Developmental regulation of placental and fetal amino acid transport


Material Information

Developmental regulation of placental and fetal amino acid transport the effects of low-protein diet-induced intrauterine growth retardation
Alternate title:
Effects of low-protein diet-induced intrauterine growth retardation
Physical Description:
ix, 207 leaves : ill. ; 29 cm.
Malandro, Marc Shane, 1966-
Publication Date:


Subjects / Keywords:
Research   ( mesh )
Diet, Protein-Restricted   ( mesh )
Fetal Growth Retardation   ( mesh )
Maternal-Fetal Exchange -- physiology   ( mesh )
Biological Transport, Active -- physiology   ( mesh )
Amino Acids -- metabolism   ( mesh )
Leucine -- metabolism   ( mesh )
Gene Expression Regulation, Developmental   ( mesh )
Rats   ( mesh )
Department of Biochemistry and Molecular Biology thesis Ph.D   ( mesh )
Dissertations, Academic -- College of Medicine -- Department of Biochemistry and Molecular Biology -- UF   ( mesh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph.D.)--University of Florida, 1995.
Bibliography: leaves 191-206.
Statement of Responsibility:
by Marc Shane Malandro.
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 - 002296914
oclc - 49808403
notis - ALQ0153
System ID:

This item is only available as the following downloads:

Full Text








I would like to thank the members of my supervisory

committee: Drs. Brian Cain, Susan Frost, Mike Kilberg, Harry

Nick, Don Novak, and Bruce Stevens. A special word of thanks

goes to my co-mentors Mike Kilberg and Don Novak. Without

them, this study would not have been possible. Thanks also

go to Mark Beveridge and Dr. Jamie Matthews for valuable

technical support. I would also like to acknowledge Drs.

Roney Laine and Bob McMahon. They both have been and will

continue to be good friends and valuable resources. Most of

all, I would like to thank my wife, Jennifer and daughter, J.

Nicole for support and love throughout this study.



LIST OF TABLES........................................... v


ABSTRACT.......... .......................................viii

CHAPTER 1 INTRODUCTION..................................... 1

Overview of Mammalian Amino Acid Transport............1
Introduction to the Placenta...........................8
Molecular Biology of Charged Mammalian
Amino Acid Transporters ..............................15

CHAPTER 2 MATERIALS AND METHODS...........................57

Materials ....... .................................. 57
Methods ..............................................58


Introduction... .................................... 66
Results............................................. 70
Discussion........................................... 87

TRANSPORTER .....................................93

Introduction....................... ...........93
Results ............................................. 103

TRANSPORT IN THE RAT ...........................138

Introduction ....................................... 138
Results ............................................. 141
Discussion ............................... .......159


TRANSPORTERS................................. 165

Introduction ........................................165
Methods....................................... ..... 168
Results .............................. .....173
Discussion............................... ........... 183


LIST OF REFERENCES.....................................190

BIOGRAPHICAL SKETCH.................................... 207


Table page

3-1. Characterization of apical and basal membrane
vesicles from rat placentas of 14 or 20 days

3-2. Inhibition of Na-dependent (System B'*) arginine
uptake in microvillous membrane vesicles from
placentas of 20 days gestation......................78

3-3. Inhibition of Na+-independent, leucine-sensitive
arginine uptake in apical and basal membrane
vesicles from placentas of 20 days gestation........79

3-4. Inhibition of Na+-independent, leucine-insensitive
(System y+) arginine uptake in apical and basal
membrane vesicles form placentas of 20 days

4-1. Transient and stable expression of rat EAAC1.......109

5-1. Maternal weight gain, fetal weight, placental
weight, and litter size for control and low-
protein animals representing three independent
experimental trials................................143

5-2. Maternal serum amino acid levels from control
and low-protein fed dams............................145

5-3. Transport of MeAIB (System A) in placental apical
and basal membrane vesicles from control and
low-protein diet-fed animals.....................151

5-4. Na+-dependent transport of glutamate (System X-)
in placental basal membrane vesicles from control
and low-protein diet-fed animals....................153


Figure page

3-1. Time-course of ['H]-arginine uptake in basal
or apical membrane vesicles from placentas of
20 days gestation ................................. 74

3-2. Inhibition of Na'-dependent ['H]-arginine uptake
with increasing concentrations of leucine in
placentas of 20 days gestation.......................76

3-3. Ontogeny of cationic amino acid transport systems
in microvillous and basal membrane vesicles.........84

3-4. Northern analysis of System y+ mRNA from
placentas of increasing gestational age.............86

4-1. Deduced amino acid sequence of rat EAAC1 cDNA.......105

4-2. Tissue distribution or rat EAAC1 mRNA ..............107

4-3. Synthesis and purification of MBP-EAAC1 fusion
protein ............................................111

4-4. Stable expression of rat EAAC1 in human 293c18
cells .......................... ................... 114

4-5. Immunoblot analysis of over-expressed rat EAAC1....116

4-6. Time-course of ['H]-glutamate uptake in basal or
apical membrane vesicles from placentas of 14 or
20 days gestation..................................119

4-7. Reverse transcriptase PCR detection of glutamate
transporters in placentas at 14 and 20 days
gestation.......................... ............121

4-8. Northern analysis of EAAC1 mRNA from placentas
of 14 or 20 days gestation........... .............124

4-9. Immunoblot analysis of EAAC1 protein in apical
and basal membrane vesicles from placentas of
20 days gestation..................................126

4-10.Immunoblot analysis of EAAC1 from placentas of
14 and 20 days gestation............................129

4-11.Uptake of glutamate in fetal, newborn, and
adult hepatocytes .................................. 131

4-12. Reverse Transcriptase PCR detection of glutamate
transporters in fetal hepatocytes................133

5-1. Immunoblot analysis of Ca2ATPase in homogenate,
apical, and basal membrane vesicles from control
and low-protein fed animals ........................149

5-2. Na-dependent and Na-independent cationic amino
acid uptake in apical and basal membrane vehicles
from control and low-protein fed animals........... 156

5-3. Northern analysis of placental System y (CAT1)
and System X-A (EAAC1) mRNA from control and
low-protein fed animals............................158

6-1. Diagram and vector map of kDR2/pDR2 ................171

6-2. Schematic diagram of library screening protocol....176

6-3. Crystal violet stain and autoradiogram of a
representative screening with EAAC1-transfected
293c18 cells........................................179

6-4. PCR detection of EAAC1 from a positive (pDR/EAAC1)
293c18 colony......................................181


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



Marc Shane Malandro

December, 1995

Chairman: Dr. Michael S. Kilberg
Major Department: Biochemistry and Molecular Biology

Gestational regulation of the placental transfer of

amino acids from maternal to fetal circulations is essential

for the proper development of the fetus. Both the cationic

and anionic amino acid transport systems of the apical

(maternal-facing) and basal (fetal-facing) membranes of the

rat placental syncytiotrophoblast were examined. Inhibition

analysis documented the presence of three kinetically

distinct cationic amino acid transport mechanisms: a single

Na+-dependent mechanism in the apical membrane which

increased in activity from 14 to 20 days gestation but was

absent from the basal membrane throughout the entire

gestational period (System B0,+); and two Na+-independent

transport systems in both membrane domains, one completely


inhibited by leucine which increased in activity in both the

microvillous and basal membrane domains, and the other

leucine insensitive which remained fairly constant in the

basal membrane and increased throughout gestation in the

microvillous membrane (System y*). Anionic amino acid uptake

is mediated by exclusively Na+-dependent systems on both the

apical and basal membranes. Northern analysis with cDNA

probes to anionic (EAAC1) and cationic (CAT1) amino acid

transporters identified specific mRNA species which increased

with increasing gestational age. Concomitant regulation of

EAAC1 mRNA, protein, and transport activity was also

observed. Induction of intrauterine growth retardation by

low-protein diet resulted in the decreased transport of amino

acid in both the apical and basal membranes. As several key

amino acid transporters have yet to be cloned, a novel

mammalian cell expression system was also developed and

characterized with a known (EAAC1) transporter.


Overview of Mammalian Amino Acid Transport

The scope of the function of amino acids lies well

beyond the role of these molecules as biosynthetic precursors

to proteins. These compounds function as carriers of

nitrogen and carbon atoms, as metabolic fuels, as

neurotransmitters, and as sensitive regulators of cell

osmolarity. These molecules must be supplied to or made in

the cells to maintain intracellular concentrations of both

essential and nonessential amino acids.

Nonessential amino acids can be supplied to cells in a

variety of ways. First, these compounds can be synthesized
de novo in the cell from precursor molecules including a-keto

acids and other amino acids. Nonessential amino acids can be

synthesized in other cell types and delivered to the cells in

the extracellular medium by a process of "interorgan

nutrition" (Christensen, 1982). Degradation of existing

cellular proteins can also supply these nutrients. On the

other hand, essential amino acids must be supplied to the

cells only by degradation of intracellular proteins or by

delivery of these nutrients to the cells from exogenous

sources. This delivery of amino acids from the diet as

protein digestion products was described as early as 1913

(Van Slyke and Meyer, 1913). Clearly, the availability of

amino acids or amino acid precursors to the extracellular

face of cells is not sufficient for the intracellular

metabolism of these compounds. These molecules must traverse

the plasma membrane which is highly impermeable to charged

and polar solutes. As such, mechanisms for delivery of amino

acids are not only absolutely necessary for cellular

function, but also, the regulation of this delivery may serve

as an additional level of control of cellular metabolism.

Introduction to Transport

The term transport refers to the transfer of a solute

across the plasma membrane, and the protein(s) that mediate

this event is referred to as a transporter. The transport

process itself can be visualized in four arbitrarily defined

steps: 1) the binding of the solute molecule to the

transporter on the extracellular face of the cell; 2) the

conformational change in the transport protein resulting in

the movement of the solute across the membrane; 3) the

release of the solute molecule on the intracellular side of

the membrane; 4) the re-orientation of the transporter such

that the solute binding site faces the extracellular medium.

Kinetically, as the steps involved in membrane transport

are mediated by proteins, the rate of solute transport does

not represent a straight line with increasing substrate

concentration, but rather, is represented by a hyperbola

suggesting saturation of the transporters as substrate

concentration increases. As such, mediated transport

activities can be examined by Michaelis-Menten parameters

analogous to those of saturable, enzyme-catalyzed reactions

(Christensen and Kilberg, 1987). Classically, the

description of the movement of amino acids by a kinetically

distinct transport activity has been termed a transport

"system." To this end, the relative activity of a particular

transport system is reflected in the maximal velocity (V.) of

transport while the relative affinity of a substrate for a

particular transport system can be represented as the

concentration of substrate leading to half-maximal velocity

(1K). Transport velocities are typically reported as amount

of substrate transported per amount of transporter,

approximated by either mg protein or number of cells, per

unit time.

Three basic mechanisms exist for the transport of

substrates across the plasma membrane. Two transport

processes mediate the accumulation of substrate against a

concentration gradient, primary active transport and

secondary active transport. The third process is driven

solely by the "downhill" movement of substrate with its

concentration gradient, facilitated transport. It should be

noted that facilitated transport is not facilitated diffusion

as the former is protein-mediated and demonstrates saturable

kinetics. Solute movement by diffusion, on the other hand,

increases in a linear fashion with increasing solute

concentration and does not show saturation.

Primary active transport is the direct coupling of an

energy-generating reaction to the transport of a solute. In

most cases, the energy used is the hydrolysis of ATP, and

this serves to accumulate solute against its concentration

gradient (chemical gradient). If the solute molecule is

charged, an electrochemical gradient is generated. This is

the case for the transport of Na+ and K' by Na+/K+ ATPase. The

hydrolysis of ATP drives the transport of 3Naa from cytoplasm

to the extracellular medium with the transport of 2K from the

extracellular medium to the cytoplasm. The transport of two

substrate molecules across the membrane in opposite

directions is referred to as antiport. The Na'

electrochemical gradient can then be utilized as an energy

source for secondary active transport. In this case, the

"downhill" flux of the Na can drive the transport of

substrate against its electrochemical gradient. The

transport of two substrate molecules in the same direction

across the membrane is referred to as symport. Facilitated

transport represents the movement of solute molecules across

the membrane utilizing only the electrochemical gradient of

the substrate molecule itself as an energy source for

transport. In the case of amino acids, these molecules are

transported by either secondary active or facilitated

transport. In mammalian cells, no transport of amino acids

has been linked to the hydrolysis of ATP, but the Na+ gradient

generated by Na/K ATPase is an important driving force for

amino acid transport.

Cationic Amino Acid Transport Systems

The identification of a transport system for cationic

amino acids distinct from neutral amino acids was defined in

Ehrlich ascites tumor cells by Christensen and Liang in 1966.

This transport system was originally defined as System Ly to

illustrate affinity for lysine. Later the transport system

was renamed System y* (Bannai et al., 1984), and the system

accepts lysine, arginine, ornithine, and histidine when

positively charged. Uptake by System y' is characterized as

Na+-independent facilitated transport of exclusively cationic

amino acids. However, in the presence of Na', this system can

interact with neutral amino acids as well (Christensen et

al., 1969). Although the driving force for this transport is

the "downhill" movement of substrate with its concentration

gradient, System y+ is electrogenic and can concentrate amino

acids to an extracellular to intracellular distribution

profile of greater than one based soley on the membrane

potential of the cell (Bussolati et al., 1987). This system

is also affected by trans-stimulation (White and Christensen,

1982); high intracellular concentrations of substrate

increase the activity of this transporter presumably by

decreasing the time required for the reorientation of the

transporter such that the substrate binding site faces the

extracellular medium. System y' mediates the transport of

cationic amino acids in virtually every cell type with the

exception of the liver (White, 1985). The cDNAs encoding

System y' activity have been cloned and are described in

detail below.

Two additional systems for the transport of cationic

amino acids have been described by Van Winkle and coworkers

(1985; 1988). The first, System B"', is a Na'-dependent

secondary active transport system that accepts not only

cationic amino acids but neutral amino acids as well (Van

Winkle et al., 1988). This system has also been described in

Xenopus oocytes (Campa and Kilberg, 1991) and in several

other cell types including human fibroblasts, porcine aortic

endothelial cells, and rat FAO hepatoma cells (Kilberg et

al., 1993). System B0' accepts neutral amino acids with bulky
side-chains and branching at both the a and B carbon atoms

(Van Winkle et al., 1988). This system is also inhibited by

bicyclic amino acids including 2-aminobicyclo[2,2,1] heptane-

2-carboxylic acid (BCH), once thought to be specific for the

Na-independent System L (Christensen et al., 1969). The

second system is a Na-independent facilitated transporter

System b0'. This system also accepts both cationic and

neutral amino acids; however, System b"'' does not accept
branching of neutral amino acids on the a or B carbon (Van

Winkle et al., 1988). This system is also present in Xenopus

oocytes and a variety of other cell types (Kilberg et al.,

1993). An analogous transport system has been identified in

erythrocyte membrane vesicles called System y+L (Devds et al.,

1992). The lack of inhibition of System y L by the sulfhydryl

modification reagent N-ethylmalemide (NEM) appears to

discriminate this activity from that of System b' (Deves et

al., 1993). Additionally, System yL mediated the uptake of

cationic amino acids in a Na+-independent manner and neutral

amino acids in the presence of Na+ (Deves et al., 1992).

System b0' mediates the uptake of both cationic and neutral

amino acids in a Na+-independent manner (Van Winkle et al.,


Anionic Amino Acid Transport Systems

A Na+-dependent, secondary active transport system has

been described for the anionic amino acids aspartate and

glutamate, System X-, (Gazzola et al., 1981). This system

transports both the D- and L- steroisomers of aspartate but

only L-glutamate and is very sensitive to membrane potential.

The mechanism for amino acid uptake through System X-

involves the symport of 2Na+ with the amino acid and the

antiport of a K' and either OH- or HCO3- for each transport

cycle (Hediger, 1995). Some cell types such as the Ehrlich

ascites tumor cell have no specific system for the transport

of anionic amino acids but rather mediate the uptake of these

substrates in the zwitterionic form through systems for

neutral amino acids such as Systems ASC and L (Garcia-Sancho

et al., 1977). In the adult rat hepatocyte, however, there

appears to be two kinetically distinguishable, Na+-dependent

systems for the uptake of anionic amino acids (Gazzola et

al., 1981). Cysteate and cysteinesulfinate effectively

inhibit both transport systems in normal hepatocytes.

However, upon transformation, cysteate and its homolog

homocysteate differentially each system (Gazzola et al.,

1981). These data were the first indication that multiple

proteins may be mediating similar transport activities, and

in fact, several cDNAs each mediating the Na*-dependent uptake

of anionic amino acids have been identified. The cDNAs

encoding this activity will be described in detail below.

A Na-independent, facilitated transporter also exists

for the uptake of anionic amino acids (Guidotti and Gazzola,

1992). System x, mediates the transport of L-glutamate and

L-cystine by trans-stimulaton of substrate coupled antiport.

In fibroblasts, a glutamine-cystine cycle exists whereby L-

glutamine enters the cell via neutral amino acid transport

Systems A and ASC (Gazzola et al., 1981). A significant

portion is converted intracellularly to glutamate (Bannai and

Ishii, 1988). High intracellular levels of glutamate are

maintained by deamination of glutamine and by Na-dependent

glutamate uptake via System X-,. Glutamate then effluxes down

its concentration gradient through x-, driving the uptake of


Introduction to the Placenta

Normal fetal growth and development requires a continuous

supply of nutrients from the maternal circulation throughout

gestation. The placenta is functionally responsible for the

uptake and transport of these nutrients from the maternal to

fetal circulations, as well as the removal of waste products

from the fetus to be metabolized in the placenta or

transported to the mother. As such, Na-dependent uptake

mechanisms and Na-independent efflux mechanisms would be

expected on both the maternal and fetal sides of the

placenta. Most of these transport mechanisms have been

localized to the apical (maternal-facing) and basal (fetal-

facing) membranes of the placental syncytiotrophoblast

(Sideri et al., 1983; Smith et al., 1992a).

The syncytiotrophoblast is a multi-nucleated, continuous

epithelial layer resulting from the fusion of terminally

differentiated cytotrophoblasts (Sideri et al., 1983). In

the placenta of rats and higher primates including humans,

the maternal surface of the syncytiotrophoblast is bathed in

blood resulting from the breakdown of the maternal

circulatory endothelium in the villous tissue. As such, the

maternal circulation is in direct apposition to the nutrient

transporters on the apical membrane of the

syncytiotrophoblast. This type of placenta is termed

hemochorial and is distinct from epitheliochorial (pig) and

syndesmochorial (sheep) placentas in which uterine tissue

lies between the maternal circulation and the trophoblast

(Munro, 1986). The only major difference between the rat and

human hemochorial placentas is that the human placenta has

only a single syncytiotrophoblast layer while the rat

placenta has three. The functional significance of these

additional layers in the rat has not yet been described.

Placental Amino Acid Transport

Christensen and Streicher first reported in 1948 that

the human placenta had the ability to concentrate many amino

acids from the maternal to fetal circulation. In addition,

the fetal to maternal serum amino acid ratio (F/M) for most

amino acids is greater than one, indicating that most amino

acids are concentrated in the fetal circulation (Bernardini

et al., 1991). The general pattern of F/M is consistent in

sheep, guinea pig, human, and rat (Phillips et al., 1978;

Marconi et al., 1989). The only notable exception is the

anionic amino acids aspartate and glutamate which show

significant fetal concentration in the rat but nearly no

concentration in all other species (Arola et al., 1982). In

general, the trophoblast concentration of amino acids is

significantly greater than that of either the maternal or

fetal circulations (Lemmons et al., 1976). This suggests

that Na-dependent concentrative transporters are required on

the maternal-facing apical membrane to maintain high

intracellular concentrations of amino acids. This also

suggests that transfer to the fetus may be mediated by efflux

of substrates down the respective concentration gradients.

Schneider and coworkers in 1979 described the

establishment of transplacental gradients for various amino

acids. Equimolar concentrations of amino acids were added to

both sides of the human placenta in a closed perfusion

system. After incubation, the F/M ratios of alanine,

glycine, and lysine increased significantly indicating

concentration by the placenta of these substrates in the

fetal circulation. At the same time, the fetal distribution

ratios of aspartate and glutamate decreased to near zero

demonstrating the clearance of these amino acids from the

fetal circulation. This fetal clearance of these substrates

was accompanied by an increase in the release from the

placenta of asparagine and glutamine leading to the first

suggestion of a glutamate/glutamine cycle described below.

Placental neutral amino acid transport. Although the

apical membrane of the syncytiotrophoblast has been shown to

contain microvilli (Booth and Vanderpuye, 1983; King, 1983),

the amino acid transporter complement represents more

ubiquitously expressed transporters rather than the

transporters associated with the microvilli of the intestine

(Stevens et al., 1984). Apical membranes isolated from

placentas of nearly all species including rat (Alonso-Torre

et al., 1992; Novak et al., in press), demonstrate Na'-

dependent neutral amino acid uptake consistent with System A

(Smith et al., 1992). This activity represents the uptake of

small neutral amino acids with no branching on the B-carbon

(Christensen et al., 1965). Other Na+-dependent components of

neutral amino acid uptake that may be present in the apical

membrane are Systems ASC and N. However, reports noting the

absence of these systems in humans have also been described

(Ganapathy et al., 1986; Johnson and Smith, 1988). Systems

A, ASC, and the Na-independent System L, have all been

described on the basal membrane of the human placenta

(Hoeltzli and Smith, 1989; Smith et al., 1992a).

Placental cationic amino acid transport. The cationic

amino acids arginine and lysine demonstrate high F/M ratios

and are avidly concentrated in the placenta and fetus

(Bernadini et al., 1991). Perfusion studies have

demonstrated transplacental flux of cationic amino acids in

guinea pig (Wheeler and Yudilevich, 1989); however, with the

exception of the human placenta, the cationic amino acid

transporters of either membrane of the trophoblast have not

been well characterized, and with regards to the rat

placenta, virtually nothing has been reported. Wheeler and

Yudelivich (1989), utilizing the perfused guinea pig model,

reported "System y+-like" transport activity within the

placenta, but the membrane surface was not identified.

Furthermore, this study found some Na-dependency to lysine

and arginine transport, more characteristic of System B0"' (Van

Winkle et al., 1985) than System y+ (Christensen and Liang,

1966). Also, the cationic amino acid transport observed was

never tested for inhibition by leucine to discriminate

between System y* and System b0'* (Van Winkle et al., 1988)

The presence of System y+ transport activity could not be

documented in human placental apical membrane vesicles (Kudo

et al., 1987), but was demonstrated by Furesz and coworkers

(1991) in basolateral membrane vesicles, along with System b0'

activity. Subsequent studies by the same laboratory

identified a Na'-independent, neutral amino acid-inhibitable

cationic transporter on the apical membrane and suggested

that it was different from that on the basal membrane (Furesz

et al., 1995). My work (Chapter 3) describes the presence of

all three cationic amino acid transport systems (y, B"', and

b"0') in both apical and basal membrane vesicles from rat


Placental anionic amino acid transport. In general, the

F/M ratios of anionic amino acids are unique from all other

amino acids in that the distribution ratio is less than one

indicating no concentration in the fetus (Battaglia, 1982).

In the rat, however, the distribution ratio is fairly high

(Arola et al., 1982). Perfusion studies in the human

(Schneider et al., 1979), guinea pig (Stegink et al., 1975),

and sheep (Lemmons et al., 1976) have all suggested that

glutamate and aspartate are extracted from both the maternal

and fetal circulation. While conversion of these anionic

amino acids to glutamine and asparagine does occur within the

placenta with subsequent efflux to the fetus, the majority of

the placental glutamate concentration is oxidixed within the

mitochondria (Vaughn et al., 1995). In human membrane

vesicles, a Na'-dependent transport system consistent with

System X-, has been described on both the apical (lioka et

al., 1985; Hoeltzli et al., 1990) and basal (Moe and Smith,

1989) membranes.

Metabolic Interrelationships Between the Placenta and Fetus

Most amino acids are supplied to the fetus in

concentrations that are well in excess of the accumulation of

these precursors in fetal protein (Lemons et al., 1976).

From this supply, the carbon skeletons of amino acids can be

used for oxidation, and this high level of metabolism is

reflected in the high level of urea production in the fetus

(Gresham et al., 1972). The placenta, also capable of amino

acid metabolism but lacking a complete urea cycle, releases a

significant amount of NH, into the fetal circulation which is

then utilized for trans-amination reactions in the fetal

liver (Holzman et al., 1977).

Glutamine/qlutamate cycle. In a single circulatory

passage in the sheep, the basal membrane of the placental

trophoblast is able to extract nearly 90% of the glutamate in

the fetal circulation (Vaughn et al., 1995). The Na'-

dependent glutamate transport mechanisms on this membrane, as

well as on the apical membrane, are responsible for the high

placental concentration of glutamate (mM) compared to either
the maternal or fetal circulation (pM) (Lemmons et al.,

1976). Metabolism of glutamine by fetal tissues is likely

the source of this high level of glutamate (Vaughn et al.,

1995). Although the majority of glutamate extracted by the

placenta is rapidly decarboxylated (Moores et al., 1994),

about 10% is returned to the fetus as glutamine by the action

of placental glutamine synthetase (Battaglia, 1992). This

glutamine is then utilized by the fetus as described above

setting up a glutamine/glutamate cycle between the fetus and


Serine/qlycine cycle. Analogous to the situation with

glutamate, serine is not concentrated in the fetal

circulation by the placenta (Cetin et al., 1991). Rather,

serine is extracted from the fetal circulation by the basal

membrane of the trophoblast (Cetin et al., 1992). The likely

source of this serine is by oxidation of glycine by the fetal

liver and extrahepatic tissues. Multiple tracer infusion

studies by Moores and coworkers (1990) with [1-'3C] and [1-'4C]

serine have demonstrated that a significant portion of the

serine extracted from the fetal circulation in the placenta

is converted to glycine and transported back to the fetus.

Again, a metabolic cycle appears to be in place for not only

the removal of a potentially toxic metabolic byproduct from

the fetal circulation but also the reutilization of this

precursor in the placenta for delivery of an essential

nutrient back to the fetus.

Molecular Biology of Charged Mammalian Amino Acid

Until 1991, no cDNAs corresponding to kinetically
defined a-amino acid transport systems had been described. It

was not until the first description of a cationic amino acid

transporter cDNA by the laboratory of James Cunningham

(Albritton et al., 1989; Kim et al., 1991) that this field

moved into a new era. As the entire course of this study has

been accomplished in this new light, a thorough treatment of

the molecular advances of charged amino acid transport is


The CAT Family for Cationic Substrates

Three members of the CAT family of amino acid

transporters have been cloned, each of which mediate the Na'-

independent transport of the cationic amino acids arginine,

lysine, and histidine when positively charged. Although

there is significant similarity in the deduced amino acid

sequence of these proteins and they share a common substrate

specificity, the tissue specific expression and apparent role

of these proteins in metabolism appears very different.

Cloning of the CAT Family

The cloning of the first member of this family by

Cunningham and co-workers resulted from the search for the

host cell protein responsible for the infection of certain

cell types by the ecotropic murine leukemia virus (Albritton

et al., 1989). The infectivity of the virus, and thus the

susceptibility of a cell, is mediated by the binding of the

virus envelope protein gp70 to a naturally occurring receptor

in the plasma membrane of host cells. DNA isolated from

susceptible murine cells (NIH 3T3 fibroblasts) was used to

stably-transfect a non-susceptible human cell line (EJ

bladder carcinoma cells). Recombinant retrovirus conferring

G418 antibiotic resistance upon infection was used to

identify EJ cells expressing murine DNA for the receptor.

Cloning of the cDNA was then accomplished by Southern blot

analysis with mouse repetitive DNA and linkage analysis to

the receptor gene.

The deduced amino acid sequence of the cDNA clone, EcoR,

revealed a 622 amino acid glycoprotein with a predicted

molecular mass of 67 kDa. Interestingly, the hydropathy

profile of the protein revealed 12 to 14 membrane spanning

regions, similar in structure to the previously cloned

facilitated glucose transporters; however, no significant

similarity to any other protein in the sequence data bases

was demonstrated.

In turn, the human gene homologous to the murine EcoR

receptor was cloned and mapped to chromosome 13 (Yoshimoto et

al., 1991). The open reading frame consists of 10 introns

and 11 exons, and codes for a mRNA transcript of

approximately 9 kb representing long untranslated regions.

The cDNA sequence predicts a 14 membrane-spanning protein of

629 amino acids with a molecular mass of 68 kDa. The murine

and human proteins share 87.6% identity.

A partial rat EcoR cDNA was also isolated (Stoll et al.,

1991), and the full-length clone followed (Wu et al., 1994).

The cDNA open reading frame codes for a 624 amino acid

protein sharing 97% sequence homology with the murine

receptor. The most recent clone (Puppi and Henning, 1995)

identified 7.0 kb of the 7.9 kb mRNA; however, the

significance of the 5 kb 3' untranslated region of the mRNA

remains to be revealed.

The identification of a homologous EcoR gene in human

cells which are not susceptible to MuLV infection suggested

another endogenous function for the receptor. However, the

function of the EcoR CDNA as a transporter remained

undescribed until two independent groups simultaneously

described the cationic amino acid transport properties of the

murine cDNA (Kim et al., 1991, Wang et al., 1991).

Upon re-evaluation of the protein sequence data bases,

Kim and coworkers (1991) identified sequence similarity

between the murine EcoR protein and the arginine and

histidine permeases of Saccharomyces cerevisiae, suggesting a

related function. Xenopus oocytes, injected with cRNA

corresponding to the murine EcoR, demonstrated significantly

higher radiolabelled gp70 binding than water-injected

controls consistent with the function of the protein as a

receptor for MuLV. The cRNA-injected oocytes also

demonstrated saturable accumulation of radiolabelled cationic

amino acids L-arginine, L-lysine, L-ornithine, and L-

histidine, while the uptake of anionic and neutral amino

acids was not different from water-injected controls. The

substrate specificity of the cRNA-mediated uptake, as well as

the apparent K. for cationic amino acids, is consistent with

the previously kinetically defined transport activity, System

y (Christensen and Liang, 1966; White, 1985).

Similar amino acid transport defining properties for

murine EcoR were described by Wang and coworkers (1991) in

Xenopus oocytes utilizing both radiolabelled substrate and

electrophysiological measurements of amino acid uptake.

While the name EcoR has been applied to the murine protein

and Rec-1 applied to the gene when describing its role in

viral infection, the cDNA has been termed mCAT1 (murine

cationic amino acid transporter 1), when considering its

amino acid transport properties

Interestingly, the cloning of the second member of this

family of amino acid transporters also came about in the

pursuit of genes unrelated to amino acid transporters, as was

the case for mCAT1. MacLeod and coworkers (1990a), in an

attempt to isolate genes involved in T-cell function and the

ability to promote tumor formation in mice, isolated several

cDNAs which were expressed at high levels in a highly

tumorigenic murine T-lymphoma clone (SL12.4) and at nearly

undetectable levels in a non-invasive sister clone. Using

probes, enriched in SL12.4 expressed mRNAs by subtractive
hybridization, and differential kgtlO cDNA library screening,

the same group isolated a cDNA (20.5) which coded for a

putative hydrophobic protein showing significant amino acid

sequence identity with the previously cloned mCAT1 (EcoR)

cDNA (MacLeod et al., 1990b). The 2.4 kb cDNA, termed Tea

(T-cell early-activation gene), predicted a 49.6 kDa protein

of 453 amino acids with seven membrane-spanning domains.

Although clearly related to the mCAT1 gene, the protein was

truncated, and the complete product of the Tea gene was not

described until the full length cDNA was isolated by the same

group (Reizer et al., 1993). The predicted amino acid

sequence added an additional 205 amino acids to the N-

terminus, contained 12 transmembrane domains, and shared 61%

amino acid identity with mCAT1. The Tea gene has been

localized to murine chromosome 8 (MacLeod et al., 1990b).

The function of the Tea gene was suggested by structural

similarity to a family of bacterial and yeast nutrient

transporters (Reizer et al., 1993), and was finally described

by Kakuda and coworkers (1993). Xenopus oocytes injected with

Tea cRNA demonstrated significantly increased uptakes of

cationic amino acids compared to water-injected controls. In

fact, of all the 20 naturally occurring amino acids, as well

as amino acid derivatives used for identification of specific

transport systems tested, only the L-cationic amino acids

elicited substrate-induced changes in membrane current

suggesting transmembrane flux. The transport elicited by the

Tea cRNA was saturable and Na- independent, similar to the

cationic amino acid transport mediated by the mCAT1 cDNA.

Although the K. for arginine determined for Tea and mCAT1

(approximately 200 pM) by Kakuda and coworkers (1993) was

two-fold that described for mCAT1 previously (Kim et al.,

1991), further analysis of amino acid transport by

electrophysiological measurements revealed a similar Km for
both transporters of approximately 150 to 250 pM. A similar

characterization of the complete Tea cDNA was also described

by Closs and coworkers (1993b). Although some confusion in

nomenclature existed due to the rapid isolation of these and

homologous transporters by independent laboratories. When

considering its amino acid transport properties, the full-

length protein product of the Tea cDNA is generally referred

to as mCAT2.

Using the partial mCAT2 cDNA provided by the MacLeod

laboratory, Closs and coworkers (1993a) cloned a third member

of the CAT family from murine liver, mCAT2a. The mCAT2 and

mCAT2a cDNAs arise from the same gene on mouse chromosome 8,

and the two sequences are identical with the exception of an

alternatively spliced region on the predicted fourth

extracellular loop. Radiolabelled amino acid uptake in

Xenopus oocytes injected with mCAT2a cRNA demonstrated

significantly higher accumulation of L-cationic substrates,

but interestingly, the apparent K, for arginine was nearly

ten-fold higher than that for either mCAT1 or mCAT2 (Closs et

al., 1993b) It should be noted that the original

designation for this cDNA was mCAT2a and the lymphocyte clone

mCAT2b (Closs et al., 1993a). However, it is generally

accepted that the liver clone is designated mCAT2a and the

lymphocyte clone, mCAT2.

Tissue Specific Expression of the CAT Transporters

In agreement with its function as the previously

kinetically defined system y, the expression of CAT1 is

fairly ubiquitous. The initial cloning of this cDNA from NIH

3T3 cells demonstrated its expression in murine fibroblasts

(Albritton et al., 1989). In murine tissues, the primary

transcript size of mCAT1 is 7.9 kb with an additional 7.0 kb

transcript, the result of differential polyadenylation

(Albritton et al., 1989), seen in some tissues (Kim et al.,

1991). Expression is seen in virtually every tissue examined

including brain, bone marrow, testis, skeletal muscle, heart,

stomach, spleen, kidney, lung, ovary, uterus, intestine, and

T-lymphocytes (Kim et al., 1991; Kakuda et al., 1993).

It rats, a primary transcript size of 7.9 kb, with an

additional, lower abundance transcript of 3.4 kb seen in some

cases, has been demonstrated by our laboratory in rat

placenta and hepatoma cells (Kilberg et al., 1993).

Transcripts are also seen in rat brain (Stoll et al., 1991),

rat intestine (Puppi et al., 1995), and primary culture

hepatocytes (Closs et al., 1993b).

Human hCAT1 mRNA expression has been reported as a 9.0

kb and 7.9 kb transcript in human T-cells (Yoshimoto et al.,

1991) and human lymphoblastic leukemia, human kidney

carcinoma, human lung carcinoma, human colon adenocarcinoma

(Yoshimoto et al., 1991), and a 7.9 kb transcript in CaCo2

human intestinal cells (Pan et al., 1995). An immunoreactive

protein, detected with an antipeptide antibody to CAT1, was

seen on the plasma membrane of human fibroblasts by our

laboratory (Woodard et al., 1994) and represents the only

published report of the detection of the endogenous CAT1


The only tissue lacking expression of CAT1 is the liver.

Consistent with these data and the lack of mRNA reported by

several groups (Kim et al., 1991; Kakuda et al., 1993;

Kilberg et al., 1993; Closs et al., 1993; Wu et al., 1994),

and by the lack of immunoreactive CAT1 protein by

histochemistry in rat and human liver (Woodard et al., 1994).

The expression of the CAT2 gene isoforms is somewhat

more limited than the expression of CAT1. The mCAT2 gene was

isolated from and is expressed at very high levels in a

murine T-lymphoma cell line capable of causing invasive

tumors in mice (MacLeod et al., 1990a). Subsequently, mCAT2

expression has been shown to be limited to Concanavalin A

activated spleen cells, thymic epithelial cells, and liver

cells (MacLeod et al., 1990b). Tissues also demonstrating

expression of mCAT2 mRNA include skeletal muscle, skin, lung,

brain, uterus and stomach (Kakuda et al., 1993).

Expression of mCAT2a mRNA is even more limited in that

detectable levels of this isoform are only detectable in

liver (Closs et al., 1993).

Regulation of CAT Transporter Expression

Expression of CAT1 seems to involve the regulation of

this transporter when cells, either by naturally occurring

developmental changes or inducible means, are shifted from a

quiescent, differentiated state to an undifferentiated,

rapidly growing state. Using several different model

systems, Meruelo and coworkers demonstrated the above

(Yoshimoto et al., 1991). First, activation of murine T-

cells by Concanavalin A, or B-cells by LPS, increased steady-

state mCAT1 mRNA expression by over 10-fold compared to that

of quiescent splenocytes or thymocytes. Splenomegaly,

induced by mink cell focus-forming virus, also increased the

expression of mCAT1 mRNA by over 10-fold. The down-

regulation of CAT1 was demonstrated in the HL-60

promyelocytic leukemia cell line as it progresses toward

terminal differentiation into granulocytes or macrophages by

either DMSO or PMA, respectively. Finally, this group also

demonstrated the elevated expression of CAT1 in tumor cells

when compared to noncancerous parental cell lines.

Yoshimoto and coworkers (1991) also made initial

observations into the mechanisms behind the regulation of

CAT1 expression by providing evidence that increased CAT1

expression is modulated by protein kinase C (PKC). Evidence

for PKC modulation of CAT1 includes 1) the increase in CAT1

expression when cells are treated with phorbol esters, 2) the

failure to increase CAT1 expression by phorbol ester analogs

that do not induce the PKC pathway, 3) the inhibition of

phorbol ester-induced CAT1 expression by PKC inhibitors, and

4) the induction of CAT1 expression by calcium ionophores.

CAT1 expression is not detectable in normal adult liver

(Kim et al., 1991; Kakuda et al., 1993; Kilberg et al., 1993;

Closs et al., 1993b; Wu et al., 1994), and consistent with

the lack of CAT1 is the inability of MuLV to infect adult

hepatocytes (Jaenisch and Hoffman, 1979). It has been

demonstrated that infection of liver cells with retroviruses

is only possible after partial hepatectomy (Jaenisch and

Hoffman, 1979), several groups have focused on the

examination of CAT1 expression in models which induce of

liver cell division, such as liver regeneration. Closs and

coworkers (1993c) found induction of mCATI expression in

murine hepatocytes after 48 hrs of culture but no detectable

CAT1 expression in liver 12 to 48 hrs after partial

hepatectomy. The latter result seems to contradict the

retroviral infection data. Although Wu and coworkers (1994)

also demonstrated no detectable CAT1 mRNA 12 hrs after

hepatectomy, there is an induction of CAT1 synthesis at 4 to

6 hrs after hepatectomy, which parallels the time course of

viral infection. CAT1 expression in liver can also be

induced by other stimuli including insulin, dexamethasone,

and excess arginine (Wu et al., 1994). Nutrient regulation

of CAT1 in liver has also been demonstrated by induction of

CAT1 synthesis in animals chronically fed a low-protein diet

(Wu et al., 1994).

Developmental regulation of CAT1 has also been

demonstrated. In a published report from the analyses

carried out in this study, our laboratory examined the

regulation of amino acid transport in rat placenta and found

an induction of CAT1 mRNA expression from 14 to 20 days

gestation, equilivent to the final trimester of human

pregnancy (Malandro et al., 1994). It is thought that this

induction in transport is to supply amino acids to

accommodate the tremendous increase in fetal growth in late

gestation. Our data was novel in that we demonstrated not

only an increase in steady-state mRNA levels but also a

parallel increase in cationic amino acid uptake, confirming

the role of CAT1 as an amino acid transporter as well.

Campione and VanWinkle (1994) have also described the

developmental regulation of the CAT genes, detecting both

mCAT1 and mCAT2 in as early as the one-cell stage of the

mouse blastocyst. The appearance of the CAT mRNAs paralleled

the appearance of the Na'-independent, murine blastocyst

cationic amino acid transport System y. Subsequently, Puppi

and Henning (1995) examined the developmental regulation of

CAT1 in rat intestine and found high levels of CAT1 in fetal

intestine that decreased upon birth and eventually returned

to high levels in adulthood.

The most well described analysis of the regulation of

the CAT2 gene isoforms led to the cloning of mCAT2. mCAT2

gene expression is undetectable in quiescent splenic T-cells,

and enhanced over 15-fold upon activation of those same cells

with Concanavalin A (MacLeod et al., 1990a).

Due to the fact that regulation of mCAT2 parallels that

of mCAT1 in activated T-cells (Yoshimoto et al., 1991),

Kakuda and coworkers (1993) examined the expression of mCAT2

in regenerating liver, a condition regulating the expression

of mCAT1 (Wu et al., 1994). No change in mCAT2 steady-state

mRNA level was seen in either control, sham-operated, or

regenerating liver at 24 hrs after partial hepatectomy

(Kakuda et al., 1993). However, it should be noted that

increased expression of mCAT1 could not be seen at 24 hrs

after hepatectomy as well (Closs et al., 1993), but enhanced

expression was seen from 4 to 6 hrs after hepatectomy (Wu et

al., 1994), a time point not examined for mCAT2. It still

remains to be seen if the transient increase in mCATI

expression is accompanied by a transient increase in mCAT2 at

these earlier time points. Interestingly, regulation of

mCAT2 was seen in skeletal muscle following either

hepatectomy, splenectomy, or fasting (Carol MacLeod, personal

communication). It is postulated that this increase in

transporter is to allow for the efflux of amino acids from

the muscle, a major source of amino acids during surgical

trauma and fasting (Souba, 1987; Rennie et al., 1990).

Genomic Regulation of CAT2 Transporter mRNA Splicing

The regulation of expression of the mCAT2 gene isoforms

involves various mRNA splicing events in both the coding and

non-coding regions of the gene. As was previously mentioned,

two different mRNA isoforms are produced from the mCAT2 gene

in a tissue specific manner. The mCAT2 and mCAT2a predicted

protein products differ in only a single stretch of 41 or 40

amino acids, respectively, in the 4th extracellular loop of

the transporter (Kakuda et al., 1993; Closs et al., 1993a).

Analysis of many cDNA clones of mCAT2 from T-lymphoma

cells by MacLeod and coworkers, led to the identification of

4 splicing isoforms which contain the entire coding sequence

but differ in 5' untranslated regions (Finley et al., in

press). An independent laboratory isolated a fifth splicing

isoform (Kavanaugh et al., 1994). Each of these mRNA

converge into a common sequence 16 base pairs upstream of the

initiation AUG start codon. The mCAT2 gene 5' untranslated

exons span a distance of over 18 kb which is thought to

contain up to 5 distinct promoters (Finley et al., in press).

The sequence upstream of the first untranslated exon contains

no TATA sequence but DNA binding motifs common to TATA-less

promoters including GC-rich sequences and SP1 binding sites,

and it is the product of this promoter that is found in all

cell types that express mCAT2. The sequence upstream of the

second untranslated exon is a classic TATA promoter. The

products of this and the other putative promoters are likely

responsible for the complex tissue specific expression of the

mCAT2 gene.

Structure and Functional Analysis of the CAT Transporters

Wang and coworkers (1991) have demonstrated in voltage

clamped Xenopus oocytes injected with mCAT1 cRNA, an

inwardly-directed, saturable current in the presence of

substrate. When the magnitude of the current is compared to

transport of radiolabelled substrate in similarly injected

oocytes, movement of one positive charge per molecule of

substrate into the cell is seen. Kavanaugh (1993) examined

the change in membrane current resulting from CAT1-induced

transport, and at membrane potentials ranging from +20 mV to

-120 mV, arginine uptake followed values predicted by

Michaelis-Menten kinetics. Interestingly, at concentrations

of arginine from 0.01 to 1 mM, influx increased exponentially

with membrane hyperpolarization rather than demonstrating

saturation kinetics. This result provided evidence that the

rate limiting step in transport, thought to be the

reorientation of the transporter to the outward facing

position, is voltage dependent. If the conformational change

resulting in the transition of the unliganded receptor to the

external membrane conformation were not associated with any

charge movement, then steady-state flux would approach a

maximum value determined by the rate-limiting voltage-

independent step for this transition, and saturation of the

transporter at hyperpolarized potentials would be observed.

These results suggest that a charge translocation occurs

during the reorientation of the unliganded receptor, and that

this charge movement significantly affects the voltage-

dependence of steady-state cationic amino acid flux mediated

by mCAT1. These data also provide a fundamental difference

in the function of mCAT1 with respect to transport mediated
by the Na+-dependent glucose and y-aminobutyric acid

transporters, both of which approach maximal rate at

hyperpolarized potentials (Kavanaugh et al., 1994).

In an effort to determine the structural features

responsible for the differences in transport activity between

mCAT1 and mCAT2/mCAT2a, Closs and coworkers (1993b) examined

the expression of chimeric proteins in Xenopus oocytes.

mCAT1, previously described in a similar system, has a high

substrate affinity and is capable of trans-stimulation by

increased concentrations of substrate on the opposite side of

the membrane (Closs et al., 1993a), similar transport

properties have been demonstrated for mCAT2 (Closs et al.,

1993b; Kakuda et al., 1993). However, mCAT2a, the liver-

specific mCAT2 isoform, has a 10-fold greater apparent K. for

arginine than mCAT1 or mCAT2 but a higher V, (Closs et al.,

1993a). The only difference between mCAT2 and mCAT2a is a

stretch of 40 or 41 amino acids contained within a putative

extracellular domain. The substitution of this region from

mCAT2a into mCAT1 resulted in a greater K. and V,, consistent

with the properties of mCAT2a. The reciprocal construct of

mCAT1 into mCAT2a resulted in a Km and V, consistent with


An examination of critical amino acid residues for the

function of mCAT1 was provided by Kabat and coworkers (Wang

et al., 1994). The authors observed the conservation of a

glutamate residue (position 107 in mCAT1) residing in a

putative membrane-spanning region of the transporter. This

residue was not only conserved when comparing mCAT1 to mCAT2

and mCAT2a, but also in the arginine and histidine permeases

from yeast. The conservative substitution of aspartate (Asp)

for glutamate (Glu) at position 107 was not thought to alter

the overall structure of the transporter. This fact was

supported by the observation that the transporter reached the

plasma membrane and bound gp70. However, the Asp-Glu

substitution led to the complete abolishment of amino acid

transport activity.

Recent reports on the function of mCAT1 and mCAT2 have

led to some interesting findings that may force investigators

to reevaluate the "one protein-one activity" paradigm. Van

Winkle and coworkers (1995) used Xenopus oocyte expression,

as well as the analysis of previously published data by other

investigators, to kinetically examine the transport activity

mediated by mCAT1 and mCAT2. Interestingly, when

electrophysiological measurements were used to assay

transport, the L-arginine uptake mediated by both mCAT1 and

mCAT2 demonstrated biphasic Eadie-Hoffstee plots indicating

at least two kinetically distinct transport activities. Both

activities could either be mediated by a single CAT protein,

or the CAT protein could be active in addition to

upregulating the activity of an endogenous oocyte

transporter. Indeed, a major transport system functioning in

the oocytes for the transport both cationic and neutral amino

acids is upregulated by certain cRNAs (described below).

The Glutamate Transporter Family

Even before the cloning of the glutamate transporter

family members, it was suggested that there would be several

members to this family as glutamate transport in synaptosomal

preparations from different regions of the brain were found

to be differentially sensitive to pharmacological agents such
as dihydrokainate and L-a-aminoadipate (Hediger, 1995).

However, unlike the CAT family where the cloning of the

initial two members of the family was by chance, the cloning

of the members of the glutamate family came from a concerted

effort involving protein purification, antibody production,

library screening, and expression cloning for the

identification of the involved cDNAs.

Cloning of the Glutamate Family

The first member of this family, GLAST, was isolated by

Stoffel and coworkers (Storck et al., 1992). During the

isolation of a galactosyltransferase from the rat brain, a

hydrophobic protein of about 66 kDa was copurified, which,

upon deglycosylation with endoglycosidase F, yielded a

protein of 60 kDa. Proteolytic fragments of this protein

were then sequenced, and oligonucleotide probes were

generated for sequencing of a rat brain cDNA library. A 3 kb

clone was isolated which coded for a protein of 543 amino

acids with a predicted molecular mass of 59.7 kDa. This cDNA

showed considerable sequence homology to the previously

cloned glutamate and carboxylate transporters of bacteria.

The predicted protein sequence also showed similar structural

features to other transport proteins including multiple

membrane spanning domains and multiple glycosylation sites.

However, the GLAST sequence shows only six well defined

putative transmembrane domains at the N-terminus. The C-

termnial portion contains six short hydrophobic stretches

which are conserved in the prokaryotic transporters as well.

This unique membrane topology is different from the CAT

transporters described above which have at least 12 well

defined predicted membrane-spanning regions, and the

description of this topology remains the subject of

investigation. To determine the function of the GLAST cDNA,

Xenopus oocytes were injected with cRNA and transport assayed

by uptake of radiolabelled substrate (Storck et al., 1992).

Transport of both glutamate and aspartate was demonstrated to

be Nat-dependent and inhibited by DL-threo-3-hydroxyaspartate,

the strongest known inbibitor of Na-glutamate transport in

brain slices (Eisenberg et al., 1984). The human homolog of

the GLAST transporter was subsequently isolated, by two

independent laboratories, from human brain by low stringency

screening of cDNA libraries by Kawakami and coworkers (hGluT,

Kawakami et al., 1994) and by Arriza and coworkers (EAAT1,

Arriza et al., 1994). The human GLAST shares 97% amino acid

sequence identity with the rat transporter. It should be

noted that in this manuscript the original clone name of

GLAST1 will be used in reference to this transporter isoform

from either rat or human not only to credit the original

isolation of this clone but also to eliminate ambiguity which

may exist particularly with respect to GluT, a designation

that has been universally established for facilitated glucose

transporter clones.

The second (GLT1) and third members (EAAC1) of this

family were isolated shortly after GLAST by Pines and

coworkers (1992) and Kanai and Hediger (1992), respectively.

The cloning of GLT1 essentially began several years before the

isolation of the cDNA with the purification to near-

homogenity of the GLT1 transport protein (Danbolt et al.,

1990), and the subsequent generation of both polyclonal

(Danbolt et al., 1992) and monoclonal antibodies (Hees et

al., 1992). The polyclonal antibody was used to immunoscreen

a cDNA expression library, and a 4.6 kb positive clone (GLT1)

was isolated (Pines et al., 1992). The cDNA predicted an

open reading frame of 1719 bp coding for 573 amino acids with

a molecular mass of 63 kDa. The hydropathy profile of GLT1

is similar to that of GLAST1 in that it predicts 6 to 8

putative membrane spanning regions at the N-terminus and

several smaller hydrophobic regions at the C-terminus. GLT1

shares homology with prokaryotic glutamate transporters and

shares 44.4% amino acid sequence identity and 67.6% homology

with GLAST1. Transport assays, to confirm the function of

the GLT1 cDNA, were carried out by the expression of the cDNA

in mammalian cells and the subsequent transport of

radiolabelled substrate. Transport of L-glutamate was Na-

dependent and stereospecific in that the uptake was

completely inhibited by L-glutamate and not D-glutamate.

Transport of other neurotransmitters was not mediated by this

cDNA as evidenced by the lack of inhibition with GABA,

dopamine, noradrenaline, and serotonin. The activity

mediated by this cDNA was dependent on internal K but was not

inhibited by the K' ionophore valinomycin, suggesting the role

of K' as a substrate rather than the K+ gradient being the

driving force for transport of glutamate. Arriza and

coworkers (EAAT2, Arriza et al., 1994) and Manfras and

coworkers (GLTR, Manfras et al., 1994) have each identified

the human homolog of GLT1. Although some small sequence

discrepancies exist, these cDNAs most likely arise from the

same gene and share 90.2% amino acid sequence identity with

the rat GLT1. Another human clone was isolated by

Shashidharan and coworkers (GLAST2, Shashidharan et al.,

1994). However, this clone had sequence variations

representing a deletion of the first 10 amino acids compared

to the rat sequence and two single base deletions which

introduces a frame-shift mutation for amino acids 253 281

in the human sequence. The remainder of the sequence is

identical to the two other human clones. These investigators

isolated 5 additional clones from two independently prepared

cDNA libraries and obtained sequences identical to the

reported cDNA confirming their original observations. As was

discussed above, mCAT2 and mCAT2a arise from the same gene

and differ in only 41 amino acids resulting from a

differential splicing event. Hence, additional

experimentation is required to determine if an analogous

situation exists with the GLT1 transporter. A mouse homolog

of GLT1 was also isolated by Kirschner and coworkers (EAAT2,

Kirschner et al., 1994) and has been localized to mouse

chromosome 2. It is interesting to note that this

transporter maps near two quantative trait loci that are

related to mouse models of epilepsy and hyperexcitability.

This result could suggest that the abnormal transport of this

neurotransmitter may be involved in these phenotypes. Again,

as with the GLAST1 transporter, the GLT1 transporter isoform

will be referred to in this manuscript as GLT1 either from

rat, human, or mouse to eliminate ambiguity.

The third member of the glutamate transport family was

identified from rabbit intestine by Kanai and Hediger (1992).

A Xenopus oocyte expression cloning system was used to

express mRNA isolated from rabbit small intestine, and a peak

mRNA size fraction was reverse transcribed and used to

generate a plasmid cDNA library. A single cRNA was

subsequently identified, injected into oocytes, and

radiolabelled substrate transport measured. A cDNA (EAAC1)

of 3.4 kb was isolated with an open reading frame coding for

a protein of 524 amino acids with a predicted molecular mass

of 57 kDa. The EAAC1 isoform shares 65.6% amino acid

sequence similarity with GLAST1 and 60.8% with GLT1.

Expression of EAAC1 resulted in the increase in saturable Na'-

dependent transport of radiolabelled L-glutamate over 1000-

fold greater than that of water injected oocytes.

Electrophysiological analysis of transport revealed high-

affinity K.'s for L-glutamate and L-aspartate in the low
micromolar range (12 to 6 (M) and also demonstrated the

inhibition of this uptake by D-aspartate and DL-threo-B-

hydroxyaspartate, two strong inhibitors of synaptosomal

glutamate uptake (Rauen et al., 1992). Subsequently, both

Arriza and coworkers (EAAT3, Arriza et al., 1994) and Kanai

and coworkers (EAAC1, Kanai et al., 1994) have identified an

identical clone as the human homolog of EAAC1 which shares

95.6% amino acid sequence similarity with the rabbit EAAC1.

Hediger and coworkers have localized the human EAAC1 to

chromosome 9p24 (Smith et al., 1994). Our laboratory in

collaboration with Bob Fremeau (Duke University) has isolated

the rat homolog of EAAC1 which shares 94.6% similarity with

the rabbit clone (Chapter 4).

Recently, a fourth member of this family has been

isolated by Fairman and coworkers (1995) using degenerate

primers from the other members of the glutamate family in low

stringency PCR from human cerebellum mRNA. The clone (EAAT4)

shares amino acid sequence identity with the other human

clones identified by the Amara laboratory on the order of 65%

(human GLAST), 41% (human GLT1), and 48% (human EAAC1).

Expression of EAAT4 in oocytes demonstrated significantly

increased, saturable L-glutamate uptake when using either

radiolabelled amino acid or electrophysiological measurements

to assay transport. Interestingly, substitution of

extracellular gluconate for chloride completely inhibited the

outwardly-directed current induced by L-aspartate at

depolarizing membrane potentials suggesting a role for

chloride in transport. However, when membrane potentials

were clamped at -60 mv, the substrate-induced current was not

significantly different either in the presence or absence of

chloride. The data are consistent with a functional model

for EAAT4 as a chloride channel activated by substrate

(glutamate/aspartate and sodium) binding. This unique

property of EAAT4 may be involved in not only the re-uptake

of glutamate from the synapse, as is the case for the other

glutamate transporters, but also may aid in the re-

establishment of membrane potential by cellular chloride


Distribution of the Glutamate Transporters

In the initial identification of rat GLAST1, mRNA

analysis by Northern blotting detected a single mRNA species

of 4.5 kb with expression exclusively in the brain (Storck et

al., 1992). In situ hybridization of frozen brain sections

was performed with the same probe and revealed a low-level,

uniform distribution across the entire cerebellum with high-

level expression seen only in the cerebellar cortex (Storck

et al., 1992). Upon further analysis, GLAST1 mRNA in the

cerebellum was seen in only the Perkinje cell layer. In situ

hybridization of rat retina by Otori and coworkers (1994)

revealed expression in retinal glial cells, specifically

MUeller cells and astrocytes. Rat GLAST protein has been

localized by immunoblotting, with a polyclonal antisera

generated from the cloned rat sequence, as a broad band with

a molecular mass of approximately 66 kDa (Lehre et al.,

1995). Expression of the protein correlates well with the

expression of the mRNA such that GLAST1 protein is seen

primarily in the molecular layer of the cerebellum with

expression being restricted to astrocytes. No GLAST1 protein

is detected in neurons. Interestingly, human GLAST1 appears

more ubiquitously expressed than rat GLAST1. A single mRNA

species of approximately 4.0 to 4.2 kb is detected not only

in brain but also in heart, lung, placenta, and skeletal

muscle (Kawakami et al., 1994; Arriza et al., 1994). In the

brain, the expression of human GLAST1 mRNA is abundant in

motor cortex, frontal cortex, hippocampus, basal ganglia, and

cerebellum (Arriza et al., 1994).

Similar to that of GLAST1, the initial descriptions of

rat GLT1 transporter mRNA expression revealed a 10 kb single

mRNA species whose expression was restricted to the brain

(Pines et al., 1992). The human homologs of GLT1, with a

similar mRNA size, were also predominantly brain-specific,

with a very weak but detectable signal in placenta

(Shashidharan et al., 1994, Arriza et al., 1994). Brain-

specific GLT1 mRNA expression was seen at high levels in the

motor cortex, frontal cortex, hippocampus, and basal ganglia

with lower levels in the cerebellum (Arriza et al., 1994).

The mouse GLT1 mRNA is expressed similarly to that of both

the rat and human homologs. However, a weak but detectable

band is also observed in the liver (Kirschner et al., 1994).

The rat GLT1 protein has been localized by immunocytochemical

methods using a polyclonal antisera to glial cells,

specifically the astrocytic processes (Danbolt et al., 1992),

and is expressed at highest levels in the hippocampus,

lateral septum, cerebral cortex, and striatum (Lehre et al.,

1995). No GLT1 protein was seen in glutaminergic nerve

terminals (Danbolt et al., 1994; Lehre et al., 1995).

Although the GLAST and GLT1 transporter isoforms have a

similar distribution in the brain, the distribution of these

two transporters in the retina, a major site of glutaminergic

nerve innervation, may significantly differ. GLAST mRNA

expression has been localized to MUeller cells and astrocytes

(Otori et al., 1994), while GLT1 expression is absent from

these cell types as detected by immunocytochemistry (Rauen

and Kanner, 1994).

By far, EAAC1 is the most widely expressed of the

glutamate transporter isoforms. The rabbit EAAC1 is

expressed as 3.5 kb and 2.5 kb mRNA species in small

intestine, kidney, brain, liver and heart (Kanai and Hediger,

1992), with a similar distribution in the human (Arriza et

al., 1994; Kanai et al., 1995) and in the rat (Velaz-

Faircloth et al., in press). The brain-specific expression

of human EAAC1 reveals nearly equal expression in the motor

cortex, frontal cortex, hippocampus, basal ganglia, and

cerebellum (Arriza et al., 1994). In situ hybridization

performed by our laboratory reveals expression of rat EAAC1

in the cerebellar granule cell layer, hippocampus, superior

colliculus, and neocortex (Velaz-Faircloth et al., in press).

In all cases, EAAC1 mRNA expression was localized to neuronal

cell bodies. These data are consistent with the suggestion

that GLAST1 and GLT1 represent glial-specific isoforms while

EAAC1 represents a neuronal-specific isoform (Kanai and

Hediger, 1992).

The EAAT4 isoform mRNA is expressed as a single mRNA

species of 2.4 kb and is expressed exclusively in the brain

and placenta (Fairman et al., 1995). Northern blot analysis

demonstrates exclusively cerebellar expression. However,

rtPCR analysis can detect low levels of EAAT4 in brain stem,

cortex, and hippocampus.

Regulation of Glutamate Transporter Expression

Developmental regulation glutamate transport in the

brain was provided before any of the transporters were

specifically identified. Utilizing Xenopus oocyte expression

of mRNA isolated from developing rat brain, Blakely and

coworkers (1991) using pharmacological inhibitors of

transport examined the regulation of transport from fetus to

adult. Glutamate transport demonstrated a region-specific

increase in transport activity throughout development. The

suggestion of different proteins mediating glutamate

transport in different regions of the brain was also provided

by the differential sensitivity of transport mediated by mRNA

from different regions of the brain. Specific mRNA sizes

were also analyzed which correspond to the subsequently

cloned transporter isoforms.

Otori and coworkers (1994) examined the regulation of

glutamate transport during ischemia in the rat retina. As

was previously mentioned, GLAST1 mRNA expression is primarily

limited in the retina to cells in the inner nuclear layer

expressing glial fibrillary acidic protein (GFAP), primarily

MUeller cells and astrocytes. Retinal ischemia, mediated by

the ligation of the central renal artery for 90 minutes then

the subsequent reperfusion of the retina for 48 hours, causes

neuronal cell death in the inner nuclear layer. However, the

glial cells expressing GLAST1 mRNA remain and express the


transporter at higher levels. Glutamate transport, possibly

mediated by GLAST1 may operate in a reverse mode under

ischemic conditions, thus allowing glutamate efflux rather

than uptake (Atwell et al., 1993). The authors suggested

that the increased GLAST1 would then be available during

reperfusion to clear the high levels of extracellular

glutamate remaining from the period of ischemia.

The regulation of glutamate transport via GLT1 was

described by Casado and coworkers (1993). In vitro, GLT1

purified from porcine brain served as a substrate for protein

kinase C leading the investigators to examine glutamate

transport activity in a rat glioma cell line (C6) and a HeLa

cell line transiently expressing the GLT1 transporter. Upon

treatment of both cell lines with phorbol ester, there is a

rapid increase (30 min) in L-glutamate transport which

follows a similar time course as the phosphorylation of the

GLT1 protein. Further evidence for the role of

phosphorylation was provided by the mutation of the single

protein kinase C consensus phosphorylation site. Site-

directed mutation of serine 113 to asparagine did not affect

the basal transport rate or transporter number assayed by

immunoblot analysis. However, the phorbol ester-mediated

increase in glutamate transport was abolished.

Plakidou-Dymock and coworkers (1993) described the

regulation of glutamate transport by amino acid deprivation

in the bovine renal epithelial cell line NBL-1. The

regluation of neutral amino acid transport via System A is

well described in many tissues and cell types (Kilberg et

al., 1993). The authors used a similar method to study the

regulation of glutamate transport by the removal of all

extracellular amino acids. Glutamate transport increased in

V, by a single kinetically defined system analogus to the

previously defined System X- (Gazzola et al., 1981). Probes

generated to the rabbit EAAC1 sequence by low-stringency PCR

identified a single mRNA species consistent with this

isoform. However, the regulation of steady-state mRNA level

does not correspond to the regulation of glutamate transport

in these cells. In fact, the steady-state mRNA level

decreases during the period of increased glutamate transport.

In this case, one would suggest either a regulation of the

transporter at the post-transcriptional level of the presence

of another of the transporter isoforms mediating the observed

glutamate transport activity.

Structure/Function Relationships of the Glutamate

All members of the glutamate transport family share a

similar structure of at least six transmembrane spanning

domains at the N-terminal portion of the protein and at least

two conserved putative glycosylation sites located on the

extracellular loop between transmembrane segments three and

four. Conradt and coworkers (1995) examined the potential

role of glycosylation of the GLAST1 transporter by

elimination of the two glycosylation sites by site-directed

mutagenesis. Comparison of the wild-type and mutant

transporters revealed that both of the putative glycosylation

sites were involved in N-linked glycosylation. However,

replacement of the asparagine residues at these sites with

threonine resulted in completely functional and properly

targeted transporter.

In an effort to elucidate the residues in the GLT1

protein essential for glutamate transport activity, Zhang and

coworkers (1994) utilized site-directed mutagenesis to change

two amino acid residues completely conserved in this family.

Lysine 298 and histidine 326, both residues in predicted

membrane-spanning regions, were substituted with small

hydrophilic and charged amino acids. Substitutions in GLT1

of arginine or histidine for lysine result in functional

proteins that retain nearly full activity. However, less

conservative substitutions of glutamine and threonine at the

same position result in a significant decrease in transport

activity. These changes were demonstrated to arise from

targeting defects of these two mutants as reconstitution of

these proteins, prepared from whole cell extracts rather than

plasma membranes, into proteoliposomes showed essentially the

same activity as the wild type in this assay. However, short

of immunofluorescence of the plasma membranes to determine if

the mutant transporters reach the plasma membrane, the

interpretation of the reconstitution data is limited. In

contrast to the lysine mutants, any substitution for the

histidine residue at position 326 resulted in a completely

non-functional protein. These data suggested that histidine

326 is essential for the function of GLT1 and possibly was

involved in substrate binding or translocation. In an

additional analysis of the GLT1 transporter, Pines and

coworkers (1995) used site-directed mutagenesis to examine

the function of conserved negatively-charged amino acid

residues located in the hydrophobic portions of the C-

terminus of the transporter. Mutation of three of the five

conserved residues (aspartate 398, aspartate 470, glutamate

404) resulted in a significant reduction in activity even

when these residues were substituted with conserved amino

acids of the same charge. Interestingly, substitution of

glutamate 404 not only reduced the glutamate transport

activity of this transporter, but changed the substrate

specificity as well in that both D- and L-aspartate were

accepted as substrates (80% of wild-type activity). As the

binding of glutamate was unaffected in this mutant, these

data suggest a possible selective impairment of glutamate

translocation or release on the opposite side of the plasma


Vandenberg and coworkers (1995), using electrophysio-

logical measurements of transport activity in Xenopus

oocytes, identified an inherent ion flux, in the absence of

glutamate, associated with the expression of human GLAST

(EAAT1) but not GLT (EAAT2). A chimeric transporter was made

by the substitution of residues 366 441 from GLT onto a

GLAST backbone. The substituted segment differs in only 18

amino acid residues from the corresponding segment of GLAST.

The chimeric transporter bound kainate with high affinity, a

property of the substituted domain of GLT exclusively. The

kainate binding also inhibited the constitutive ion flux of

the chimeric transporter. This activity was attributed to an

uncoupled flux of cations across the plasma membrane.

Differential activity of the GLAST versus GLT transporters is

also seen in the response of each transporter to arachidonic

acid, a messenger molecule involved in ischemia and released

upon activation of glutamate receptors (Zerangue et al.,

1995). Arachidonic acid inhibits glutamate transport via

GLAST by decreasing the transport velocity by about 30%. In

contrast, arachidonic acid stimulates transport via GLT by

increasing the affinity of the transporter for substrate by

about two-fold. Taken together, these data suggest that

although the glutamate transporter family is highly

homologous, up to 80% in some regions, differential

regulation and modulation of transport activity is essential

for normal brain function.

Associated Members of the Glutamate Transporter Family

The Kilberg laboratory, in collaboration with Fremeau

and coworkers at Duke University (SATT), and Amara and

coworkers (ASCT1) independently isolated a cDNA from human

brain that shares near 40% sequence identity with the members

of the glutamate transport family (Shafqat et al., 1993;

Arriza et al., 1993). Although some sequence discrepancy

existed, publication of a revised sequence indicated that the

two clones were in fact identical (Kilberg et al., 1994).

The cDNA identified an open reading frame of 532 amino acids

with a molecular size of 56 kDa. Analogous to GLAST, GLT,

and EAAC1, the ASCT1 sequence predicts six well defined

transmembrane sequences near the N-terminal of the protein

and up to six additional hydrophobic stretches near the C-

terminal. However, Arriza and coworkers (1993), again

utilizing Xenopous oocyte expression and subsequent

electrophysiological assays of transport activity,

demonstrated Na'-dependent uptake not of negatively charged

amino acids but of neutral amino acids including alanine,

serine, and cysteine. The data presented utilizing transient

expression of the transporter in mammalian cells by

recombinant vaccinia virus and subsequent assay of transport

activity by radiolabelled substrate, gave similar results

with the exception of cysteine (Shafqat et al., 1993), a

discrepancy that was later resolved (Kilberg et al., 1994).

These data were consistent with the previously kinetically

defined System ASC (Christensen et al., 1967). System ASC

activity has been identified as a high activity transporter

in nearly every cell type tested. The mRNAs (approximately

5.0, 3.5, 2.2 kb) for ASCT1 is expressed in high levels in

only the brain, skeletal muscle and pancreas. This may

suggest an analogous situation to the CAT transporter family

described above, in that a similar activity may be mediated

by different gene products in different tissues.

A unique property of the kinetically defined system ASC

is the ability to accept negatively charged substrates at low

pH (Makowske and Christensen, 1982; Vadgama and Christensen,

1984). When System ASC activity is assayed at neutral pH,

uptake of primarily neutral amino acids are observed.

However, decreasing the extracellular pH results in

competitive inhibition by anionic amino acids as well. Use

of anionic amino acid analogs cysteate or cysteinesulfinate,

with pK values near 1, eliminates titration of the substrate

to neutral charge. Tamarappoo and coworkers (in press),

utilizing expression of ASCT1 in HeLa cells, observed that

ASCT1-mediated activity was inhibited at low pH by anionic

amino acid analogs further confirming the assignment of

System ASC activity to the ASCT1 clone.

The rBAT and 4F2hc Family of Transporters

One of the most unique families of transport proteins

are those of the rBAT and 4F2hc family. These proteins up-

regulate the transport of some neutral and positively charged

amino acids when their cRNAs are injected into Xenopus

oocytes. However, the predicted structure of this family is

very different from the other transport families in that

these proteins are predicted to have only one or four

membrane spanning regions instead of the 10 to 12 predicted

in the other transporter families. This has led some

investigators to suggest these proteins function as

regulators or subunits of transport activity, and the

expression of these proteins in oocytes serves to induce an

already present oocyte transporter. Given either the role of

transport or regulatory protein, it is clear that a defect in

this protein has profound effects in the kidney leading to

the disease cystinuria.

Cloning of the rBAT/4F2hc Family

As was mentioned above, the members of this family were

identified by expression cloning strategies in Xenopus

oocytes. The initial observations from several laboratories

of the induction of cationic and neutral amino acid transport

in oocytes expressing size-fractionated kidney mRNA, laid the

foundation on which the members of this family were cloned

(Tate et al., 1989; Bertran et al., 1992b; Magagnin et al.,

1992). Expression cloning from three independent

laboratories led to the identification of a new class of

proteins identified in the kidney from rat (NAA-Tr, Tate et

al., 1992; D2, Wells and Hediger, 1992), rabbit (rBAT,

Bertran et al., 1992) and human (rBAT, Bertran et al., 1993;

D2, Lee et al., 1993). It should be noted that the human

sequence reported by Lee and coworkers (1993) appears to lack

22 amino acids of the C-terminus present in the other

proteins from rat and rabbit and a similar human clone. Each

of the proteins share 80 to 85% sequence identity and a

similar structure. We will use the nomenclature of Palacin

(1994) by referring to all of these clones as rBAT (related

to Basic Amino acid Transport).

The rBAT cDNA predicts an open reading frame of 2049

nucleotides coding for a protein of 683 amino acids with a

molecular mass of 77 to 79 kDa. The rBAT protein contains

seven putative glycosylaton sites, and a leucine-zipper motif

is seen in the rat protein from amino acids 548 to 569 (Wells

and Hediger, 1993). Most notably, however, is the unique

structure of this protein. Hydrophobicity profiles of rBAT

revealed either four (Tate et al., 1992) or one (Wells and

Hediger, 1992; Bertran et al., 1992; Bertran et al., 1993;

Lee et al., 1993) transmembrane domains. Given the high

degree of identity between these proteins, the apparent

discrepancy most likely arises from the hydropathy computer

program used to predict transmembrane segments rather than

from a fundamental difference in protein structure. Several

investigators are using molecular approaches to address the

topology of this family of proteins and will be discussed

below. This structure is in striking contrast to all other

nutrient transporters described in which 10 to 12

transmembrane domains are expected. Even in the glutamate

transporter family in which only the first six transmembrane

segments are well placed, there are at least four to six

other hydrophobic streches in the proteins that may possibly

be membrane bound. The only class of proteins that share

structural features with rBAT is the ligand-gated ion

channels including the nicotinic acetylcholine and GABA

receptors. These proteins, however, have an N-terminal

signal sequence which is cleaved from the mature protein such

that the N- and C- termini of these proteins are located

extracellularly. As such, the rBAT protein appears to define

a new class of membrane proteins.

The gene for the human rBAT was localized to chromosome

2-p21 (Lee et al., 1993; Yan et al., 1994). The functional

promoter of the rat rBAT gene has also been identified (Yan

et al., 1994).

Expression of the rBAT cRNA in oocytes results in

significantly increased the Na-independent uptake of both

neutral (including alanine, phenylalanine, leucine, and

cysteine) and cationic amino acids (lysine and arginine) over

water-injected controls. From these data, the Tate and

coworkers (1992) concluded that rBAT mediated transport

similar to the previously kinetically defined system L

(Oxender and Christensen, 1966). However, several features

of the cRNA-mediated amino acid uptake are not consistent

with this system. First, BCH, a model substrate for system

L, does not inhibit neutral amino acid uptake in the injected

oocytes. Second, there is no increase in the transport of

tryptophan, a high affinity substrate for system L, and

lastly, there is significant inhibition of neutral amino acid

uptake by cationic amino acids. These data are more

consistent with uptake similar to the kinetically defined

system b0',, a transporter of both neutral and cationic amino

acids (Van Winkle et al., 1988). This system is expressed at

high levels in the oocyte and serves as the major upake

mechanism for endogenous cationic amino acid uptake in

oocytes (Campa and Kilberg, 1991). This fact, along with the

unique structure of rBAT, has led some investigators to

suggest that the rBAT protein may be interacting with an

endogenous oocyte transporter, inducing system b0o, uptake

rather than functioning as a transport protein itself. Short

of purification of the rBAT protein to homogeneity and

reconstitution into proteoliposomes, this question will

remain unanswered.

Two laboratories identified sequence similarities

between the rBAT proteins and a previously cloned protein

4F2hc. The 4F2hc protein was identified in a human T-cell

line as a heterodimer composed of an 85 kDa glycosylated

heavy chain, which spans the membrane a single time, in

disulfide linkage with a 45 kDa nonglycosylated light chain

(Eisenbarth et al., 1980; Hemler et al., 1982). The cDNA was

subsequently isolated from both mouse and human (Quackenbush

et al., 1987; Teixeira et al., 1987; Parmaceck et al., 1989),

and the 4F2 heavy chain (4F2hc) was found to share 15% amino

acid sequence identity with the rBAT protein. The similarity

with the rBAT proteins led Bertran and coworkers (1992b) and

Wells and coworkers (1992) to investigate amino acid

transport activity associated with 4F2hc. Injection of 4F2hc

cRNA into oocytes resulted in the induction of Na'-independent

cationic amino acid uptake over water-injected controls.

Neutral amino acid transport was also induced but in a Na'-

dependent rather than Na'-independent manner. Also of note is

the fact that the expression of 4F2hc cRNA did not induce the

transport of cysteine, suggesting a fundamental difference

between the activities of rBAT and 4F2hc. The activities

induced by 4F2hc more closely resemble the previously

kinetically defined y4L identified in human erythrocytes

(Deves et al., 1992). System yL transport is characterized

as high affinity, Na'-independent cationic and Na-dependent

neutral amino acid uptake. However, 4F2hc could also be

interacting with System y', which mediates Na'-independent

cationic amino acid transport and has been shown to interact

with neutral amino acids in the presence of Na' (Christensen

et al., 1969b). Interestingly, the expression of 4F2hc in T-

cells and the possible interaction with cationic amino acid

transporters could suggest further examination of the

inducible expression of mCAT2 and its possible interaction

with 4F2hc.

Tissue Distribution of rBAT

rBAT mRNA is expressed in various tissues as a 2.2 to

2.5 kb species and a 3.5 to 4.4 kb mRNA species. Both of the

mRNA species identified and expressed in oocytes from rabbit

kidney mRNA (Markovich et al., 1993). Both transcripts can

induce transport in a similar fashion, and the difference in

size was demonstrated to be differential polyadenylation of

the transcripts. rBAT mRNA, as assayed by Northern analysis,

is expressed primarily in the kidney and cultured kidney

cells regardless of the species (Tate et al., 1992; Yan et

al., 1992; Wells and Hediger, 1992; Bertran et al.,1992b).

Northern analysis detects rBAT mRNA in both the cortex and

medulla of the kidney. In situ hybridization localized the

expression of rBAT mRNA in rat kidney primarily to the S3

portion of the proximal tubule (Kanai et al., 1992). Low

level expression is also seen in the S1 and S2 segments of

the tubule. This localization is consistent with the role of

rBAT in amino acid absorption in the kidney, as

microperfusion experiments of the proximal straight tubule

(Sl and S2) have identified high-affinity cysteine transport

in this area (Voelkl et al., 1982; Schafer et al., 1984).

High level expression is also seen in all segments of the

small intestine (duodenum, jejunum, and ileum). A related

transcript also appears in the brain and heart but at

slightly larger mRNA sizes (Yan et al., 1992). The identity

and role of these related transcripts is unclear.

rBAT protein has been detected by polyclonal antibodies

as a 84 to 87 kDa species in epithelial tissue from rat

kidney and small intestine (Mosckovitz et al., 1993).

Immunohistochemistry and immunoelectronmicroscopy have more

discretly localized the rBAT protein to the epithelial

membrane lining the rat kidney proximal tubule and to rat

jejunal microvilli (Pickel et al., 1993; Furriols et al.,


Functional Analysis of rBAT

Electrophysiological analysis of rBAT injected oocytes

has provided some interesting insight into the function of

this protein. Two independent laboratories demonstrated the

rBAT cRNA mediated electrogenic uptake of both neutral and

cationic amino acids (Busch et al., 1994; Coady et al.,

1994). Most interesting, however, was the observation that

neutral amino acid uptake mediated by rBAT resulted in the

generation of outwardly-directed currents, a process expected

to be electroneutral with simply the uptake of uncharged

substrate. The authors found that preloading of the oocytes

with cationic amino acids increased the uptake of neutral

amino acids and vise versa. However, significant inhibition

of uptake was seen when the preloaded amino acid had a

similar charge as the transported substrate. These data are

consistent with a model of "trans-stimulation" of uptake

where the uptake of charged amino acids is enhanced by the

outward flux of neutral amino acids and vise versa.

Role of rBAT in Cystinuria

The location of rBAT and its role in cysteine uptake in

the kidney made this protein a likely candidate for

cystinuria. This autosomal recessive genetic disease is

characterized by high levels of cystine and cationic amino

acids in the urine of these patients (Segal and Thier, 1989).

The high levels of cystine cause deposits in the urinary

tract and lead to infection and renal insufficiency.

Investigators suggested a likely cause of this disease was

aberrant reabsorption of cystine and cationic amino acids in

the kidney and small intestine (Rosenberg et al., 1969).

Sequencing of rBAT mRNAs expressed from several patients, and

subsequent sequencing of genomic DNA, revealed six point

mutations in the rBAT gene specific to cystinuria and

represented 30% of the cystinuric patients analyzed (Calonge

et al., 1994). The most commonly occurring mutation was the

substitution of threonine for methionine at position 467.


This mutation resulted in the decrease of nearly 80% in the

transport activity of rBAT expressed in oocytes for all

substrates tested (cystine, arginine, leucine). Additional

support for the identification of rBAT in cystinuria was

provided by linkage analysis which localized the cystinuria

gene to chromosome 2P (Pras et al., 1994), the location of

rBAT (Lee et al., 1993; Yan et al., 1994).

__ _


This chapter describes the general methods and protocols

used throughout this study. Some methods used in Chapters 4

and 6 differ considerably from the methods utilized in the

majority of this study described in this chapter. As such,

these methods will be described in the respective chapters in

which they appear.


L-arginine [2,3,-3H] and L-glutamic acid [2,3,-3H] were

obtained from DuPont New England Nuclear (Boston, MA).

Deoxycytidine 5'-[a-32P] triphosphate triethylammonium salt

was obtained from Amersham (Arlington Heights, IL). 2-

(methylamino) isobutyric acid, [2-methyl-3H], was obtained

from ARC (St. Louis, MO). Nitrocellulose filters (0.45 pm)

were used for transport assays (Millipore, Bedford, MA). All

other reagents and chemicals were of the highest quality

available and were obtained from either Sigma Chemical (St.

Louis, MO) or Fisher Scientific (Pittsburgh, PA). The MCAT1

(System y*) cDNA insert was a generous gift from Dr. James

Cunningham at the Brigham and Women's Hospital (Boston, MA).

The rat EAAC1 cDNA was cloned from a rat hippocampal library

as described in Chapter 4. The cathepsin B cDNA was a

generous gift of Dr. Harry Nick at the University of Florida

(Gainesville, FL). Timed pregnant Sprague-Dawley rat dams

were obtained from Zivic-Miller Company (Zellienople, PA).


Animal feeding. Animals were housed in a temperature-

controlled room with twelve hour light/dark cycles and given

water ad libitum. Animals in gestational studies were given

food similar in composition to the control diet, described

below, ad libitum. Animals in the maternal protein

restriction study were weight-matched and separated into

either control (C) or low-protein (LP) groups. Both the

control diet (19.3% protein, 10.0% fat, 4.3% fiber, 60.6%

carbohydrate) and the low-protein diet (4.6% protein, 10.0%

fat, 5.7% fiber, 75.3% carbohydrate) were prepared by Purina

Mills (St. Louis, MO), and the diets were made isocaloric

with the addition of sucrose. In order to ensure that the LP

and C groups consumed equal calories throughout the study, LP

and C animals were approximately weight matched and

subsequently pair-fed from days 6 to 19 gestation. The LP

animals were given the low-protein diet ad libitum, and the

amount of food consumed by each animal each day was recorded.

The C animals were then each given the control diet equal in

weight to the amount of food consumed by the pair-matched LP

animal the previous day. This feeding protocol continued

until sacrifice on day 20 gestation. These studies were

approved by the Laboratory Animal Medical Ethics Committee of

the University of Florida.

Apical and basal membrane vesicle isolation. At day 20

gestation, the dams were sacrificed and placentas removed and

weighed. Isolation of rat apical membrane vesicles was

performed by a modification of a human preparation (Booth et

al., 1980). Minced and washed placental villous tissue from

4 to 8 litters was incubated for 1 hour at 40C and then passed

through 210 pm nylon mesh. All steps were performed at 40C

unless otherwise noted. The filtrate was centrifuged at 800

x g for 10 min. The pellet was discarded and the supernatant

centrifuged at 10,000 x g for 10 min. The pellet was again

discarded. Crude apical membranes were then obtained by

centrifugation at 150,000 x g for 25 min, resuspension in

Tris-mannitol buffer (300 mM mannitol, 2 mM Tris-base, pH

7.0), and vesicle formation induced by 12 strokes in a glass

homogenizer with Teflon pestle. An equal volume of Tris-

mannitol buffer was added to the homogenate and MgCl2 added to

a final concentration of 10 mM. An additional 12 stroke

homogenization was performed, and the membranes incubated for

10 min. The solution was then centrifuged at 2,200 x g for

12 min and the pellet discarded. The supernatant was

centrifuged at 150,000 x g for 25 min to pellet the apical-

enriched membrane vesicles. The membranes were then

resuspended in HEPES-sucrose buffer (300 mM sucrose, 10 mM

HEPES-Tris-base, pH 7.4). An aliquot of the vesicle

preparation was used immediately for total protein and marker

enzyme analysis and the remainder stored frozen in liquid

nitrogen until use. Storage for several months was shown not

to affect transport activity.

Basal membrane vesicles were prepared utilizing the

tissue remaining on the 210 pm nylon mesh filter from the

apical membrane vesicle preparation. This material was

washed on the filter three times with ice-cold PBS and then

three times with 50 mM Tris-HCl (pH 6.9). The membranes were

then removed from the filter by gentle scraping, resuspended

in 50 mM Tris-HCl (pH 6.9), and sonicated. The membranes

were then re-filtered and washed 3 times with 5 mM Tris-HCl

(pH 6.9). The membranes were then incubated with stirring

for 30 min, filtered and washed. The washed membranes were

then added to an equal volume of Tris-sucrose-EDTA buffer

(300 mM sucrose, 2 mM Tris-base, 10 mM NaEDTA, pH 6.9). After

an additional incubation at room temperature for 10 min, the

solution was again filtered through the 210 pm nylon mesh.

The membranes were added to 10 mM EDTA, incubated at room

temperature for another 20 min, then sonicated for 20 sec.

The suspension was filtered through 6-ply gauze, centrifuged

at 4,000 x g for 10 min, and then at 10,000 x g for 20 min.

The pellet was discarded, and membrane vesicles pelleted by

centrifugation at 150,000 x g for 25 min. The final basal

membrane pellet was resuspended in HEPES-sucrose buffer and

treated as discussed above for the apical membrane vesicles.

Both vesicle preparations were analyzed for total membrane

protein as determined by the Lowry method using bovine serum

albumin as a standard (Lowry, 1951). Subsequent analysis of

vesicle preparations included assays for dihydroalprenolol

(DHA) binding, a marker for the basal membrane (Kelley et

al., 1983), and alkaline phosphatase, a marker enzyme for the

apical membrane (Bowers and McComb, 1966; Glazier et al.,


Vesicle transport assay. Timed uptakes of radiolabelled

amino acids were measured by a nitrocellulose filter assay

described previously (Novak et al., 1989). In brief, uptakes

were measured at 37"C and performed in uptake buffer (10 mM

HEPES-KOH, pH 7.5, 10 mM MgC2,, 0.2 mM CaCl2) with either 125

mM Na+ or K' thiocyanate (SCN). All uptake solutions were

adjusted to isoosmolarity (310 mosm) with the addition of

sucrose. Radiolabelled substrate was then added in tracer

amounts to unlabelled substrate to achieve 5-50 pM final

concentration, except where otherwise noted, and where

indicated, unlabelled amino acid inhibitors were added to the

uptake media at the concentrations noted. In each assay, a

20 pl aliquot of vesicle preparation corresponding to 20 70

pg protein was incubated at 370C for 2 min. After

temperature equilibration of the membranes, 80 pl of uptake

solution described above was added, and the mixture was

immediately vortexed. Transport was terminated by the

addition of 3 ml of ice-cold stop solution (100 mM NaSCN or

KSCN, 10 mM HEPES-Tris pH 7.4, 100 mM sucrose). The entire

mixture was then rapidly passed over a 0.45 pm nitrocellulose

filter and subjected to vacuum filtration. The uptake tube

was washed with an additional 3 ml aliquot of ice-cold stop

buffer, and this was poured over the filter as well. The

filter was washed two additional times with 3 ml stop buffer

and then placed in 4.5 ml Scintiverse Bio-HP scintillation

cocktail (Fisher Scientific, Pittsburgh, PA) to determine the

bound radioactivity. Membrane binding of substrate was

assessed by adding ice-cold stop solution to the

radiolabelled substrate prior to mixing with membranes and

subsequent filtration. These "blank" values were subtracted

from each assay. The uptake velocities are reported as

substrate' protein-unit-' time and are given as

the averages SEM for 3 4 assays.

Fetal hepatocyte isolation. Fetal hepatocytes were

isolated by the method of Leffert et al. (1979). Timed-

pregnant rat dams were sacrificed on the appropriate day

gestation, and the fetuses were removed to 370C perfusion

buffer (25 mM Na2HPO,, 3 mM KC1, 119 mM NaC1, 11 mM glucose, 1

mg/ml BSA, pH 7.4). Fetal livers were removed by dissection

and placed in sterile 37C perfusion buffer until all livers

were isolated. The buffer was aspirated and replaced with 1

mg/ml collagenase (Sigma, Type I) in perfusion buffer and

incubated for 15 minutes at 37C with vigorous stirring.

Dispersed cells were aspirated and placed into 50 ml 370C

arginine-free minimal essential medium (Arg--MEM) supplemented

with 4% fetal bovine serum (FBS). The cells were pelleted by

centrifugation at 128 x g for 2 minutes. The medium was then

removed and replaced with 25 ml of same. To the remainder of

the liver tissue, an additional aliquot of 1 mg/ml

collagenase was added and incubated again at 370C. The cells

from both incubations were pooled and washed five times in

Arg--MEM/4% FBS. Visual inspection of the pellet for the

absence of red blood cells was performed, and if

contamination by these cells was suspected, the cells were

washed additionally in the same manner until removal of

contaminants. Cells were counted and viability estimated by

exclusion of trypan blue. Hepatocytes isolated by this

procedure had greater than 95% viability. The cells were

then plated on culture plates pre-coated with 0.02 mg/ml

collagen (Sigma, Type III) at a density of 500,000 cells per

well of a 24-well cluster tray (Costar)(Kilberg, 1989).

Cells were cultured for only 24 hours to allow the cells to

attach as it has been demonstrated that after this time in

culture, the hepatocyte amino acid transporter complement is

significantly altered (Closs et al., 1993).

Whole cell transport assay. Amino acid uptake of

adherent hepatocytes was measured by the cluster tray method

of Gazzola et al. (1981) with modifications by our laboratory

(Kilberg et al., 1989). Hepatocytes were plated on 24-well

Costar trays as described above. To deplete the

intracellular pools of amino acids to minimize trans-effects

on transport (Kilberg et al., 1989), cells were incubated at

370C for 30 minutes in Na'-Krebs-Ringer phosphate (NaKRP)(119

mM NaC1, 25 mM Na2HPO4, 5.9 mM KC1, 1.2 mM MgSO,, 1.2 mM KHCO3,

5.6 mM glucose, 0.5 mM CaCl2, pH 7.4). Radiolabelled

substrate and the appropriate inhibitors, were then added to

0.25 ml of NaKRP, or to measure Na-independent uptake, to

0.25 ml choline-Krebs-Ringer phosphate

(cholKRP)(corresponding Na-salts substituted with choline

chloride and choline phosphate at the same concentration, pH

7.4). After transport, cells were washed 4 times with 3 ml

ice-cold cholKRP and subsequently air-dried. Cells were then

solubilized in 0.2 ml 0.2 N NaOH/0.2% SDS for 15 minutes. A

0.1 ml aliquot was added to vials with 5 ml scintillation

cocktail (Scintiverse, Fisher Scientific) to determine

radioactivity. The remaining 0.1 ml was used for estimation

of total protein as determined by the method of Lowry (1951).

SDS-PAGE and immunoblotting. SDS-PAGE (7.5%) was

performed by the method of Laemmli (1970). Ater

electrophoresis, the proteins were electrotransferred to a

0.45 yM nitrocellulose membrane (Schleicher & Schuell, Keene,

NH). Primary antibody was diluted 1:250-1:1000 in blocking

buffer (5% non-fat dry milk in 10 mM Tris-Cl, pH 7.5, 150 mM

NaC1), and incubated with the blot for 1 h at room

temperature with agitation. Horseradish peroxidase-

conjugated Protein A was used at 1:20,000 dilution in

blocking buffer for detection of the immunoreactive band by

visualization with the Enhanced Chemiluminescence Kit

(Amersham, Arlington Heights, IL). Densitometric analysis

was performed and presented as relative absorbance units per

microgram of protein.

RNA Isolation and Northern Analysis. Dams were

sacrificed, and for each isolation, one gram of placental

tissue was snap-frozen and ground to a powder in a mortar and

pestle under liquid nitrogen. The placental powder was then

added to denaturing solution (4 M guanidinium thiocyanate,

0.5% N-lauryl sarcosine, 25 mM sodium citrate pH 7.0, 100 mM

B-mercaptoethanol) to inhibit RNAse activity. The solution

was homogenized with 10 strokes in a glass homogenizer with a

motor-driven Teflon pestle. Total RNA was isolated by the

method of Chomczynski and Sacchi (1987), and poly A+ selected

mRNA was isolated from 500 yg total RNA using the Poly ATract

System (Promega, Madison, WI). Equal amounts of poly A+

selected mRNA were loaded per lane (3 pg) and subjected to 1%

agarose gel electrophoresis in the presence of 0.02 M

formaldehyde. The RNA was capillary transferred to nylon

membrane and hybridized with 32P-labeled cDNA probe prepared

by random priming extension (GibcoBRL, Gaithersburg, MD). The

resulting autoradiogram was quantified by densitometry. To

correct for loading differences, the levels of mRNA were

normalized to the constituatively expressed cathepsin B mRNA.

Experiments having loading differences of more than 10% per

lane were not used to quantify mRNA content. Preliminary

experiments were performed to ensure that the

autoradiographic exposures were within the linear range of

the film and densitometer.



The placenta serves as the functional interface between

the maternal and fetal circulations and is capable of

concentrative amino acid delivery to the fetus via

transporters in both the apical (maternal-facing) and basal

(fetal-facing) membranes of the syncytiotrophoblast (Smith et

al., 1992). Of all the amino acids in humans, those

demonstrating the highest fetal to maternal serum ratios are

the cationic amino acids arginine and ornithine (Economides

et al., 1989). Additionally, arginine is unique in that the

fetomaternal serum ratio increases with increasing

gestational age (Bernardini et al., 1991). As the

nutritional needs of the developing fetus change, it is not

unreasonable to suspect that the placental delivery of

nutrients, and therefore the transport systems responsible

for their delivery from the maternal circulation, change in

response to those needs.

Arginine plasma membrane transport in mammalian cells is

mediated by at least four kinetically distinct transport

systems. With the exception of the liver, in the adult

mammal, the primary route of cationic amino acid uptake is

via the Na -independent, high affinity System y' (White,

1985). However, the normal adult liver lacks System y' and

instead contains a low-affinity, liver-specific transporter

with similar substrate specificity to System y' (Kim et al.,

1991). The corresponding cDNAs (mCAT1 and mCAT2a) have been

isolated (Kim et al., 1991; Closs et al., 1993). Two

additional transport systems, first described in mouse

blastocysts, transport not only cationic amino acids but

small neutral amino acids as well. These activities are the

Na+-dependent System B and the Na'-independent System b0'

(Van Winkle et al., 1985; Van Winkle et al., 1988). Another

Na+-independent transporter of both cationic and neutral amino

acids described in the erythrocyte, with substrate

specificity and kinetic properties very similar to System b' ,

is System y+L (Dev4s et al., 1992). The primary kinetic

difference between these two systems is the inhibition by

neutral amino acids. In System b ', neutral amino acids

inhibit the uptake of cationic amino acids in the presence

and absence of Na+ (Van Winkle et al., 1988) while System y+L

requires the presence of Na+ for neutral amino acid inhibition

of cationic uptake (Deves et al., 1992). However, since

System y' has been shown to interact with neutral amino acids

in the presence of Na+ (Christensen and Handlogten, 1969), the

identity of System y.L remains ambiguous. In addition to

these activities, the intestinal brush border has both Na'-

dependent and Na-independent cationic amino acid transport

systems which resemble Systems B and b respectively, with

a somewhat narrower substrate specificity (Cassano et al.,

1983; Harvey et al., 1993; Wolfram et al., 1984). Injection

of intestinal mRNA into oocytes induced System b0+ activity,

the primary route of endogenous transport in Xenopus oocytes

for neutral and cationic amino acids (Campa and Kilberg,

1991), by the synthesis of a Type II membrane protein (rBAT)

with a structure that is dissimilar to those of other known

transporters (Magagnin et al., 1992).

The trophoblast of human placenta has been well

characterized as to the complement of nutrient transporters

by several laboratories using many different methods

(Yudilevich and Sweiry, 1985; Smith et al., 1992). This

subject has been reviewed in the Introduction. However, few

reports exist on the kinetic discrimination of the cationic

amino acid transport systems in the placenta.

Wheeler and Yudelivich (1989) described the trans-

placental flux of lysine and alanine in an in situ dually-

perfused placental system in the guinea pig. Saturable

transport systems for lysine were identified on both the

apical and basal membranes, with a significantly higher flux

at the maternal (apical) membrane. Additionally, an

increased efflux of lysine was detected on the basal

membrane. These data are consistent with the presence of an

increased lysine transporter number on the apical membrane

for the extraction of this nutrient from the maternal

circulation and delivery to the fetus. This lysine uptake

and efflux was inhibited strongly by other cationic amino

acids (arginine, ornithine) and more weakly by histidine, an

amino acid with about 0.1% cationic species at physiologic pH

(Mann et al., 1984). While the neutral amino acid alanine

significantly inhibited the uptake of lysine consistent with

either System y+L or y' in the presence of Na+ (Christensen and

Handlogten, 1969), the reciprocal inhibition was not noted

leaving system assignment difficult.

In purified membrane preparations from term human

placentas, cationic amino acid uptake was seen in both apical

and basal membrane vesicles (Kudo et al., 1987; Kudo and

Boyd, 1990). However, the extremely high inhibitor

concentrations used (30-50 mM), 100-fold to 10,000-fold over

the published Km for these transport systems (White, 1985),

makes discrimination of transport systems difficult. It

should be noted that in this study, inhibitor concentrations

of this magnitude show noncompetitive inhibition of nearly

all transport systems.

Furesz and coworkers (1991), examined cationic amino

acid uptake in basal membrane vesicles derived from human

placenta. Two systems were identified in basal membrane

vesicles with similar affinity for both arginine and lysine
(100-200 pM). Inhibition of each system by neutral amino

acids in the absence of Na', gave kinetic results consistent

with the presence of Systems y+ and b' Subsequently, the

same group of investigators described cationic amino acid in

human apical membrane vesicles (Furesz et al., 1995).

Similar results to that of the basal membrane were described

with the exception of neutral amino acid inhibition of

cationic amino acid uptake. Inhibition of the apical

cationic transport systems by neutral amino acids required

the presence of Na* more consistent with System yL than b' .

There have been no reports of the characterization of

cationic amino acid transport activity in the rat placenta,

and no reports to address the effect of the needs of the

developing fetus on this activity with the exception of the

manuscript published from the data described in this chapter

(Malandro et al., 1994). This chapter describes changes in

cationic amino acid transport occurring with increasing

gestational age in a rat placental model. Competitive

inhibition analysis using L-arginine as a substrate was used

to identify the transport systems responsible for cationic

amino acid uptake at 14 to 20 days gestation. Vesicle

preparations enriched for either apical or basal membranes

were used to establish differences in transport activities

between plasma membrane domains. The presence of System y+

was further demonstrated by the analysis of mRNA content

throughout gestation.


Characterization of apical and basal membrane vesicle

preparations. To ensure that uptake measurements from

vesicles isolated from placentas of different gestational

ages would not be affected by variations in size, vesicle

volumes and diameters were examined at the two age extremes,

Table 3-1. Characterization of apical and basal membrane
vesicles from rat placentas of 14 or 20 days gestation.

14 Day 20 Day 14 Day 20 Day
Apical Apical Basal Basal

Vesicle Diameter 0.45 0.46 0.43 0.36
Marker Enzyme
Phosphatase 48.3 23 43.0 12 3.3 1 2.4 1
-DHA Binding 2.7 1 2.7 1 15.1 5 16.3 7

Vesicle diameter was determined by direct sizing by laser-
light diffraction analysis (Nicomp Model 270; Particle Sizing
Systems Inc., Goleta, CA). Vesicle diameters are given as
Im. Marker enzyme data are given as -fold enrichment over
homogenate and represent the mean standard deviations for
at least 6 independent vesicle preparations.

14 and 20 days gestation. For both apical and basal membrane

vesicles, volume, as determined by the steady-state

distribution of "H-alanine (Kletzien et al., 1975), did not

significantly differ with gestational age in either

preparation. Average vesicle diameters estimated by laser-

light diffraction analysis (Niven et al., 1990) of each

membrane preparation were similar for both gestational ages

as well (Table 3-1).

The alkaline phosphatase enrichments of the apical

membrane preparation (Table 3-1) were comparable to or

greater than those reported by several investigators for

human placenta (Alonso-Torre et al., 1992; Karl et al., 1989;

Moe and Smith, 1989) and for rat placenta (Glazier et al.,

1993). The enrichment of DHA binding seen in the basal

membrane preparation was not as high as that reported for the

human basal vesicle preparation of Kelley, Smith and King

(1983), but was comparable to enrichments reported by other

investigators (Bravo et al., 1993, Marin et al., 1990).

Although the basal membrane preparation is contaminated

with a small percentage of apical membranes, several lines of

evidence exist suggesting a functional difference between the

two preparations. First, the Na+-dependent arginine

transport activity clearly present in apical membrane

vesicles was not detectable in the basal membrane preparation

after 14 days gestation (Figure 3-3). It is not likely that

the Na+-dependent System B'+ was inactivated in the

preparation of basal membrane vesicles due to the fact that

other Na+-dependent amino acid transporter systems such as

Systems X- and A can be detected in basal membrane vesicles

(Chapters 4 and 5, respectively). Second, the different

gestational related changes in System y+ and leucine-sensitive

arginine transport in the two preparations suggest a

functional separation of the two membrane surfaces (Figure 3-

3). The contamination of the basal membrane preparation with

apical membranes if anything, would cause an under-estimation

the magnitude of these gestational changes.

Characterization of arqinine transport systems of the apical

and basal membranes. To establish that 3H-arginine associated

with the membrane vesicles was due at least in part to

transport and not merely non-specific binding, steady state

(10 min) arginine uptake was examined after incubation in

buffer containing increasing extravesicular osmolarity

(Lever, 1980). The arginine accumulation decreased with

increasing extravesicluar osmolarity demonstrating transport

into an osmotically active space.

As arginine is a substrate for at least four distinct

transport systems, three of which have been described in more

than one tissue, we characterized arginine uptake through

each of these systems. Low affinity Na-independent cationic

amino acid transport mediated by mCAT2a is considered liver

specific in its distribution (Closs et al., 1993). To

confirm the absence of mCAT2a in the placenta, Northern blot

analysis was performed with an mCAT2a cDNA, and rtPCR

analysis with mCAT2a specific primers. No detectable

50 -

40 -


20 -
--O-- Na
10 Basal K

S 80
70 -
S60 -
D. 50
30 -
10 t Apical
0 10 20 30 40 50 60 70 80


Figure 3-1. Time-course of [3H]-arginine uptake in basal or
apical membrane vesicles from placentas of 20 days gestation.
Transport of 10 pM [3H]-arginine was performed at 37"C in uptake
buffer containing a final concentration of either 100 mM NaSCN
(open circles) or 100 mM KSCN (closed circles). Equilibrium
points were assayed in both basal (Na+= 42 + 1.9, K'= 40 + 0.6)
and apical (Na+= 47 0.9, K = 41 2.2) membranes. Values are
reported as the means standard deviations for triplicate
determinations. Where not shown, the standard deviation bars
are contained within the symbol.

hybridization or PCR product was observed with either total

RNA or poly (A)+ selected mRNA isolated from placentas at 14

through 20 days gestation when total or poly (A)+ mRNA from

rat liver was used as a positive control.

Na-dependent cationic amino acid transport occurs in

mouse blastocysts (Van Winkle et al., 1985), Xenopus oocytes

(Campa and Kilberg, 1991), and several other cell types

(Christensen and Kilberg, 1987; Kilberg et al., 1993) by an

activity termed System B0". This system also is capable of

neutral amino acid transport. A time course of total L-

arginine uptake (10 pM) in either apical or basal membrane

vesicles from placentas of 20 days gestation was examined in

the presence of 100 mM NaSCN or KSCN (Figure 3-1). Subsequent

experiments showed arginine uptake was approximately linear

to 10 sec, thereafter approaching equilibrium. Given the

technical difficulties for measuring transport at shorter

times, 10 sec uptakes were used to approximate initial rates

in all the uptake assays reported for both membrane

preparations. The virtual superimposition of the Na+ and K+

time course curves in basal membrane vesicles (Figure 3-1)

documents the absence of Na+-dependent arginine uptake and

hence, the absence of System B0"' in this preparation. In

contrast, as shown in Table 3-2, Na+-dependent uptake was

noted in apical membranes which was nearly completely

inhibited by arginine or lysine; the neutral amino acids,

homoserine (83%), leucine (56%), and the branched chain

analog 2-aminobicyclo [2,2,1] heptane-2-carboxylic acid


System b0,+

20 v

E 10-

E i

0 250 500 750 1000 1250 1500

[Leucine], [tM

Figure 3-2. Inhibition of Na*-independent [3H]-arginine uptake
with increasing concentrations of leucine in placentas of 20
days gestation. Uptake of 10 pM [3H]-arginine was performed for
10 sec at 37"C in uptake buffer containing a final concentration
of 100 mM KSCN and the indicated concentration of leucine. All
uptake solutions were adjusted to isoosmolarity (310 mosm) with
the addition of sucrose. Values are reported as the means
standard deviation for triplicate determinations. Where not
shown, the standard deviation bars are contained within the

(BCH)(49%) also showed some inhibition. Each of these amino

acids are known substrates for System B"' in blastocysts (Van

Winkle et al., 1985).

The Na+-independent uptake of cationic amino acids can

occur through several transport systems including Systems y+,

b0' and y+L. System y is present in most mammalian cells and

at physiologic substrate concentrations appears to mediate

the transport of cationic amino acids primarily (White,

1985). However, in the presence of Na', neutral amino acids

at non-physiologic concentrations can competitively inhibit

System y+ (Christensen and Handlogten, 1969). System b'*,

first described in mouse blastocysts (Van Winkle et al.,

1988) and later in Xenopus oocytes (Campa and Kilberg, 1991)

and basal membrane vesicles from human placenta (Furesz et

al., 1991), accepts both neutral and cationic amino acids.

System b"' activity is typically measured as leucine-

inhibitable lysine or arginine transport or vice versa (Van

Winkle et al., 1988). Another transport system recently

described in human erythrocytes, System y+L, is similar in

properties to System b"'* and can be assayed in a similar

fashion (Dev4s et al., 1992). In the rat apical membrane

vesicles prepared from placentas of 20 days gestation,

arginine uptake was partially inhibited by the presence of

leucine at 50-1500 yM (Figure 3-2). A Dixon plot of the data

from a separate experiment measuring arginine uptake (10 -

100 pM) at increasing concentrations of leucine, indicated

competitive inhibition with an estimated Ki value of

Table 3-2. Inhibition of Na+-dependent (System Bo,+) arginine
uptake in microvillous membrane vesicles from placentas of 20
days gestation.

Na+ K+ Na+-dependent %Inhib

No Inhibitor 60.2 5.1 41.9 2.1 18.3
Arginine 3.6 1.0 3.5 1.3 0.1 99
Lysine 2.2 0.7 2.0 0.3 0.2 99
Homoserine 32.9 2.4 29.7 2.5 3.2 83
Leucine 37.0 4.5 29.0 1.2 8.0 56
Proline 30.7 3.1 30.3 1.1 0.4 98
BCH 39.4 0.5 30.0 1.5 9.4 49

The uptake of 10 pM 3H-arginine was assayed for 10 sec at 37"C
in either Na or K+ -containing buffer as described in
Materials and Methods (Chapter 2). Inhibitors were included
at a concentration of 2.0 mM, and solutions were brought to
isoosmolarity (310 mosm) by the addition of sucrose. The
values represent the means (n = 3) S.D. in pmol-mg-'
protein*10 sec'. All inhibitors were significantly (P <
0.005) different from control (no inhibitor) (BCH = 2-
aminobicyclo [2,2,1] heptane-2-carboxylic acid)

Table 3-3. Inhibition of Na+-independent, leucine-sensitive
arginine uptake in apical and basal membrane vesicles from
placentas of 20 days gestation.


-Leu +Leu Leu-sensitive %Inhib

No Inhibitor 38.4 1.3 29.1 2.5 9.3 --
Arginine 3.5 1.3 4.1 1.1 -- 100
Lysine 2.0 0.3 2.4 0.1 -- 100
Homoserine 29.7 2.1 22.5 2.2 7.2 23
Proline 30.3 1.1 20.3 2.0 10.0 0
BCH 30.0 1.5 19.2 2.1 10.8 0


-Leu +Leu Leu-sensitive %Inhib

No Inhibitor 54.0 3.1 32.5 1.6 21.5
Arginine 8.7 0.5 11.3 1.2 -- 100
Lysine 4.5 1.8 6.2 1.5 -- 100
Homoserine 31.8 1.1 14.7 2.2 17.1 20
Proline 41.6 5.5 15.4 1.6 26.2 0
BCH 36.7 3.8 13.8 1.5 22.9 0

The uptake of 10 pM 3H-arginine in the presence or absence of
10 mM leucine was assayed for 10 sec at 370C as described in
Materials and Methods. Inhibitors were included at a
concentration of 2.0 mM, and solutions were brought to
isoosmolarity by the addition of sucrose. The values
represent the means (n=3) S.D. in pmol-mg-1 protein*10 sec1.
(BCH = 2-aminobicyclo [2,2,1] heptane-2-carboxylic acid)

approximately 500 pM. Thus, the level of leucine used as an

inhibitor in subsequent assays, at least 2.5 mM, should be

adequate to block most if not all of the leucine-sensitive

component of arginine uptake. Analysis of the leucine-

sensitive arginine uptake in both the basal and apical

membranes (Table 3-3), showed complete inhibition by the

cationic amino acids arginine and lysine, and partial

inhibition by the neutral amino acid homoserine (54% basal

and 76% apical). Additionally, no significant inhibition was

seen with BCH or proline, previously shown to be poor

substrates for System b0,' as well as System y+L (Devds et al.,

1992; Van Winkle et al., 1988).

Even at high concentrations of leucine (Figure 3-2), not

all arginine uptake was inhibited indicating the presence of

a leucine-insensitive component. To further characterize

this portion of arginine uptake, inhibitor studies were

performed (Table 3-4). Only a small amount of inhibition was

seen with homoserine, proline, and BCH. A similar degree of

inhibition with these amino acids was noted by other

investigators in basal membrane vesicles prepared from human

term placentas (Furesz et al., 1991). In contrast,

significant inhibition was noted with arginine (91% apical

and 73% basal) and lysine (92% apical and 86% basal).

Inhibition of leucine-insensitive arginine uptake by

homoserine in the presence of Na* was also seen (87% apical

and 92% basal). These data indicate that the Na'-independent

leucine-insensitive component of arginine uptake is mediated

Table 3-4. Inhibition of Na+-independent, leucine-insensitive
(System y+) arginine uptake in apical and basal membrane
vesicles from placentas of 20 days gestation.

Apical Basal


No Inhibitor -- --
Arginine 91 73
Lysine 92 86
Homoserine 23 16
Proline 12 12
BCH 19 17

Uptake of 10 yM 3H-arginine was assayed in the presence of 10
mM leucine for 10 sec at 370C as described in Materials and
Methods. Inhibitors were included at a concentration of 2.0
mM, and solutions were brought to isoosmolarity by the
addition of sucrose. The results are the average of at least
three assays, and the standard deviations were typically less
than 10%. The control values (no inhibitor) for apical and
basal membranes were 29.1 2.5 and 32.5 1.6 pmol-mg-1
protein*10 sec-1, respectively. (BCH = 2-aminobicyclo [2,2,1]
heptane-2-carboxylic acid)

by a cationic amino acid-preferring system similar to System

Ontogeny of cationic amino acid transport systems with

increasing gestational age. As the fetal to maternal serum

ratio for arginine increases with increasing placental

development (Bernardini et al., 1991, Economides et al.,

1989), the contribution of each cationic amino acid transport

system to arginine uptake at 14 to 20 days gestation was

examined. The transport activities, defined as described

above, were examined in both the apical and basal membranes

(Figure 3-3). System B0' activity on the basal membrane was

detectable at 14 days but, in agreement with Figure 3-1, was

absent from 16 to 20 days gestation. In contrast, System B0'-

mediated arginine uptake in the apical membrane increased by

4-fold, from 3.0 0.2 to 12 1.5 pmol-mg-' protein*10 sec-'

over the same gestational period. The Na-independent,

leucine-sensitive component of arginine uptake increased with

increasing gestational age from 14 to 20 days in both the

apical (0.8 0.4 to 6.6 2.0 pmol-mg' protein*10 sec-') and

basal (1.8 0.3 to 15 2.6 pmol-mg-' protein.10 sec-1)

membrane. The Na-independent System y' component of arginine

uptake increased from 14 to 16 days gestation in the basal

membrane (2.3 0.3 to 8.2 1.0 pmol*mg-' protein-10 sec-')

and then remained fairly constant throughout the gestational

period. A steady increase of nearly 10-fold in System y+

activity was seen in the apical membrane over the same

Figure 3-3. Ontogeny of cationic amino acid transport systems
in microvillous and basal membrane vesicles. Cross-hatched bars
represent uptake by microvillous membrane vesicles and the
filled bars, represent uptake by basal membrane vesicles. A)
Na+-dependent uptake of 10 yM [3H]-arginine (System B0'") was
assayed for 10 sec at 37C. Values were calculated from the
difference in uptake in 100 mM NaSCN and KSCN buffers. B) The
Na-independent, leucine-sensitive uptake of 10 pM [3H]-arginine.
Values were calculated from the difference in uptake in KSCN
buffer and KSCN buffer containing 2.5 mM leucine. C) The Na+-
independent, leucine-insensitive uptake of 10 pM [3H]-arginine
(System y). Values were calculated from the difference in
uptake in KSCN buffer containing 10 mM leucine and KSCN buffer
containing 10 mM unlabelled arginine. Values represent the
means standard deviations for triplicate determinations from
2 independent preparations at each day gestation.

21 A
18 -
12 -* Apical
6 -15 Basal
6 6-
(D 3
o 0 MEML--

. 21-
- 18
S 15

O 21L -
E 18- C


14 16 18 20
Days Gestation

Figure 3-4. Northern analysis of System y' mRNA from placentas
of increasing gestational age. Total placental RNA (30 pg per
lane) was separated by 1% agarose gel electrophoresis and
transferred to nylon membrane. The filter was hybridized with
32P-labelled System y cDNA probe and washed at 650C (375 mM
NaC1, 40 mM Na2HPO,, pH 7.2, 1 mM EDTA, 1% SDS). A single mRNA
species was detected at 7.4 7.9 Kb in each sample (Top).
Densitometric analysis of the autoradiogram was performed
(Bottom), and adjusted for any loading differences (< 10%) by
normalization to the 28S ribosomal RNA, quantified by
densitometry of the photographic negative of the ethidum bromide
stained gel. Values are representative of four independent
experiments and expressed in relative absorbance units.


20 18 16 14


c 5
L 4

Q) 2 -

20 18 16 14
Days Gestation

gestational period (2.8 0.3 to 20.0 2.0 pmol-mg-'

protein*10 sec').

Northern analysis of System y+ mRNA levels. A mouse cDNA

probe (mCAT1) for the System y+ transporter (Kim et al., 1991)

was used to correlate the level of activity with the cellular

mRNA level (Figure 3-4). Consistent with the mCAT1 mRNA size

in other tissues and species, an mRNA species of

approximately 7.4 7.9 kb was detected in total RNA isolated

from placentas of 14 to 20 days gestation. No molecular

probes have been described for either System B'" or bO'.

Densitometric analysis revealed a significant increase in the

steady-state content of System y+ mRNA over the gestational

period. These data parallel the observed increase in

arginine transport via System y' in both the apical and basal

membranes (Figure 3-3).


The rat provides an excellent model system for the study

of amino acid transport in the developing placenta. The rat

has a hemochorial placenta similar to the human, the major

difference being an additional two trophoblast layers between

the maternal and fetal circulations. Alkaline phosphatase

has been shown by cytochemistry to be localized to the apical

membrane of the human trophoblast and to the apical membrane

of rat trophoblast layer II (Glazier et al., 1993). Rat

trophoblast layer I is fenestrated with large gaps allowing

for the nearly unrestricted passage of large molecules, such

as horseradish peroxidase, injected into the maternal

circulation (Metz et al., 1978). These molecules are nearly

totally excluded by trophoblast layer II. Therefore, it is

likely that the first functional barrier encountered by a

solute in the maternal circulation is trophoblast layer II,

and as such, the apical membrane preparation from the rat

placenta is enriched for a membrane population analogous to

that of the human trophoblast apical membrane. We cannot

definitively assign the B-adrenergic receptor, and thus DHA

binding, to a specific basal membrane of trophoblast layers

II or III.

In the original description of the basal membrane

preparation from human placenta (Kelley et al., 1983), the

association of the basal membrane with the basal lamina is

critically important, and it is this interaction exclusively

that prevents the removal of this membrane under conditions

necessary for the removal of the apical membrane and the

cytoplasmic contents of the trophoblast. The adjustment of

the EDTA concentration and the additional sonication then are

responsible for the stripping of this membrane from the basal

lamina. Only the basal membrane of trophoblast layer III is

in association with the basal lamina in the rat trophoblast.

As both layers II and III are syncytial, the conditions

described for the removal of the apical membrane and

syncytial cytoplasm in the human basal membrane preparation

should remove these membranes as well leaving the basal

membrane of layer III attached to the basal lamina. As these

observations have never been tested, the basal membrane

preparation could have arisen from either trophoblast layers

II or III. We would suggest that even if the membrane

preparation was to arise from the basal membrane of layers II

or III, this would still represent a functional basal

membrane barrier between the apical membrane and the fetal

circulation. It should be pointed out that the transporters

we have assigned to the rat basal membrane also have been

assigned to the basal membrane of the human trophoblast

(Furesz et al., 1991) lending even more credence to the

analogy between the human basal membrane and our basal

membrane preparation. In addition, Ca2'ATPase localized to

the basal membrane of the human trophoblast has been

localized by immunohistochemistry to the basal membrane of

trophoblast layer III in the rat placenta (Borke et al.,

1989). These data also support the functional similarity

between the human trophoblast basal membrane and the rat

basal membrane of trophoblast layer III.

The current study demonstrates that cationic amino acid

uptake in the rat placental syncytiotrophoblast is mediated

by an asymmetric distribution of both Na+-dependent and

Na+-independent transport systems in the apical and basal

membranes resulting in the net flux of amino acids from the

maternal to fetal circulations. The uptake of amino acids

from the maternal circulation into the syncytial cytoplasm

takes place at the apical membrane, in direct apposition to

the maternal blood supply. Our data show that System B0', a

known secondary-active Na'-dependent transporter of cationic

as well as neutral amino acids, to be present within this

membrane domain. By virtue of the inwardly directed Na'

gradient in vivo, it contributes to the concentrative

transfer of cationic amino acids across the trophoblast. The

concentrated levels of arginine and lysine in the human

trophoblast, 0.32 and 0.51 mM respectively, compared to the

maternal serum levels of these amino acids, 0.043 and 0.10 mM

respectively (Economides et al., 1989), would support this

hypothesis. However, both Systems y' and the neutral amino

acid-inhibitable system (System b0'), are also present on the

apical membrane. These Na+-independent non-concentrative

systems presumably play an important role in the regulation

of the syncytial cytoplasmic levels of these amino acids by

permitting transfer in the reverse direction if necessary.

The basal membrane of the syncytiotrophoblast, in

contrast to the apical membrane, has little or no significant

Na-dependent cationic amino acid uptake. The presence of

System B0"' on the fetal-facing membrane could result in the

non-specific and perhaps detrimental clearance of neutral as

well as cationic amino acids from the fetal circulation. In

contrast, the high expression of both Nat-independent systems

on the basal surface is important to mediate the transfer of

amino acids to the fetus by facilitative transport down their

concentration gradient. However, when net flux of cationic

amino acids from maternal to fetal circulation is considered,

the role of membrane potential (Bussolati et al., 1987), as

well as multiple transport systems with overlapping substrate

specificities must be evaluated and better understood.

As was previously mentioned, the fetal to maternal

plasma ratio of arginine increases with increasing

gestational age (Bernardini et al., 1991). Our data

demonstrate that this may be due in part to the concurrent

developmental changes in the transporter complement of the

apical and basal trophoblast membranes. During 14 to 20 days

gestation there is a significant increase in rate of

secondary-active transport by System Bo' on the apical

membrane. This probably participates in the increased

clearance of cationic amino acids from the maternal

circulation; the subsequent release of these into the fetal

circulation would occur by one or both of the Na-independent

cationic amino acid transporters on the basal membrane, both

of which increase with increasing gestational age as well.

With respect to the placental transport of amino acids,

previous studies have included only activity measurements to

investigate changes during development. Our study is the

first to use a cDNA probe to provide further evidence for the

existence and regulation of a particular placental amino acid

transporter. Northern analysis demonstrates that the

developmental regulation of System y activity is accompanied

by an increase in the steady-state mRNA level for the

transporter as well. Thus, either transcriptional control or

mRNA stabilization must play a role in regulating cationic

amino acid transport during placental development.

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EJCEHNEAW_U7T8N0 INGEST_TIME 2012-02-29T17:24:31Z PACKAGE AA00009033_00001