The human Type 5, tartrate-resistant acid phosphatase


Material Information

The human Type 5, tartrate-resistant acid phosphatase purification, characterization and molecular cloning
Physical Description:
xiv, 165 leaves : ill. ; 29 cm.
Ketcham, Catherine Mary, 1959-
Publication Date:


Subjects / Keywords:
Acid Phosphatase -- isolation & purification   ( mesh )
Isoenzymes -- biosynthesis   ( mesh )
Biochemistry and Molecular Biology thesis Ph.D   ( mesh )
Dissertations, Academic -- Biochemistry and Molecular Biology -- UF   ( mesh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph.D.)--University of Florida, 1988.
Bibliography: leaves 150-164.
Statement of Responsibility:
by Catherine Mary Ketcham.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001031877
oclc - 20429296
notis - AFB4055
System ID:

Full Text








My sincere thanks are extended to the members of my supervisory

committee for their guidance and encouragement: Dr. R. Michael Roberts,

Dr. Harry S. Nick, Dr. Daniel L. Purich, Dr. Peter M. McGuire, Dr.

Michael S. Kilberg, and Dr. Fuller W. Bazer. I especially thank Dr. R.

Michael Roberts for convincing me that I could earn a Ph.D., and for his

continued support over all of the years that we worked together. I am

indebted to Dr. Harry S. Nick for allowing me to be a part of his

laboratory. I would like to express my gratitude to Dr. Richard P.

Boyce, Graduate Coordinator--a graduate student's best friend.

Warm thanks are extended to the many helpful scientists from

laboratories of Dr. Roberts and Dr. Bazer, especially Dr. George A.

Baumbach, Dr. William C. Buhi, Dr. Phillipa T.K. Saunders and Dr. Mary

K. Murray. I am also indebted to my fellow graduate students in the

laboratory of Dr. Nick. A number of undergraduate assistants and

laboratory technicians have been helpful to me. I would specifically

like to express my appreciation to Joy Whitman, Samantha Roberts,

Penelope Edwards and Carol Ann Harper.

I am grateful to Dr. Raul Braylan for providing the human spleens;

and to Dr. Rosalia C. M. Simmen for the generous gift of two porcine

uteroferrin clones. I sincerely thank Liz Dampier, who cheerfully typed

every word of this dissertation.

A special word of thanks goes to my parents, Alfred and Antoinette

Scopa, and to my husband's parents, Sigurd and Valerie Martin. Their

loyalty and approval have always been very important to me.

I wish to express my greatest appreciation to my dear husband,

Glenn. His patience and unfailing love have been the keys to my success.



ACKNOWLEDGMENTS ................................................. ii

LIST OF TABLES ................................................ vii

LIST OF FIGURES .................................................. ix

ABBREVIATIONS .................................................... xi

ABSTRACT ...................... ................................ .xiii


1 INTRODUCTION ............................................... 1

The Tartrate-resistant Acid Phosphatases .................. 1
Uteroferrin ...................................... ........ .. 1
Uteroferrin is Synthesized by the Uterine
Endometrium Under the Influence of Progesterone ..... 2
The Purification of Uteroferrin from Uterine
Secretions .......................................... 3
Uteroferrin is an Iron-containing Acid
Phosphatase ......................................... 5
Uteroferrin Contains Two Iron-binding Sites ........... 6
A Proposed Reaction Mechanism for Uteroferrin ......... 8
Uteroferrin Contains at Least One High
Mannose Oligosaccharide Chain ....................... 9
The Function of Uteroferrin .......................... 11
Other Purple, Iron-containing Acid Phosphatases ............ 12
Human Acid Phosphatases .................................... 15
Characterization of the Human Acid Phosphatases ....... 15
Hairy Cell Leukemia ................................... 17
Tartrate-resistant Acid Phosphatase Levels in
Other Diseases ...................................... 18
Purification and Characterization of the
Type 5 Isozyme ...................................... 19

WITH PORCINE UTEROFERRIN ................................... 22

Introduction ............................................... 22

Materials and Methods ......................................
Materials .............................................
Methods ...............................................
Results ....................................................
Purification of Uteroferrin .............................
Purification of the Human Type 5 Acid Phosphatase
from Hairy Cell Spleen ................................
Purification of the Type 5 Isozyme from Normal
Human Spleen ..........................................
Purification of the Type 5 Phosphatase from
Human Placenta ........................................
Purification of the Bovine Spleen, Bovine Uterine
and Rat Spleen Phosphatases ...........................
pH Optima of the Type 5 Phosphatases ....................
Glycoprotein Nature of the Type 5 Phosphatases ..........
Iron Content of the Type 5 Phosphatases .................
Activation by Reducing Agents ...........................
Effects of Inhibitors on Phosphatase Activity ..........
Substrate Specificity for Phosphatase Activity ..........
Immunological Cross-reactivity ..........................
The Production of Monoclonal Antibodies against
Porcine Uteroferrin ...................................
The Production of Monoclonal Antibodies against
the Hairy Cell Spleen Enzyme ..........................
The Binding of the Monoclonal Antibodies to Other
Tartrate-resistant Acid Phosphatases ..................
Inhibition of Enzymatic Activity by the Binding
of Monoclonal Antibodies ..............................
Discussion .................................................

IN LEUKEMIA CELLS ..........................................

Introduction ...............................................
Materials and Methods ......................................
Materials ...............................................
Methods .................................................
Results ....................................................
The Identity of the Positive Clones Obtained by
Immunoscreening the Human Placenta cDNA Library .......
Dot Blot Analysis of Uteroferrin and Fibronectin
with Anti-Uteroferrin .................................
Results from Screening the Mouse Spleen
cDNA Library ..........................................
Results from Screening the Porcine Uterine
Endometrium cDNA Library ..............................
The Molecular Cloning of a cDNA Coding for
the Human Type 5, Tartrate-resistant
Acid Phosphatase ......................................
cDNA Sequences ..........................................
Characteristics of the Deduced Amino Acid Sequence ......







Northern Analysis ....................................... 128
Induction of Tartrate-resistant Acid Phosphatase
by TPA ................................................ 133
The Effects of Hemin on Acid Phosphatase Expression ..... 135
Discussion ................................................. 135

4 CONCLUSIONS AND FUTURE DIRECTIONS .......................... 143

REFERENCES ..................................................... 150

BIOGRAPHICAL SKETCH ............................................ 165

Table page

2-1 Tartrate insensitivity and activation by
2-mercaptoethanol of acid phosphatase activity
in spleen homogenates from a patient with hairy
cell leukemia and from an individual with a
normal spleen ........................................... 41

2-2 Purification of tartrate-resistant acid
phosphatase from spleen of a patient with
hairy cell leukemia ..................................... 46

2-3 Levels of tartrate-resistant acid phosphatase
activity in various tissues ............................. 52

2-4 pH optima for the tartrate-resistant acid phosphatases .. 53

2-5 The iron content of the tartrate-resistant
acid phosphatases ....................................... 55

2-6 Activation of purified human hairy cell phosphatase
by 2-mercaptoethanol .................................... 57

2-7 Action of various potential inhibitors on the
hairy cell acid phosphatase, uteroferrin, bovine spleen
and bovine uterine acid phosphatases .................... 58

2-8 Comparison of substrate specificities of hairy cell
phosphatase, uteroferrin, bovine spleen and
bovine uterine phosphatases .............................. 62

2-9 Dissociation constants (KD values) for binding of
monoclonal antibodies to uteroferrin, the human spleen
enzyme, the rat spleen enzyme, the bovine spleen and
uterine enzymes .......................................... 73

2-10 Inhibition of acid phosphatase activity of uteroferrin by
monoclonal antibodies .................................... 75

3-1 The clones employed for generation of the complete
sequence of the human tartrate-resistant
acid phosphatase .........................................124

3-2 Tartrate-resistant acid phosphatase levels in K562 and
JURKAT cells maintained on 10-8M TPA and K562 cells
maintained on 60pM hemin ................................. 134

3-3 Potential epitopes common to porcine uteroferrin and
human fibronectin ........................................ 137


Figure page

2-1 Purification of uteroferrin from allantoic fluid
of a day 67 pregnant pig ................................. 39

2-2 Purification of human spleen phosphatase ................. 44

2-3 SDS-polyacrylamide gel electrophoresis of purified
uteroferrin and hairy cell phosphatase ................... 47

2-4 Polyacrylamide gel electrophoresis of acid
phosphatases at pH 5.4 in f-alanine buffer ............... 48

2-5 SDS-polyacrylamide gel electrophoresis of uteroferrin,
the phosphatase from human spleen and the phosphatase from
placenta ................................................. 51

2-6 Lineweaver-Burk plots of uteroferrin and hairy cell
phosphatase activities in the presence of 2-mercap-
toethanol and various inhibitors ........................ 61

2-7 Solid phase radiobinding assay of whole mouse
antiserum from mice immunized with human spleen
phosphatase (upper panel) or uteroferrin (lower panel) ... 65

2-8 Competitive binding of anti-uteroferrin monoclonal
antibodies to uteroferrin adsorbed to flexvinyl
microtiter wells ........................................ 69

2-9 Binding of four monoclonal antibodies to uteroferrin
and to the immunoaffinity purified phosphatases
from human, bovine and rat spleens ...................... 72

3-1 The nucleotide sequences of the cDNA clones coding for
porcine uteroferrin used for screening the human placenta
cDNA library and their deduced amino acid sequences ...... 90

3-2 The sequences of the redundant oligonucleotides used for
Southern blot analysis and their deduced amino
acid sequences ........................................... 96

3-3 Nitrocellulose filters containing immunoreactive fusion
proteins produced by recombinant lambda gtll phage from
a human placenta cDNA library which was screened with
anti-uteroferrin antibodies .............................106

3-4 Agarose gel electrophoresis of DNA from six positive
clones obtained by immunoscreening human placenta and
mouse spleen cDNA libraries and Southern analysis of
the clones probed with HP6.1. ............................107

3-5 Southern blot analysis of clone HP 6.1 probed with
radiolabeled oligonucleotide KM28 ........................109

3-6 Polyacrylamide gel electrophoresis of clone HP 6.1
restriction fragments obtained with Alu I and Hae III ....110

3-7 The nucleotide and inferred amino acid sequences of
clone HP 6.1, which codes for human fibronectin...........113

3-8 Dot blot analysis of the binding of polyclonal and
monoclonal antibodies against porcine uteroferrin to
uteroferrin and human fibronectin ........................116

3-9 Two positive clones, 2a and 6a, identified by screening
the human placenta cDNA library with the cDNAs coding for
porcine uteroferrin ......................................119

3-10 Agarose gel electrophoresis of EcoR I digests DNA from
two positive clones, 2a and 6a ...........................120

3-11 Polyacrylamide gel electrophoresis of the restriction
fragments obtained by the digestion of clones 2a and 6a
with Alu I, Hae III and Msp I .......................... 122

3-12 Subcloning and sequencing strategy for clone 6a, the
cDNA encoding the human Type 5, tartrate-resistant acid
phosphatase ..............................................123

3-13 The nucleotide sequence of the cDNA clone 6a and the
deduced amino acid sequence of the human Type 5,
tartrate-resistant acid phosphatase ......................127

3-14 Comparison of the deduced amino acid sequence of the
human tartrate-resistant acid phosphatase with the
amino acid sequences of porcine uteroferrin and the
bovine spleen acid phosphatase............................130

3-15 Expression of tartrate-resistant acid phosphatase mRNA
and its induction by TPA ..................................132


bp, base pairs
BSA, bovine serum albumin
cDNA, complementary DNA
CM-cellulose, carboxymethyl cellulose
ConA, Concanavalin A
DME, Dulbecco's Modified Eagle's Medium
DME-HAT, DME with hypoxanthine, aminopterin and thymidine
DNase, deoxyribonuclease
DTT, dithiothreitol
EBV, Epstein-Barr Virus
EDTA, ethylenediaminetetraacetic acid
GlcNAc, N-acetylglucosamine
GTC, guanidinium thiocyanate solution
IgG, immunoglobulin G
i.p., intraperitoneally
IPTG, isopropylthio-f-D-galactopyranoside
kb, kilobases

KD, dissociation constant
MOPS, 3-(N-morpholino)propanesulfonic acid
Mr, relative molecular mass
PAGE, polyacrylamide gel electrophoresis
PBS, phosphate buffered saline
Phosphodiesterase, N-acetylglucosamine-l-phosphodiester a-N-acetyl-
Phosphotransferase, UDP-N-acetylglucosamine:lysosomal enzyme
N-acetylglucosamine 1-phosphate transferase
PMSF, phenylmethylsulfonylfluoride
pNPP, p-nitrophenylphosphate
RNase, ribonuclease
SDS, sodium dodecyl sulfate

SDS-PAGE, polyacrylamide gel electrophoresis with sodium dodecyl sulfate
SP2/0, SP2/0-Agl4
SSC, standard sodium citrate buffer
TE, Tris-HCl with EDTA
TPA, 12-0-tetradecanoylphorbol 13-acetate
TR-AP, tartrate resistant acid phosphatase (Type 5 isozyme)
Tris, tris(hydroxymethyl)aminomethane
Uf, uteroferrin
X-gal, 5-chloro-4-bromo-3-indolyl f-D-galactopyranoside

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



Catherine Mary Ketcham

August, 1988

Chairman: Dr. R. Michael Roberts
Major Department: Biochemistry and Molecular Biology

The spleens of patients with hairy cell leukemia contain high levels

of a tartrate-resistant cationic acid phosphatase, known as the Type 5

isozyme. This enzyme is also present in small quantities in normal

spleen and term placenta. This phosphatase has been purified from these

three sources by a procedure which involves only two chromatographic

steps: CM-cellulose chromatography and immunoaffinity chromatography on

a column of anti-uteroferrin antibodies. Uteroferrin is an abundant

purple, iron-containing acid phosphatase that can be easily purified

from porcine uterine secretions. Like uteroferrin, the human enzyme is

an iron-containing glycoprotein of apparent molecular weight 34,000.

The human phosphatase and uteroferrin also resemble each other closely

in electrophoretic mobility, substrate specificity, and response to a

variety of activators and inhibitors. Three anti-uteroferrin monoclonal

antibodies which bind with high affinities to distinct sites on the

uteroferrin molecule also recognize the human spleen enzyme, but bind to


it with much lower affinities. These antibodies also recognize purple,

iron-containing acid phosphatases from bovine and rat spleens and the

bovine uterus. A 1412-base pair cDNA has been cloned from human

placenta which encodes the entire human tartrate-resistant, Type 5

isozyme. This cDNA contains an open reading frame of 969 base pairs,

corresponding to a protein of 323 amino acids. A putative signal

sequence of 19 amino acids and two potential glycosylation sites are

present. The deduced amino acid sequence of the human enzyme is 85%

identical to that of porcine uteroferrin and 82% identical to the

corresponding regions of a partial amino acid sequence of the bovine

spleen enzyme. Northern blotting techniques employing a radiolabeled

cDNA probe coding for the human enzyme revealed the presence of a 1.5 kb

transcript in leukemic hairy cells, an Epstein-Barr virus-transformed

B-cell line, and the erythroleukemia cell line K562. Culture of K562

cells in the presence of 10-8M 12-0-tetradecanoylphorbol 13-acetate

enhanced tartrate-resistant acid phosphatase activity about 30-fold and

led to a corresponding increase in Type 5 isozyme mRNA levels.


The Tartrate-resistant Acid Phosphatases

The tartrate-resistant acid phosphatases have been studied

extensively by various research groups with different objectives. The

most abundant enzyme in this family, porcine uteroferrin, has been well

characterized physically and enzymatically, yet much of the research on

this protein has involved the determination of its role in iron

metabolism in the mid-pregnant pig. The beef spleen acid phosphatase

was first identified over 25 years ago, and while its physical

characteristics as a metalloprotein are well known, its function remains

obscure. The human tartrate-resistant acid phosphatase was first

described by researchers interested in the clinical significance of

phosphatase levels in normal and pathological tissue. Few reports have

been published on the characterization of this human enzyme. It is

clear, however, that all of the above proteins share several unusual

features, and that the clinically relevant human enzyme is worthy of

further study.


The uterine secretions of pigs contain a purple, basic, iron-

containing glycoprotein with acid phosphatase activity known as

uteroferrin. Uteroferrin has been well characterized since it was first

described in 1972 by Murray et al. Its enzymatic characteristics have

been studied by our laboratory, and its interesting spectral properties

have been the subject of numerous publications by many different

researchers. Our laboratory has also determined the likely function of

uteroferrin in the pregnant pig.

Uteroferrin is Synthesized by the Uterine Endometrium Under the
Influence of Progesterone

One of the major goals of our laboratory has been to purify and

characterize the proteins secreted by the porcine uterus. The quantity

and variety of proteins produced depend on the hormonal state of the

animal. Knight et al. (1974) described the collection of uterine fluids

from ovariectomized gilts which had been treated with estradiol 17-f,

progesterone, progesterone plus estradiol 17-f, or corn oil (control),

for 15 days. It was demonstrated that progesterone or progesterone plus

estradiol treatment caused a highly significant increase in the amount

of total protein that could be recovered relative to corn oil control or

estradiol treatment alone. Estradiol alone promoted neither increased

protein accumulation nor the appearance of novel proteins, but it acted

synergistically with progesterone to stimulate uterine secretary

activity when given at low dosage. A variety of proteins were thus

shown to be secreted by the porcine uterus under the influence of

progesterone. A protein now known as uteroferrin accounted for about

10-15% of the total protein in the uterine secretions of progesterone

treated animals (Schlosnagle et al., 1974).

Simmen et al. (1988) compared the biosynthetic rate of uteroferrin

to uteroferrin mRNA levels throughout pregnancy in the pig. Synthesis

of uteroferrin begins to increase markedly after day 30 of pregnancy in

association with a decline in maternal estrogen levels. Uteroferrin

biosynthesis reaches a maximum between days 60 and 75 (Basha et al.,

1979), and then declines towards term, which is about day 115 (Bazer et

al., 1975). The progesterone to estrogen ratios are highest between

days 35 to 75, which coincides well with maximal uteroferrin production

(Roberts and Bazer, 1980). Levels of uteroferrin in allantoic fluid

during pregnancy reach a maximum at a similar time (Roberts and Bazer,

1980). In comparison, highest levels of mRNA coding for uteroferrin

were detected during mid- and late-pregnancy. The mRNA levels at days

45-60 were comparable to those at days 90-110 (Simmen et al., 1988).

Thus, the amount of uteroferrin mRNA as determined by Northern analysis

does not appear to reflect the rate of uteroferrin biosynthesis as

determined by the incorporation of radioactive leucine. The regulation

of uteroferrin synthesis may therefore involve transcriptional and post-

transcriptional control. These results support the earlier theory that

although progesterone is essential for uteroferrin production, estrogens

may have an important modulating role, due to the fact that uteroferrin

synthesis is highest when the ratio of progesterone to estrogen is

highest (Roberts and Bazer, 1980).

The Purification of Uteroferrin from Uterine Secretions

Uteroferrin is most conveniently purified from the uterine flushes

of pseudopregnant animals (Basha et al., 1980), or from the allantoic

fluid of mid-pregnant pigs (Bazer et al., 1975). Pseudopregnancy can be

induced in sows by administering estradiol-17P daily between days 11-14

after estrus (Frank et al., 1978). The injected estradiol appears to

mimic an estrogen burst from the blastocyst which is believed to be the

signal to the sow which indicates she is pregnant (Bazer and Thatcher,

1977). Secretions of a day 60 pseudopregnant animal are comparable in

quality and quantity to those of an ovariectomized animal given daily

doses of progesterone for two months. Uteroferrin can be purified from

allantoic fluid after day 30 of pregnancy. However, uteroferrin in

sterile allantoic fluid is rapidly degraded, even when stored at 4*C.

Therefore, the allantoic fluid must be frozen after collection or used

immediately, and the purification of uteroferrin must be carried out


Because uteroferrin is available in gram quantities, and because of

its purple color, the development of a purification protocol for

uteroferrin was quite straightforward. Since uteroferrin is a very

basic protein, it binds to carboxymethyl cellulose even at high pH

values. A combination of ion exchange chromatography and gel filtration

on Sephadex G-100 is sufficient to yield pure uteroferrin of Mr 35,000,

as determined by SDS polyacrylamide gel electrophoresis (Chen et al.,

1973). Interestingly, uterine flushes and allantoic fluid obtained from

animals during early pregnancy (days 45-60) also contain a high

molecular weight (Mr-80,000), pink form of uteroferrin (Bazer et al.,

1975). It is now known that this protein consists of a molecule of

uteroferrin non-covalently associated with another, antigenically

unrelated protein (Baumbach et al., 1986). The reason for the

association of uteroferrin with this second protein is unknown.

A number of other interesting proteins, though none as abundant as

uteroferrin, can also be purified from porcine uterine secretions (see

Roberts and Bazer, 1988). These proteins include a family of

trypsin/plasmin protease inhibitors ( Mullins et al., 1980; Fazleabas et

al., 1982), and a group of retinol binding proteins (Adams et al.,


Uteroferrin is an Iron-containing Acid Phosphatase

Schlosnagle et al. (1974) demonstrated that the abundant, basic,

purple glycoprotein from porcine uterine secretions was an iron-

containing acid phosphatase. They reported a pH optimum of 4.9 and a Km

of 2.2mM for the enzyme's preferred substrate, p-nitrophenylphosphate.

Weak reducing agents, such as 2-mercaptoethanol enhanced activity 2- to

4-fold, without changing the Km. Uteroferrin's phosphatase activity

could be inhibited by fluoride, arsenate, phosphate and molybdate

(Schlosnagle et al., 1976). Sodium dithionite caused an immediate loss

of enzymatic activity and bleaching of color, although activity could be

restored with Fe3+ salts (Schlosnagle et al., 1976). Oxidizing agents,

and reagents which interact with sulfhydryl groups also inhibited

uteroferrin (Schlosnagle et al., 1976). Uf was not inhibited by

tartrate, an unusual feature in an acid phosphatase. The substrate

specificity of uteroferrin was also unusual. It readily hydrolyzed

compounds with strong leaving groups, such as p-nitrophenylphosphate and

ATP, but had little or no activity towards aliphatic phosphates such as

f-glycerol phosphate, or hexose phosphates such as D-glucose 6-phosphate

(Schlosnagle et al., 1974). Uteroferrin also displayed phosphoprotein

phosphatase activity (Roberts and Bazer, 1976).

Uteroferrin, as purified, is predominantly purple and enzymatically

inactive. This purple coloration originates from a broad absorption

band centered around 545nm. However, when uteroferrin is activated by

2-mercaptoethanol, the extinction maximum changes from 545nm (purple) to

508nm (pink) (Schlosnagle et al., 1974). The reverse reaction (pink to

purple) is promoted by oxidizing agents. When purified, uteroferrin

consists of an equilibrium mixture of the reduced (pink) and oxidized

(purple) forms, with the latter predominating. The equilibrium between

the pink and purple forms may depend on the redox state of the fluid

sample from which the protein was recovered. In contrast, the stable

pink, high molecular weight form of uteroferrin does not require

reducing agents for activity, it is fully active as purified (Baumbach

et al., 1986). It is not clear how non-covalent association with a

second, unrelated protein stabilizes the active form of the enzyme.

Uteroferrin Contains Two Iron-binding Sites

The amount of iron bound to uteroferrin has been the subject of

controversy. Early reports (Roberts and Bazer, 1980) indicated that

uteroferrin contained a monoferric site. However, it is now clear that

each molecule of uteroferrin can bind up to two iron atoms, although

only one of these iron atoms is necessary for the purple/pink

coloration. It has been demonstrated that the iron at the more labile,

non-chromophoric site can be replaced by a variety of other metals

including zinc, copper and mercury (Beck et al., 1984). The apparent

lower content of iron (less than two atoms per molecule) consistently

noted in some samples of uteroferrin (see Baumbach et al., 1986) may

have resulted from a partial absence of iron at the non-chromophoric

site and its replacement by other metals.

A number of methods have been employed to study the iron binding

sites in uteroferrin and the related beef spleen purple phosphatase.

EPR, Mossbauer, NMR and magnetic susceptibility studies are consistent

with a model in which the iron atoms occupy two distinct but adjacent

sites which are sufficiently close so that the two atoms are

magnetically coupled (Davis and Averill, 1982; Debranner et al., 1983;

Lauffer et al., 1983). Purple, enzymatically inactive uteroferrin is

believed to contain two ferric ions. This form of uteroferrin gives no

EPR signal, presumably because the spins of the two ferric ions are

antiferromagnetically coupled, with a net spin of zero. When purple

(Emax=545nm) uteroferrin is treated with mild reducing agents and turns

pink (Emax=508nm), it contains a ferric-ferrous iron pair, which gives

rise to an intense g'-1.74 EPR signal at liquid helium temperatures

(Antanaitis and Aisen, 1983). The naturally occurring, high molecular

weight pink form of uteroferrin gives rise to the same g'-1.74 EPR

signal without the addition of reducing agents (Baumbach et al., 1986).

At present there is no adequate model to describe the iron binding

sites of uteroferrin. Resonance-Raman studies have shown that the

purple-pink color most likely arises from coordination of iron to

tyrosine residues (Gaber et al., 1979). The iron atom involved is

probably the less labile one at the Fe(III) site on pink reduced

uteroferrin, since the molar extinction coefficient does not change when

the protein is reduced from its purple to its pink form. However, the

color change from purple to pink most likely represents disulfide

reduction rather than reduction of iron (Schlosnagle et al., 1976).

Histidine has also been implicated as a ligand for the iron atoms

(Lauffer et al., 1983), and an oxygen atom has been suggested to bridge

the iron sites (Mockler et al., 1983). Orthophosphate is coordinated

with the iron center on purple uteroferrin (Antanaitis and Aisen, 1983).

The amino acid sequence of uteroferrin revealed that the iron binding

sites must be non-identical, because the protein contains no internal

repeats with histidine and tyrosine residues in identical spacial

relationships. It is not clear which of the 10 histidine and 10

tyrosine residues which are conserved in uteroferrin and the beef spleen

enzyme could be ligands for the metals (Hunt et al., 1987).

A Proposed Reaction Mechanism for Uteroferrin

Unlike the alkaline phosphatases where the role of the Zn(II) is

fairly well understood, the function of the iron atoms in catalytic

activity of uteroferrin and the other purple acid phosphatases is

unknown. Competitive inhibitors such as phosphate and arsenate interact

with the paramagnetic centers on these proteins, which suggest that the

metal contributes to substrate binding (Davis and Averill, 1982, Kawabe

et al., 1984). Such substrate-iron complexes would probably be stable

around pH 5.0, and may explain the acid pH optimum. The iron, by virtue

of its strong electron withdrawing properties, could therefore promote


It is well documented that the reaction mechanism of the zinc-

containing alkaline phosphatases involves a phosphoryl-enzyme

intermediate (Levine et al., 1969). Previous experimental evidence

obtained with a variety of acid phosphatases also indicated the

formation of a phosphoryl-enzyme intermediate, with a double

displacement reaction mechanism (Hickey and VanEtten, 1972; Igarashi et

al., 1970; Feldman and Butler, 1969). Kinetic evaluation of bovine

purple acid phosphatases indicated that these uteroferrin-like enzymes

undergo a pseudo Uni Bi hydrolytic transfer reaction mechanism (Lau et

al., 1987; Davis et al., 1981). This theory is supported by the fact

that transition state analogs of phosphate such as vanadate, molybdate

and fluoride are potent inhibitors of uteroferrin and the other purple,

iron-containing acid phosphatases. Lack of inhibition by p-nitrophenol

suggests that the hydrolysis reaction proceeds via sequential release of

p-nitrophenol and phosphate (Davis et al., 1981):


E + pNPP ___ E ppNPpNPPE P pPE-Pi= EE + Pi

Uteroferrin Contains at Least One High Mannose Oligosaccharide

Chen et al. (1973) reported that the progesterone-induced purple

protein of the porcine uterus was a glycoprotein, and that the

oligosaccharide contained significant amounts of mannose. The structure

of the triantennary 5 or 6 mannose oligosaccharide has been determined

(Saunders et al., 1985), for uteroferrin purified from allantoic fluid

and uterine flushes from pseudopregnant animals.



> cxl,6

cal, 3

Man Man P1,4 -GlcNAc _1,4) GlcNAc

(Man ac1.2) Man

The biosynthesis of uteroferrin's oligosaccharide chain has been

studied in vitro with explants of uterine endometrium. When [32p] was

provided to such cultures, the label became incorporated into the high

mannose oligosaccharide chains of uteroferrin (Baumbach et al., 1984).

Approximately one-third of the oligosaccharide chains cleaved from

uteroferrin by endoglycosidase H had phosphorylated carbohydrate groups:

Thus, uteroferrin, a secreted acid phosphatase, carries the lysosomal

recognition marker, mannose 6-phosphate, considered to be responsible

for the intracellular targeting of acid hydrolyses from the Golgi

complex to the lysosomes (Shephard et al., 1983). These results implied

that uteroferrin is a substrate in vivo for the first of the two enzymes

responsible for the addition of phosphate to terminal mannosyl residues

on lysosomal enzymes. UDP-N-acetylglucosamine: lysosomal enzyme

N-acetylglucosamine 1-phosphate transferase (phosphotransferase) adds

N-acetylglucosamine 1-phosphate from the nucleotide sugar UDP-N-acetyl-

glucosamine to the 6-position of terminal mannose residues. Then

N-acetylglucosamine-l-phosphodiester o-N-acetylglucosaminidase

(phosphodiesterase) cleaves the resulting phosphodiester bond adjacent

to the N-acetylglucosamine residue (Goldberg and Kornfeld, 1983).

Receptors for mannose 6-phosphate then transport these enzymes from the

Golgi to lysosomes (Shephard et al., 1983). It was also demonstrated

that uteroferrin is an excellent substrate for the phosphotransferase in

vitro (Lang et al., 1984), in fact, it was the best substrate of those

tested so far.

Why is uteroferrin secreted rather than directed to lysosomes?

There are a number of possible explanations. Although uteroferrin is a

good substrate for the phosphotransferase, the covering N-acetyl-

glucosamine is not efficiently removed by the phosphodiesterase, thus

masking the mannose 6-phosphate (Baumbach et al., 1984), and such

phosphodiesters are poor substrates for both mannose 6-phosphate

receptors (Hoflack et al., 1987). The presence of only one

phosphorylated mannose residue on uteroferrin's oligosaccharide may also

make it a poor ligand for the mannose 6-phosphate receptors (S.

Kornfeld, personal communication). The most likely explanation,

however, is that the rate of uteroferrin synthesis in the pregnant

uterus outstrips the intracellular targeting mechanism. Finally, it is

also possible that the primary amino acid sequence of uteroferrin

contains targeting information that directs it to secretary granules

despite the fact that it carries phosphorylated mannose residues.

The Function of Uteroferrin

Roberts and Bazer (1980) have proposed that the primary function of

uteroferrin is not phosphate ester hydrolysis, but in iron metabolism.

Uteroferrin's high Km, unusual substrate specificity and lack of

activity at the pH of the fluid environment in which it is found argues

against an enzymatic role. The uterus of a mid-pregnant pig can produce

more than two grams of uteroferrin per day (Basha et al., 1979), an

amount far exceeding that which could conceivably be used for enzymatic

purposes. Since large quantities of uteroferrin are known to cross the

pig placenta, and the iron from [59 Fe] labeled uteroferrin is used for

fetal erythropoiesis (Buhi et al., 1982a; Ducsay et al., 1982; Renegar

et al., 1982), the role of uteroferrin in the pregnant pig is proposed

to be in iron transport.

The pig has a non-invasive type of placentation, and at no time

during pregnancy is the uterine epithelium eroded. The concepts

therefore relies upon the secretion of macromolecular products by the


mother for a large part of pregnancy. Using a variety of immunochemical

and radiotracer techniques, the path of uteroferrin movement from the

mother to the fetus has been determined in the pig (Buhi et al., 1982a;

Ducsay et al., 1982; Renegar et al., 1982; Ducsay et al., 1984; Roberts

et al., 1986b). Uteroferrin is synthesized in the uterine glands

(Renegar et al., 1982), each of which is covered by a special region of

the chorion (fetal placenta) called an areola. This structure is made

up of specialized absorptive cells filled with large endocytotic

vacuoles. Uteroferrin has been identified in these structures by

immunogold staining (Roberts et al., 1986a) and within the venous

drainage of the placenta. It is cleared from the bloodstream in part by

the fetal liver, which is the major site of fetal hematopoiesis, and

also by the kidney where it enters the urinary filtrate and is voided

into the allantoic sac (Renegar et al., 1982). Uteroferrin in the

allantoic fluid loses its iron to fetal transferring, which in turn

transports the iron to the fetal liver (Buhi et al., 1982a).

Uteroferrin's synthesis in the porcine uterus and its turnover in

the fetal liver could theoretically supply developing fetuses with

enough iron until day 75 of pregnancy. However, uteroferrin synthesis

appears to fall in late pregnancy, just when iron demand for

erythropoiesis is at a maximum (Roberts and Bazer, 1980). Therefore,

other mechanisms of iron transport may be necessary in order to supply

these needs.

Other Purple. Iron-containing Acid Phosphatases

Several other tartrate-resistant acid phosphatases have been

purified from a variety of sources. A uteroferrin-like purple acid

phosphatase has been purified from the uterine secretions of mares in

prolonged diestrus (Zavy et al., 1978; McDowell et al., 1982). An acid

phosphatase has also been partially purified from the uterine secretions

of cows treated with progesterone (Dixon and Gibbons, 1979). These

discoveries were not unexpected, because the pig, the cow and the mare

have similar types of non-invasive placentation.

In 1960, a basic, purple phosphoprotein phosphatase was identified

in beef spleen (Revel and Racker, 1960; Glomset and Porath, 1960) and

was later shown to contain iron (Campbell and Zerner, 1973). Like

uteroferrin, this enzyme showed intense absorption in the range 510-

545nm, and contained two iron binding sites (Davis et al., 1981). Davis

et al. also demonstrated that the substrate specificities, sensitivity

to inhibitors and activation by reducing agents exhibited by the beef

spleen enzyme were similar to those of uteroferrin. However, the beef

spleen enzyme (Mr=40,000) was reported to be made up of two subunits

(Mr-24,000 and 15,000), while uteroferrin is a monomeric enzyme (Mr-

35,000). The beef spleen enzyme has been well characterized with regard

to its iroT binding sites, which are virtually identical to those of

uteroferrin (discussed earlier). Thus, the bovine spleen enzyme's iron

cluster is an asymmetrical complex containing two Fe(III) ions, one of

which is reduced to Fe(II) by reductive activation of the enzymes (Davis

and Averill, 1982; Antanaitis and Aisen, 1983; Debranner et al., 1983;

Antanaitis and Aisen, 1982).

The function of the beef spleen enzyme has been addressed only

recently. Schindelmeiser et al. (1987) have localized the enzyme to

lysosomal-like organelles of cells of the reticulo-phagocytic system.


The phagocytosing cells that contained the phosphatase were frequently

found in close contact with clusters of aged and deformed erythrocytes.

The authors hypothesized that the enzyme is involved either in the

breakdown of erythrocyte membrane and cytoskeletal phosphoproteins, or

in the metabolism of iron from these erythrocytes.

There have been reports of a bovine skeletal tartrate-resistant acid

phosphatase, assumed to be distinct from the bovine spleen enzyme (Lau

et al., 1985, 1987). This bovine skeletal enzyme resembles the bovine

spleen enzyme and porcine uteroferrin in its molecular weight, pH

optimum, substrate specificities and sensitivity to inhibitors. Like

uteroferrin, and unlike the bovine spleen enzyme, the bovine skeletal

enzyme is reported to be monomeric. The authors chose to stress the

phosphotyrosine phosphatase activity of the enzyme and postulated that

the skeletal acid phosphatase functions in the regulation of

proliferation and differentiation. The authors have not yet described a

physiologic substrate for this phosphoryrosyl-specific protein

phosphatase, and have not demonstrated that it plays an important role

in cell growth. Thus, the function of the acid phosphatase in bone

remains unknown.

Purple, iron-containing acid phosphatases have been purified from

rat spleen (Hara et al., 1984) and rat bone (Kato et al., 1986; Anderson

and Toverud, 1986), and an immunologically related phosphatase was

purified in small quantities from rat epidermis (Hara et al., 1985).

These monomeric enzymes also resembled uteroferrin in molecular weight,

substrate specificities and sensitivity to inhibitors. Hara et al.

(1984, 1985) believe that the spleen and epidermis enzymes may play a


role in nucleotide metabolism and that the bone enzyme may be acting as

a phosphoprotein phosphatase in vivo. Although Anderson and Toverud

(1986) have indicated that studies were underway to determine the

function of the rat bone enzyme, the functions of the rat enzymes also

remain obscure.

Human Acid Phosphatases

The human acid phosphatases were first characterized by researchers

interested in the clinical significance of acid phosphatase levels in

normal and pathological tissue. One of the human enzymes, a cationic,

tartrate-resistant enzyme known as the Type 5 phosphatase, was of

interest to us because it seemed as though it might be related to the

well characterized abundant acid phosphatase, porcine uteroferrin.

Characterization of the Human Acid Phosphatases

Li et al. (1970) described electrophoretic separation of human

leukocyte acid phosphatases on 7.5% polyacrylamide gels at pH 4.0, which

were then stained for acid phosphatase activity. They detected six

isozymes of acid phosphatase, varying in electrophoretic mobility. Band

O did not enter the gel. Band 1 was the slowest migrating form, band 5

the fastest and most basic. The various isozymes could also be

distinguished by their substrate specificities and sensitivity to


Lam et al. (1973) described in detail the distribution of acid

phosphatases in normal human tissue. Isozymes 1 and 3 were the most

widely distributed and most active forms. Higher type 2 levels were

characteristic of the prostate. Type 5 was the least common and least

abundant isozyme, detected in only small quantities in spleen, kidney

and liver.

These researchers noticed differences in acid phosphatase levels and

in relative isozyme ratios in normal plasma and leukocytes versus those

from patients with certain diseases. In normal leukocytes, band 1 was

the strongest, bands 2, 3 and 4 were of moderate strength and band 5 was

very weak (Yam et al., 1971). Patients with infectious mononucleosis,

for example, exhibited elevated levels of isozymes 1, 3 and 5 in

leukocytes (Li et al., 1973). Isozyme 2 was found to be elevated in the

plasma of patients with prostatic cancer (Lam et al., 1973). Perhaps

the most striking isozyme pattern was evident in the leukocytes and

spleen homogenate (but not plasma) of patients with hairy cell leukemia.

The leukemic white cells of these patients produced very small amounts

of isozymes 1 through 4, and large quantities of isozyme 5 (Yam et al.,


Yam et al. (1971) also determined that while acid phosphatase

isozymes 1,2,3 and 4 were completely inhibited by 50mM L-(+)-tartrate,

isozyme 5 was unaffected by this reagent. (Recall that uteroferrin is

resistant to inhibition by tartrate). This observation made it easier

to detect the presence of elevated levels of the Type 5 isozyme in hairy

cell leukemia and allowed cytochemical localization of only this isozyme

in leukocytes and tissue sections. The presence of tartrate-resistant

acid phosphatase activity in the hairy cells in blood smears is the

major method of diagnosis of hairy cell leukemia today (Nanba et al.,

1977; Janckila et al., 1978; Braylan et al., 1979).

Hairy Cell Leukemia

The origin of hairy cells has been a subject of controversy for a

number of years. Surface marker studies done with monoclonal antibodies

have revealed hairy cells with properties common to both B-cells and

T-cells (Burns et al., 1980; Worman et al., 1983; Armitage et al.,

1985). Hairy cells have been shown to phagocytose latex particles,

properties which indicated a common origin with monocytes or histiocytes

(Rosenszajn et al., 1976). However, most researchers now agree that

hairy cells have originated from B-lymphocytes, possibly splenic white

pulp marginal zone lymphocytes (Van der Oord et al., 1985). More

specifically, hairy cells may originate from a lymphocyte in a terminal

stage of differentiation, at the point of switching from an IgM-bearing

small lymphocyte to a mature plasma cell (Jansen et al., 1979).

Recently it was proposed that the hairy cell may represent the

terminally differentiated B-lymphocyte (Robinson et al., 1985; Gazitt

and Polliack, 1987).

Clinically, hairy cell leukemia is represented by a slow insidious

onset, splenomegaly, anemia and the appearance of leukemic hairy cells

in the spleen, bone marrow and peripheral blood (Nanba et al., 1977).

Significant quantities of tartrate-resistant acid phosphatase are

contained in the reticulum and epithelioid cells of the patients'

spleens. In the past splenectomy was often performed in order to reduce

the tumor load. More recently, it has been demonstrated that

c-interferon treatment is successful in causing remission in hairy cell

leukemia patients (see Porzsolt, 1986).

Tartrate-resistant Acid Phosphatase Levels in Other Diseases

Tartrate-resistant acid phosphatase levels are elevated in several

types of leukemia other than hairy cell leukemia but not in a consistent

manner. Drexler et al. (1985, 1986) have studied acid phosphatase

levels in a variety of human leukemia cell lines and in fresh leukemia

cells obtained directly from patients. These researchers found that the

few leukemia cell lines which did express the phosphatase were those

which represented lymphocytes arrested late in differentiation. Many

researchers now believe that tartrate-resistant acid phosphatase may be

useful as an enzymatic marker for B-lymphocyte differentiation (Drexler

et al., 1985) and that measurement of its activity may have prognostic

value (Pieters and Veerman, 1987).

Measurement of serum acid phosphatase levels may also be of clinical

significance. While the Type 5, tartrate resistant acid phosphatase is

not detected at high levels in the plasma of patients with hairy cell

leukemia, plasma levels of this isozyme are high in patients with

diseases which involve bone reabsorption and regeneration, and in

patients with malignancies metastasized to bone (Lam et al., 1978).

These diseases include Paget's disease, which is a type of bone tumor,

and osteoporosis (Li et al., 1973). Tartrate-resistant acid phosphatase

activity is also elevated in the plasma of children during bone growth

(Chen et al., 1979). Plasma levels of the Type 5 phosphatase are also

high in patients with Gaucher's disease, which is an inherited

glycosphingolipid storage disease characterized by f-glucocerebrosidase

deficiency. The phosphatase in Gaucher's disease is believed by some

researchers to be of osteoclast origin, and most likely a secondary

phenomenon indicative of bone involvement (Choy, 1985).

It is interesting to note that while the acid phosphatase of hairy

cells is intracellular, confined to discrete inclusion bodies resembling

lysosomes (Lam et al., 1976), osteoclasts secrete this enzyme under the

influence of parathyroid hormone (Miller, 1985; Braidman et al., 1986;

Chambers et al., 1987). It is not clear whether the secreted osteoclast

and lysosomal spleen Type 5 phosphatases are distinct enzymes.

Purification and Characterization of the Human Type 5 Isozyme

In all of the studies described above, acid phosphatase levels were

measured by histochemical methods or by enzymatic staining of non-

denaturing polyacrylamide gels. There are a few reports in the

literature of attempts to purify and characterize the human tartrate-

resistant acid phosphatase. Although these reports contain some

contradictory information, one fact becomes apparent: This human Type

5, tartrate-resistant acid phosphatase has a number of properties

reminiscent of porcine uteroferrin.

An attempt was made by Lam and Yam (1977) to purify this human Type

5 acid phosphatase from the spleen of a patient with hairy cell

leukemia. They were not able to purify the enzyme to homogeneity.

Although the authors reported an Mr of 64,000, which is almost twice

that of uteroferrin, many of the characteristics they reported for the

enzyme were reminiscent of the characteristics of uteroferrin and the

other purple, iron-containing acid phosphatases. The human hairy cell

enzyme, like uteroferrin, was a very basic protein. This human enzyme,

also like uteroferrin, hydrolyzed p-nitrophenylphosphate, ATP and

pyrophosphate but was virtually inactive towards hexose phosphates,

B-glycerol phosphate and AMP. The human enzyme was also activated by

mild reducing agents and very sensitive to inhibition by molybdate.

A tartrate-resistant acid phosphatase was also purified from a

Gaucher's disease spleen (Robinson and Glew, 1980; 1981). This enzyme

was believed to be distinct from the hairy cell spleen enzyme. Like

uteroferrin and the other purple, iron-containing acid phosphatases,

this enzyme was reported to have an Mr of 33,000. It bound avidly to

Concanavalin A Sepharose, indicating its glycoprotein nature. It, too,

was sensitive to inhibition by molybdate, fluoride and dithionite.

However, this enzyme had a number of features which were unlike the

hairy cell enzyme. It was reported that the Gaucher enzyme was made up

of two subunits of Mr 20,000 and 16,000, did not hydrolyze

pyrophosphate, and was not activated by reducing agents. Thus, it was

not clear whether the Gaucher spleen and hairy cell spleen enzymes were

the same protein. Efstratiatis and Moss (1985) partially purified an

Mr-37,000 tartrate-resistant, molybdate-sensitive acid phosphatase from

lung, and detected an apparently identical enzyme in osteoclasts. It is

not known whether the lung/osteoclast tartrate-resistant acid

phosphatase is identical to either of the spleen enzymes.

Although it was once believed that uteroferrin was uniquely

associated with the uterus (Roberts and Bazer, 1980), it is now clear

that there is a growing family of purple, iron-containing acid

phosphatases. The human tartrate-resistant acid phosphatase may belong

to this enzyme family. Because of the clinical significance of the

human Type 5 phosphatase, it seemed to be a subject worthy of further

study. In Chapter 2, the purification of the human tartrate-resistant

acid phosphatase to homogeneity is described, and its physical

properties are clearly defined. The enzymatic characteristics of the

phosphatase are described in detail, and the immunologic relationship of

the human enzyme to the purple, iron-containing acid phosphatases is

investigated. Chapter 3 deals with the molecular cloning of a cDNA

coding for the human tartrate-resistant acid phosphatase, the

characteristics of the deduced amino acid sequence, and the expression

of the enzyme in leukemia cells.



Measurement of serum and tissue levels of acid phosphatase can prove

useful in the diagnosis of several human disease states (Gutman et al.,

1936; Yam et al., 1980; Drexler and Gaedicke, 1983; Yam et al., 1983;

Allhoff et al., 1983). At least six isozymes of acid phosphatase have

been identified by their relative electrophoretic mobilities towards the

cathode at low pH. The most basic, a minor isozyme known as Type 5, is

elevated in the circulating white cells and enlarged spleens of patients

with hairy cell leukemia (Lam et al., 1980; Yam et al., 1983; Yam et

al., 1971; Lam and Yam, 1977) and in spleens and sera of patients with

Gaucher's disease (Choy, 1983). This same isozyme also appears to be a

secretary product of osteoclasts (Minkin, 1982; Chen et al., 1979;

Stepan et al., 1983), and of osteoclastic bone tumors (Chen et al.,

1979; Tavassoli et al., 1980). The Type 5 phosphatase can be

distinguished from the other isozymes, not only by its basic nature, but

also because it is not inhibited by L-(+)-tartrate. Indeed, the

cytohistochemical demonstration of a tartrate-resistant acid phosphatase

within leukemic hairy cells is currently the most usual method of

diagnosing this rare form of leukemia (Nanba et al., 1977; Janckila et

al., 1978; Braylan et al., 1979).

The enzyme produced by the spleen histiocytes of patients with

Gaucher's disease has been purified in very small quantities by Robinson

and Glew (1980). This phosphatase was activated by reducing agents and

had high catalytic activity towards a variety of substrates, both

artificial and natural, which had strong leaving groups, but had very

low activity towards aliphatic phosphate esters. It was also a

phosphoprotein phosphatase. A similar enzyme has been isolated from the

spleen of a patient with hairy cell leukemia (Lam and Yam, 1977).

However, the latter enzyme appeared to differ from the Gaucher

phosphatase in a number of respects, including its response to reducing

agents, its molecular weight and its substrate specificity. Moreover

its specific activity was about 10-fold lower than that of the Gaucher


The enzymatic properties of the Type 5 human isozyme are very

similar to those of a well characterized, abundant acid phosphatase

known as uteroferrin. This unusual molecule is a purple colored, iron-

containing glycoprotein that was first purified from uterine secretions

of pigs (Chen et al., 1973). Its synthesis, which occurs within the

glandular epithelium of the uterus (Renegar et al., 1982), is under

progesterone control (Roberts and Bazer, 1980). During pregnancy

approximately 2g of the protein are produced daily (Basha et al., 1979).

It has been shown to be taken up by the overlying placenta and to enter

the fetal blood stream (Renegar et al., 1982). Its iron is then rapidly

utilized in fetal erythropoiesis (Ducsay et al., 1982; Buhi et al.,

1982a). A role for uteroferrin in transplacental iron transport has

been proposed (Roberts and Bazer, 1980; Buhi et al., 1982a). However,

the fact that uteroferrin, particularly when activated with mild

reducing agents such as ascorbate or mercaptoethanol, is a potent acid

phosphatase (Schlosnagle et al., 1974; 1976) has raised questions as to

whether iron transport is the major or only function of the protein

(Davis et al., 1981). In this regard, acid phosphatases which are

clearly similar to uteroferrin in their spectral properties, molecular

weight, and iron content have been isolated from a variety of sources

(Antanaitis and Aisen, 1983). In addition, there are several

descriptions of acid phosphatases, including the Type 5 human isozyme,

which resemble uteroferrin in their enzymatic properties, but which have

either not been purified or else have been isolated in such small

quantities that they have not been fully characterized (Roberts and

Bazer, 1976). For the above reasons it has been proposed that

uteroferrin may be but one member of a broad class of acid phosphatases

which has a wide distribution (Roberts and Bazer, 1980; 1976).

Chapter 2 contains the description of a simple method for the

purification of the enzyme from human spleen by making use of antibodies

to the readily available uteroferrin. It is demonstrated conclusively

that this human isozyme belongs to the class of iron-containing acid

phosphatase best illustrated by uteroferrin. Monoclonal antibodies have

been prepared against both uteroferrin and the human enzyme which appear

to have broad reactivity towards this class of enzyme.

Materials and Methods


CM-cellulose was obtained from Whatman, Sephadex G-100 and

Sepharose-4B from Pharmacia, and Triton X-100 from Rohm and Haas Co.


All substrates, activators, and inhibitors (except anions) used for acid

phosphatase assays, cyanogen bromide, protein standards, Fast Garnet GBC

salt, buffers and protease inhibitors were purchased from Sigma.

Reagents for the Bradford protein assay were purchased from Bio-Rad.

Isozyme standards from serum (for human acid phosphatase activity) were

obtained from Calbiochem-Behring. All other chemicals (reagent grade or

better) were obtained from Fisher.


Purification of uteroferrin. Uteroferrin was purified from the

uterine secretions of pseudopregnant pigs or from allantoic fluid as

previously described (Schosnagle et al., 1974; Roberts and Bazer 1980).

Allantoic fluid was dialyzed overnight at 4C against 0.01M Tris-HCl,

pH 8.2. A slurry of carboxymethyl-cellulose (CM-cellulose; 100ml of

settled resin per 4L allantoic fluid) was added in the same buffer.

This mixture was stirred for 1 hour, then packed into a column and

washed with the above buffer. The bound, basic proteins were eluted in

one step by a high salt buffer (0.01M Tris-HCl, pH 8.2, 0.5M NaCl). The

proteins were dialyzed against 0.01M Tris-HC1, pH 8.2 and applied to a

Sephadex G-100 column (5 x 90 cm) which had been previously equilibrated

in 0.1M sodium acetate, pH 4.9 and 0.33M NaCI. (Uterine secretions of

pseudopregnant animals were loaded directly onto the Sephadex G-100

column after centrifugation at 10,000xg at 40C for 30 minutes.) The

pink, high molecular weight uteroferrin peak, which immediately preceded

the purple peak, was pooled and used without further purification. The

uteroferrin peak, purple in color, was pooled and dialyzed against 0.01M

Tris-HCl, pH 8.2. The uteroferrin was then loaded onto a column of

CM-cellulose and eluted with a linear salt gradient (0.01-0.5M NaC1).

Uteroferrin eluted as a symmetrical peak between 0.20 and 0.25M NaCI.

The final preparations of uteroferrin had A280/A545 ratios of less than

14 and gave a single band of apparent Mr=35,000 upon polyacrylamide gel

electrophoresis (PAGE) in the presence of sodium dodecyl sulfate (SDS).

Purification of acid phosphatase from human spleen. Spleen tissue was

used from a male aged 52 who was diagnosed as having hairy cell

leukemia. The weight of the spleen was 2.3kg. The normal control

spleen from an accident victim weighed 0.22kg. Tissue was cut into

slices, the outer fascia removed, and the slices diced into small cubes.

These were homogenized at 40C in extraction buffer [0.05M Tris acetate,

pH 7.5, 2% (v/v) Nonidet P-40, 0.02% (w/v) sodium azide, ImM

phenylmethanesulfonyl fluoride (PMSF), ImM EDTA, and aprotinin (1

trypsin inhibitory unit/ml)] by means of a Teckmar Model SDT homogenizer

(half-full setting) with the largest probe until a homogeneous

preparation was obtained. A volume of 150ml of extraction buffer was

used for each 100g of tissue. The homogenate was centrifuged (30,000xg;

20 minutes) and the supernatant fraction dialyzed overnight against

0.01M Tris-HCl buffer, pH 8.2. A slurry (50ml) of CM-cellulose was

added to the retentate and stirred for 1 hour. This material was

collected on a Whatman No. 4 filter paper on a Buchner funnel. After it

was washed several times with 0.01M Tris-HC1 buffer, pH 8.2, the CM-

cellulose was suspended in the same buffer and loaded into a glass

column. Proteins which had remained bound were eluted with 0.5M NaCI.

In initial experiments, the eluted material was dialyzed and loaded onto

a column of CM-cellulose (1 x 5cm) in 0.01M Tris-HC1 buffer, pH 8.2.


Elution was performed by means of a gradient (180ml; 0.01-0.5M) of NaCl

at pH 8.2. Fractions which contained the enzyme were pooled, dialyzed,

and freeze-dried, and then subjected to gel filtration on a column (2.5

x 80cm) of Sephadex G-100 which was equilibrated with 0.01M sodium

acetate buffer, pH 4.9, containing 0.33M NaCl. Elution of the protein

was followed by its absorbance at 280nm and by measuring acid

phosphatase activity (see below). Fractions which contained the

phosphatase were pooled and loaded directly onto an anti-uteroferrin

immunoaffinity column (see below).

After the immunoaffinty column had been washed thoroughly with

loading buffer (0.01M Tris-HCl, pH 8.2, containing 0.3M NaC1), the bound

enzyme was eluted with 0.05M glycine-HCl buffer containing 0.15M NaCI,

pH 2.3. It was collected in Iml fractions. Each fraction was

immediately neutralized with 0.lml of 1M Tris-HCl, pH 8.2.

More recently, enzyme which had been eluted form the CM-cellulose

column with 0.5M NaCI was loaded directly onto the immunoaffinity column

without including intervening ion exchange and gel filtration steps.

Purification of acid phosphatase from human placenta. A normal term

placenta, weighing 630g was employed to purify the acid phosphatase by

exactly the same procedure described for human spleen. An alternative

method of enzyme purification was employed for a second normal term

placenta weighing 650g. The second placenta was homogenized in 1300ml

of a lysis buffer which consisted of 0.3M KC1, 2% (v/v) Nonidet P-40,

ImM PMSF and 0.2% (w/v) sodium azide. After centrifugation at 30,000xg

for 20 minutes at 4C, protamine sulfate was added to the supernatant

fraction to 0.1% (w/v). The solution was stirred for 1 hour at 40C,


then centrifuged at 30,000xg for 15 minutes. The resulting supernatant

fraction was dialyzed overnight against 0.01M Tris-HC1, pH 8.2. A

slurry of CM-cellulose (50ml) was added and the mixture stirred for 1

hour at 4*C. The bound proteins were eluted from the CM-cellulose as

described for uteroferrin. The entire pool of basic proteins was then

subjected to immunoaffinity chromatography as described for the human

spleen enzyme.

Purification of the bovine spleen acid phosphatase. The procedure

described for the human spleen enzyme was also employed for the bovine

spleen enzyme. Approximately 200g of beef spleen tissue were used.

Purification of the rat spleen phosphatase. The spleens of three

rats (approximately 2g) were homogenized as described for the human

enzyme. After centrifugation at 30,000xg, the supernatant fraction was

loaded directly onto the immunoaffinity column. Elution was performed

as described for the human enzyme.

Purification of the bovine uterine acid phosphatase. Bovine uterine

fluids were obtained from the ligated uterine horn of day 270

unilaterally pregnant cattle (Bartol, 1983). Approximately 400ml of

bovine uterine fluids were dialyzed overnight against 4 liters of 0.01M

Tris-HCl, pH 8.2. The solution was centrifuged (30,000xg; 20 minutes)

to remove particulate matter and passed through a column (1 x 5cm) of

CM-cellulose equilibrated in the same buffer. After washing the column,

proteins which had bound were eluted with 0.5M NaCl in 0.01M Tris-HCl,

pH 8.2. These basic proteins were then subjected to gel filtration on a

column (2.5 x 80cm) of Sephadex G-100 which had been equilibrated with

0.01M acetate buffer, pH 4.9 containing 0.33M NaCl. Protein


concentration was monitored by absorbance at 280nm, and acid phosphatase

activity monitored by p-nitrophenylphosphatase activity (see below).

Measurement of p-nitrophenylphosphatase activity. The standard

colorimetric assay utilized an appropriately diluted sample of enzyme,

20mM p-nitrophenylphosphate and 0.1M sodium acetate buffer (pH 4.9 for

uteroferrin, pH 5.3 for the human and rat enzymes, pH 6.0 for the beef

spleen enzyme and pH 4.5 for the bovine uterine enzyme) in a final

volume of Iml. The enzymes were assayed in the presence of 0.1M

2-mercaptoethanol except where stated in the text. The reaction was

allowed to proceed at 37C for 10 minutes and stopped by the addition

of Iml of 1M NaOH. The absorbance was then read at 410nm. Activity is

expressed as units, where one unit is the release of lpmol p-nitro-


Hydrolysis of other substrates. With the exception of phosvitin,

other substrates were assayed at a final concentration of 5mM. The

release of orthophosphate was measured spectrophotometrically by the

method of Bartlett (1959), which involves the reduction of

phosphomolybdate to a blue color (Emax-830nm). Blanks without enzyme

were run for each substrate. The release of orthophosphate from

phosvitin (lOmg/ml) was measured as described by Roberts and Bazer


pH optima. The pH optima of the acid phosphatases were determined

with the use of 0.1M sodium acetate, pH 3.0-5.4, 0.1M Tris acetate, pH

5.6-6.5, and 0.1M Tris-HCl, pH 7.0-7.5 as buffers. The substrate used

was p-nitrophenylphosphate.


Protein determination. Protein concentration was determined by the

method of Lowry et al. (1951) with crystalline bovine serum albumin as a

standard. In order to measure low amounts of protein (less than 10g)

the method of Bradford (1976) was employed with lysozyme as a standard

(see Bio-Rad Bulletin 1069, February 1979).

Determination of iron. Iron was measured by the method of Cameron

(1965) as modified by Campbell and Zerner (1973). All glassware was

soaked in 4M HC1 and rinsed several times with iron-free water to remove

metal contaminants. Protein samples, blanks and ferrous ammonium

sulfate standards were treated with 70% (v/v) perchloric acid for 30

minutes at room temperature to release iron from the protein. A

solution of 10% (w/v) hydroxylamine hydrochloride was added, and the

samples were left at room temperature for 30 minutes. Finally, batho-

phenantroline solution (4mg/ml), pyridine, and iron free water were

added successively to each sample and the solutions mixed thoroughly.

The absorbance of the solutions was measured at 536nm. The

concentration of iron in the protein was obtained from a plot of

absorbance versus iron concentration of the standards, which was linear

over a range of 0-4 ig.

Immunoaffinity chromatography. Sepharose CL-4B, purchased in its

cyanogen bromide activated form, was prepared according to the

instructions of the manufacturer. Protein was coupled to the matrix in

0.1M sodium carbonate, pH 8.5 and 0.5M NaCl at 4C on a tube turner.

Alternatively, Sepharose 4B was activated for 30 minutes at 0C in IM

potassium phosphate, pH 12, using cyanogen bromide (Cuatrecacas et al.,

1968) dissolved in N', N' dimethyl formamide. The gel was washed with


0.1M sodium bicarbonate, pH 8.9. Proteins were coupled at 4C overnight

on a tube turner in a buffer of 0.1M sodium bicarbonate, pH 8.9 and 0.5M

NaCl. Unreacted sites on the matrices were blocked with 1M

ethanolamine, pH 8.9 for 4 hours at 4C, after washing out unbound

protein with the coupling buffer. Finally, the gel was washed with

0.05M sodium acetate, pH 4.0, plus lM NaCl, followed by 0.05M sodium

borate, pH 8.9, plus 1M NaCl.

Uteroferrin (10mg/ml) purified as described above was coupled to the

matrix with a 30-50% coupling efficiency. When stored at 4C in the

presence of 0.1% (w/v) sodium azide, it retained its purple coloration

and acid phosphatase activity for over a year. The uteroferrin-

Sepharose matrix was employed to purify anti-uteroferrin polyclonal

antibodies (see below). Anti-uteroferrin polyclonal antibodies were

then coupled to the matrix at a concentration of 10mg protein per ml of

gel. This anti-uteroferrin immunoaffinity matrix was employed to bind

the various spleen enzymes.

Preparation of polvclonal antibodies against uteroferrin. Samples

of uteroferrin (up to 10mg) were mixed with 5ml of Freund's Complete

Adjuvant to form an emulsion and injected intradermally over the

shoulder, rear leg and abdominal regions of one side of a female lamb.

Seven days later the procedure was repeated on the other side of the

lamb. A third injection was given seven days after the second. One

month after the first injection, serum was collected by jugular

venipuncture. The serum was allowed to clot at 40C overnight.

Following centrifugation at 1000xg for 20 minutes, the serum was frozen

in small aliquots.

Polyacrylamide gel electrophoresis. SDS-polyacrylamide gel

electrophoresis (SDS-PAGE) was carried out in 12.5% (w/v) gels according

to the method of Laemmli (1970). Protein samples were solubilized in 5%

(w/v) SDS, 5mM Tris-HCl, pH 6.8, and 5% (v/v) 2-mercaptoethanol. Gels

were stained with 0.125% (w/v) Coomassie Brilliant Blue.

Electrophoresis was carried out under non-denaturing conditions

according to the method of Reisfield et al. (1962) as modified by Basha

et al. (1979). Basic proteins were resuspended in 5mM Tris-HCl, pH 6.8

and 10% (v/v) glycerol. The samples were subjected to electrophoresis

in 7.5% (w/v) polyacrylamide gels, pH 4.0 (which did not contain SDS) in

a P-alanine buffer, pH 4.5. Proteins with acid phosphatase activity

were visualized with o-naphthylphosphate substrate and fast Garnet GBC

as the coupling dye.

Binding to Concanavalin A-Sepharose (Con A-Sepharose). Enzyme

samples in 0.3M NaC1 were adjusted to pH 6.5 and loaded onto a column of

Con A-Sepharose (approximately 5ml bed volume). The column was washed

with 3 volumes of 0.1M Tris-HCl, pH 8.2, 0.3M NaC1, ImM CaCl2, 1mM MnC1

and 0.02% (w/v) sodium azide. Bound protein was eluted with warm (50-

60*C) 0.01M cc-methyl-D-mannoside or 0.1M acetic acid (Baumbach et al.,


Immunization of mice. Female Balb/c mice at 8-48 weeks of age were

injected with 17.5pg of antigen intraperitoneally (i.p.; Katzmann et

al., 1981) with antigen emulsified in 100pl Freund's Complete Adjuvant.

A subsequent injection of 17.5 pg was performed as above with Freund's

Incomplete Adjuvant, 10 days later. On day 17, a final injection was

administered in normal saline (0.9%, w/v) via the tail vein. The


antigens employed were as follows: For fusion 5, high molecular weight

uteroferrin was employed. This fusion was outlined in detail by

Baumbach (1984) and was carried out in collaboration with George

Baumbach. For fusion 6, uteroferrin (purple, Mr-35,000) was employed.

For fusion 13 the human hairy cell acid phosphatase was used.

Generation of monoclonal antibodies. The SP2/0-Ag 14 (SP2/0;

Schulman et al., 1978) non-secreting myeloma cell line was cultured in

Dulbecco's Modified Eagle's medium (DME) containing 10% (v/v) each of

heat inactivated fetal calf serum and agammaglobulinemic horse serum;

antibiotic solution (10pg/ml penicillin, 10g/ml streptomycin,0.25pg/ml

amphotericin B), and 2mM glutamine. One day-old (conditioned) medium

from logarithmic cultures was harvested and used for hybridoma medium


The method of splenocyte fusion and propagation of the resulting

hybrid cells was by established methodologies (Kennet et al., 1980; Oi

and Herzenberg, 1980). Two mice were anaesthetized with ether and the

spleens excised aseptically, trimmed of fat and then minced with sterile

scalpel blades in 20ml of DME. The tissue was then triturated several

times and the large debris allowed to settle. The splenocytes in the

supernatant fluid were then collected by centrifugation at 1000xg for 5

minutes at room temperature, washed in 15ml of DME, and resuspended in a

final volume of 5ml DME and kept at room temperature. Approximately 108

SP2/0 cells were mixed and centrifuged at 1000xg for 5 minutes. The

mixture of cells was resuspended in lml of 50% (w/v) polyethylene glycol

1000 in DME for 1 minute, then diluted to 31ml by the addition of 6ml

each minute for 5 minutes of DME-HAT (DME with hypoxanthine, 0.1mM;

aminopterin, 0.8jM; thymidine, 16juM; Littlefield, 1964) supplemented

with 10% (v/v) each of fetal calf serum and heat-inactivated

agammaglobulinemic horse serum, 40% (v/v) conditioned medium and 2mM

glutamine. Aliquots (100pl) were seeded into 96 well microtiter culture

dishes and incubated in an atmosphere of 5% CO2 at 37*C.

Two weeks later, the culture supernatant fluid was withdrawn and

assayed for specific binding to uteroferrin using a solid phase binding

assay (see below). Positive cultures were resuspended in 500pl of DME-

HAT supplemented as described above and placed in 48 well cluster

dishes. When a culture was confluent, the cells were transferred to

25cm2 culture flasks and maintained in the medium described for SP2/0

cells. When cultures were growing well, cell densities were counted

with a hemacytometer and the cells cloned by limiting dilution (Kennet

et al., 1980). The cells were diluted to yield an average of 5, 1 or 0.1

cells per 100pl. Thirty-two wells were seeded with each dilution, and

the initial cell density which subsequently had less than 33% of the

wells with growth was rescreened for antibody production. Hybridoma

cells were maintained as described for SP2/0 cells and were frozen in

the standard growth medium with 30% (v/v) fetal calf serum and 10% (v/v)

glycerol or 10% (v/v) dimethyl sulfoxide.

Purification of mouse antibodies. Immune sera from mice were

collected immediately prior to splenectomy by cutting the descending

aorta aspirating the blood into a hypodermic syringe. Nonimmune

(normal) mouse sera were collected from uninjected, healthy,

anaesthetized animals.


Culture medium from growing hybridoma cells was used as a source of

monoclonal antibody in some instances. In other cases ascites fluid was

prepared by growing cloned hybridoma cells (1 x 107) in the peritoneal

cavity 5 days after 2 i.p. injections of 0.5ml pristane (2,6,10,14-

tetramethylpentadecane) spaced 10 days apart. The ascites fluid was

harvested after the tumor had grown sufficiently, about 10-14 days. The

sample was allowed to clot at 40C, centrifuged at 1000xg for 20 minutes,

and sodium azide was added to 0.02% (w/v).

Todination of antibodies. lodination was performed according to the

method of Markwell and Fox (1978) and Markwell (Pierce Chemical Company

bulletin, 1978). Glass tubes were coated with 100pg lodogen (1,3,4,6-

tetrachloro-3x-6x-diphenylglycolluril) by evaporation of methylene

chloride. Mouse antibodies against uteroferrin or the human enzyme

(100pg) in Iml of buffer (0.02M sodium barbital, pH 7.5, 0.4M NaC1) were

added to the lodogen coated tube. Carrier-free Na125I (100pCi) was

added and the tube shaken gently every minute for 15 minutes. The

iodinated protein was separated from unreacted 1251 by dialysis (0.01M

Tris-HCl, pH 8.2) or gel filtration on a column of Sephadex G-50 (0.01M

Tris-HCl, pH 8.2, 0.2M NaC1).

Assay of antibodies. Antigen was dialyzed overnight against

phosphate-buffered saline (PBS; 0.14M NaC1, 1.5mM KH2PO4, 8mM Na2HP04,

3mM KC1, 0.5mM MgC12, ImM CaCl2, pH 7.4) and adjusted to a concentration

of 50pg/ml. Antigen was passively adsorbed overnight at 4C to

polyvinylchloride microtiter flex plates as outlined by Tsu and

Herzenberg (1980). The plates were washed three times with PBS

containing 1% (w/v) bovine serum albumin (BSA) and 0.02% (w/v) sodium


azide. Hybridoma culture medium (0.05ml) or diluted mouse antisera were

added and allowed to bind for 1 hour at room temperature. The plates

were washed as above and 50,000cpm of 125I-labeled anti-mouse IgG in

50ml were added and allowed to bind for 1 hour. The plates were washed

as above, the wells cut out and bound radioactivity measured in a gamma

counter. Immune and non-immune sera were replicate tested six times,

and culture media were tested in duplicate.

Affinities of antibodies. Monoclonal antibodies against uteroferrin

or the spleen enzyme were affinity purified from either cell growth

medium or from ascites fluid on uteroferrin immobilized on Sepharose 4B.

The procedure was identical to that described earlier. The purified

antibodies were serially diluted in PBS containing BSA (Img/ml) over a

range of concentrations extending from 10-6 to 10-10M. Uteroferrin or

the spleen acid phosphatases were plated out into wells on flexvinyl

plates as described in the previous section. After the wells were

washed and the remaining adsorption sites were blocked with BSA, the

plates were dried. Appropriately diluted solutions of antibody (0.05ml)

were then added to the wells and after 1 hour they were withdrawn and

the plates were washed. The amount of antibody that bound at each

dilution was then measured by addition of a constant amount of a second

labeled antibody. This antibody was 125I-labeled sheep anti-mouse IgG

(see below). Its specific activity was approximately 106 cpm/pg. After

1 hour, this solution was withdrawn, the plates were washed and each

well was analyzed for radioactivity separately in a gamma counter. All

determinations were carried out in duplicate. Controls were described

in the previous section. The binding affinities of each monoclonal

antibody to the different antigens were calculated by the method of

Scatchard (1949).

Competitive binding of monoclonal antibodies. In order to determine

whether various monoclonal antibodies bound to the same or different

sites on uteroferrin, a competitive binding assay was employed.

Uteroferrin was adsorbed to the wells of flexvinyl plates as described

earlier. Monoclonal antibodies were iodinated with 125 to a specific

activity of approximately 106 cpm/pg. Samples of antibody (0.05ml;

50,000cpm) were added to each well in presence of increasing

concentrations (10-6-10-10M) of the second unlabeled antibody. If this

second antibody failed to compete or competed poorly with the labeled

antibody for binding to uteroferrin, it was assumed to occupy a distinct

site on the protein. Three monoclonal antibodies (5.122.10, 6.21.2, and

6.22.1), which bound to uteroferrin with high affinity, appeared to

compete only poorly with each other in the above assay.


Purification of Uteroferrin

Figure 2-1A shows the fractionation on Sephadex G-100 of the basic

proteins obtained from one sample (4 liters) of allantoic fluid obtained

from a day 67 pregnant pig. Three major protein fractions were apparent

and designated Fill, FIV and FV by previous convention (Murray et al.,

1972; Chen et al., 1973). The average molecular weights of these

fractions were calculated to be approximately 80,000, 35,000 and 15,000

respectively. Fraction V contained lysozyme (Roberts et al., 1976) and

a group of trypsin/plasmin inhibitors (Mullins et al., 1980; Fazleabas

et al., 1982) and will not be discussed further. Fraction IV, purple in

0 0 4 4 4z
r-4O r. 4)O 0Q)
*e o oi o p. a
r4 4 r e-4

r- o u dX 0 o,

4-4 0 a ) a a CP
4 0 U 0 -4 40 4) 4 3
4- '* 0 ) -i > )O 4 d
Ol C U. ti B C -0

Q p oa) r. ta 10 Co
-AC 4 Ca.04 0 4 )

0 ( o 0 4
*1-4 15 0 4 I 1 cU o)
S43 U O r ;: 44
40 CU bO >M Ct 4-4 4 w
r-4 r ;o Cu P3 od Pa

od C o ( B B c( o1 04

0 M 4 0 ) C ) 4) C

-4 r-4 4 W 4 4) Cn
*r c u 1 0 wi 4 bo It
* 4 4 0 It4
c 4 a u ) CC c a 4-

0 > 0 h .14

B43 tB O 4 ) 41 0 4->r
a ia0T W34) 0jJ4
W- N 1 3- CA O4J r-

441 0 r. > w CU) cd
0 -.0 C 1 0 ) 0d p 0 C
4) a o a w 4 to b Q
10 0 aa (i u o t3

0i co *O 41 P. 4d x 0 z
-A 4 0 ( ) 4) c ) Co O
P 0 W O 4- O
0 c a 0 o4 *( C

ca ) 1 t- r.-04 0 wc
,4o Ou 0 r r 0 m

44 *i T or> 0 .0
4-H 4) Ni Oc o
p tf irl C ) 3 aQ Z

Ij co 00 0 0
03 bq V 43 S O aC C

.rj 01 0 0 1 cd C
06b o l 00 r4 u z o
*-i a4 o a8 4 (U U 0 4 -

a c o 00 oi 1 u o 0
oaMcr 4 < 0t 1- 3 40

a9 aa )44 O ON
e4l -1 4z 0

Wi Cl r4l 0 4) 1 0 0
11L Iu a a <1 gn E ou 4 u

Eb Q.<)- r UicM OQ M 00 U43J

(0) SgS V

S6 0



(6) 0' t

0 0

- d 0 0




ot v

color, contained uteroferrin. When FIV was pooled, dialyzed, loaded

onto a column of CM-cellulose, and eluted with a linear sodium chloride

gradient, uteroferrin eluted as a single symmetrical peak (Fig. 2-1B).

Samples with an A280/A545 ratio of less than 14.0 were pooled and

considered to be pure uteroferrin (Buhi, 1982). In this typical

preparation of uteroferrin from allantoic fluid, about 135mg of pure

uteroferrin were obtained, which is considered to be an average yield

(Roberts and Bazer, 1980). Fraction III had an obvious pink color and

was demonstrated to be a high molecular weight form of uteroferrin,

consisting of uteroferrin non-covalently associated with a second

protein whose function is unknown (Baumbach et al., 1986). The yield of

the high molecular weight uteroferrin, pooled immediately after the gel

filtration step, was 47mg. The amount of Fill obtained varies greatly

between animals, with uterine secretions from early pregnancy or

pseudopregnancy containing more Fill than those from late pregnancy or

pseudopregnancy (Baumbach et al., 1986). The specific activities of

uteroferrin and high molecular weight uteroferrin from this preparation

were shown to be approximately 230 units/mg protein and 106 units/mg

protein, respectively. The total recoverable acid phosphatase activity

from the allantoic fluid of this day 67 pregnant pig was greater than

3.6 x 104 units.

Purification of the Human Type 5 Acid Phosphatase from Hairy Cell Spleen

Levels of acid phosphatase in spleens from a patient with hairy cell

leukemia and from a normal individual are presented in Table 2-1. Note

that the spleen of the patient with hairy cell leukemia was greatly

enlarged. In addition, the specific activity of the extract from the


Tartrate insensitivity and activation by 2-mercaptoethanol of acid
phosphatase activity in spleen homogenates from a patient with hairy
cell leukemia and from an individual with a normal spleen

Weight of

Addition to

Patient Spleen Phosphatase Assay Specific Activity
kg units/mg protein
1 2.3 None 0.453
0.01M NaK tartrate 0.483
0.1M 2-mercaptoethanol 0.662
None 0.011
2 0.22 0.01M NaK tartrate 0.007
0.1M 2-mercaptoethanol 0.012

Results from spleen 1 (weight 2.3kg) were obtained from 30,000xg
supernatant fraction of a slice (250g) of tissue that had been frozen
(-80*C) for 2 weeks prior to assay. Results from spleen 2 were obtained
from the 30,000xg supernatant fraction from freshly homogenized tissue
of the entire organ which weighed 0.22kg. NaK tartrate is Rochelle's
salt and corresponds with the L-(+)-form of tartrate.

large hairy cell spleen was about 40-fold higher than that from the

normal spleen. Moreover, the acid phosphatase activity from this spleen

was unaffected by the addition of 10mM tartrate but was stimulated by

the presence of 0.1M 2-mecaptoethanol. By contrast, tartrate caused

almost a 40% inhibition of the total acid phosphatase activity in the

extract of normal spleen, and there was only a slight activation with


Two procedures have been employed for purification of the tartrate-

resistant acid phosphatase in the extract from the hairy cell spleen.

The first is illustrated in Fig. 2-2 and employed a gradient salt

elution from a column of CM-cellulose (Fig.2-2A), gel filtration on

Sephadex G-100 (Fig. 2-2B) and, as a final step, immunoaffinity

chromatography (Fig. 2-2C). The acid phosphatase, which is very basic,

was eluted from CM-cellulose at a salt concentration of 0.25M. Upon gel

filtration the enzyme emerged as a peak of apparent Mr-34,000, about one

fraction behind the elution position of highly purified uteroferrin.

The fractions containing the enzyme were pooled and loaded directly onto

an immunoaffinity column consisting of anti-uteroferrin antibodies

covalently linked to Sepharose 4B. All of the phosphatase activity

bound and could be quantitatively eluted using low pH buffer.

In the second procedure, the dialyzed extract of the spleen was

applied batchwise to CM-cellulose at pH 8.2 and the acid phosphatase

activity was eluted with 0.5M NaCl. Rather than subjecting the enzyme

which had bound to further CM-cellulose ion exchange chromatography and

gel filtration, the preparation was loaded directly onto the anti-

uteroferrin immunoaffinity column. This abbreviated procedure provided

4) Ln tri t
NO Vo a> C
N C00 0
O .0 M 4 H0
44l o 4 z C () eM
o *4 0 0 0 ? ,C A 4)m
0 4 1 -) ( P 4-)

S( 1 41) () M 44 o 0 P. W 4) 1 o >'i
o ) iLj r. C4 c 3 44i o 4J CO C 4 a
-4 4- U -l 0 0 bO M
S$ -r O P64 b P -4 r. c
-i or >t o 0o a 0o -Hc
,rl -4 ol c cd C 0) 4- 0 0
o u M o bO) 0 CO *U A 44 > E -, >-H
* n d 4CU C -H- 4J ) 0 3 4-J bO 4)
i 4 44) 4 0 Pd () .4 0) C

UCN l O3 S OP- 30U ue-4 -)
*4) 4 0o opca s 4 u P -
r. $- -I 0 'O Bil z e o w N' '0 4 u
N 0i a O r bO r. w 0d 4 -
ca (L ) o O o -4 4) co wY i
- 4 4 U -l 4 C U N 0 M
1C'4 00 00, )
cc 3 41) a) -4 B) rl (B 4> co
Sw 3 0 a U E r 1 lC
CL 4 > r- o bO (
Sr-4 c 4-r4 4) r4 -) q 0
0 G- Mo c a O r. C -
SA* o c 0 Q ) *-iu, a) N
O 0 M v (0Cd 0 W 44 W
4C C ()d a 4 a r-4 4 a) W'-
00 O Q9 ri wr-4 a4 O..C 44 Z
o 41 N C .4 oo0 i- a) 0o 0 i
r I) *C 4) ma CM M4 w 1
0 o.0 r- C C ) 0) to O l
mn 4-1 4 0l 4 3 O) Ca P. U 441 4> 4JG
H M -H 04 0 1o 3 *H-i
s 0 r. o Z r. u
C*a r- 0 ,) u "4 >0 0
tf vl i o o ,f Bo u c-i 4 iO
oM A 4- o C' cl 04 4 r O
MP4 ) 1-1 O C 0 M 0
4 MW 4 U C D C
0 j 10 U 4 44 HCU
0 A 4) 0 4 Q -A 0 H 0 I 0B

-i 0 a ) l 0 o4- 0 -i 4 4- i )
CO W- -r-4 bO-4 U N 4J- C P. f X U 0 3 01)
*U 40 C 3 0 O B o r-4 4
4-4 04- 4-U r- 3 O 4 c)
- cUfl O r-4 0) 4-4 44 Q 0 r- -4

,C 4J )-l U 4J > Q )

Ui U) -r-4 00 bO 0 4J W 0 M 0
a)o 4 L C CU -
b 40 ) r.4 0 r
C%4 r. r 1 C "4 0 4) 03 0 C
iOr-l N 0 3 -

i4 x 0C 00 0cB 0 0 0 0 0 4 ,-4
WC 0 Z O'.- o uM4 Cm U 0

c~ii~~ O S.2~Nf


A Ni


OZy 4

(C) 3WAZN3 J0 SiINn

r p_
'0 .i~ -*
jo e 0 C

(*)3WNZN3 j0 SlINn

"l [wi 1)eN





a greater than 500-fold purification from the extract of the hairy cell

spleen within 24 hours (Table 2-2). The final specific activities have

varied between about 100 and 500 units/mg of protein.

Major losses of activity occurred whichever procedure was employed.

Collection of the enzyme on CM-cellulose was particularly inefficient.

However, if this step was repeated several more times (results not

shown) the percent of recovery of enzyme from the original extract could

be increased to over 40%. The phosphatase activity in the 30,000xg

supernatant fraction from the spleen extract was relatively unstable.

Storage at 4C led to a complete loss of activity within 48 hours.

Samples kept at -20C lost over 60% of their activity in 20 days. The

purified enzyme, on the other hand, appeared to relatively more stable

when frozen at -20*C, and samples have been stored successfully for up

to 2 months with retention of most of their original activity.

When the immunoaffinity-purified enzyme was analyzed by SDS-PAGE it

gave a major protein band of apparent Mr=34,000 and two minor bands of

about Mr-20,000 and 14,000 (Fig. 2-3). The relative amounts of the

higher molecular weight species and the two lower molecular weight bands

varied between preparations from the same spleen (results not shown).

The purified enzyme was also examined by electrophoresis under

nondenaturing conditions (Fig. 2-4). The gels were then "stained" for

acid phosphatase activity. The hairy cell enzyme (lane 3) gave two

bands of activity, a minor fast migrating species and a dominant slower

band. Uteroferrin also gave 2 bands, but the slower migrating band in

this case was scarcely visible (lane 2). A commercial standard for


Purification of tartrate-resistant acid phosphatase from
spleen of a patient with hairy cell leukemia

Purification Total Total Specific Fold
Step Protein Activity Recovery Activity Purification
mg units % units/mg X
30,000xg 5,060 3,230 100 0.64
CM-cellulose 3.4 674 20.9 198 309
Affinity column 0.41 145 4.2 358 560

A 212g piece of a 2.4kg spleen was homogenized in lysis buffer. The
homogenate was centrifuged at 30,000xg and the enzyme was purified from
the supernatant fraction. Activities represent enzyme activated with
0.1M 2-mercaptoethanol at pH 5.3 for 10 minutes.

1 2 3





Fig. 2-3 SDS-polyacrylamide gel electrophoresis of purified
uteroferrin and hairy cell phosphatase. The scale on the left is
molecular weight x 10-3. Lane 1, molecular weight standards; lane 2,
purified uteroferrin; lane 3, hairy cell phosphatase following
immunoaffinity chromatography. The gel was 10% (w/v) polyacrylamide; it
was stained with Coomassie Blue. Note that the hairy cell enzyme gave
major bands at 34,000, 20,000 and 14,000.

1 2 3

Fig. 2-4 Polyacrylamide gel electrophoresis of acid phosphatases at pH
5.4 in P-alanine buffer. Gels were stained with phosphatase substrate,
c-naphthylphosphate, and hydrolysis product coupled with Fast Garnet GBC
salt. Lane 1, Type 5 isozyme electrophoretic standard from Calbiochem-
Behring; lane 2, uteroferrin; lane 3, hairy cell enzyme immunoaffinity
purified from spleen. Note that a standard isozyme 5, uteroferrin, and
the hairy cell phosphatase gave similar, but not identical,
electrophoretic patterns. Migration (towards the cathode) was from the
top to bottom.

human isozyme 5 is shown in lane 1. Its pattern resembled that of


Purification of the Type 5 Isozyme from Normal Human Spleen

An isozyme with a high cathodal mobility has been detected in

extracts of normal spleen by others (Lam et al., 1973). In addition,

the extract of spleen described in Table 2-1 contained significant

amounts of a tartrate-resistant acid phosphatase activity. Therefore an

attempt was made to isolate the enzyme from the total extract of a

normal spleen by means of procedure 2. Approximately 16 units of enzyme

activity bound to CM-cellulose at pH 8.2. This represented less than 5%

of the total acid phosphatase activity in the extract. A total of 6

units were recovered from the immunoaffinity column. This final

preparation was insensitive to L-(+)-tartrate and was activatable by 2-

mercaptoethanol. Upon electrophoresis in nondenaturing gels it gave two

cathodally migrating bands of activity. These were identical to the

bands of activity seen with the hairy cell enzyme (results not shown).

The results confirm that the Type 5 phosphatase is present as a minor

component of spleens of normal individuals.

Purification of the Type 5 Phosphatase from Human Placenta

Because of the difficulty in obtaining human spleens, it was

necessary to determine whether other tissues could be used as a source

of the human tartrate-resistant acid phosphatase. Human placenta can be

obtained fresh, in large quantities, at any time. Therefore, this

tissue was tested for the presence of the enzyme. Two placentas were

studied, and a slightly different purification protocol was used for

each one. A tartrate-resistant acid phosphatase was purified from each

placenta. Figure 2-5 shows that the placenta enzyme, like the spleen

enzyme, has an apparent Mr of 34,000. Interestingly, the two minor

bands found associated with the placenta enzyme are of different

apparent Mr values than those associated with the spleen enzyme. Human

placenta can contain about as much Type 5 phosphatase as normal human

spleen (Table 2-3) when measured per gram of tissue. Note that human

hairy cell spleen contains at least 20-fold more Type 5 isozyme per gram

of tissue than the other sources of the human enzyme.

Purification of the Bovine Spleen. Bovine Uterine and Rat Spleen

Table 2-3 also shows the yields of the purple, iron-containing

phosphatases from other sources. Davis et al. (1981) claim that they

can purify up to 2mg of purple phosphatase per 2kg of beef spleen, but

that the yield varied greatly between animals. The yields obtained in

this study were consistently lower than 0.5mg/kg (data not shown). The

yield of the bovine uterine enzyme was considerably less than that of

the porcine enzyme on a per animal basis (Table 2-3). The yield of the

rat spleen enzyme was quite high (Table 2-3) and comparable to that of

Hara et al. (1984).

pH Optima of the Type 5 Phosphatases

The pH optima for the various Type 5 phosphatases are listed in

Table 2-4. The values range from pH 4.2 for the bovine uterine enzyme

to pH 6.0 for the bovine spleen enzyme.

Glycoprotein Nature of the Type 5 Phosphatases

Porcine uteroferrin is known to contain a single high-mannose

oligosaccharide chain (Saunders et al., 1985). The rat spleen enzyme

(Hara et al., 1984), and the bovine spleen enzyme (Davis et al., 1984)

1 2






3 4




Fig. 2-5 SDS polyacrylamide gel electrophoresis of uteroferrin, the
phosphatase from human spleen and the phosphatase from placenta. The
molecular weight standards are shown in lane 1. Note that the spleen
enzyme (lane 2) gave major bands at 34,000, 20,000 and 15,000, while the
placenta (lane 4) gave major bands at 34,000, 16,000 and 13,000.
Uteroferrin (lane 3) has an apparent Mr of 35,000. Approximately 20pg
of protein were present in each lane. Gels were stained with Coomassie
blue. The gel was composed of 12.5% (w/v) polyacrylamide. The scale on
the left is molecular weight x 103.


Levels of tartrate-resistant acid phosphatase
activity in various tissues

Units of Enzyme Units of Enzyme Activity
Source of Enzyme Activity per 100g Tissue

placenta 1 (630g) 19 3.0

placenta 2 (650g) 7 1.1

Normal human
spleen (220g) 6 2.7

Hairy cell spleen
(212g)a 145 68.3

Beef spleen (200g)a 4.5 4.5

Rat spleen (2g)b 0.74 37

Bovine uterine fluids
(400ml)c 224

Porcine allantoic
fluids (4L)c 36,400

aApproximately one-tenth the total tissue mass
bObtained from three animals
cObtained from one animal

Purified enzymes were incubated with 0.1M 2-mercaptoethanol and
assayed at their pH optima with 20mM p-nitrophenylphosphate. One unit
of enzyme hydrolyzes the release of lmol of p-nitrophenol per minute.


pH optima for the tartrate-resistant acid phosphatases

Enzyme pH Optimum

Porcine uteroferrin 4.9
Human hairy cell enzyme 5.3
Bovine spleen enzyme 6.0
Bovine uterine enzyme 4.2
Rat spleen enzyme 5.0-5.8
a Hara et al., 1984

Enzymes were assayed with 0.1M sodium acetate, pH 3.0-5.4, 0.1M
Tris acetate, pH 5.6-6.5, 0.1M Tris-HC1, pH 7.0-7.5, buffers in the
presence of 0.1M 2-meracaptoethanol with 20mM p-nitrophenylphosphate as

were demonstrated to be glycoproteins. In this study it was

demonstrated that the human spleen and placenta enzymes bind avidly to

Con A-Sepharose. The enzymes could be eluted with 10mM c=-methyl-

mannoside at 50-600C or with 0.1M acetic acid. These results indicate

that the human Type 5 phosphatase, along with the purple, iron-contain-

ing phosphatases, contain high mannose or hybrid-type oligosaccharide


Iron Content of the Type 5 Phosphatases

It has been demonstrated that the bovine spleen enzyme (Davis et

al., 1981) and rat spleen enzyme (Hara et al., 1984) contain two atoms

of iron per molecule of protein. However, Buhi et al. (1982b)

consistently demonstrated that certain preparations of uteroferrin bound

about 1 mol Fe/mol protein. Table 2-5 compares the iron contents of

pink, high molecular weight uteroferrin, purple uteroferrin, the bovine

uterine enzyme, the rat spleen enzyme, the bovine spleen enzyme and the

human spleen enzyme. While pink, high molecular weight uteroferrin

contains two atoms of iron per molecule of uteroferrin, purple

uteroferrin preparations contain considerably less iron. Triplicate

determinations on a single purified sample of the human enzyme revealed

that it contained 2.2 + 0.4 atoms of iron/molecule of protein.

Activation by Reducing Agents

Reports on the effect of reducing agents such as ascorbate and

2-mercaptoethanol on the activities of the Type 5 isozyme of human

spleen have been at variance (Lam et at., 1977; Robinson and Glew,

1980). The degree of activation has been unpredictable throughout the

course of this study. With partially purified enzyme


The iron content of the tartrate-resistant acid phosphatases

Source of Enz me

mmnl Fp/mmn1 Prntpjn

d65 allantoic fluid,
purple porcine Uf

d65 allantoic fluid,
pink Mr-80,000 Uf

dll0 pseudopregnant fluid,
purple porcine Uf

dllO pseudopregnant fluid,
pink Mr-80,000 Uf

Human hairy cell spleen

Bovine spleen

Bovine uterine fluid

Rat spleen

aBaumbach et al., 1986
bDavis et al., 1981; Campbell
CHara et al., 1984

et al., 1978

The iron content was determined by the method of Cameron (1985) as
modified by Campbell and Zerner (1973). Approximately 50-100pg of
protein were used.






Source~~... ofEn emml --o Prot-ei-


(post-CM-cellulose) a 2-3-fold activation was usually observed when

enzyme was incubated with 0.1M 2-mercaptoethanol for 10 minutes at

pH 5.3 before being assayed for activity. The degree of activation of

crude enzyme (Table 2-2) was usually less than two-fold. However, some

preparations of affinity-purified enzyme have shown negligible activity

in the absence of reducing agent while other, more inherently active

preparations were only stimulated about two-fold. The results with one

purified preparation are summarized in Table 2-6. The extent of

activation was clearly dependent upon the concentration of

mercaptoethanol. Results with ascorbic acid (not shown) were similar.

Activation by mercaptoethanol leads to an increase in V rather than

a change in Km for both the human spleen enzyme and uteroferrin (Fig. 2-

6A). Note that the Km values for both enzymes at their pH optima (pH

5.3 and 4.9, respectively) are similar. Values for Km between 0.75 and

3 mM have been reported for the spleen enzyme. V is also variable, the

more active preparations giving values greater than 500 units/mg protein

when pretreated with 2-mercaptoethanol.

Interestingly, pink, high molecular weight uteroferrin does not

require pretreatment with 2-mercaptoethanol in order to exhibit maximal

activity (data not shown).

Effects of Inhibitors on Phosphate Activity

The effects of a range of inhibitors on the human spleen enzyme and

uteroferrin are very similar (Table 2-7). Both enzymes are completely

inhibited by 0.1M dithionite, a reagent which is known to cause release

of Fe from uteroferrin (Schlosnagle et al., 1976). Hydrogen peroxide

and agents which interact with -SH groups also inhibit. The chelators


Activation of purified human hairy cell phosphatase by 2-mercaptoethanol

2-Mercaptoethanol Activity
M % control
0 100
0.01 240
0.05 300
0.10 360
0.25 440

Enzyme was incubated with 2-mercaptoethanol for 20 minutes at pH 5.3
in 0.1M sodium acetate buffer prior to assay.


Action of various potential inhibitors on the hairy cell acid
phosphatase. uteroferrin. bovine spleen and
bovine uterine acid phosphatases

Compound Concentration Acid Phosphatase

Human Porcine Bovine Bovine
Hairy Cell Uteroferrin Spleen Uterine
% control

Na dithionite 10-1 M 0 0 12 9
H202 0.01%(v/v) 75 73 37 67
Na iodacetate 10-3M 72 85 NDa ND
Na iodoacetamide 10-3M 70 78 60 73
p-mercuribenzoate 10"4M 42 57 ND 50
FeC12 10-5M 45 58 ND ND
NaEDTA 10-3M 100 100 66 52
bipyridine 10-3M 100 100 ND ND
o-phenanthroline 10-3M ND 100 60 52
NaK tartrate 10-2 100 100 95 100
NaK tartrate 10-3M 100 100 ND ND
Na phosphate 5x10-3M 5 5 ND ND
Na phosphate 10-3M 40 37 43 28
Na phosphate 10-4M 73 71 ND ND
Na arsenate 10-3M 18 8 ND ND
Na arsenate 5x104M 36 19 ND ND
Na arsenate 5x10-5M 83 74 67 29
Na molybdate 10-4M 8 0 ND ND
Na molybdate 10-5M 4 12 ND ND
Na molybdate 5x10-7M 38 63 76 41
Na fluoride 10-3M 60 67 53 ND
Na fluoride 5x10"4 80 78 ND ND
a ND, not determined

The enzymes were assayed in triplicate with p-nitrophenylphosphate as
substrate (concentration 20mM). The enzymes were assayed at their pH
optima (see Table 2-4). All activities are presented as per cent of a
control assay performed simultaneously with no added inhibitor.

EDTA and bipyridine have no effect. Tartrate does not inhibit either

phosphatase. Inhibition is observed with phosphate, arsenate,

molybdate, and fluoride, however. Molybdate is the most effective,

inhibiting strongly even at concentrations below 10-6M; it appears to

inhibit both enzymes in a noncompetitive manner (Fig. 2-6B).

Arsenate and phosphate are competitive inhibitors (Fig. 2-6C).

Inhibition with 10-3M fluoride (results not shown) is complex, and

curvilinear Lineweaver-Burk plots are observed.

The effects of these inhibitors on pink, high molecular weight

uteroferrin (data not shown) and the bovine uterine and spleen enzymes

were quite similar to the results obtained for uteroferrin (Table 2-7).

Substrate Specificity for Phosphatase Activity

The substrate specificities of the human spleen enzyme and

uteroferrin are quite similar (Table 2-8). p-Nitrophenylphosphate was

the best substrate tested, but o-naphthylphosphate, pyrophosphate, and

nucleotide tri- and di-phosphates are also hydrolyzed. Both enzymes

hydrolyze ADP more effectively than ATP, whereas AMP is a poor

substrate. Hexose phosphates are only very slowly hydrolyzed, if at


Both enzymes are phosphoprotein phosphatases and can release

orthophosphate from the egg yolk protein phosvitin. The human spleen

enzyme (lpg) had the ability to release 2.1nmol of orthophosphate from

phosphvitin (lOmg/ml) per minute at 37*C. The value for uteroferrin

assayed identically was 1.5nmol of orthophosphate released per minute.

The substrate specificities of pink, high molecular weight

uteroferrin (data not shown) and the bovine emzymes (Table 2-8) were

1 t)

4C4) 0 0 ) > N 0 t
p) a wi I- 4) '0o mco ) M (i
M 4-)A H p ( 0 m u) x bo d 4- co
SP) c d a) 4
cal U 00) 4 U ) 0 0)3 m
13 w1 ) 0c o4 Q a > 4 )#i
p3. 0 a n t0 4> 4) '0o $. w0
aM 4 rI J a )O 40 a0 O
a..,-i Na) ,o as p-4 4 aN4
.. 0 C Q t 44- a) 03 0 to C M i-1
-4 c 4) 4-) 4 o o (B 4 a) U P,4 4
4)' a) r. -3 ) ca O4t C$ r r- 4 4J
S"A 0 Q C c a) U-( ct
:1 0 W 4)or4 C J 4J ta, ..i
o I n a m a a w m u, a
*-10 p 1 a) a ) n 4)
W -r4 r) 4) C -i U4 S. 0oV) 0 p 4-) -H a >
" p .44 N P- CO ) (0U 4J 0 0- >' W) a d3
0 a) c o QO 0 O > 0 P4
r3 > 0 w H -H T0 4 0! 0 W n a-
$ ( () .4-) 4 C4-*. toUS )
aJ 0 0 `4 U U UOl 0 4-4 4 41
C rS.- -4 9:r.l c > 1) V T ) (lU a) 0 0 *,4
CO C 4) O --- p i S A 0 r a ) *-1
4-) u a o)00PUS 4'a -4oA

p-1 w 41 r-~ .c,^ u)4 .0 14. 4) > g-
C 0 N 4- ) --
e) C3 3 o C 0 r a .w 0 41 8

0 ) O 0Ii O )-r4 V) 0 4 0
w0 0 w 0e 4o H (
a r. Cd 0 .3 W C 4 ) P 4- >-tZf r-a 4)
4-a I.4 Cd 0. 0 >-a ) 4-U W -H 0. M
S O r-4.-4C 4i A4 44 C m W 0 4 0
44 w ir C C H > "-1 "4 0 44 Ca 4 "4
0 0 W) 3t 4 4 4' W- 0 W- 4.) 93
a4J AL 3 0 1) 0 uJ a4 wa u A)
n 1 P 4- 44 P. 0 4) 4 4)) *0

4 a3. CO a) -Ua> )C -4 U4 00 U
0 0 0 P P O4 O9 r
4UN al 4- 4 N 4 ) q 0
P 0 W Ia C co C. A 0 WU
: C a 3a M U S P 4.)
s ha I o 44) Z3 44i n
) C4 0 >oP 1 41 0 3

S 1-)-4t o 3 amB 0 C4 1-4 P. 0 t
okam 0 0 Cd > aU
S4-) 0 0 U t) 0 4 ) 41> U )
r3 Z 44-l W C W aU.CL 0 0 4-) H -3
-H r.C 4 41 +-) 4) ao : UO)
c ) () w 0 3 ) X 4) 0 T-1 4J WH 4J
U) r. e > d (A "4 %-" -i 3 .0
) O 30 0 10 4 r-4 0 CB P O 0 oa
%. l -H r N 41 0 0 4<1 P a 4B < )
0 -H (fa (l 4) 4) 4-> ) Iz uB o3 ) 4-)
> >> >Na w P C r.D -Hl3
*- *4 *4 CB- 4- B 0 *4 ) ) 4 ) 0
S1 4-> 4) ( "4 U) i- a>) I 0 -4 M W WI -r4 U
*4 U u O 4 w C 0 0 0 C i aU) 4 3 C3 a)
4 Cd ct I) C3 41> a > 41 &0. tB 2 n 4 U Q













0 *


Comparison of substrate specificities of hairy cell phosphatase.
uteroferrin. bovine spleen and bovine uterine phosphatases

Substrate Acid Phosphatase
Human Hairy Porcine Bovine Bovine
Cell Uteroferrin Spleen Uterine
% control
p-nitrophenylphosphate 100 100 100 100
c-naphthylphosphate 71 26 23 84
sodium pyrophosphate 24 52 7 45
ATP 49 36 20 3
ADP 92 81 18 87
AMP 4 2 3 10
D-glucose 6-phosphate 3 0
D-mannose 6-phosphate 0 3
D-fructose 6-phosphate 4 0
All compounds were assayed at a concentration of 4mM at the pH
optima of the enzymes.

generally similar to the results obtained with uteroferrin. However,

the bovine uterine enzyme did not hydrolyze ATP well, and the bovine

spleen enzyme was virtually inactive against pyrophosphate. The fact

that each bovine enzyme was tested at its pH optimum (pH 6.0 for the

spleen enzyme and pH 4.2 for the uterine enzyme) rather than at the same

pH, may have influenced the results.

Immunological Cross-reactivity

The human spleen enzyme clearly shows at least some immunological

cross-reactivity with uteroferrin since it bound to anti-uteroferrin

immunoglobulins immobilized on Sepharose (Fig. 2-2C). The extent of

this cross-reactivity has been measured in a solid phase immunobinding

assay which employed whole antiserum from mice, which had been immunized

with either spleen enzyme (Fig. 2-7A) or uteroferrin (Fig. 2-7B). The

antiserum to the spleen enzyme clearly bound to uteroferrin but binding

was reduced by 50% when the antiserum was diluted 1:80. No binding

could be detected below a 320-fold dilution. The titer towards the

spleen enzyme was clearly much higher, with half-maximal binding

occurring at an antiserum dilution of 1:1280.

When whole mouse antiserum raised against uteroferrin was employed,

little antibody binding to the spleen enzyme was detectable at an

antiserum dilution below 1:320. By contrast, binding to uteroferrin

remained high down to an antibody dilution of 1:10,240. Therefore, the

two proteins show only partial immunological cross-reactivity.

The Production of Monoclonal Antibodies Against Porcine Uteroferrin

The mice for fusion 5 were injected with high molecular weight pink

uteroferrin. The results of this fusion, which was done in

Fig. 2-7 Solid phase radiobinding assay of whole mouse antiserum
from mice immunized with human spleen phosphatase (upper panel) or
uteroferrin (lower panel). Antibody was diluted serially and
tested for binding to either uteroferrin (0), human spleen
phosphatase (0), or bovine serum albumin (&). Binding was
measured by means of affinity purified 12 I-sheep anti-mouse IgG.

o spleen phosphatase
a bovine serum albumin

40- Mouse anti splei



x 0
" Mouse anti Uf


* 80

$1 60 -

10 40 160 640 2560 10240
Antiserum dilution

collaboration with George Baumbach, are outlined in detail in Baumbach


The mice for fusion 6 were injected with pure uteroferrin. Upon

initial screening with uteroferrin as the antigen, 26 out of the 199

colonies screened were identified as positive. After each of these

positive colonies was cloned by the limiting dilution method, 11 of the

original colonies contained positive clones.

Three clones from fusion 5 and three from fusion 6 were chosen for

further studies. These hybridomas, 5.58.1, 5.122.10, 5.127.3, 6.21.2,

6.22.1 and 6.37.4, were used to produce large quantities of antibody in

mouse ascites fluid or culture media.

These antibodies were tested for binding to uteroferrin and pink,

high molecular weight uteroferrin. All of the antibodies bound with

high affinity to pink, high molecular weight uteroferrin. Only one

antibody, 5.58.1, did not bind to uteroferrin. It appeared that 5.58.1

recognized the second, antigenically unrelated protein in Fill. This

antibody will not be discussed further.

In order to determine which of these antibodies recognized the same

epitope on the uteroferrin molecule (i.e., would compete with one

another for binding to uteroferrin) and which recognized different

epitopes (did not compete with one another for binding to porcine

uteroferrin), competitive radiobinding assays were carried out.

Each of the five monoclonal antibodies was iodinated to a specific

activity of approximately 106 cpm/pg. Each iodinated monoclonal

antibody was tested in a competitive binding assay with the same,

unlabeled antibody. This was done to ensure that the binding abilities

of antibodies were not destroyed by iodination. Each labeled antibody

was then tested with each remaining unlabeled antibody. The results of

several of the competitive binding assays are shown in Fig. 2-8.

Antibodies 5.122.10, 5.127.3 and 6.37.4 compete with one another for

binding to porcine uteroferrin, even at very low concentrations of

unlabelled antibody. Thus, these three monoclonal antibodies presumably

interact with the same site on the uteroferrin molecule. Monoclonal

antibodies 6.21.2 and 6.22.1 do not compete with each other, nor do they

compete with 5.122.10, for binding to porcine uteroferrin. Each of

these antibodies presumably recognizes a different site on the

uteroferrin molecule.

The Production of Monoclonal Antibodies Against the Hairy Cell Spleen

Hybridoma colonies derived from a mouse which had been immunized

with the human spleen enzyme were screened for the production of

antibodies which were able to bind both the spleen enzyme and to

uteroferrin. From 200 wells containing live cells, 44 positive colonies

were selected. When these positive colonies were rescreened, 29

remained positive. Only two of these colonies, 13.15 and 13.122,

produced antibodies which bound as well or better to uteroferrin than to

the human spleen enzyme. The dissociation constants for the binding of

antibody 13.15 to the human enzyme and uteroferrin, respectively, were

22 and 16nM. The dissociation constants for the binding of antibody

13.122 to the human enzyme and uteroferrin, respectively, were 40 and

12nM. These two antibodies did not compete with one another for binding

the porcine uteroferrin.

.r -4
1 o ,o-1 0
> ,C r-4

\,0 0 4) 1-1 0n b0-A4 4-1 r-4lr r-
U) 4 4 M .* '0
m 0 o 7 c -
00 0 r- 44 ) 4 1
,M Ln 41 > o ,4
*-rl -' 4- > r-4 c bO
C ) > 44 Cf 0o *4
r 4 r-4 -I 0 t C 4 10
-4 P4 40 0 r-14 0 0

04 W --0 1 r-1 vI
r- ) < 4 -* a 4 0
0 0 4 0 1 0
0 41 *rd CA 0 l 41 bO
Cr :) 0 0 C0
0 4 *
S 4) P 0 *b0 *
Coi t a o a o r IN
C0 04) -A -r4 C-4
A r4-4 r-4 r-4 rd -r 4 r-4
P3 r-1 Ia. 0 4" 0C
S4) 0 : m -4 4) UL
w r-4 4O *r-4 l 0

4o-4 0 P4 a 4
S4) 4p 0 o 4-4 4a) 41

PU 4 '0 0 b u P 40
r. r 44 t- L t) C r--I *H 0
4c> a o B op
44 r4 ) 0 C Ui c s-r (c
Sr-I D 4 r- 4-)
*r0 -4 O r-4 )O C) cN
4 > r-4 > 4) -4

Cf 0 0' e i (lfl' r*
S34)4 ) 04 o C *s-4
In () Q 0 r C) 44$
0 0 .14 r 4 0 Cd'
)4 l ca 0
> 0 0 *o -
*r4 o r. ) -4 -4
W O(1 4) 3cn O D -
S 0 *p 0 r-4 C4 l0
4) O 3 ) .-4
ul m --I '0 (1) 1 N C 4
0 0 *rl *f T} c tf l
0 40 c- N04 d CU 4 CM -
0 m 0 0 O C0 *-

w 0 4 ,0 0) 0 C- QO

P Lr) ri *i 4 0 4) C O r-I B1 41 r-4 Cfl 4)
*4 *1 ;:. r 00 r. *

0 1



0 I


(wdCOL X) CNOs i,,,

The Binding of the Monoclonal Antibodies to Other Tartrate-resistant
Acid Phosphatases

In Fig. 2-9 the three monoclonal antibodies raised against porcine

uteroferrin have been compared on the basis of their binding to

uteroferrin, the human enzyme, and the purple, iron-containing rat

spleen and bovine spleen acid phosphatases. All three antibodies bound

to uteroferrin with very high affinity (Fig. 2-9 and Table 2-9) but

bound relatively weakly to the beef spleen and human enzymes. The rat

enzyme had an intermediate affinity (KD=33nM) for antibody 6.21.2 and a

relatively high affinity (8nM) for antibody 6.22.1. The binding of the

anti-uteroferrin antibodies were also tested for binding to pink, high

molecular weight uteroferrin and the bovine uterine enzyme. There is no

difference in the binding of these monoclonal antibodies to uteroferrin

and pink, high molecular weight uteroferrin (data not shown). The

bovine uterine enzyme had a high affinity for antibody 5.122.10 (13nM)

and low affinities for the remaining monoclonal antibodies. The bovine

uterine enzyme clearly differs from the bovine spleen enzyme in its

affinity for antibody 5.122.10 (13nM versus 120nM). Thus, the five

phosphatases showed some immunological crossreactivities when tested in

this manner, but the epitopes recognized by any one of the antibodies

were clearly not identical.

Antibody 13.122, raised against the human spleen enzyme, was tested

for its binding to the five phosphatases. Figure 2-9 and Table 2-9 show

that this antibody had a slightly lower affinity towards the human

enzyme (the antigen against which it had been raised) than towards

uteroferrin or the rat spleen enzyme. The antibody bound as well to the

bovine uterine enzyme as it did to the human enzyme, and bound most

w 4-4
SO 3 0 0 C

0 c 0 0| k 44 a-4 4)
C n l ) ) ) -, U 4
rC C N 0 C co 0 (U ,C
o ) ) r. cd 1- 4) aM 4
4 0) r-4 Cd ) 41 c b0
rl 44 bO -4C ,
0 4 o r3 O w > U 4,
cd o #0 -1 C r a) co
4-1 ) r-i z 4) 4)
.r4l W ll *r- -A M, 4 0 C U) r C
4 r- C4 I 0 0 0 0

0 oo 0 41 o 0 a
Wa) ,C 0) 4CS 4.) O Z C
)C ( a '0 4 o ;, C
0 A0 0o a 31 0 0 4 o to

z bo O (n r--I & 4-) C) 0 C
U) 0 0 CO :$ 0 H) C M CO t
o M o -4 -ri r. r -4

"Ac 0 0a O Pd .)
Uo) c Q) c M ) cc 4 e ac

-0 to ad 00 --
4J 4 l W4 CO C= C

r4 C C0- O 4 C C ) c
cof '0 4 >N r. 4 4

04 U 0l M3 4J 4) "

3 0 -- **-i 4 0 a )n
0 m -cd W *4 M 4 C0

o- t Co w 0 0 C3
U) i -4) rC 0 UC WC P CD R
0 co w 41 : 0 .
4-1 ,O Al r4 x4 0 0 -
0 C.CO0C0 0 ) ) P -A O0-4
1-o V 0 c o to Q w r-c
0fl s- I r. 0 O a r-t o 4 o 4) cqi
*dl c -d U)0 ) c.4
., -4 .) r- (3 H ( a* 34

C34 a4 C o C -
0d 0 <-- 0 co i O
0 V) 0 0 0 W3 0
4-41 o 4 C 4)
O) -3 C 0 r. () Cr -
* 3 i- o Co a>0 6 CM

0 4 -- 0- Ua 0
*0 C 'd "A o cc C-
0 a n b O I U0 V)-" ( o

00 "a wo -w 0
-AO L 04 r4 > ) m 41
44 C O C 0 B 4J Ur 0

M 44 p 9 (D w M 4)
N0 )d4 ) ) c "A
4 --1 P : C 4) 2t ,C

W ( e :O cB 4 V) W 4 16 0

( 0I x 6r) punoq KpoqW uV




Dissociation constants (KD values) for binding of monoclonal antibodies
to uteroferrin. the human spleen enzyme, the rat spleen enzyme.
the bovine spleen and uterine enzymes

Antigen Monoclonal Antibody

5.122.10 6.21.2 6.22.2 13.122
KD, nM

Uteroferrin 1.0 1.2 2.7 12
Enzyme from hairy cell spleen 145 600 270 40
Enzyme from rat spleen 33 8 13
Enzyme from beef spleen 120 130 125 80
Enzyme from cow uterus 13 117 75 40

Monoclonal antibodies 5.122.10, 6.21.2 and 6.22.1 were produced by
hybridoma clones which originated from a mouse immunized with
uteroferrin. They were selected by their ability to bind uteroferrin
with high affinity in a solid phase radiobinding assay. Monoclonal
antibody 13.122 was produced by a hybridoma clone which originated from
a mouse immunized with the enzyme purified from human spleen. It was
selected because of its ability to bind uteroferrin and the human spleen
enzyme with approximately similar affinities.

weakly to the beef spleen enzyme (Table 2-9). However, the KD values

for all five phosphatases were not markedly different, a result which

suggests that antibody 13.122 recognized a relatively conserved site.

Inhibition of Enzymatic Activity by the Binding of Monoclonal Antibodies

The acid phosphatase activity of uteroferrin was measured (without

the addition of 2-mercaptoethanol) after each of the monoclonal

antibodies had been allowed to bind to the enzyme at pH 7.0 for 1 hour.

Table 2-10 demonstrates that antibody 5.127.3 effectively inhibited acid

phosphatase activity when a 2-4-fold molar excess of antibody was added.

Monoclonal antibody 6.22.1 inhibited the enzyme activity slightly at

lower concentrations, while antibody 6.21.2 did not inhibit activity at

all. Monoclonal antibody 13.122 inhibited the phosphatase activity

about 30% at all concentrations tested. None of these antibodies were

able to inhibit the human, rat or bovine spleen enzymes at the

concentrations tested (data not shown).


Tartrate-resistant acid phosphatases from various tissue sources can

be purified by a procedure which closely resembles the fast, convenient

purification protocol for porcine uteroferrin. This is particularly

important in the case of the human enzyme, where previous attempts at

purification to homogeneity were either not successful (Lam et al.,

1977) or involved a large number of time-consuming steps (Robinson and

Glew, 1980).

Three properties allowed the hairy cell phosphatase to be purified

readily from extracts of spleen. The high isoelectric point ensured

that it bound to CM-cellulose at high pH. Its immunological

TABLE 2-10

Inhibition of acid phosphatase activity of uteroferrin
by monoclonal antibodies

Antibody Molar Ratio, Acid Phosphatase Activity
Antibody: Uteroferrin
% control

5.127.3 0.5:1 53
2:1 22
4:1 24
8:1 42

6.21.2 0.5:1 97
2:1 86
4:1 109
8:1 138

6.22.1 0.5:1 71
2:1 68
4:1 90
8:1 111

13.122 0.5:1 79
2:1 73
4:1 73
8:1 72

Uteroferrin (0.06mg) was incubated for 1 hour at room temperature
with the antibody specified, then assayed in triplicate for acid
phosphatase activity with 20mM p-nitrophenylphosphate as substrate and
without the addition of 2-mercaptoethanol. All activities are presented
as percent of a control assay performed simultaneously.

cross-reactivity with anti-uteroferrin antibodies permitted it to be

adsorbed to the immunoaffinity column. Finally, it was relatively

stable to the conditions of elution, namely a brief exposure to pH 2.3

glycine buffer. It is conceivable that some of the losses that occurred

at the immunoaffinity step did result from the elution conditions.

Uteroferrin, although relatively stable to low pH, does begin to lose

its phosphatase activity and presumably its bound iron if maintained

below pH 3.0 (Campbell and Zerner, 1973). However, the major loss of

yield occurred during the binding to CM-cellulose, a step which rarely

provided more than 20% recovery of enzyme. This failure of the cation

exchanger to bind all of the enzyme was not the result of its binding

capacity being exceeded nor to the presence of enzyme variants of lower

isoelectric point. Neither was it related to the presence of detergent.

Possibly the enzyme remains completed with some soluble, strongly

anionic component in the homogenate. Whatever the basis of the

phenomenon, repeated treatment with CM-cellulose does provide a much

improved yield, with up to 40% of the total phosphatase activity of the

spleen extract eventually binding.

When enzyme was purified from human placenta, protamine sulfate was

added in order to bind various anionic components in the homogenate.

This procedure was first employed by Glomset and Porath for purification

of the bovine spleen enzyme (1960) and was also used by Anderson and

Toverud in purification of the rat bone enzyme (1986). Protamine

sulfate precipitation of anionic components is now routinely empolyed by

our laboratory for the purification of tartrate-resistant acid

phosphatases (see Allen et al., in press).

Despite the fact that the spleen was homogenized in a buffer which

contained a wide range of protease inhibitors, there was evidence to

suggest that much of the enzyme may have been cleaved into subunitss"

during or prior to purification. Activity was lost, for example, if the

homogenate was not rapidly processed, and the relative yield of 34,000

monomer and of the two subunits varied from preparation to preparation.

Moreover, the final enzyme, although exhibiting a symmetrical peak of

protein and enzyme activity of apparent Mr=34,000 during gel filtration

on Sephadex G-100 always displayed some subunits during SDS-PAGE.

Robinson and Glew (1981), working with the closely similar Gaucher

spleen enzyme, have shown that the presence of the two lower molecular

weight polypeptides could be minimized, but not prevented, if a protease

inhibitor was included in the homogenizing buffer. In addition, when

uteroferrin was treated with low concentrations of trypsin or

chymotrypsin (Buhi, 1981) or incubated in allantoic fluid (Buhi et al.,

1982b) an active enzyme with subunits of apparent Mr=20,000 and 15,000

was produced. Presumably both proteins possess a region of peptide

which is sensitive to proteases. The reports that the closely similar

beef spleen enzyme, which we have shown to crossreact immunologically

with uteroferrin, possesses two such subunits (Davis et al., 1981;

Campbell et al., 1978) implies that the bovine enzyme might also become

cleaved proteolytically either during normal in vitro processing or

during its isolation from the spleen homogenate. In contrast, the

bovine bone enzyme (Lau et al., 1985, 1987) and the bovine uterine

enzyme do not seem to be as sensitive to this proteolysis, they appear

as intact monomers after purification. The human bone enzyme, unlike

the human spleen and placenta enzymes, is also easily purified without

evidence of proteolytic products (Allen et al., in press).

The earlier report that the hairy cell enzyme is not activated by

reducing agents (Lam et al., 1977) can probably be explained by the

observation that activation was transient and varied between

preparations. Similar observations have been made with uteroferrin.

The differences in molecular weight (64,000 versus 34,000) can in turn

by attributed to the fact that hemoglobin was employed as a standard in

the earlier work (Lam et al., 1977) and assumed to maintain the

molecular weight of the tetramer during sucrose density gradient

centrifugation. The variability in kinetic parameters (particularly Km

and V) are probably best explained by the sensitivity of these enzymes

to a variety of inhibitors, including sulfhydryl agents, anions, heavy

metals, and oxidizing agents. The properties of the hairy cell enzyme

do, however, differ somewhat from those reported for the tartrate-

insensitive phosphatase isolated by Robinson and Glew (1980) from

spleens of patients with Gaucher's disease. That phosphatase, despite

the fact that it appeared to crossreact immunologically with the hairy

cell enzyme (Lam et al., 1981), failed to utilize pyrophosphate as a

substrate and hydrolyzed ATP considerably better than ADP (Robinson and

Glew, 1980). It is still not clear, therefore, whether or not the

Gaucher phosphatase and the one from hairy cell spleen are identical


The experiments with whole antiserum and with the different

monoclonal antibodies to uteroferrin and the human spleen enzyme

emphasize that the two proteins must have cross-reacting and presumably

structurally related determinants (epitopes) on their surfaces.

Monoclonal antibody 13.122, for example, was selected for its ability to

bind to both uteroferrin and to the spleen enzyme. Its binding

characteristics towards both proteins and towards the rat spleen, bovine

spleen and bovine uterine enzymes appeared fairly similar, suggesting

that the epitope it recognized was relatively conserved. On the other

hand, the monoclonal antibodies 5.122.10, 6.21.2 and 6.22.2 bound to the

human, rat, and bovine spleen enzymes and the bovine uterine enzyme with

much lower affinity than to uteroferrin. It can be concluded that all

of the enzymes carry structurally homologous epitopes on their

respective surface, but these sites are not equally conserved.

Nevertheless, the ability to generate both polyclonal and monoclonal

antibodies which have broad immunological crossreactivity should prove

useful in further studies on the biosynthesis and localization of this

class of acid phosphatase. These antibodies may have diagnostic value

both for identifying leukemic hairy cells immunocytochemically and for

measuring the level of the Type 5 isozyme in plasma by immunoassay. The

latter procedure may be particularly useful for assessing osteoclast

activity in individuals exhibiting abnormal bone metabolism (Minkin,

1982; Chen et al., 1975; Stepan et al., 1983) or in patients with

osteoclastic tumors (Tavassoli et al., 1980).

The results of this study clearly demonstrate that uteroferrin, the

purple-colored acid phosphatases from rat spleen, bovine spleen and the

bovine uterus, and the enzyme from the spleens of human patients with

hairy cell leukemia are closely related proteins. In addition to their

immunological cross-reactivity, they are similar in kinetic properties,


in their substrate specificities, and in their sensitivities to a broad

range of inhibitors. They also resemble each other in molecular size,

isoelectric point, glycoprotein nature, and iron content. One obvious

difference, however, between uterine enzymes and the spleen enzymes

studied in this chapter is that the uterine enzymes are secreted

glycoproteins whereas the others are principally intracellular.

Nevertheless, even this paradox may be resolved as there is now

considerable evidence that uteroferrin may represent a hypersecreted

lysosomal enzyme (Baumbach et al., 1984). It is, for example, a

substrate for the transferase which is specifically responsible for the

selective phosphorylation of mannose residues on lysosomal enzymes (Lang

et al., 1984). Uteroferrin is also secreted with the so-called

lysosomal recognition marker, mannose 6-phosphate, present on its

carbohydrate chain (Baumbach et al., 1984). Possibly all members of

this class of acid phosphatase have the properties of lysosomal enzymes.

Certainly many of them appear to reside intracellularly in inclusion

bodies resembling lysosomes (Baumbach et al., 1984; Yam et al., 1971;

Schindelmeiser et al., 1987).

The two bovine enzymes which were employed for comparative purposes

in this study should perhaps be studied in more detail. Under the

influence of progesterone, the bovine uterus secretes a uteroferrin-like

protein which appears to be quite distinct from the bovine spleen

enzyme, which is lysosomal. It has been demonstrated that the bovine

uterine and spleen enzymes differ in a number of ways: pH optima (4.2

and 6.0, respectively), Mr (32,000 and 40,000), ability to hydrolyze

pyrophosphate, and cross-reactivity with monoclonal antibody 5.122.10

(KDS 17 and 130nM). It would be interesting to determine whether

differential targeting is due to variation in the enzymes' primary

sequences. An alternative approach to study targeting of the purple

phosphatases would involve the isolation and characterization of a

porcine spleen enzyme which would presumably be lysosomal. It appears

that such an enzyme exists, and is distinct from uteroferrin (M. Kazemi

and R.M. Roberts, unpublished results).

In conclusion, it is proposed that the Type 5 human isozyme belongs

to the growing class of iron-containing phosphatases of which

uteroferrin and the phosphoprotein phosphatase of beef spleen are the

best characterized. It remains to be determined whether the human

enzyme possesses the purple color and characteristic g'=1.74 ESR iron

signal which is exhibited by uteroferrin (Antanaitis et al., 1983) and

the bovine spleen phosphatase (Davis and Averill, 1982). Why the Type 5

isozyme accumulates in the leukemic hairy cell and in the spleen

histiocytes of patients with Gaucher's disease is unknown. Both these

cell types are capable of phagocytosing erythrocytes (Robinson and Glew,

1980; Nanba et al., 1977), and the Gaucher cell is known to accumulate

large amounts of iron (Robinson and Glew, 1980). Possibly the elevation

of the Type 5 phosphatase is a reflection of abnormal iron metabolism in

these cells.



There are at least six distinct types of acid phosphatase in human

leukocytes which can be distinguished by electrophoretic and other

characteristics (Li et al., 1979; Lam et al., 1973; Li et al., 1973).

The Type 5 isozyme is the most cationic of the acid phosphatases and is

the only isozyme insensitive to inhibition by L-(+)-tartrate. It has

been detected in spleen, lung, liver and bone as a minor isozyme (Lam et

al., 1973; Yam et al., 1971). However, it can become the dominant

isozyme in certain pathological states. High tartrate-resistant acid

phosphatase levels are often found within the spleen (Robinson and Glew,

1980) and monocytes (Troy et al., 1985) of patients with Gaucher's

disease; the splenocytes and circulating white cells of patients with

hairy cell leukemia (Yam et al., 1971); the spleens of patients with

Hodgkin's disease (Drexler et al., 1986); and the sera of individuals

undergoing active bone turnover (Yam, 1974). Elevated levels of this

isozyme are also associated with various B-cell and T-cell leukemias

(Drexler et al., 1986). In systematic studies of leukemia cell lines,

it appeared that the tartrate-resistant, Type 5 acid phosphatase was not

generally expressed in immature lymphoid cells, but rather by cells

arrested in later stages of differentiation (Drexler et al., 1985,


In the previous chapter the purification and characterization of

this Type 5 acid phosphatase from human hairy cell spleen, normal spleen

and placenta were reported. It was demonstrated that this human enzyme

is remarkably similar to an abundant tartrate-resistant acid phosphatase

secreted by the porcine uterus under the influence of progesterone,

known as uteroferrin. Both the human enzyme and uteroferrin are iron-

containing glycoproteins, and resemble each other closely in

electrophoretic mobility, substrate specificity and sensitivity to a

variety of activators and inhibitors. The two proteins are also

immunologically related. While uteroferrin is believed to function in

iron metabolism (see Roberts and Bazer, 1985; 1988), the function of the

human enzyme is unknown.

In view of the clinical significance of the human tartrate-resistant

acid phosphatase, further studies on its transcriptional control seemed

warranted. A human placenta cDNA library was screened with polyclonal

antibodies against porcine uteroferrin. However, these antibodies

failed to detect a cDNA coding for the human tartrate-resistant acid

phosphatase. Accordingly, two short cDNA clones coding for porcine

uteroferrin were employed to isolate a cDNA coding for the human enzyme.

In this chapter, the molecular cloning of a 1412 bp cDNA which covers

the complete coding region of the human tartrate-resistant acid

phosphatase is described, and the deduced amino acid sequence of the

human enzyme is compared to the amino acid sequences of porcine

uteroferrin and a related bovine spleen acid phosphatase. The

expression of this human phosphatase in human leukemia cells is

demonstrated, and its regulation studied in the human erythroleukemia

cell line K562.

Materials and Methods


The human placenta cDNA library, mouse spleen cDNA library and human

placenta poly (A)+ RNA were obtained from Clontech. Goat anti-rabbit

and goat anti-mouse antibodies coupled to horseradish peroxidase were

from Bio-Rad. All restriction endonucleases, the large fragment of E.

coli DNA polymerase, T4 DNA ligase, and the RNA ladder were from

Bethesda Research Laboratories. The [o-32P]-dATP was purchased from New

England Nuclear Corp. The 2'-deoxynucleotide triphosphate/2'3'-

dideoxynucleotide triphosphate sequencing mixes were from New England

BioLabs. Heparin-Agarose, hemin and TPA were from Sigma. Fetal bovine

serum and RPMI 1640 were obtained from Gibco-BRL. Oligodeoxyhexamer for

random primer extension probes was from Pharmacia. All other chemicals,

reagent grade or better, were from Sigma or Fisher.


Preparation of antisera for immunoscreening. Adult New Zealand

White rabbits were immunized with 0.2 to 0.5mg of uteroferrin mixed with

Freund's complete adjuvant and then injected subcutaneously at multiple

sites (Herbert, 1973). Booster injections of 0.2 to 0.5mg of antigen in

incomplete adjuvant were administered monthly. Bleedings of 20 to 40ml

were taken 10 days after injections. The samples were allowed to clot

overnight at 4C and then centrifuged to separate the serum from the

clot. In some cases the antisera were affinity purified on anti-

uteroferrin Sepharose as outlined in Chapter 2. In other cases the

antisera were run through a column of E. coli proteins which had been

coupled to Sepharose, in order to remove antibodies against E. coli

proteins (Huynh et al., 1984).

Screening of the lambda gtll libraries with anti-uteroferrin

antibodies. A human placenta cDNA library prepared by Millan (1985) was

purchased from Clontech. This library contains 1 x 106 independent

clones, with 96% recombinant phage. The insert sizes range from 0.8-

3.6 kb, and the average insert size is 1.8 kb. A mouse spleen cDNA

library was purchased from Clontech. This library contains 1 x 106

independent clones, with 93% recombinant phage. The average insert size

is 1.0 kb. The porcine uterine endometrium cDNA library was prepared by

George Baumbach and Patrick Gillevet at the University of Florida. The

method used for screening was developed by Young and Davis (1983a,b) and

Huynh et al. (1984) and modified by deWet et al. (1984).

E. coli strain Y1090 was grown overnight in LB medium [1% (w/v)

tryptone, 0.5% (w/v) yeast extract, 0.5% (w/v) NaCl] plus 0.2% (w/v)

maltose and 50pg/ml ampicillin at 370C with good aeration. The next

morning, the bacteria were centrifuged at 1500xg at room temperature for

10 minutes, and resuspended in one-fifth original volume of SM buffer

[0.01M Tris-HCl, pH 7.5, 0.01M MgC12, 0.1mM EDTA, 0.2% (w/v) gelatin].

An appropriate number of phage were mixed with either 75pl of bacteria

(for 100mm dishes) or 300pl bacteria (for 150mm dishes). The bacteria

and phage were incubated at 37C for 15 minutes to allow the phage to

adhere to the bacteria. The phage and bacteria were then mixed with

either 3 ml (for 100mm dishes) or 7.5ml (for 150mm dishes) of melted LB

containing 0.7% (w/v) agarose, which had been kept at 45C. The mixture

was poured onto plates containing LB with 1.5% (w/v) agar, which had

been pre-warmed to 370C. The plates were then placed in an incubator at

42*C for 3 to 4 hours, or until plaques began to appear on a confluent

lawn of bacteria. Nitrocellulose filters, which had been soaking in

10mM isopropyl P-D-thiogalactopyranoside (IPTG) were then placed on the

plates. The plates were incubated at 37C for 4 hours which allowed a

f-galactosidase fusion protein to be formed, which adhered to the


After the filters had been removed, the plates were saved at 4C

until needed. The filters were washed three times with Tris buffered

saline (TBS; 0.1M Tris-HCl, pH 8.0, 0.15M NaC1) and placed in antibody

solution containing a 1:100 dilution of antisera in TBS with 3% (w/v)

bovine serum albumin (BSA). The filters were incubated in antibody

solution overnight at 40C on a rocking platform at low speed.

The next day, the antibody solution was removed, filter sterilized

through a Millipore 0.2pm filter, and stored at 4C. (Each antibody

preparation was used up to five times.) The filters were washed

thoroughly with three changes of TBS. The additional protein binding

sites on the nitrocellulose were blocked by incubation in TBS plus 3%

(w/v) BSA. The filters were then treated with goat anti-rabbit

antibody coupled to horseradish peroxidase diluted in TBS plus 3% (w/v)

BSA at dilutions 3- to 5-fold more concentrated than the manufacturer's

suggested dilution for Western blotting. The filters were incubated in

the second antibody solution for 2 hours at room temperature on a

rocking platform at low speed. The antibody was then removed and

discarded, and the filters washed thoroughly as before.