
cm 3 484 46 10
cm 4 552 49 10
spn 1 785 60 8
spn 2 768 51 7
liv 1 848 57 8
liv 2 713 50 7
liv 3 880 63 8
hyp 1 358 41 12
hyp 2 419 45 11
rud l 51 1 52 12
pal l 453 49 12
we 1 459 47 10
spe 2 550 47 9
LSD”, 189 11
fl

 

 

 

57

Table 5. Total, dialyzable (ug'g"dry wt.) and percent
dialyzable Fe in leaves of 12 Amanthus species grown in

the field.
Species Total Dialynble %Dialyzable

 

 

 

Fe Fe Fe
tri 705 52 8
dub 546 56 10
hyb 569 52 10
can 554 51 10
vir 581 43 8
cm 53 l 49 10
spn 776 55
liv 814 57
hyp 388 43 12
rud 51 1 54 12
pal 453 49 12
spe 504 47 9
LSD.01 174 7 3

 

 

 

 

58

Table 6. Total, dialyzable (rig-g'l dry wt.) and percent
dialyzable Fe in leaves of 24 lines of Arnaranthus gown

 

 

 

 

 

 

 

in greenhouse. j _
Line Total Dialyzable % Dialyzable
Fe Fe Fe
hi 1 82 38 47
hi 2 89 41 46
tri 4 101 39 38
tri 7 75 38 49
tri 11 96 41 43
hi 12 79 38 48
dub 1 75 31 42
dub 2 82 35 43
dub 3 93 35 39
dub 4 84 34 41
dub 5 89 37 43
dub 6 81 33 41
liv 1 109 45 42
liv 2 113 51 45
liv 3 107 48 45
liv 4 118 48 41
liv 5 123 51 42
liv 6 110 45 41
hyp 1 60 27 45
hyp 2 73 33 45
hyp 3 61 28 46
hyp 4 61 28 46
hyp 5 64 28 45
hyp 6 55 24 44
LSD g! 28 10 as

 

59

Table 7. Total, dialyzable (pg-g" dry wt.) and percent
dialyzable Fe in leaves of four Arnaranthus species
grown in tlfigreenhouse.

Species Total Dialyzable %Dialyzable

 

 

Fe Fe Fe
hi 87 39 45
dub 84 34 41
liv 1 13 48 43
hyp 62 28 45
LSD”, 9 2 as

 

 

 

 

60

Literature Cited
Abadia J. 1992. Leafresponses to iron deficiency: A review. J. Plant Nutr. 15: 1699-1714.

Abadia, J. E. Monge, L. Montafies and L. Heras. 1984. Extraction of iron fiorn plant leaves
by iron (II) chelators. J. Plant Nutr. 7 2777-784.

Alder, P.R. and GE. Wilcox. 1985. Rapid perchloric acid digest methods for analysis of
major elements in plant tissue. Comm Soil Sci. Plant Anal. 16(1 1):1 153-1 163.

Chandra, M. and P. Gupta. 1990. Reduction of iron by leaf extracts and its sigrificance for
the assay of Fe2+ iron in plants. Plant Phys. 93:1017—1020.

Chawla, S., A. Saxena and S. Seshadri. 1988. In vitro availability of iron in various green
leafy vegetables. J. Sci. Food Ag. 46: 125-127.

Chweya, J .A. 1985. Identification and nutritional importance of indigenous leaf vegetables
in Kenya. Acta Hort. 153:99-104.

DeKock, PC, A Hall and RH.E. Inkson. 1979. Active iron in plant leaves. Ann. Bot.
43:737-740

Deutsch, J .A 1978. Genetic variation of yield and nuhitional value of several Mus
species used as leafy vegetables. Diss. Abstr. Intl. B. 38:3969-B.

Elias, J. 1977. Food composition tables for comparative nuhient composition of amaranth
greens and seeds. First Amaranth Seminar Proc. (Emmaus, PA: Rodale Press Inc.)
pp. 16-35.

El-Sherifi AR, AZ. Osman, MK. Sadik, and SM. Shata. 1984. Determination of ferrous
and ferric iron ratio in spinach plants and their relation to iron application. J. Plant
Nutr. 7 2767-776.

Forbes, A., C.E. Adams, M.J. Arnaud, C.O. Chichester, J.D. Cook, BM. Harrison, RF.
Hurrell, S.G. Kalm, ER Morris, J.T. Tanner, and P. Whittaker. 1989. Comparison
of in vitro, animal and clinical determinations of iron bioavailability: International
Nuhitional Anemia Consultative Group Task Force report on iron availability. Amer.
J. Clin. Nutr. 49:225-238.

Gillooly, M., T.H. Bothwell, J.D. Torrance, AP. MacPhail, D.P. Derrnan, WR Bezwada,
W. Mills, and R.W. Charlton. 1983. The effects of organic acids, phytates and
polyphenols on the absorption of iron fi'om vegetables. Brit. J. Nutr. 49:331-342.

Grubben, G.J.H. 1976. Cultivation of amaranth as a tropical leaf vegetable. Koninklijk
Institute voor de Tropen. Comm. #67 234-41.

61

Guiernot, ML. and Y. Yi. 1994. Iron: nutrition, noxious and not readily available. Plant
Phys. 104:815-820.

Hazell, T. and LT. Johnson. 1987. In vitro estimation of iron availability fi'om a range of
plant foods; influence of phytate, ascorbate and citrate. Brit. J. of Nutr. 57:223-233.

Huang, RC. 1979. Study of the taxonomy of edible amaranth: An investigation of amaranth
both of botanical and horticultural characteristics. Proc. of Second Amaranth
Conference. (Emmaus, PA: Rodale Press, Inc.) pp. 142-150.

Ifon, ET. and O. Bassir. 1978. The efliciency of utilizing the iron in leafy geen vegetables
for hemoglobin synthesis in anemic rats. Nutr. Rep. Intl. 18(4): 481-486.

Kapsokefalou, M. and DD. Miller. 1991. Efl‘ects of meat and selected food components on
the valence of nonheme iron during in-vitro digestion. J. Food Sci. 56(2):352-355,
358.

Kigel, J. and B. Rubin. 1985. ”Amaranthus.” In CRC Handka QfFleering, Vol. 1, ed.
AH. Halevy. (Boca Raton: CRC Press, Inc). pp. 427-433.

Kulakow, PA and SK Jain. 1990. Grain amaranths: Crop species, evolution and genetics.
Proc. of the Fourth National Amaranth Symposium: Perspectives on Production,
Processing and Marketing. (Minneapolis: Minnesota Extension Service). pp. 105-
1 14.

Latunde-Dada, G. O. 1991. Some physical properties of ten soybean varieties and effects
of processing on iron levels and availability. Food Chem. 42:89-98.

Layrisse, M., J.D. Cook, C. Martinez, M. Roche, I.N. Kuhn, RB. Walker and CA Finch.
1969. Food Iron Absorption: A Comparison of Vegetable and Animal Foods. Blood.
33(3):430-443.

Lee, K and FM Clydesdale. 1981. Efi'ect of thermal processing on endogenous and added
iron in canned spinach. J. Food Sci. 46:1064-1068, 1073.

MacPhail, AP. and TH. Bothwell. 1989. Fortification of the diet as a strategy for
preventing iron deficiency. Acta. Paediatr. Scand. Suppl. 361 :1 14-124.

Makus, DJ. 1984. Evaluation of amaranth as a potential geens crop in the mid south.
HortSci. 19(6):881-3.

Martin, F .W. and R. Ruberte. 1977. Selected amaranth cultivars for geen leaves. Proc. for
the First Amaranth Seminar. (Emmaus, PA: Rodale Press Inc.). pp. 105-111.

62

Miller, DD. and BK Schricker. 1982. ”In Vitro Estimation of Food Iron Bioavailability.”

In W ed. C. Kies. (New York: American Chem
Society). pp. 11-26.

Monsen, ER 1988. Iron nutrition and absorption: dietary factors which impact iron
bioavailability. J. Amer. Diet. Assn. 88(7): 786-790.

National Research Council. 1989. Recommmdod Dietm Allowances. Food and nutrition
board subcommittee on the 10th edition of the RDA's. (Washington DC: National
Academy Press). pp 195-205.

Oyejola, AD. and O. Bassir. 197 5. Physiological availability of the iron content of some
Nigerian leafy vegetables. Plant Foods for Man. 1:177-183.

Reddy, NS. and KS. Kulkami. 1986. Availability of iron from some uncommon edible
geen leafy vegetables determined by in vitro method. Nutr. Reps. Intl. 34(5):859-
861.

Reddy, 8. and V.G. Malewar. 1992. Bioavailability of iron from spinach cultivated in soil
fortified with g'aded levels of iron. Plant Foods Human Nutr. 42:313-318.

Reykdal, O. and K. Lee. 1991. Soluble, dialyuble and ionic calcium in raw and processed
skim milk, whole milk and spinach. J. Food Sci. 56(3):864-866.

Schricker, BR, D.D. Miller, R. Rasmussen and D. Van Campen. 1981. A comparison of
in vivo and in vitro methods for determining availability of iron fi'om meals. Amer.
J. Clin. Nutr. 34:2257-2263.

Scrimshaw,N.S. 1991. Iron deficiency. Sci. Amer. Oct:46-52.

Thompson, BB. 1988. Ch. 5. ”Iron.” In Tm Minegls in Foods, ed. K.T. Smith. (New
York: Marcel Dekker, Inc) pp. 157-208.

Welch, RM. and W.H. House. 1984. Ch. 3. "Factors Afl‘ecting the Bioavailability of
Mineral Nutrients in Plant Foods.” eds. RM. Welch and W.H. Gabelrnan. Mos

Sooroes of Notfients for Humans. Proceedings of a symposium. Nov 28,-Dec 3,
1982. Agonomy Society of America Special Publication #48.

Wien, E.M., D.R Van Campen, J.M. Rivers. 1975. Factors afi‘ecting the concentration and
bioavailability of iron in turnip greens to rats. J. Nutr. 1052459-466.

Van Campen, DR and RM. Welch. 1980. Availability to rats of iron fiom spinach: efl‘ects
ofoxalic acid. J. Nutr. 110(2): 1618-1621.

Chapter 2
Changes in Iron Concentration and Bioavailability
with Amaranth Leaf Development

Abstract
If attempts are made to improve the iron bioavailability from Arnaranthrts, it is important to
standardize plant sampling by understanding the changes in iron content and bioavailability
fiom the leaves at different stages of plant and leaf development. An initial screening of 18
lines of amaranth harvested over three days indicated that total iron concentration of leaves
may fluctuate over time, but the amount of bioavailable iron within the harvested leaves
remained relatively constant over the duration of the experiment. A subsequent geenhouse
study was conducted to examine the changes in total, soluble and bioavailable (dialyzable)
iron over the period of leaf development. The total iron concentration was highest in the
indexed leaf just prior to firll leaf expansion The soluble iron concentration, after the in vitro
digestion, was maximum during the early, rapid expansion phase of leaf development, but it
decreased sigrificantly as the leaf approached full expansion. The bioavailable iron remained
constant across all harvest dates. Despite the fluctuations in total iron concentrations
observed in both studies, the amount of bioavailable iron remained relatively constant.
However, the amount of soluble and potentially available iron changed, depending on leaf age
and maturity. Characterization of this soluble iron fi'action may ofl‘er indications of forms of

iron which may be bioavailable and which may be enhanced through plant breeding.

63

64

Introduction

Researchers examining iron (Fe) content or nutritional quality of Arnaranthus species
and other GLVs at difi‘erent harvest dates oflen have suggested optimum harvest times for
maximizingthesequalities. Whentheharvestdatesweredelayed from day49 to day 63 after
planting the Fe concenhation doubled in leaves of Arnaranthus hybridus L. (Stafford et al.,
1976) and increased 20% in four lines of celosia(Celosia argentea L.)(Omueti, 1982). In
three separate experiments evaluating Fe accumulation in several amaranth lines harvested
from day 21 to day 46 after planting, all lines accumulated the highest levels of total Fe
around day 32 (Deutsch, 1978). The total Fe content of sweetpotato stem tip cuttings (used
as a GLV) decreased over three harvest dates (spanning 30 days), but the amount of Fe
available for absorption did not vary with harvest dates (Pace and Bonsi, 1987). However,
conclusions about nutritional quality remain problermtic, due to the diverse environments and
methods of production and difl‘erences in leaves sampled for analysis in these studies. The
efi‘ects of environmental conditions on Fe content and bioavailability were not explored in
these studies.

Although these studies described changes in Fe content over harvest dates, few
provided detailed information on leaves sampled, and none have compared the amount of
bioavailable Fe from the same leaves over time. Within individual sweetpotato vines, 'older’
leaves (10 to 20 cm basipetal to the apical meristem) accumulated more total Fe than
‘younger‘ leaves (within 10 cm of the meristem) (Pace et al., 1985). An analysis of nuhitional
and anti-nuhitional qualities of individual leaves of an A. tricolor line indicated leaves seven
and eight fiom the base of the plant accumulated the highest protein and carotenoid levels

(Prakash and Pal, 1991). Oxalate concentations increased as the location of leaves

6S

progessed toward the apical meristem, whereas nitrate levels were similar at all positions.
Fe content was not examined. Turnip geens accumulated geater Fe concentrations in
younger leaves than in the four outermost, older leaves (“Wren et al., 1975). Analysis of a
single leaf of Cucumis melo L. cv. Galia for total and soluble (F e“) concentrations and
activities of Fe-containing enzymes revealed changes in these parameters with phenological
stage and aging of the leaf (Valenzuela et al., 1992). In addition, changes from a vegetative
to reproductive state may affect the forms of Fe in plants (Seckback, 1982). However, the
efl‘ect of leaf development and maturation on Fe bioavailability has not been studied.

In the previous chapter, several lines and species of amaranth were examined for
differences in Fe bioavailability and content due to genetic variation. If attempts are made
to improve the Fe bioavailability from this GLV through plant breeding, it is important to
understand the difi‘erences in this trait within a plant in order to standardize plant sampling.
Two experiments were conducted to explore changes in Fe content and bioavailability from
Amarwrthus species with changes in leaf maturity and age. In an initial field experiment, 18
lines of Amaranthus were screened for Fe content and bioavailability on three harvest dates.
A subsequent geenhouse study focused on the changes in total, soluble and bioavailable Fe

content of one indexed leaf, over five harvest dates.

Materials and Methods
Field Study. Eighteen lines of Amaranthus (Table 8) were selected fiom the Plant
Introduction Station collection at Ames, Iowa. This selection included eight species,
representing both cultivated grain and vegetable types and wild types. These lines were
direct-seeded on July 10, 1990 in a randomized complete block design with four replications

66

at the Horticulture Teaching and Research Center, Michigan State University. A field site
with a sandy loam soil (Marlette fine sandy loam) was fertilized with 60 lb nitrogen per acre
as urea Plants were thinned to a six-inch in-row spacing one week after planting. Irrigation
supplemented rainfall to provide a minimum of one inch of water per week. No pesticides
were applied, and all weeding was done by hand.

Five to eight plants were selected at random and harvested fi'om each line on dates
28, 35 and 42 after seeding. These dates will be referred to as harvest 1, 2 and 3 in the text.
On dates 1 and 2, no flowering was observed in any of the lines. By date 3, most lines
showed some flower inititation. Harvesting occurred at 7 :30 AM, to minimize field heat.
The upper whorls of new expanding leaves up to the most firlly expanded leaf were removed
fi'om each plant, bagged and placed in a cooler for transport. Samples were stored at 4° C,

for up to five hours prior to processing.

Greenhouse Study. In order to examine the changes in Fe content and bioavailability of one
leaf over time, a geenhouse study was conducted using A. species 'Blood Red' (Ames 5366)
fiom Bangladesh and A. cruentus (Ames 1968) fiom Ghana. These species were selected
based on their popularity as GLVs and their single stem, non—branching gowth habits, which
allowed for specific leaf sampling. Seeds were germinated in flats, and after the appearance
of the first true leaves, uniform seedlings were transplanted into 15-cm diameter clay pots,
using a 1:1 ratio of a sandy loam soil and a commercially available peat-based soil medium.
Plants were fertilized at every watering to soil saturatation (approximately 500 ml), using 150
ppm nitrogen in a formulation of 20-10-20 N-P205-KZO. Five days before the first harvest,
plants from each line were sorted into three blocks by the size of the smallest leaf and the

number of leaves, to increase accuracy of sampling for an index leaf.

67

Plants were harvested on dates 14, 20, 24, 29, and 35 alter h'ansplanting. In the text,
these dateswill be referred to as harvest 1 through 5. The index leafharvested on each day
was leaf4, counting up fiorn the soil (not including the cotyledons and the first true leaves).

At each harvest, six to ten plants were selected at random from each block. More plants
wereutilized inthefirsttwo harvestsbecauseofthe small size ofthe index leaves. The total
number of leaves geater than 2 cm in length were counted, to verify that the index leaves
were at similar developmental stages. The lengths of the index leaves were recorded and the

leaves were harvested, bagged, and placed in a cooler until processing.

Leaf processing and Fe quwrnficatiorrs. To remove surface soil contamination, leaves were
washed as follows: three rinses with distilled water, a 15-second rinse in 0.1N llINO3 and
three more distilled water rinses. Excess water was removed from leaves by spinning in a
salad spinner, and samples were held at -20°C prior to fleece-drying. The dried samples were
gound in a stainless steel Wiley Mill, with a 40-mesh screen.

Total and bioavailable Fe were determined in duplicate, for each sample. Total Fe
was determined using atomic absorption spectrophotometry (Video 12, Instumentation
Laboratory AA Spectrophotometer, Andover, MA). Samples were wet-ashed with
perchloric acid and hydrogen peroxide, diluted to 10 ml, and Fe concentration was determined
by comparison of absorption at 248.3 nm to a standard curve (Alder and Wilcox, 1985).
Bioavailable Fe was estimated using the in vitro assay described by Kapsokefalou and Miller
(1991). The assay was modified as described in Chapter 1. The percent dialyzable
(bioavailable) Fe was calculated fiom the values obtained fiom total and in vitro dialyzable

Fe.

68
Soluble Fe was determined for the leaf samples hour the greenhouse study. Afier the

removal of the dialysis tubing in the bioavailable Fe assay, the remaining digest was
centrifuged (1500 rpm, 15 min) and Fe concentration of the supernatant was determined,
using the AA protocol. This Fe fraction included those soluble Fe compounds which may
have molecular weight geater than 8000, the size cutoff for the dialysis tubing (Miller and
Schricker, 1982).

Analysis of variance was performed for total, dialyzable, soluble, and percent of total
Fe dialyzable or soluble where appropriate, for each experiment. The geenhouse experiment
was analyzed as a factorial design for cultivar and harvest date.

Results
Field study. A. tricolor accumulated the highest level of total Fe on three dates, and A.
hflrochondriacus accumulated the lowest total Fe in harvested leaves on all three dates.
Statistical analysis of the complete data set indicated sigrificant interactions between harvest
date and line. Examination of species averages for total Fe over the three harvest dates
indicated tlmt most species had the highest levels of total Fe at harvest 2, whereas A. tricolor
displayed progessively increasing levels of total Fe over the course of the experiment (Table
9). When A. tricolor lines were removed fi'om the data set, no sigrificant interactions
between harvest and line were observed for total and dialyzable Fe. Sigrificant differences
(P<.01) for total and dialyzable Fe existed among harvests and lines (Table 9). Total Fe
concentrations were sigrificantly higher at harvest 2 than at harvest 1 or 3, values ranging

fi'om 358 to 756 ppm for the lines on date 2.

69
DialyzableFewashighestatharvest l anddecreasedtoaconstantleveloverthe other

two harvest dates. The range of dialyzable Fe was relatively narrow at each harvest: day 1,
48-62 ppm; day 2, 41-62 ppm; day 3, 4075 ppm Total Fe concentrations varied to a much
larger extent than the dialyzable Fe for all lines tested. A sigrificant but low correlation was
detected between dialymble and total Fe (r =0.44). As the total Fe concentration increased,

the percent of Fe that was dialyzable decreased (Table 9).

Greenhouse Study. The length of the index leaf 4 increased for both lines over the period of
the experiment (Figure 1). By harvest 3 (24 days alter transplanting), the index leaves of
both specieswereapproachingmaxinmmlength Ateachharvest date, plants fromthe same
species had equivalent numbers of total leaves greater than 2 cm in length (Figure 1).
Although A. cruentus accumulated significantly (p<.01) lower levels of total, dialyzable and
soluble Fe in leaf 4 thanA. species at each harvest date, the patterns of Fe accumulation over
the five dates were similar for both species (Table 10). Total Fe concentrations for leaf 4
increased until harvest 3 after which the levels decreased (Table 10). Soluble Fe
concenhations were not sigrificantly different between the first two harvests, but decreased
to a sigrificantly lower level at subsequent harvests. Dialyzable Fe remained constant over
the five harvests, averaging 54 i 5 ppm for A. species and 39 i- 7 ppm for A. cruentus (Table
10).

The percent soluble Fe and percent dialyzable Fe were calculated from the values for
total, soluble and dialyzable Fe. The percent soluble Fe decreased after harvest 2 to a steady
level for the remaining harvests, and the percent dialyzable Fe decreased over the harvest

dates (Figure 2). A. cruentus had significantly (p<.01) lower percent dialyzable Fe than A.

70

species on all harvest dates , however the percent soluble Fe of the two lines was similar

(Table 10).

Discussion

Results fortotal, dialyzable and percent dialyzable Fe from the examination of 18 lines
gown in the field and harvested over three dates were similar to the results observed with the
35-line genetic evaluation (Chapter 1). Generally, those lines which accumulated high levels
of total Fe had high amounts of Fe in a dialyzable form (Table 9). A. tricolor accumulated
the highest concenhation of total Fe on all harvest dates. When examined over three harvest
dates, changes in total Fe concentrations of a line did not lead to proportional changes in
bioavailable Fe, and the actual amounts of dialymble Fe were not highly variable among these
dates. The average dialyzable Fe levels decreased slightly at harvest 2, despite the large
increases in total Fe detected on this day. Chawla et al. (1988) also found that total Fe in
GLVs did not correlate well with in vitro available Fe. These results were in agreement with
a study of sweetpotato tips, in which bioavailable Fe remained constant, whereas total Fe
concentrations changed over harvest dates (Pace and Bonsi, 1987).

On the second harvest (day 35) all lines except those of A. tricolor accumulated the
highest Fe content of the three harvests. If the fluctuations in Fe concentration observed in
other species were due to changes in the physiological state of the plants and not
environmental variation, a later harvest date may have demonstrated a decrease in Fe
concentration for leaves of A. tricolor, as observed in other research (Deutsch, 1978). In
three separate field evaluations, with different seasonal starting dates, amaranth lines

dernonsh'ated a similar pattern of increasing in Fe concentration of young, expanding leaves

71

followed by a decrease in levels later in plant development (Deutsch, 197 8). Deutsch
explained the results as being related to environmental conditions, despite the similar pattern
observed across the three trials. Environmental conditions may have contributed to the
increased Fe accunmlation at harvest 2 in the field experiment. An increased growth rate in
amaranth (a C4 plant) in response to the high light and temperature conditions of late July,
may have contributed to the fluctuations in Fe concentration of leaves harvested on date 2.

A second potential explanation for the decrease in Fe content of young leaves detected
at harvest 3 may have been in response to a change in the physiological state of the plant.
Newly expanding leaves of the plant were harvested at each date for analysis, but there were
not the same leaves on all three days. Flowering and seed development was observed in most
of the lines (except A. tricolor) by the third harvest, which would represent a stong sink for
inorganic ions (Valenzuela et al., 1992). In soybean plants, Fe concentration of the young
leaves was observed to increase to a maximum concentration just prior to the initiation of
flowering, and then to decrease (Drossopoulos et al., 1994). Amaranth seeds have a high
Fe concentration (Ologunde et al., 1991). In a study examining ferritin deposition in Fe-
depleted Xanthiunr pensylvanicum plants, floral induction decreased total Fe uptake and
ferritin deposition in expanding leaves when compared to non-induced plants (Seckback,
1982). Further examination of the Fe fluxes in amaranth leaf tissue with the transition from
vegetative to reproductive states would be required to determine the effect of this transition
on leaf Fe concentration and bioavailability.

In the field study, several leaves from the same vertical position on the plant were
pooled, but these leaves were not necessarily at the same physiological state. This especially

would be true for the young and expanding leaves. Any difi’erences in Fe content and

72
bioavailability among the leaves within these groups were masked by the sampling method.
From a practical perspective, food-harvesting methodology would pool several leaves fi'om
the same location on the plant. Therefore, harvesting all of the young and expanding leaves
was considered adequate in the initial design. However, the efi‘ects of leaf expansion and
maturity on Fe bioavailability could not be determined with pooling of leaves.

To explore further the changes in bioavailability with leaf expansion and maturation,
the greenhouse study focused on a single leaf over several harvest dates. In both of the
amaranth species examined, the total Fe concentration of leaf 4 increased until harvest 3 and
then decreased. This increase in total Fe corresponded with the leaf expansion to full size,
and the decrease in Fe concentration occurred over subsequent harvests. The Fe
requirements for a leaf are firlfilled during early gowth phases of the leaf, prior to full
expansion (Bienfait, 1989). The concentration of nutrient elements in a leaf may decrease
with time, due to the increasing dry matter accumulation (Tinker, 1981). However, leaf 4
exhibited decreased percentage dry matter over the full period of the study (Table 10).
Although Fe is not considered to be a mobile plant nutrient (Marschner, 1986), these results
may support reports of Fe remobilization fiom mature or senescing leaves to other regions
of the plant (Vasilas, 1987; Valenzuela et al., 1992; Drossopoulos et al., 1994). However,
the leaves sampled on each harvest date were not identical. To establish if Fe is being
translocated fi'om senescing amaranth leaves into other portions of the plant, firrther research
(perhaps using radiolabelled Fe) would be required to observe the Fe fluxes of the same leaf
over time.

Soluble Fe levels in leaf 4 did not fluctuate with total Fe concentrations. The

decrease in both the amount of soluble Fe and the percent soluble Fe observed with leaf

73

maturity at later harvests coincided with the approach of full leaf expansion. The higher
percentage of soluble Fe observed in the young expanding leaves may be due to the presence
of Fe in low-molecular-weight, chelated forms (i.e. Fe citrate). As leaves approach full
expansion, soluble Fe may be bound into cell structures or compounds which resist digestion
in the in vitro assay. The majority of plant Fe is found in the chloroplast (Terry and Low,
1982). Full photosynthetic capacity and chloroplast development was detected just prior to
firll leaf expansion, as leaves make the transition fiom a metabolic carbon sink to a carbon
source for other developing tissues (Turgeon and Webb, 1973). With development of fill]
photosynthetic competence, the Fe forms which were soluble at earlier leaf development
stages may become complexed or bound into other cell structures (Abadia, 1992).

Other plant compounds, such as tannins, may accumulate after the leaf achieves full
expansion (Mooney et al., 1983), and these may affect leaf Fe bioavailability. These
compounds may bind Fe into insoluble compounds, decreasing Fe solubility in the in vitro
assay (Gillooly et al., 1983). The solubilities of added FeSO, and purified heme (both soluble
and highly available) were decreased dramatically, and the heme complex was destroyed in
the presence of a spinach puree (Lee and Clydesdale, 1981). No explanation of the
observations was given. It would be of interest to determine Fe reactivity in plant material
of difi‘erent ages and to determine if endogenous plant compounds accumulate at a particular
leaf maturity, thus contributing to decreased Fe solubility. Identification of those compounds
which may decrease Fe solubility in GLVs would improve efl‘orts to enhance Fe bioavailability
fi'om these species Although soluble Fe cannot be correlated directly with bioavailable Fe,
increasing sohrble Fe may nuke a contribution to enhancing forms of Fe which may be more

bioavailable (Thompson, 1988).

74

Dialyzable Fe did not change to any large extent due to environment of production
or leafage, even with fluctuations in total Fe. Dialyzable Fe levels were similar between the
field study (range 45 to 60 ppm) and the geenhouse study (range 38 to 59 ppm). From a
practical perspective, the narrow range of dialyzable Fe that was detected may indicate a
stable fraction of Fe that is dialyzable, regardless of the total Fe concentration. The
interaction with other plant compounds, as described above, cannot be discounted. In the
sweetpotato study, authors proposed that although the amount of available Fe did not change,
the fiaction of Fe available increased, perhaps due to some undetermined enhancing factor in
older greens or to changes in the form of Fe available (Pace and Bonsi, 1987). The similarity
of dialyzable Fe values observed between geenhouse and field studies may have been related
to the relative accumulation of other compounds in plants grown in these two environments.

The results from these field and geenhouse studies indicated that fluctuations in the
total and soluble Fe content of leaves of amaranth may occur over the gowing season and
could be related to changes in plant physiological state or leaf maturity. Despite these
observed fluctuations in total Fe and soluble Fe concentrations, the dialyzable Fe fi'action of
leaves remained relatively constant in both the field and the geenhouse. The effect of
enhancing the amount of soluble leaf Fe on the level of bioavailable Fe needs to be explored.
Very little information exists as to the chemical structure of the soluble and available Fe
compounds in leaves (Welch and House, 1984). Further research is needed to determine
what forms of plant Fe comprise this soluble Fe fraction, if this fiaction could be increased
in size and if the chemical forms could be modified to be more digestible and consequently
more bioavailable. Breeding efforts could then focus on enhancing the levels of these specific

Fe compounds in GLVs.

75

Table 8. Amaranthus accessions evaluated for total and bioavailable Fe under field

conditions on three harvest dates.
—

 

 

Line P.I. Numberl Species and cultivar Origin

hi 1 Ames 5147 A.tricolor 'Red leaf Taiwan

hi 2 Ames 2154 A.tricolor 'Red leaf Taiwan

hi 3 PI 173837 A.tricolor 'Chulai' India

hi 4 Ames 51 13 A. tricolor 'Duradera' Taiwan

dub l Ames 5114 Adubius 'Stubby' Taiwan

hyb 1 P1482049 Ahybridus Zimbabwe
hyb 2 PI 210995 A.hybridus Afghanistan
cau 1 PI 166045 A.caudatus 'Chua' India

cau 2 PI 490609 A.caudatus Equador
cm 1 Ames 5598 A.cruentus Benin

cm 2 PI 433228 A.cruentus Guatemala
cm 3 Ames 8269 A.cruentus United States
cm 4 PI 511715 A.cruentus Guatemala
hyp 1 P1477917 A.h}pochondriacus Mexico

hyp 2 Ames 2171 A.h}pochondriacus Nepal

pal l Ames 5306 Apalrnerii Senegal
spe 1 PI 511752 Aspecies Peru

8 2 PI 337611 Aspecies __ Uganda

1National Plant Introduction Station accession number

 

 

 

76

Table 9. Total, dialyzable (DialXug-g" dry weight) and percent dialyzable (%Dial) Fe
in leaves of 18 lines of Amaranthus gown in the field and harvested on three dates.

Harvestl Harvest2 Harvest3

 

Line Total Dial %Dial Total Dial %Dial Total Dial %Dial

 

 

hi 1 618 62 l l 641 52 8 709 62 9
hi 2 552 60 12 575 49 9 674 61 10
hi 3 509 49 10 805 62 8 879 75

hi 4 589 52 9 756 45 6 806 57

dub l 478 56 12 560 54 10 490 59 12
hyb l 478 54 12 596 53 9 299 52 17
hyb 2 408 50 12 450 52 12 238 45 20
can 1 443 60 13 615 55 10 506 54 12
can 2 448 54 13 492 48 10 321 52 17
crul 415 52 13 595 51 9 319 50 17
cm 2 369 49 13 495 48 l l 197 40 21
01113 313 53 17 484 46 10 255 45 18
cm 4 456 S6 13 552 49 10 304 48 16
hyp l 279 48 19 358 41 12 190 45 24
hyp2 331 55 17 419 45 11 183 44 24
pal l 383 53 14 453 49 12 249 43 17
spe l 420 52 13 459 47 10 392 58 15
spe 2 390 49 13 550 47 9 224 43 20
LSD.“ l 15 6 3 128 6 3 175 9 4

-

 

 

 

 

 

77

 

 

2 z a 2 S a 3.8..
ma 8 a on 2 an. M: 5. mm m
3 8. 8 a... E X. t 8 a ..
2. 2. ea c. 3 e: S a 8 m
8.: x: 3 an 9: n2 8. R om N
e... 8. E a. no. 2. 8. a z .

on on cm on on om
29:. .8... 8328 23.85 29.. .38. 832% 2.8.15 38 .82....

.4 c.8838 EV i— ..qufiae §~§§V

   

.826 28.8.5“. o>m co 8.822 as. 8.6

  

2: fi :30

a8... 55...... as. e :8. 5 Easy .8....... 5 .88 85.383 5 .23 s. .3... es. 2.....3 £98.35 .2 as:

78

Figure 1. Change in total number of leaves (> 2 cm)
and length of leaf 4 on two amaranth lines grown in

the greenhouse and harvested on five dates.

 

20-

16—

12—

Number of leaves (> 2 cm)

+A.:p¢cies
+A.cruenm

 

 

18

 

 

15-4

12—

Length of leaf 4 (cm)

 

 

 

Days After Transplanting

4O

79

Figure 2. Change in the percent of soluble Fe and percent available
Fe of leaf 4 from two amaranth lines grown in the greenhouse and

 

harvested on five dates.
1 00
+ A. species - % soluble
90 _ -O— A. species - % available
63* ‘l— A. cruentus - % soluble
g

 

‘D— A. cruentus - % available

 

 

 

 

 

80—
E
:2 70..
g
E"
5:5 60-‘
g
as 50-
an
40“ ‘
"
3o-
20 l I I I
o 5 10 15 20 25

Days After Transplanting

80

Literature Cited

Alder, P.R. and GE. Wilcox. 1985. Rapid perchloric acid digest methods for analysis of
major elements in plant tissue. Comm. Soil Sci. Plant. Anal. 16(1 1): 1 153-1 163.

Bienfait, F. 1989. Prevention of stress in iron metabolism of plants. Acta Bot. Neerl.
38:105-129.

Chawla, S., A Saxena and S. Seshadri. 1988. In vitro availability of iron in various green
leafy vegetables. J. Sci. Food Agr. 46:125-127.

Deutsch, 1A 197 8. Genetic variation of yield and nutritional value in several Amaranthus
species used as a leafy vegetable. Diss. Abstr. Intl. B. 38:3969-B.

Drossopoulos, J.B., D.L. Bouranis, B.D. Bairaktari. 1994. Patterns of mineral nutrient
fluctuations in soybean leaves in relation to their position. J. Plant Nutr. ”:10”-
1035.

Gillooly, M., T.H. Bothwell, J.D. Torrance, A P. MacPhail, D.P. Derrnan, W.R Bezwada,
W. Mills, and RW Charton. 1983. The efi‘ects of organic acids, phytates and
polyphenols on the absorption of iron from vegetables. Brit. J. Nutr. 49:331-342.

Kapsokefalou, M and DD. Miller. 1991. Effect of meat and selected food components on
the valence on nonheme iron during in vitro digestion. J. Food Sci. S6(2):352-355,
358.

Lee, K and FM Clydesdale. 1981. Efl‘ect of thermal processing on endogenous and added
iron in canned spinach. J. Food Sci. 46:1065-1069, 1073.

Marschner, H. 1986. n N 'ti f Hi Pl . (New York: Academic Press).
p.674.

Miller, DD. and BK Schricker. 1982. ”In Vitro Estimation of Food Iron Bioavailability."

In Nutritional Bioavailabilig of Iron, ed. C. Kies. (New York: American Chem.
Society). pp. 11-26.

Mooney, HA, S.L. Gulrnon, and ND. Johnson. 1983. "Physiological constraints on plant
chemical defenses.” In Plant Resistance to Insects ed. P.A Hedin. Amer. Chem.
Soc. Syrnp. No. 208. (Washington D.C.:Amer. Chem. Soc). pp. 21-34.

 

Ologunde, M.O., R.L. Shepard, O.A Afolabi, O.L. Oke. 1991. Bioavailability to rats of
iron from fortified grain amaranth flour. Intl. J. Food Sci. and Tech. 26:493-500.

Omueti, O. 1982. Effects of age on the elemental nutrients of celosia cultivar. Exp. Agric.
18:89-92.

81

Pace, RD. and C. Bonsi. 1987. Iron content and iron availability of sweet potato tips
collected at difl‘erent times during the season. Nutr. Reps. Intl. 35(6):] 151-1156.

Pace, RD., T.E. Sibiya, B.R Phills, and 6.6. Dull. 1985. Calcium, iron and zinc content
of 'Jewel' sweet potato greens as affected by harvesting practices. J. Food Sci.
50:940-941.

Prakash, D. and M Pal. 1991. Nutritional and antinutritional composition of vegetable and
grain amaranth leaves. J. Sci. Food Agr. 57:573-583.

Seckback, J. 1982. Ferretting out the secrets of plant ferritin- A review. J. Plant Nutr.
52369-394.

Stafl‘ord, W.L., J .S. Mugema, R Bwabye. 1976. Efl‘ects of methods of cooking, application
of nitrogen fertilizer and maturity on certain nutrients in the leaves of Amaranthus
hybridus subspecies hybridus (Green Head). Plant Foods for Man. 227-13.

Terry, N. and G. Low. 1982. Leaf chlorophyll content and its relation to the intracellular
location of iron. J. Plant Nutr. 5:301-310.

Turgeon, RT. and J .A Webb. 1973. Leafdevelopment and phloem transport in Cucurbita
pepo: transition from import to export. Planta. 1132179-191.

Welch, RM. and W.H. House. 1984. Ch. 3. ”Factors AFfecting the Bioavailability of

MineralNutrientsinPlant Foods.” In as ofN tri f rH eds.
RM. Welch and W.H. Gabelrnan. Proceedings of a symposium. Nov 28,-Dec 3,

1982. Agronomy Society of America Special Publication #48.

Valenzuela, J .L., A Sanchez, A. del Rio, 1. Lopez-Canterero, and L. Romero. 1992.
Influence of plant age on mature leaf iron parameters. J. Plant Nutr. 15:2035-2043.

Vasilas, BL. 1987. Efl‘ect of planting date and growth on secondary and micronutrient
content of soybean tissue. J. Plant Nutr. 10:113-127.

Chapter 3
Evaluation of Iron Bioavailability from Amaranthus Species
Determined by Hemoglobin Repletion in Anemic Rats

Abstract
Initial screening of 35 lines from 12 species of Amaranthus indicated wide variation in total
iron, and small but significant differences in bioavailable iron when estimated by an in vitro
assay. To verify if the differences in bioavailable iron detected by the in vitro assay were
biologically significant, three lines of amaranth, fiom the species A. tricolor, A.
hwochondriacus and A. cruentus, were evaluated using a hemoglobin repletion assay with
anemic rats. The anemic rats were fed treatment diets in which all iron was provided by the
amaranth lines, and the hemoglobin gains were compared to a those of rats fed control FeSO4
diets. Both slope-ratio and hemoglobin repletion eficiency analyses were applied to
determine the iron bioavailability. Although A. tricolor contained a higher concentration of
total iron (690 ppm), the relative bioavailability of this iron was lower (42%) than A.
hwochondriacus (66%) or A. cruentus (63%). However, when the same amount of
amaranth was added to each diet, A. tricolor supported the largest hemoglobin gain of the
three lines. These relative difi‘erences in iron bioavailability estimated by in vitro assay were
confirmed with the rat bioassay, supporting the potential use of the in vitro assay in breeding

programs focused on improving the iron nutritional quality of green leafy vegetables.

82

83
Introduction

Iron (Fe) deficiency anemia is one of the most prevalent nutritional deficiencies
afi‘ecting the world's population today, especially among women and children of developing
countries. Most nutrition programs aimed at decreasing the incidence of Fe deficiency utilize
fortification of the diet with various Fe compounds. An alternative approach may be to
enhance the Fe nutritional quality of the diet, through improvement of Fe bioavailability fi'om
commonly consumed foods. Green leafy vegetables (GLVs), such as amaranth (Amaranthus
spp.), are one such food group. Amaranth is an inexpensive, widely available GLV cultivated
in many regions of the world where Fe deficiency anemia is a problem, such as Central
America, Asia and Afiica. This crop accumulates a higher level of Fe than many other
commonly cultivated greens (Ifon and Bassir, 1978; Chawla et al., 1988; Makus, 1984).
Enhancement of the nutritional quality of this vegetable may be a very cost-efi‘ective and
easily available method to provide populations with a good source of Fe in their diet.

Most studies on Fe bioavailability of GLVs have examined spinach (Spinacea
oleracea L.), a temperate GLV. A few researchers have screened popular local tropical
greens. However, it is difiicult to compare these studies due to the different Fe
bioavailability assessment protocols and the difi‘erent methods for collecting and preparing
the GLVs for analysis. In human feeding studies evaluating spinach, bioavailable Fe ranged
fiom 1.3-13% of total Fe (Moore and Dubach, 1951; McMillan and Johnson, 1951; Moore
and Dubach, 1956; Layrisse et al., 1969). In feeding studies using anemic and Fe-replete
rats, the bioavailable Fe in spinach ranged fiom 41-70% or 26-59% total Fe, respectively (Pye
and MacLeod, 1946; Van Campen and Welch, 1980; Gordon and Chao, 1984; Zhang et al.,

1985; Miller, 1987; Zhang et al., 1989). In a comparison of ten tropical GLVs using anemic

84
rats, Ifon and Bassir (1978) found a range of Fe bioavailabilities from 7 .7-36.2%, one line of

amaranth having the highest percent bioavailable Fe of all species tested. Using various
modifications of an in vitro simulation of gastrointestinal digestion, researchers found
between 7.5 and 53.6 % total Fe in GLVs to be bioavailable (Reddy and Kulkami, 1986;
Chawla et. al., 1988; Duhaiman, 1988; Reddy and Malewar, 1992).

In order to develop specific cultivars of a GLV with improved Fe nutritional quality,
differences in Fe bioavailability due to plant genetic variation must be established. In most
previous studies, GLVs were collected from a market or fi'om non-cultivated areas,
preventing any experimental control for environmental variation. Initial investigations
compared 35 lines of amaranth for Fe content and bioavailability (Chapter I). These lines
were selected fi'om 12 species and included both wild and cultivated types of amaranth.
When grown in a replicated field trial, lines of amaranth accumulated from 360 to 880 ppm
total iron. Utilizing an in vitro assay which simulates gastrointestinal digestion
(Kapsokefalou and Miller, 1991), the relative Fe bioavailability of these lines varied fi'om 6-
12%, and significant differences were detected among lines and species.

Although differences were observed in Fe bioavailability with the in vitro assay, these
may not be significant in an animal, due to other intrinsic and extrinsic factors affecting Fe
bioavailability. In order to verify the significance of the difl‘erences observed with the in vitro
assay, iron bioavailability fiom three amaranth lines was estimated by hemoglobin (Hb)
regeneration in anemic rats (Forbes et al., 1989). Preliminary screening of a food product
with in-vitro analysis followed by the evaluation of Hb repletion in rats has been
recommended as a method of choice for predicting Fe bioavailability (Forbes et al., 1989).

The confirmation of relative differences in Fe bioavailability of this GLV is essential prior to

85
attempting to enhance the nutritive quality through plant breeding. In addition, verification
of in vitro results would increase the applicability of the in vitro method of estimating Fe
bioavailability to other plant breeding programs interested in enhancing the mineral nutritional

quality of vegetable foods.

Materials and Methods

Dietary Test Ingredients. Based on previous genetic evaluations, three lines of amaranth
were identified (Table 11) that displayed large differences in total Fe accumulation when
grown under field conditions (Chapter 1). These lines were direct-seeded into the field
(Horticulture Research Center, Michigan State University) and harvested alter 35 days.

After removing stems, leaves were washed in three rinses of distilled water, soaked in 0.1N
HNO3 for 15 sec and rinsed three more times in distilled water to remove surface soil
contamination Leaves then were steam-blanched and fieeze—dried. The dried material was
ground to l rmn mesh size and was stored at -20° C. After wet digestion with perchloric acid
(Alder and Wilcox, 1985), total Fe was estimated using atomic absorption spectrophotometry
(Video 12 Instrumentation Laboratory AA Spectrophotometer, Andover, MA) and in vitro
dialyzable Fe was evaluated using the modification of the Kapsokefalou and Miller (1991)
assay (Chapter 1). The neutral detergent fiber content of the plant materials was determined
(Goering and Van Soest, 1970) and the protein content was estimated by the Kjeldahl

method.

Animals card diets. Male weanling Sprague Dawley rats (40—50g)(Harlan Sprague Dawley,

Inc., Indianapolis, Ind.) were housed in individual, suspended, stainless steel cages with wire

86
mesh floors to prevent coprophagy. Light was regulated to provide 12-h light and dark

cycles and temperature was maintained at 25° C throughout the study. Food and distilled,
deionized water were provided ad libitum Rats were fed the AlN-93G (Reeves et al., 1993)
semi-purified diet with no added Fe (Dyds, Inc., Bethlehem, Pa.). Afier 28 days, blood was
taken from the tail artery, and Hb was determined colorimetrically (Sigma Cherrrical Co., St.
Louis, Mo., diagnostic kit 525-A). Rats then were sorted into treatment groups (8 animals
per group) such that all treatment groups had similar average Hb levels (4.07 i .57 g-dL")
and weight (205 i 12 g) prior to feeding treatment diets.

The composition of the treatment and control diets are presented in Table 12. One
group of anirmls confirmed to receive the basal low Fe did (C0). Three groups received the
basal did supplemented with either 5, 15, or 25 ppm added FeS04-7H20 (referred to as C5,
C15, and C25). All Fe in the test diets was provided by amaranth leaf material. Six of the
groups received diets containing Fe fiom A. tricolor (ATS, AT15, AT25) or A.
hwochondriacus (AHS, AHlS, AH25) substituted for corn starch, casein and fiber in the
diets to provide 5, 15 or 25 ppm of added Fe. The one remaining group received a did
containingA. cruentus (AC7) added at a level of 30 g'kg" did, providing 7.5 ppm added Fe.
This final treatment was included to compare the Hb repletion in rats fi'orn each of the
amaranth lines when provided at the same quantity to the did (AC, AHS, and AT25 in Table
11). Due to the high concentration of A. W required to achieve 25 ppm added
Fe in did AH25, the amount of added fiber was decreased using the conservative estimate
of 10% fiber from the plant materials (actual NDF fiber content averaged 22%, Table 11).

Fe and protein contents of the diets were determined as described for the plant materials. The

87
total energy (kcal-100g") was determined using bomb calorimetry (Parr 1241 Adiabatic
Calorimder, Moline, 111.).
All treatment and control diets were provided ad libitum for 14 days, and didary
intakes were recorded. At the end of the repletion period, animal weights were recorded,

and blood Hb levels were analyzed as previously described.

Determination of bioavailable Fe and statistical analyses. The graded treatment levels
provided data for a slope ratio analysis (SRA) of the eficiency of Hb repletion by the
different amaranth lines. The statistical analysis for slope ratio analysis of the Hb gain during
therepldionperiodwasconductedasoutlinedbyFinney(1964). In ordertouse SRA, three
criteria must be satisfied : 1) a linear response to increasing treatment levels must exist, 2) the
equations describing these linear responses must not be significantly different for Y
interception, and 3) intersections do not vary significantly from the Hb gain of the no-Fe
treatment (F inney, I964; Amine and Hegstead, 1974). Over the range of the levels tested,
the relative Fe bioavailability, or relative biological value (RBVSM), of each treatment
compared to the FeSO4 standard can be ddermined fi'om the ratio of the slopes of the linear
regression equations describing the treatment and control, multiplied by 100.

Another measure of relative biological value (RBVHRE) can be determined from the
Hb repletion efiiciency (HRE) (Mahoney and Hendricks, 1982):

mgHbFe= bodywt(g) x .067mlblood/gbodywt
xgHb/ml blood x 3.35 mg Fe/gHb

HRE= mg flb-Fgfinal) - mg Hb-ngnitifl) x 100
mg Fe consumed

RBVm= (HRE treatment/ HRE FeSO4) x 100

88
TheRBVmwerecalculatedbydividing the treatment HRE bythe control HRE containing
the same added Fe level. The HRE and RBVm values were calculated for rats receiving
the three diets containing 30 g-kg" did of the three amaranths (AC7, AHS, and AT25) to
ddernineifdifl‘erencesinthepalatability, Fecontentandotherplant-related factors may have
contrilartedtotreatmentefi‘ects. Hbgain, weightgain, total did consumed, andHREwere

analyzed for statistical differences using a one-way analysis of variance.

Results

The three amaranth lines tested in this study, the total Fe content, in vitro dialyzable
Fe ofthese lines, and fiber and protein content are listed in Table 11. The lines had similar
protein (32%) and neutral detergent fiber (22% of dry matter) contents.

Results from analyses of the treatment dids for Fe, protein, and total energy are
presentedinTable 13. ActualmeasuredFecontentaveragedabout 5 ppm higherthanadded
Fe values, due to small Fe contributions fi'om other did ingredients. Dietary protein levels
averaged 18%, and total energy averaged 452.4 kcal-100g" (Table 13).

Rat weight gain, total did consumed and Hb gain ofthe treatment dids are presented
in Table 14. Rat weight gain and total did consumed were lower for rats fed the low Fe
dids CO, AH5 and ATS than for rats fed the other treatments. For other dids, weight gain
ranged fi'om 67 to 87 g, and did consumed ranged from 233 to 285 g over the study. The
magnitude oftheI-Ib gainvaried by Fe level and amaranth line. As didary Fe concentration
increased, Hb gain increased, regardless of the Fe source in the did. The highest observed
Hb gain, 7.19 g-dL", occurred in rats fed the 25 ppm control did (C25). Over the repletion

period, rats fed A. tricolor achieved less than one half of the Hb gain of rats receiving A.

89

hypoclronrb-iacus at each of the added Fe levels, and Hb gain from the 15 ppm A.
lgpochonrb'icrcus did (A111 5) exceeded the gain fiom the highest Fe A. tricolor did (AT25).

The Hb repletion emciency (HRE) and the associated RBVm for each did are
presented in Table 14. Although the basal did (C0) contained no added Fe, and rats fed this
did showed a decrease in Hb, the HRE was positive. Rats continued to gain weight on this
treatment; thus some Fe was absorbed from the basal did or animal internal Fe stores to
support cellular Fe requirements for gowth (Table 14). Rats fed dids containing A. tricolor
had the lowest average HRE (26.9) compared to the HRE fi'orn the A. lupochoncb'iacus dids
(42) and the FeSO, controls (63.9). Assuming that the Fe provided by the FeSO. control
dids was maximally absorbed (100% RBVHRE), then the RBVm of the A. lopochondriacus
diets averaged 66%, the RBVm for A. tricolor averaged 42%, and the RBVm of the A.
cruentus did was 63%.

The dids AH5, AT25 and AC7 contained 30 g amaranth-kg did. Weight gain of rats
fed did AH5 was significantly lower than rats fed dids AT25 and AC7. When the same
amount of amaranth was added to each did, A. tricolor supported the highest Hb gain, 1.78
gdL", followed by A. cruentus andA. Inpochona'rracw at 0.94 and 0.38 g-dL", respectively
(Table 14). However, the HREs for these three dids were not significantly different.
AlthoughA. lopoclrorxbraau supported the lowest Hb gain, Fe was absorbed from this line
with efiiciency similar to the other two lines.

Prior to using slope ratios for estimating Fe bioavailability, an analysis of variance of
Hb gain was conducted to confirm that the conditions for the slope ratio analysis were
satisfied (Finney, 1964). No sigrificant deviation fiom regession due to curvature,

intersection, or blanks was detected (Table 15), satisfying the conditions for slope ratio

90
analysis. The RBVm (compared to FeSO,) for A. lopochondriacus was 61%, and for A.

tricolor was 44% (Table 16). No RBVM could be calculated for A. cruentus since only one

Fe level was tested.

Discussion

Previous analyses of amaranths for both total Fe and in vitro dialyzable Fe indicated
that difi‘erences were geater among amaranth species than among lines within individual
species (Chapter 1). Therefore, lines fiom three species which accumulated different levels
of total Fe were selected for this study. These three lines also represent the difi‘erent uses
of amaranth as a GLV or gain crop. A. tricolor (Ames 5113), a cultivated GLV called
'Duradera' in Taiwan, accumulated the highest total Fe (Table 11). A. hypochondriacus
(Ames 2171), a gain amaranth from Nepal, and A. cruentus (PI 433228), a gain/GLV type
fiom Guatemala, accumulated less than one halfthe total Fe of A. tricolor.

In order to provide 25 ppm added Fe using the low-Fe amaranth line, did AH25
contained 14.7% A. lmrochondriacus leaf material. Despite this high concentration of
amaranth, consumption of this did did not differ sigrificantly fiom the C25 control did.
Dids containing A. tricolor were consumed slightly less than other treatment diets, but no
correlation was found bdween Fe source and consumption of did. This line was added at
low concentrations to the treatment dids to provide the desired levels of Fe (Table 12). As
a cultivated GLV, this line would not be expected to contain 'ofllflavors', although negative
flavors to the rat cannot be discounted. A significant but weak correlation was found
bdween did Fe content and did consumption (r = 0.46), which may explain the lower didary

intake and weight gain observed in animals fed dids CO, AHS, and ATS compared to other

91

treatments. Rats fed the C5 did did not have significantly difi‘erent weight gain or did
consumption than other higher-Fe-containing dids, presumably due to the higher
bioavailability of the FeSO, in this did.

Slope ratio analysis (SRA), originally described by Finney (1964), is a dose-response
method to assess the relative 'potency' of a test material; in this case, the relative Fe
bioavailability fi'om leaf material. Results fiom SRA of Hb repletion in anemic rats have been
demonstrated to have good correlation with human clinical measurements of Fe
bioavailability, making it a widely accepted method (Forbes et al., 1989). As a standard
protocol, results from this study may be compared with results fiom other research on food
Fe sources using the same protocol. SRA, using multiple levels of each treatment, will
indicate if treatment levels have saturated biological uptake capacity or if other didary factors
which decrease diet palatability are present. Either condition may contribute to an
underestimation of Fe bioavailability. Using a single-dose mdhod for estimating Fe
bioavailability from GLVs (Ifon and Bassir, 1978) may result in difficulties in interpretation
if large difi‘erences in consumption occur. This study is the first reported to use SRA for
testing Fe bioavailability from a GLV. The data for the control and treatment dids satisfied
the conditions for linearity of Hb gain with increasing levels of Fe (fiom amaranth) in the did
(Table 15), providing no evidence of interactions of did components at the levels of amaranth
added. The slope of the FeSO4 control treatments equalled 0.277, similar to the slope values
found by other researchers using the same protocol (Amine and Hegstead, 1974; Forbes et
al., 1989). FeSO4, the control Fe source in this study, is considered to give the maximum
Hb response per unit Fe and is the standard Fe source to which other compounds are

compared. The RBV for A. hwochondriacus was 61% and for A. tricolor was 44% when

92

compared to FeSO4 (Table 16). Similar RBV values were calculated using HRE values, 44 -
66% (Table 14).

The HRE calculation method for estimating Fe bioavailability is based on the
efficiency of conversion of did Fe to Hb Fe. The Hb gain is corrected for difl‘erences in
animal weights and didary intake during the repldion period (Mahoney and Hendricks,
1982). The HRE value, which is a percent of ingested Fe converted to Hb Fe, may be
compared to other studies which used single-dose Fe bioavailability tests or to in vitro Fe
bioavailability analyses. In this study, the HREs for amaranth ranged from 26.9 - 42%, and
HREs observed for the control treatments (552-73.4%) were similar to those observed in
other research (Zhang d al., 1989). These HRE values were similar to results of other
research on Fe availability of GLVs using Hb repldion of anemic rats. Pye and MacLeod
(1946) as well as Gordon and Chao (1984) found 26% (RBV=67 and 63%) of the Fe in
spinach available to rats. In two separate studies, Zhang and coworkers (1985, 1989)
examined spinach Fe bioavailability to anemic rats and found HREs of 34 and 37% (RBV=
51%). However, no information was provided in any of these studies as to the cultivars of
GLV tested.

When the same concentration of amaranth was present in the did (30 g-kg"),
differences in Hb gain depended on the total Fe concentration of the diet (or the plant
material). A. tricolor, which contnbuted 25 ppm Fe to did AT25 supported geater Hb gain
than either A. hypoclroncb'iacus~ or A. cruentus, which contributed 9 or 13 ppm, respectively,
to diets AH5 or AC7. The HRE values indicated no significant difference in the Fe
absorption emciency from these three lines, although the HRE for A. tricolor was slightly

lower than the other two lines (Table 14). Ifon and Bassir (1978) supplemented diets with

93
20 g-kg‘l of 10 GLVs and found a range of HREs fiom 7.7 to 36.2%. The highest HRE was

from a line of A. Intbridus and was equivalent to the HREs observed in the current study for
A. lopochondriacus, 37.4%, and A. cruentus, 35.8%.

The similar HRE values observed in response to the three diets containing 30 g-kg'l
amaranth suggested that the presence of other plant compounds had minimal efi‘ect on Fe
bioavailability. Had other plant compounds been influencing the Fe bioavailability fi'om the
dids, the HREs would not have been equal, and Hb gain would not have been proportional
to the Fe content of the did. Due to the small quantity of these plant materials added to the
dids (3% of total did), the effect of other plant compounds on Fe bioavailability in the meal
would be expected to be minimal. Gordon and Chao (1984) examined different combinations
of spinach components (lignin, cellulose, oxalate) at levels found in the leaves and found no
decrease in Fe bioavailability to anemic rats due to these compounds. Although these
combinations were tested separate fi'om the plant material, these results suggest that the
chemical form of Fe in GLVs may be relatively more important than the presence of
secondary compounds which might decrease meal Fe bioavailability.

The relative differences in Fe bioavailability ddermined by in vitro analyses were
confirmed in the anemic rat Hb repldion assay, supporting the use of the in vitro screening
mdhod for dderrnination of the Fe nutritional quality of GLVs. A. lypochondriacus and
A. cruentus had a geater percent bioavailable Fe than A. tricolor in both in vitro and animal
bioassays. Estimates from the in vitro analyses were more conservative than those generated
in the rat assay. In vitro analysis indicated 26% Fe bioavailable fi'om A. hypochondriacus,
21% from A. cruentus and 9% Fe bioavailable fiom A. tricolor. Another study used an in

vitro assay to estimate percent bioavailability Fe fi'om four different amaranth species and

94
found a range from 15.2 - 53.6% (Reddy and Kulkami, 1986). In contrast to in vitro results,

the HREs (or percent bioavailable Fe) for the A. lnpochondriacus and A. cruentus averaged
42% and for A. tricolor, the HRE averaged 27%. Anemic rats have high Fe repletion
efficiencies and tend to be less sensitive to didary factors which may decrease Fe
bioavailability in other systems (Reddy and Cook, 1991).

The total Fe concentration of these GLVs was not correlated with the RBVs.
Although A. tricolor contained the highest level of total Fe, a lower percentage of this Fe is
bioavailable. Despite the lower percent of available plant Fe in this species, the absolute
amount of available Fe contributed to the did was larger. When the three lines were
provided in equal amounts to the rat dids, A. tricolor supported the largest Hb gain. Future
research on the bioavailable Fe in GLVs should focus on the chemical forms of Fe in GLVs
which are bioavailable. Once identified, breeding efi‘orts could focus on these specific
bioavailable Fe compounds. Initial screening with the in vitro assay for Fe bioavailability
followed by confirmation of improved lines using Hb repletion in anemic rats is recommended
as an emcient screening mdhod for plant breeding efforts. Until the chemical forms of
bioavailable Fe in GLVs are identified, selection for high Fe-accumulating GLVs would be
the most emcient mdhod to increase the amount of Fe provided to the human did by this

vegdable food.

95

«3.2 2.2.9.. .5 8.... .

.5... 8...... .......z .

.. 5.9.5 .85.: .5. 3.2.2 .. 8.8.... a a. .3238... .o 8.2.... e... .. .
9.2.8.... commnoooe 5.2% 5.838.... .5... .3282 .

 

 

. H o. .. H 2. .N . H m. N. a :N , 8.... .. 38.5....

.H m. .. Ho... . Tu... .38. 238...... 8.8.... ....

3... a“... 8 .11.... .. HE. E. 8...... 3.38.... .V
o.

.52.... 2.32.... .9.

8.5 .\. he... .\. 2.8.3.. o. .3... .352... 8.8...

 

.8... 0.80.... c. .5382... .. :50 082— .. 8.3.56
.3... £535... 8....82... 1.2.3.... .5. .. 2.3%... as... .5. .23 e. 3...... .532... .3... .: 2....

96

98.. ...—3250.. so... .
.93.. .5823 .8. ...»... ...... 8...... 8 ...... ...... 13...... 08-2....
...0 8.8.96.0 ......00 .3838... .m.D .838.
.§§Sa§a=..fi.. ...... 83.3.8.
.88.. ...—30.53 so... .395 So... 33:95.. 983 .88.. ....a 088..» £25.. ....8 . .Omo... 8558:... ..o 8598.8 5.3 8.8.3.»... 8.9

so... ...... ...a....> 08.22.

 

 

c8. 8... 8... .8. .8. 8o. 8... c8. .8. .8. 8o. .38.
... 3 .... 3 2 n. 3 ... ... ... ... 33...... 8......0
m m . m . . . m m m m 8.88...
... ... ... o. ... ... ... ... o. ... ... ...... ......s.>
8 8 8 8 8 n. 8 8 8 8 8 ...... 88.2883.
... .... 8 8 .... ... 2.. 8 8 8 8 ...3...
2 2 2 2 2 2 2 2 2 2 2 .8538.
8. 8. 8. 8. 8. 8. 8. 8. 8. 8. 8. .088...
8. 8. 8. 8. 8. 8. 8. 8. 8. 8. 8. 82.2..
8. 8. 8. 8. n... 8. 8. 8. 8. 8. 8. ...oau
38 .8. 28 ...... n2. ...... 28 I... 2.8 28 28 ......8 88
2... n8. 2... 0.... .08..
GM 35.29% .V
2 2 ... 8.8... ...
.... 8 2 3.8.88.8. ...

8< a... a: 3... a... 2.2 ...... no 20 8 8

flag... .8633 .88.»...

.88 0.80.... ... ......o.

 

 

 

.....o 2...... .... .88.... ...... 982.... ......82 ... 2......

97
Table 13. Analysis of diet compositions. Added and actual Fe values are represented

as a!" d_rz wmgt.

 

Diet Description Added Fe Actual Fe' % Crude Protein” Total Energy’
(130140084)
C0 Basal(NoFe) 0 4:2 17.4iO.7 459:5
C5 Control (F680) 5 8 i 1 18.4 i 0.7 449 i 1
C15 Control (F680) 15 1712 18311.0 453:1
C25 Control (F6800 25 31 i 1 18.1 10.5 455 i 5
AHS A. hwochondfiacus 5 9 i 1 18.7 1 0.4 451 i 5
A1115 A. Mochandn’acus 15 22 i 2 19.2 i 0.6 447 i 2
AHZS A. Iopochona‘rlacus 25 31 :1 17.9:05 461 :9
ATS A. tricolor 5 12 i 1 17.6 :1; 1.4 450 i 2
ATIS A. tricolor 15 18 i 3 18.7 i 0.8 452 i 5
AT25 A. tricolor 25 25 i 1 18.1 i 0.4 452 1' 3
AC7 A. cruentus 7 13:1 19.0:0.1 449:4

 

     

' Values for 101:.ch represem " 1'—'_' fierepvlicates ' " ____ ‘W " v
2"/oneldahl N * 6.25. Values for percent crude protein represent mean 3 SD. M 3 replicates.
3Values for total energy represent mean 1 SD. from 2 replicates

98

.253. 333» 2.. a 33.82 .8 22» 8< .23.. >9. .2 353m
8:32: .22.... 83235833353583.32393893028.55338280233322:
2.. .8< 8% 3.2 >9. 2:0 8383 3. 3.2 .3058. 2. £32. 5 3:33 .... 3283 as 99: 81> 332m 233. .
.8238 2.5 4:548 3 Bacon... 8 32.28 a: @5680 83.22 583252 N
.89 a sang 5.8.2.? 2. be». £56.23 05... 2a B 3.28.52 .8 «5.38 85 a 8:3 _

 

 

 

8 .... 2. H ”.2 .a. H :2. .4. H 4H ...... cm H an ..o H S H 35.5 ... B<
2. ... 3 H 4.8 .. S. H a: . n. H G ...: H a: .e H E a .288 ... «S.
R .. 3 HM: L6H 2:. 3. H2. ..2 H mm 3H2 2 .2881 :.<
2. _. 2.2 H :3 . S. H 8... 1. H 2 .... 2 H as ... a H s H 8288 a. 2k
8 .... 3. H ad. .. 2. H 8... .. n. H 2 3.2 H SN .1. H a a a§§§§§ ... $2
3 ... S H 3:. . a. H a.” _. n. H 2 ... a H ”a .3 H 2 2 3:239:83 «. :2
8 .... a: H in a 3. H and _ N. H 2 . «N H m2 .0 H an n sustensge a. 22
82 ... «H H as . mm. H 2 s . .... H 2 . : H 22 . n H a a .880 m8
2: . 2: H12 .. 8. H 3.4 . ... H 3. an H e: ...a H S 2 .35 2o
82 ... 93 H «.2 .. 2. H a... _ N. H 3 a: H SN ...; H S m 6580 8
£2ng .mmHas- 2H2 ...NHE 3H8 o 23m 8

333 8.5 E cognac 3 28.5
$5 43.2mm: a.» .E 9.35 a .05 .56 ..3 on 3?. SEE 35
.3958

um .52.. ea 95¢ 8:: 3522.. 0238 as £25 .6868» :22? sneeze: 333.8 as .32 :22?
5.: 2: $.56 3% .858: .5 .228 3 ea .2 335 on 2.5 35:28 we 5... =32» omeo>< .3 2.3

99

Table 15. Analysis of variance of hemoglobin gain in rats to test for
satisifaction of en'ten'a for slope ratio analysis. Actual diet Fe levels were
used in the analysis.

Source of Variation Degrees of Sums of Mean Squares F Value

 

freedom Squares
Due to regression 3 375.26 125.09 286.03”
Duetoblanks l 1.19 1.19 2.72
Due to curvature 3 3.76 1.25 2.86
Due to intersection 2 .09 .04 .10
Error 66 28.86 .44
Total 75 409.16

 

 

"Statistically significant (P<.01) '

100

Table 16. Linear regression equations and calculations of relative Fe bioavailability,
using slope ratio analysis, of two amaranth species compared to the FeSO‘ control Fe

 

 

source.
Fe Source Regression Regression Relative Bioavailable
Coeffieent Equation Fe‘('/o)
FeSO4 .941 y = .277 x - 1.089 100
A. hypochondriacus .800 y = .170 x - 1.030 61
A. tricolor .543 y = .123 x - 1.295 44

 

' Relative Bioavailable Fe calculated by dividing slope of the treatment regression equation by slope
of FeSO. control and multiplying by 100.

101

Literature Cited

Amine, E.K. and D.M. Hegstead. 1974. Biological assessment of available iron in food
products. J. Ag. Food Chem. 22(3): 470-476.

Chawla, S., A Saxena and S. Seshadri. 1988. In vitro availability of iron in various green
leafy vegetables. J. Sci. Food Ag. 46:125-127.

Duhaiman, A.S. 1988. Total iron content and bioavailability from liver, meat and
vegetables. Nutr. Rep. Intl. 37(3):645-651.

Finney,D.J. 1964. ”Ch. 7. SlopeRatioAssays." In ' ' M h inBil '
Amy. (New York: Hafner Publishing Co). pp. 187-213.

Forbes, A., C.E. Adams, M.J. Amaud, C.O. Chichester, JD. Cook, B.M. Harrison, RF.
Hurrell, S.G. Kahn, ER Morris, J.T. Tanner, and P. Whittaker. 1989.
Comparison of in vitro, animal and clinical determinations of iron bioavailability:
International Nutritional Anemia Consultative Group Task Force report on iron
availability. Amer. J. Clin. Nutr. 49:225-238.

Gordon, D.T. and L. Chao. 1984. Relationship of components in wheat bran and spinach
to iron bioavailability in anemic rat. J. Nutr. 114:526-535.

Goering, H.K. and R]. Van Soest. 1970. Forage fiber analysis. USDA Agr. Hdbk. 379.

Ifon, ET. and O. Bassir. 1978. The emciency of utilizing the iron in leafy green
vegetables for hemoglobin synthesis in anemic rats. Nutr. Rep. Intl. 18(4): 481-
486.

Kapsokefalou, M. and DD. Miller. 1991. Efl‘ect of meat and selected food components
on the valence of nonheme iron during in vitro digestion. J. Food Sci. 56(2):352-
355, 358.

Layrisse, M., J.D. Cook, C. Martinez, M. Roche, I.N. Kuhn, RB. Walker and CA Finch.
1969. Food iron absorption: a comparison of vegetable and animal foods. Blood.
33(3):430—443.

McMillan, TI. and FA Johnston. 1951. The absorption of iron from spinach by six
young women and the effect of beef upon the absorption. J. Nutr. 44:383-398.

Makus, DJ. 1984. Evaluation of amaranth as a potential greens crop in the mid south.
HortScience. 19(6):881-883.

102

Mahoney, A W. and D. G. Hendricks. 1982. ”Efi‘eciency of hemoglobin repletion as a
method of assessing iron bioavailability 1n food products." In Nutritional
WW. ed. C. Kies. (Washington, D. C.: American Chemical
Society). p. 1-11.

Miller, J. 1987. Bioavailable iron in raw and cooked spinach and broccoli. Nutr. Rep.
Intl. 36(2):435-440.

Moore, CV. and R Dubach. 1951. Observations on the absorption of iron from foods
tagged with radioiron. Trans. Assoc. Amer. Physicians. 64: 245-256.

Moore, CV. and R Dubach. 1956. Metabolism and requirements of iron in the human.
J. Amer. Med. Assoc. 162:197-204.

Pye, O.F. and G. MacLeod. 1946. The utilization of iron from different foods by normal
young rats. J. Nutr. 32:677-687.

Reddy, MB. and J .D. Cook. 1991. Assessment of dietary determinants of nonheme iron
absorption in humans and rats. Amer. J. Clin. Nutr. 54:723-728.

Reddy, NS. and KS. Kulkarni. 1976. Availability of iron from some uncommon edible
green leafy vegetables determined by in vitro method. Nutr. Reps. Intl.
34(5):859-861.

Reddy, S. and V.G. Malewar. 1992. Bioavailabilitiy of iron from spinach cultivated in
soil fortified with graded levels of iron. P1. Foods Human Nutr. 42:313-318.

Reeves, P.G., F.H. Nielsen, and G. C. Fahey. 1993. AIN-93 Purified diets for laboratory
rodents: final report of the American Institute of Nutrition Ad Hoc Writing
Committee on the reformulation of the AIN-76A rodent diet. J. Nutr. 123:1939-
195 l .

Van Campen, D. and RM. Welch. 1980. Availability to rats of iron from spinach: Efl‘ects
of oxalic acid. J. Nutr. 110(2):1618-1620.

Zhang, D., D. G. Hendricks, A W. Mahoney, D. P. Comforth. 1985. Bioavailability of
iron in green peas, spinach, bran cereal and cornmeal fed to anemic rats. J. Food
Sci. 50:426-428.

Zhang, D., D. G. Hendricks, A. W. Mahoney, D. P. Comforth. 1989. Bioavailability of
total iron from meat spinach and meat spinach mixtures by anemic and nonanemic
rats. Brit. J. Nutr. 61 1331-343.

Summary

If green leafy vegetables, such as Amanthus species, are to be promoted as sources
of Fe in the human diet, the bioavailability from these plant foods needs to be improved.
Research was initiated to explore the genetic variation among species of amaranth for Fe
bioavailability and Fe content. Previous research suggested that genetic variation in total Fe
(Deutsch, 1978; Makus, 1984; Reddy and Kulkami, 1986) and bioavailable Fe (Reddy and
Kulkami, 1986) exists among species of amaranth. The genus Amaranthus is known for its
ability to form interspecific hybrids readily (Kulakow and Jain, 1990), which would allow
genetic manipulation of Fe bioavailability if variation among species were confirmed.

Initial experiments involved screening of 46 lines of amaranth, fiom 12 species, for
genetic variation in Fe bioavailability in two growing environments (Chapter 1). An initial
field study was followed by a greenhouse study to determine if relative difl‘erences in Fe
bioavailability were maintained when plants were grown in a controlled environment. High
Fe-accumulating species, such as A. tricolor L. and A. Iividus L., tended to have higher
bioavailable Fe. Within species, differences were detected only in these high Fe-accumulating
species. Significant differences in the percent of bioavailable Fe also were detected among
species of amaranth, with a range of 6 to 12% of total Fe measured as bioavailable. Species
accumulating lower levels of total Fe tended to have a higher percentage of the total Fe in

bioavailable forms.

103

104

The relative difl‘erences in total Fe and the amount of bioavailable Fe among species
were maintained when grown in both environments. However, when grown in the
greenhouse, all lines contributed a much higher percentage of total Fe as bioavailable, and no
significant differences were detected for the percent bioavailable Fe. Field-grown plants
accumulated greater amounts of total Fe into compounds which were not solubilized during
the in vitro assay. These results indicated that field-screening of GLVs for improving Fe
nutritional quality is essential due to the much larger total Fe concentrations and potential
interactions of other plant compounds or Fe forms which accumulate in field-grown material.

If attempts are made to improve the Fe bioavailability fi'om amaranth using plant
breeding, it is important to understand how this trait changes with harvest time and leaf
selection, in order to standardize plant sampling. Two experiments were conducted to
investigate these changes (Chapter 2). An initial field examination of 18 lines harvested over
three days indicated that large changes in total Fe did not lead to similar changes in
bioavailable Fe. The amount of Fe mwsured as bioavailable remained relatively constant over
the three harvest dates. This stability in the dialyzable Fe fi'action also was observed in a
greenhouse study that focused on the changes in Fe bioavailability over the development of
one selected index leaf. However, the amount of Fe solubilized by the in vitro digestion
changed with leaf development. Maximal solubilization of leaf Fe occurred during the early
phases of leaf growth and expansion, followed by decreased levels as the leaf approached firll
expansion. Characterization of this soluble fraction may indicate the presence of forms of
Fe which could be enhanced through plant breeding.

The significant differences observed in Fe bioavailability among amaranth species were

based on estimates fi'om an in vitro assay. However, the variation detected may not have

105

been significant in an animal, due to other intrinsic and extrinsic factors afl‘ecting Fe
bioavailability. In order to verify the significance of the difi‘erences observed, three lines of
amaranth were fed to anemic rats, and Fe bioavailability was determined by measuring the
extent of hemoglobin regeneration. Use of the in vitro assay for initial screening of food
material, followed by slope-ratio analysis in rats has been recommended as a method of choice
in establishing food Fe bioavailability (Forbes et al., 1989). A. tricolor L. had lower levels
of percent of total Fe bioavailable to rats, when compared to either A. lopochondriacus L.
or A. cruentus L. However, when the same amount of amaranth was added to the diet, the
line containing the highest Fe concentration (A. tricolor L.) supported the largest gain in
hemoglobin. The efficiency of Fe absorption from all the lines was equivalent, suggesting
that any differences in other plant compounds had a minimal efl‘ect on Fe bioavailability.

These results confirmed the relative relationships observed in the in vitro assay and supported
the potential use of the in vitro assay in breeding programs focused on improving the Fe
nutritional quality of amaranth and other green leafy vegetables. In addition, selection for

high Fe-containing amaranth would be an eflicient means to enhance Fe content to diets.

Limitations to Approach

The underlying assumption in this research was that if the Fe bioavailability fiom a
green leafy vegetable could be improved via plant breeding, greater amounts of Fe would be
available to those who consume the improved food. However, the bioavailability of Fe to
humans is also dependent on interactions of Fe with other meal components (Bothwell et al.,
1979; Hallberg, 1981; Latunde-Dada and Neale, 1986; Clydesdale, 1988). In many of the

regions of the world where the population consumes significant amounts of amaranth, diets

106
are based predominately on cereals or dry beans (Oyejola and Bassir, 1975; Monsen, 1988a;

Thompson, 1988; Latunde Dada, 1991). The foods have been reported to decrease the Fe
bioavailability in meals (Hallberg, 1981; Walker, 1982; Latunde-Dada and Neale, 1986 ).
Once incorporated into a cereal-based diet, the availability of Fe from an improved GLV will
be prone to the diverse interactions of the meal. However, for the Fe-deficient individual,
the additional Fe in the diet would still be significant (Thompson, 1988).

This research focused only on difi‘erences in bioavailability of endogenous leaf Fe.
Other researchers have reported that certain unidentified compounds present in leaf material
may decrease Fe availability from other meal Fe sources (Kojima et al., 1981; Lee and
Clydesdale, 1981). Interactions which occur between endogenous leaf Fe and other plant
compounds were not explored in our study. Enhancement of the Fe bioavailability from
green leafy vegetables may not be important if other secondary plant compounds significantly

decrease total meal Fe bioavailability.

Future Research

In order to improve the eficiency of plant breeding efl‘orts, future research should
focus on elucidating the forms of Fe from GLVs which are highly available. Although
selecting for higher total Fe concentration would be fairly simple, higher total Fe
concentrations were observed to result in only small incremental increases in available Fe.
The relationship between the 'active' Fe fi'action identified in plant Fe biochemistry research
and bioavailable Fe needs to be explored. Highly available forms of Fe might be identified
from these types of investigations. In addition, the changing levels of soluble Fe observed

with leaf development also might provide clues to the identity and nature of potential forms

107
of Fe to be enhanced. Although higher soluble Fe does not equate with higher available Fe,

the higher levels of soluble Fe could be significant to the deficient individual (Thompson,
1988). Once identified, plant breeding efl‘orts then might focus on enhancing the
concentrations of these compounds.

In addition, more carefirl characterization of the irrteracfions of other plant compounds
on Fe bioavailability from GLVs is needed. Of particular interest would be endogenous
tannins and phosphorous compounds, both of which are present in leaf tissue and reported
to be negatively correlated with leaf and meal Fe bioavailability (Disler et al., 1975; Reddy
et al., 1993). The effect of enhanced citric or ascorbic acid on GLV Fe bioavailability also
might be explored, although there has been some question in the literature as to their
correlation with leaf Fe bioavailability (Reddy et al., 1993). Any investigations should focus
on the efi‘ects of changes in the levels of these other plant compounds on interactions with
both endogenous leaf Fe and meal Fe. However, changing the concentrations of the other
plant compounds may change the susceptibility of these green leafy vegetables to other insect
and disease organisms. The trade-ofl‘ in production losses with enhanced Fe bioavailability
must be weighed carefirlly.

Until further information is available on the chemical forms of bioavailable Fe in
GLVs, breeding efl‘orts should focus on enhancing the levels of total Fe within these plant
materials. The variability which was observed for total Fe concentration would facilitate
selection for improved lines. Lines from Amaranthus Iividus L. and Amaranthus tricolor L.,
two popular cultivated vegetable amaranth species, contained a wide range of Fe

concentrations, indicating some genetic variation. Enhancement of foliar Fe concentration

108
in desirable cultivars of these two species could provide very significant sources of dietary Fe

to populations suffering from nutritional Fe deficiency.

Conclusions

Genetic variability in Fe bioavailability was identified within the genus Amaranthus.
The variability in the percent of total Fe which was bioavailable was confirmed using both
in vitro simulation of gastrointestinal digestion and a hemoglobin regeneration assay with
anemic rats. The amount of bioavailable Fe in a leaf was observed to be relatively constant
over several harvest dates and in two environments. Total Fe concentrations did not
correlate well with dialyzable Fe, but lines accumulating a high total Fe tended to have higher
levels of available Fe. Much of the Fe from lines with high total Fe concentrations was
sequestered in insoluble, unavailable forms. The Fe which was solubilized by in vitro
digestion was observed to decrease with leaf expansion and maturity. Future research on Fe
bioavailability from GLVs should focus on the nature of these Fe compounds. Until
information is available on the chemical forms of bioavailable Fe in GLVs, breeding efforts

should focus on enhancing the levels of total Fe within these plant materials.

APPENDICES

109

APPENDIX A

Table 17. Recommended Daily Allowances of Iron"2

 

Category Age (years) or Condition Iron (mg)
Infant .0-.5 10
.5-1.0 10
Children 1-10 10
Males 1 1-18 12
19-50 10
50+ 10
Females 1 1-18 15
19-50 15
50+ 10
Pregnant 30
Lactating lst 6 months 15
2nd 6 months 15

 

 

lThe allowances are expressed as average daily intakes over time, and
are intended to provide for individual variations among most normal
persons living in the United States under usual environmental stress.
”Table was adapted from National Research Council. Recommendfl
DiMAllowancestood and Nutrition Board Subcommittee on the
thh Edition of the RDA's. (Washington DC: National Academy
Press, 1989).

 

110

APPENDIX B

Table 18. Total, dialyzable (pg-g“), and percent dialyzable Fe in leaves of A. tricolor
(tri) (Ames5113), A. lopochoncb'iacus (hyp) (Ames 2171) and A. cruentus (cm)
'I 433228 ;_4 the fiel different ears.

    

 

 

 

LeafFe Fraction Species Year 11 Year 2

 

 

 

Year 22
Total Fe tri 755 i 131 506 i 33 "
hyp 419: 112 170: 47 "
cm 495 :192 217 i 89 **
Average 557 i 202 298 i 165 **
LSD”, 228 7a
Dialyzable Fe tri 45 i 4 70 j; 9 **
hyp 45 :11 44 j; 5
cm 46 i 5 55 i ll *"‘
Average 45 _+_- 6 56 i 14 "
LSD_os -- 12
Percent Dialyzable Fe tri 6 j: 1 13 i 2 **
hyp 11: 2 27: 5 *"‘
cm 10: 5 27: 5 **
Average 9: 4 22 i 8 u
LSD.05 - 7

 

 

‘Year1=l990, Year 2 =1993.
2For each row, '* *' indicates significant differences (5% level) between Year 1 and

Year 2.

 

LIST OF REFERENCES

LIST OF REFERENCES

AbadiaJ. 1992. Leafresponsestoirondeficiency: Arevie w. J. PlantNutr. 15:1699-1714.

Ajayi, S.O., S.F. Oderinde and O. Osibanjo. 1980. Vitamin C losses in cooked fi'esh leafy
vegetables. Food Chem. 52243-247.

Amine, EK. and D.M. Hegstead. 1971. Effect of diet on iron absorption in iron-deficient
rats. J. Nutr. 101:927-936.

Andrews, S.C., P. Arosio, W. Bottke, J.F. Briat, M. von Darl, P.M. Harrison, F.P. Laulhere,
S. Levi, S. Lobreaux, S.J. Yewdall. 1992. Structure, firnction and evolution of
ferritins. J. Inorg. Biochem. 17:161-174.

Akpapunam, M. 1984. Effects of wilting, blanching and storage temperature on ascorbic
acid and total carotenoids of some Nrgerian fresh vegetables. Plant Foods Human
Nutr. 34: 177-180.

Albrecht, J.A, H.W. Schafer, E.A Zottola. 1991. Sulfydryl and ascorbic acid relations in
selected fruits and vegetables. J. Food Sci. 56: 427-431.

Bear, F., S.J. Toth, and AL. Prince. 1948. Mineral composition of vegetables. Soil Sci. Soc.
Amer. Proc. 13:380-395.

Berner, LA. and DD. Miller. 1985. Effects of dietary protein on iron bioavailability: A
review. Food Chem. 18:47-69.

Bienfait, F. 1987. Ch 18. "Biochemical Basis of Iron Efiiciency Reactions in Plants." In
Iron Trmsmrt in Migrgms, Plants and Animals, eds. G. Winkelrnann, D. van der
Helm, and 1B. Neilands. (New York: VCH Publishers) p.339-349.

Bienfait, F. 1989. Prevention of stress in iron metabolism of plants. Acta Bot. Neerl.
38:105-129.

111

112

Bloem, MW., M. Wedel, E.J. Van Agtrnaal, AJ. Speck, S. Saowakontha, W.H.P. Schreurs.
1990. Vitamin A intervention: Short term effects of a single oral massive dose on
iron metabolism. Amer. J. Clin. Nutr. 51:76-79.

Bothwell, T.H., RW. Charlton, J.D. Cook, C.A Finch. 1979. Ch. 12. ”Iron Absorption.”
In Mn Metabolism in Man. (OxfordzBlackwll Scientific Publications) p. 576.

Brune, M, L. Rossander and L. Hallberg. 1989. Iron absorption and phenolic compounds:
Importance of difl‘erent phenolic structures. Euro. J. Clin. Nutr. 43:547-558.

Chawla, S., A. Saxena and S. Seshadri. 1988. In vitro availability of iron in various green
leafy vegetables. J. Sci. Food Agr. 46:125-127.

Chewya, J .A 1985. Identification and nutritional importance of indigenous green leafy
vegetables in Kenya Acta Hort. 153:99-108.

Chidambaram, M.V., M.B. Reddy, J .L. Thompson and G.W. Bates. 1989. In vitro studies
of iron bioavailability: Probing the concentration and oxidation-reduction reactivity
of pinto bean iron with ferrous chromogens. Biol. Trace Element Res. 19:25-40.

Clydesdale, FM 1988. "Mineral Interactions in Foods.” In Em hams-3103s, eds. C.E.
Bodwell and J.W. Erdman Jr. (New York: Marcel Drekker). pg. 73-113.

Clydesdale, F .M., Chi-Tang Ho, C.Y. Lee, NJ. Mendy, RL. Shewfelt. 1991. The efl'ects
of postharvest treatment and chemical interactions on the bioavailability of ascorbic
acid, thiarnin, vitamin A, carotenoids and minerals. Crit. Rev. Food Sci. Nutr.
30:599-638.

Consul, JR and K. Lee. 1983. Extrinsic tagging in iron bioavailability research: A critical
review. J. Agr. Food Chem. 31:684-389.

Cook, JD. 1983. Determinants of nonheme iron absorption in man. Food Tech. Oct.1124-
126.

Cook, J.D. and ER Monsen. 1976. Food iron absorption in human subjects: 111.
Comparison of the effect of animal proteins on nonheme iron absorption. Amer. J.
Clin. Nutr. 29:859-867.

DeKock, P.C, K. Commisiong, V.C. Farmer, RHE. Inkson. 1960. Interrelationships of
catalase, perosidase, hematin and chlorophyll. Plant Phys. 35:737-740.

Deutsch, J.A 1978. Genetic variation of yield and nutritional value in several Amanmtlms
species used as a leafy vegetable. Diss. Abstr. Intl. B. 38:3969-B.

113

Devadas, RP. and S. Saroja. 1980. Availability of iron and B-carotene fiom amaranth to
children Proc. of Second Anmral Amaranth Conference. (Emmaus, PAzRodale Press
Inc.) pp. 15-21.

Disler, P.B, S.R Lynch, R.W. Charlton, J.D. Torrance, T.H. Bothwell, RB. Walker and F.
Mayet. 1975. The efi‘ect of tea on iron absorption. Gut. 16: 193.

Duhaiman, AS. 1988. Total iron content and bioavailability from liver, meat and vegetables.
Nutr. Rep. Intl. 37:645-651.

El-Sherifi AR, AZ. Osman, M.K. Sadik, and SM. Shata. 1984. Determination of ferrous
and ferric iron ratio in spinach plants and their relation to iron application. J. Plant
Nutr. 7:767-776.

Fairweather-Tait, 8., Z. Piper, S. Fatemi, G. Moore. 1991. The effect of tea on iron and
aluminum metabolism in the rat. Brit. J. Nutr. 65:61-68.

Forbes, A., C.E. Adams, M.J. Amaud, C.O. Chichester, J.D. Cook, B.M. Harrison, RF.
Hurrell, S.G. Kahn, ER Morris, J .T. Tanner, and P. Whittaker. 1989. Comparison
of invitro, animal and clinical determinations of iron bioavailability: International
Nutritional Anemia Consultative Group Task Force report on iron availability. Amer.
J. Clin. Nutr. 49:225-238.

Fritz, J. C., G.W. Pla, T. Roberts, J. W. Boehne and E. L. Hove. 1970. Biological
availability in animals of iron fiom common dietary sources. J. Agric. Food Chem.
18:647-651.

Gillooly, M., T.H. Bothwell, J.D. Torrance, AP. MacPhail, D.P. Derrnan, W.R Bezwada,
W. Mills, and RW. Charlton. 1983. The efi‘ects of organic acids, phytates and
polyphenols on the absorption of iron fiom vegetables. Brit. J. Nutr. 49:331-342.

Gordon, D.T. and L. Chao. 1984. Relationship of components in wheat bran and spinach to
iron bioavailability in anemic rat. J. Nutr. 114:526-535.

Grubben, G.J.H. 1976. Cultivation of amaranth as a tropical leaf vegetable. Koninklijk
Institute voor de Tropen. Comm. 67 :34-41.

Grunes, D.L. and W.H. Allaway. 1985. Ch 17. ”Nutritional Quality of Plants in Relation to

FertilizerUse.” InF il' rT hn l . 3rd ed. (Madison: Soil Sci. Soc.
Amer). pp. 589-619.

Guerinot, ML. and Y. Yi. 1994. Iron: Nutritious, noxious and not readily available. Plant.
Phys. 104:815-820.

114

Gupta, U. 1992. Characterization of the iron status in plant parts and its relation to soil pH
on acid soils. J. Plant Nutr. 15: 1521-1540.

Hallberg, L. 1974. The pool concept of food iron absorption and some of its implications.
Proc. Nutr. Soc. 33:285-291.

Hallberg, L. 1980. ”Food Iron Absorption.” In M h inH 1 ed. J.D. Cook.
(London: Churchill Inc.).

Hallberg, L. 1981. Bioavailability of dietary iron in man. Annu. Rev. Nutr. 1:123-147.

Hallberg, L. and L. Rossander. 1982. Effect of soy protein on nonheme iron absorption in
man. Amer. J. Clin. Nutr. 36:514-520.

Hazell, T. and LT. Johnson. 1987a. Efi‘ects of food processing and fiuit juices on in vitro
estimated iron availability fi'om cereals, fiuits and vegetables. J. Sci. Food Agr.
38:73-82.

Hazell, T. and LT. Johnson. 1987b. In vitro estimation of iron availability fi'om a range of
plant foods; influence of phytate, ascorbate and citrate. Brit. J. of Nutr. 57:223-233.

Hewitt, E.J. 1983. ”Essential and Functional Metals in Plants." In W

Mi n ri nt ' if i 11 Pl eds. D.A Rabb,W.S. Pierpoint.
(New York: Academic Press). pp. 277-323.

Howard, L., M. Buchowski, B. Wang, and DD. Miller. 1993. Bioavailability of electrolytic
iron in fortified infant cereal determined by hemoglobin repletion in piglets. Nutr.
Res. 13:287-295.

Hunt, SM. and IL. Grofl‘. 1990. WWW
(Minneapolis: West Publishing Co.) p. 517.

Ifon, ET. and O. Bassir. 1978. The efiiciency of utilizing the iron in leafy green vegetables
for hemoglobin synthesis in anemic rats. Nutr. Rep. Intl. 18: 481-486.

Kabata-Pendias, A and H Pendias. 1992. El i Soil P1 . (Ann Arbor:
CRC Press). p. 265.

Kapsokefalou, M. and DD. Miller. 1991. Efi‘ect of meat and selected food components on
the valence of nonheme iron during in vitro digestion J. Food Sci. 56:352-355, 358.

Kelly, J .F . and BB. Rhodes. 197 5. The potential for improving the nutrient composition of
horticultural crops. Food Tech. May: 134-140.

115

Kojima, N., D. Wallace and G.W. Bates. 1981. The effect of chemical agents, beverages and
spinach on the in vitro solubilization of iron from cooked pinto beans. Amer. J. Clin.
Nutr. 34:1392-1401.

Latunde-Dada, GO. 1990. Effect of processing on iron levels in and availability from some
Nigerian vegetables. J. Sci. Food Agr. 53:355-361.

Latunde-Dada, GO. 1991. Some physical properties of ten soybean varieties and effects of
processing of iron levels and availability. Food Chem. 42:89-98.

Latunde-Dada, GO. and RJ. Neale. 1986. Review: Availability of iron fiom foods. J. Food
Tech. 21:255-268.

Laulhere, JP. and JP. Briat. 1993. Iron release and uptake by plant ferritin, effects of pH,
reduction and chelation. Biochem. J. 290:693-699.

Layrisse, M., J.D. Cook, C. Martinez, M. Roche, I.N. Kuhn, RB. Walker and CA Finch.
1969. Food iron absorption: A comparison of vegetable and animal foods. Blood.
33:430-443.

Lee, K and FM. Clydesdale. 1980. Chemical changes of iron in food and drying processes.
J. Food Sci. 45:711-715.

Lonnerdal, B. 1988. Ch. 6. "Vitamin-Mineral Interactions.” MW eds.
C.E. Bodwell and J .W. Erdman. (Marcel-Dreker, Inc: New York). pp. 163-186.

Lynch, SR and AM Covell. 1987 . Iron in soybean flour is bound to phytoferritin. Amer.
J. Clin. Nutr. 45:866.

McMillan, T]. and FA Johnston. 1951. The absorption of iron from spinach by six young
women, and the effect of beef upon the absorption. J. Nutr. 44:383-398.

van der Mark, F.M.L., van der Briel, J .W.AM. van Oers, and HF. Bienfait. 1982. Ferritin
in bean leaves with constant and changing Fe status. Planta. 156:341-344.

Martinez-Torres, C., I. Leets, P. Taylor, J. Ramirez, M. de Valle Camacho, M. Layrisse.
1986. Heme, ferritin and vegetable iron absorption in humans from meals denatured
of heme iron during the cooking of beef. J. Nutr. 116: 1720-1725.

Marschner, H 1987. Minml Numg' 'gn ngighet Plants. (New York: Academic Press). p.
674.

MengeL K. 1979. Influence of exogenous factors on the quality and chemical composition
of vegetables. Acta Hort. 93:133-151.

116

Mengel, K. and EA Kirkby. 1979. Prineiples efPlgrt Nam 'en. (Berne, Switzerland:
International Potash Institute). pp. 425-439.

Miller, DD. and B.R Schricker. 1981. In vitro method for estimation of iron availability
fi'om meals. Amer. J. Clin. Nutr. 34:2248-2256.

Miller, DD. and RR Schricker. 1982. ”In Vitro Estimation of Food Iron Bioavailability.”
N riti n Bi v '1 ili f Iro ed. C. Kies. (New York: American Chem.
Society). pp. 11-26.

Miller, J. 1987. Bioavailable iron in raw and cooked spinach and broccoli. Nutr. Rep. Intl.
36:435-440.

Monsen, E.R. 1988a. Iron nutrition and absorptionzDietary factors which impact iron
bioavailabilility. J. Amer. Diet. Assoc. 88:786-790.

Monsen, ER 1988b. "Protein Iron interactions: Influence on Absorption, Metabolism and
Status." In W eds. C.E. Bodwell and J .W. Erdman. (New York:
Marcel-Dreker, Inc). pp. 149-162.

Monsen, ER and JD. Cook. 1976. Food iron absorption in human subjects IV: Effect of
calcium and phosphate salts on the absorption of nonheme iron. Amer. J. Clin. Nutr.
29: 1 142.

Monsen, ER and JD. Cook. 1979. Food iron absorption in human subjects. V. Efl‘ects of
the major dietary constituents of a semi synthetic meal. Amer. J. Clin. Nutr. 32:804-
808.

Moore, C.V. and R Dubach. 1951. Observations on the absorption of iron fi'om foods
tagged with radioiron. Trans. Assoc. Amer. Physicians. 64:245-256.

Moore, CV. and R Dubach. 1956. Metabolism and requirements of iron in the human. J.
Amer. Med. Assoc. 162:197-204.

National Research Council. 1989. Reeemmengg Dietary Allevgaeees; PM gd Nutritien
mmi n 1 ° ° 11 f h RDA' . (Washington D.C.:National
Academy Press). pp 195-205.

Neilands, J. B. Konopka, B. Schwyn, M. Coy, RT. Francis, B.H. Paw, A Bag. 1987. Ch
1. " Comparative Biochemistry of Microbial Iron Assimilation. " in Iren Tmtt in

Mierebes, Plants agd Animfls, eds. G. kaelmann, D. van der Helm, and JR
Neilands. (New York: VCH Publishers) p.532.

Omueti, O. 1982. Efl‘ects of age on the elemental nutrients of celosia cultivar. Exp. Agr.
18:89-92.

117

Onayemi, O. and G.I.O. Badifir. 1987. Efl‘ect of blanching and drying methods on the
nutritional and sensory quality of leafy vegetables. Plant Foods Human Nutr. 37:291-
298.

Oyejola, AO. and O. Bassir. 1975. Physiological availability of the iron content of some
Nigerian leafy vegetables. Plant Foods for Man. 12177-183.

Pace, RD. and C. Bonsi. 1987. Iron content and iron availability of sweet potato tips
collected at different times during the season. Nutr. Reps. Intl. 35(6):1151-1156.

Pace, RD., T.E. Sibiya, B.R Phills, and G.G. Dull. 1985. Calcium, iron and zinc content
of 'Jewel' sweet potato greens as afl‘ected by harvesting practices. J. Food Sci.
50:940-941.

Pirie, NW. 1985. Comments on the importance of green vegetables. Plant Foods for
Human Nutr. 35:73-80.

Prakash, D. and M Pal. 1991. Nutritional and antinutritional composition of vegetable and
grain amaranth leaves. J. Sci. Food Agr. 57:573-583.

Prasad, AN. and C. Prasad. 1991. Iron deficiency: Nonhematological measurements. Prog.
Food Nutr Sci. 15:255-283.

Pye, OR and G. MacLeod. 1946. The utilization of iron fiom different foods by normal
young rats. J. Nutr. 322677-687.

Quarterman, J. 1973. Factors which influence the amount and availability of trace elements
in human food plants. Plant Foods Human Nutr. 1(3): 171-190.

Rao, N. and T. Prabhavathi. 197 8. An in vitro method for predicting the bioavailability of
iron fi'om foods. Amer. J. Clin. Nutr. 312169-175.

Reddy, M., M.V. Chidambaram, J. Fonesca, G.W. Bates. 1986. Potential role of in vitro
iron bioavailability studies in combatting iron deficiency: A study of the effects of
phosphovitin on iron mobilization from pinto beans. Clin. Physiol. Biochem. 4:78-86.

Reddy, M. and J .D. Cook. 1991. Assessment of dietary determinants of nonheme iron
absorption in humans and rats. Amer. J. Clin. Nutr. 54:723-728.

Reddy, NS. and KS. Kulkarni. 1986. Availability of iron fiom some uncommon edible
green leafy vegetables determined by in vitro method. Nutr. Reps. Intl. 34(5):859-
861.

Reddy, S. and V.G. Malewar. 1992. Bioavailabilitiy of iron fi'om spinach cultivated in soil
fortified with graded levels of iron. Plant Foods Human Nutr. 42:313-318.

118

Reddy, S., C.V. Sondge, and TN. Khan. 1993. In vitro bioavailability of iron fi'om spinach
(.Soirracea oleracea L.) cultivated in soil fortified with graded levels of iron and zinc.
Plant Foods Human Nutr. 442241-247 .

Reinhold, J .G., J .S. Garcia, and P. Garzon. 1981. Binding of iron by fiber of wheat and
maize. Amer. J. Clin. Nutr. 34:1384-1391.

Romheld, V. 1987. Ch. 19. ”Existence of Two Difi‘erent Strategies for the Acquisition of

Iron In Higher PlantS" In W eds G
W'mkelmann, D. van der Helm, and J. B. Neilands. (New York: VCH Publishers)

pp.353-374.

Romheld, V. and H. Marschner. 1986. Mobilization of iron in the rhizoshpere of different
plant species. Adv. Plant Nutr. 22155-204.

Rossander- Hulthen, L., A Gleerup, and L. Hallberg. 1990. Inhibitory efi‘ect of oat products
on non-heme iron absorption in man. Euro. J. Clin. Nutr. 44:783-791.

Schmidt, DR 1971. Comparison of yield and composition of eight tropical leafy vegetables
grown at two soil fertility levels. Agr. J. 63:546-550.

Schricker, B.R, D.D. Miller, R Rasmussen and D.Van Campen. 1981. A comparison of in
vivo and in vitro methods for determining availability of iron from meals. Amer. J.
Clin. Nutr. 34:2257-2263.

Scrimshaw, NS. 1991. Iron deficiency. Sci. Amer. 265(10):46-52.

Seckback, J. 1968. Studies on the deposition of plant ferritin as influenced by supply to iron
deficient beans. J. Ultrastr. Res. 222413-423.

Seckback, J. 1982. Ferrating out the secrets of plant ferritin: A review. J. Plant Nutr.
52369-396.

Simpson, K.M., ER Morris, and JD. Cook. 1981. The inhibitory effect of bran on iron
absorption in man. Amer. J. Clin. Nutr. 3421469.

Slatkavitz, C. and FM. Clydesdale. 1988. Solubility of inorganic iron as afl‘ected by
proteolytic digestion. Amer. J. Clin. Nutr. 1988. 472487-95.

Smith, RN. 1984. Iron in higher plants: Storage and metabolic role. J. Plant Nutr. 7 :7 59-
766.

Smith, I.F. 1982. Leafy vegetables as sources of minerals in southern Nigerian Diets. Nutr.
Rep. Intl. 26(4):679-688.

119

Smith, KT. 1983. Effects of the chemical environment on iron bioavailability measurements.
Food Tech. 10:115-120.

Smith, MC. and L. Otis. 1937. Hemoglobin regeneration in anemic rats in relations to iron
intake. J. Nutr. 132573-582.

Solomons, NW. 1986. Competitive interaction of iron and zinc in the diet: Consequences
for human nutrition J. Nutr. 116:927-935.

Stafl‘ord, W.L., J .S. Mugerwa, R Bwabye. 1976. Efl‘ects of methods of cooking, application
of nitrogen fertilizer and maturity on certain nutrients in the leaves of Amataathus
Midas subspecies hybridus (Green Head). Plant Foods for Man. 227-13.

Suzuki, T., F.M. Clydesdale, T. Pandolf. 1992. Solubility of iron in model systems
containing organic acids and lignin. J. Food Prot. 55(1 1)2893-898.

Tarnpus, G. 1993. "Nutrition problem is next target." Mm’la Chrenicle. 9-16-1993.
Terry, N. and J. Abadia. 1986. Function of iron in chloroplasts. J. P1. Nutr. 9:609-646.

Terry, N. and G. Low. 1982. Leaf chlorophyll content and its relation to the intracellular
location of iron. J. P1. Nutr. 52301-310.

Tirnperley, M H, RR Brooks and P. J. Peterson. 1973. The distribution of nickel, copper,
zinc and iron in tree leaves. J. Exp. Bot. 24(82):889-895.

Tinker, PB. 1981. Levels, distribution and forms of trace elements in food plants. Phil.
Trans. Royal Soc. Lond. B. 294241-55.

Van Campen, D. and E. Gross. 1969. Effect of histidine and certain other amino acids on
absorption of iron” by rats. J. Nutr. 99:68-74.

Van Campen, D. and RM. Welch. 1980. Availability to rats of iron from spinach: Efl‘ects
of oxalic acid. J. Nutr. 110(2):]618-1620.

Van Campen, D. 1983. Iron bioavailability techniques: An overview. Food Tech. 10: 127-
132.

Walker, A F. 1982. Physiological effects of legumes in the human diet: A review. J. Plant
Foods. 425-14.

Wapnir, RA. 1990. Ch. 6. "Nutritional Factors, Proteins and the Absorption of Iron and

Cobalt.” In Pretm’ Ntan_t1' 'en and Mineral Abgzrptien. (Boca Raton: CRC Press) pp.
99-129.

120

Welch, RM. and D. Van Campen. 1975. Iron availability to rats fiom soybeans. J. Nutr.
105:253-256.

Welch, RM. and W.H. House. 1982. Availability to rats of zinc fi'om soybean seeds as
affected by maturity of seeds, sourse of dietary protein and soluble phytate. J. Nutr.
112:879-885.

Welch, RM. and W.H. House. 1984. Ch. 3. ”Factors Affecting the Bioavailability of

Mineral Nutrients in Plant Foods.” In Craps as Sources otiNutrifls fer Hams,
eds. RM Welch and W.H. Gabelrnan. Proceedings of a Symposium. Nov 28.-Dec

3, 1982. Agronomy Society of America Special Publication #48.

Welch, RM and W.H. House, and D. R Van Campen. 1977. Effects of oxalic acid on the
availability of zinc from spinach leaves and zinc sulfate supplied to rats. J. Nutr.
107:929-933.

West, TS. 1981. Soil as the source of trace elements. Phil. Trans. R Soc. Lond. B.
294219-39.

Wien, EM, DR Van Campen, J.M Rivers. 1975. Factors effecting the concentration and
bioavailability of iron in turnip greens to rats. J. Nutr. 1052459-466.

Williams, RJ.P. 1990. Ch. 1. ”An Introduction to the Nature of Iron Transport and

Storage.” In W eds. P. Ponk, H.M. Schulrnan, RC.
Woodsworth. (Boca Raton: CRC Press). pp 1-16.

Zhang, D., D. G. Hendricks, A W. Mahoney, D. P. Comforth. 1985. Bioavailability of iron
in green peas, spinach, bran cereal and cornmeal fed to anemic rats. J. Food Sci.
50:426-428.

Zhang, D., D. G. Hendricks, A W. Mahoney, D. P. Comforth. 1989. Bioavailability of total
iron from meat spinach and meat spinach mixtures by anemic and nonanemic rats.
Brit. J. Nutr. 61:331-343.

HICHIGQN STRTE UNIV. LIBRRRIES
llllflllll 11111111111 11111111111me
31293014201663