CHARACTERIZATION OF THE RABBIT RENAL H,K-ATPASES
W. GRADY CAMPBELL
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIRENIENTS FOR THE DEGREE
OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
This thesis is dedicated to my father, William Maxwell Campbell, and my mother, Clara
I would like to acknowledge the contributions of a number of people who helped me
in this work. First I would like to thank my committee, the chair Dr. Brian Cain, Dr.
Charles Wingo, Dr. Susan Frost, Dr. Harry Nick, and Dr. Michael Kilberg. Also Dr. David
Weiner was very generous to share his lab and microscopes. Dr. Jill Verlander directed a
very interesting line of research to answer some questions that arose out of our own
I would also like to thank my fellow lab members for putting up with me all this time.
Thanks go to Kim McCormick, Philip Hartzog, Abbe Stack, Jim Gordon, James Gardner,
Paul Sorgen, Tammy Caviston, and Regina Perry. James, Paul, Tammy, Regina, and I
shared many years together as a particularly cohesive lab family. Also I want to thank my
labmates in my lab home away from lab home, Jeanette Lynch, Amy Frank, and Robin
Moudy. They really went out of their way for me, and it was a pleasure working with
them. Mary Handlogten was also especially helpful in getting me started in protein work. I
want to acknowledge the contribution of the Yang lab, including Mike Litt, Chien Chen,
and Sue Lee. Having an adjacent lab with open doors and shared space between them has
turned out to be a great design, there has been a high level of camaraderie and exchange of
ideas between the labs.
I also appreciate the support and encouragement of my family, my sister Diane, her
children Emory and Lisacole, my brother-in-law Ken, and the newest addition to our
family Samma. The companionship and support of Ruth has helped me throughout my
time here. And most of all, I appreciate the enthusiastic support of my father and late
mother during my years here.
TABLE OF CONTENTS
ACKNOWLEDGMENTS ............................................. iii
LIST OF TABLES ............................................ ......... vii
LIST O F FIG U R E S .......... ......... .................................. viii
ABBREVIATIONS .............................. ........................ x
ABSTRACT ....................... .............................. xiii
BACKGROUND AND SIGNIFICANCE ................................... 1
Overview .................. .......................... ............ 1
H,K-ATPase in the Kidney ............................................ 4
H,K-ATPase Structure and Function ................................... 17
R enal Tissue Culture Cells ............................................. 23
Summary ............................ .............................. 26
MATERIALS AND METHODS ............ ........................... 27
M olecular Biology .......................... ......................... 27
Biochemistry ........................... ............................ 36
Fluorescence Microscopy ............................................. 40
H,K-ATPASE 3 SUBUNITS IN THE RABBIT RENAL MEDULLARY
COLLECTING DUCT .......................................... .. 45
Introduction ...... ......................... ...................... 45
Renal Medulla HK1 mRNA Variant ................... ............. 46
D iscu ssio n . . . . . . . . . . . . . . 4 9
H,K-ATPASE a SUBUNITS IN THE RABBIT RENAL CORTICAL
COLLECTING DUCT ............................................... 52
Introduction ......... ........................................... 52
M multiple H,K-ATPase a Subunits in the Kidney ............................ 53
Alternative Splicing ofH,K-ATPase a Subunits in the Kidney ................ 69
Expression ofH,K-ATPase a Subunits in the Kidney ...................... 72
D discussion .................. ......................... ......... 76
H,K-ATPASE ACTIVITY IN A RABBIT KIDNEY CORTICAL
COLLECTING DUCT CELL LINE ...................................... 81
In tro d u ctio n . . . . . . . . . . . . . . 8 1
Detection ofH,K-ATPase in RCCT-28A Cells ............................. 82
D iscu ssio n .. . .. . . . . . . . . . .. . 10 0
PERSPECTIVE AND FUTURE DIRECTIONS ............................. 106
Multiplicity of H,K-ATPase Isoforms in the Kidney ........................ 106
Cell Type Specificity of H,K-ATPase in the Kidney ........................ 112
F future studies ....................................................... 113
R E FE R E N C E S ........................................................ 117
BIOGRAPHICAL SKETCH .......................................... 129
LIST OF TABLES
2-1. PCR primer pairs ............................................ 32
2-2. Solutions for determination of pH ................................. 41
LIST OF FIGURES
1-1. Schematic diagram of H,K-ATPase .............................. 18
3-1. Northern analysis showing presence of HK3 mRNA in renal cortex, renal
medulla, and stomach ....................................... 47
3-2. 3' and 5' RACE reactions to amplify HK3 cDNAs using rabbit renal
medulla as template ................................... ...... 48
3-3. HKP3 and 3' subunit mRNAs ........................... ..... .. 50
4-1. Design of degenerate primers for RT-PCR of novel P-type ATPases ....... 56
4-2. RT-PCR product amplified from rabbit renal cortex RNA using degenerate
p rim e rs . . . . . . . . . . . . 5 7
4-3. BLAST search using sequence of419 bp fragment of HKa2 ............ 58
4-4. Cloning of HKot2a and HKX2c cDNAs ............................... 59
4-5. GenBank accession records for rabbit HKoc2 sequences .................. 61
4-6. Northern analysis showing presence of HKc*2, in distal colon and renal
cortex ...................................................... 67
4-7. Distance analysis of selected HKat and NaKca subunit coding ............. 70
4-8. Rabbit HKIc2 gene sequence at the 5' end ............................ 71
4-9. Western analysis showing presence of HKc2,, and IHKci, protein in renal
cortex .................................................... 73
4-10. Immunohistochemistry by Dr. Jill Verlander and Ms. Robin Moudy ........ 75
5-1. Southern blots ofH,K-ATPase subunit mRNA in RCCT-28A cells ........ 84
5-2. HKP3 and HKKc, subunit mRNA in RCCT-28A cells .................... 86
5-3. HK(a2 subunit mRNA in RCCT-28A cells ............................. 88
5-4. Western analysis showing presence of HKot2. protein in RCCT-28A cells ..... 90
5-5. pHi recovery from an acid load by RCCT-28A cells in the absence ofNa* .... 96
5-6. pH, recovery from an acid load by RCCT-28A cells in the presence of EIPA 98
5-7. Summary of the rates ofpH, recovery from an acid load by RCCT-28A cells 102
acetoxymethyl ester of BCECF
ethylene diamine tetraacetic acid
fetal bovine serum
N-2- hydroxyethylpiperazine-N'-2-ethanesulfonic acid
H,K-ATPase a- (catalytic) subunit
H,K-ATPase a, subunit
H,K-ATPase U 2, subunit
H,K-ATPase cL2b subunit
H,K-ATPase ct2 subunit
H,K-ATPase o3 subunit
HKKc4 H,K-ATPase i4 subunit
HK3 H,K-ATPase (3 subunit
HK3' H,K-ATPase 3' subunit
kb thousand base pairs
kDa thousand Daltons
MW molecular weight
NaKa Na,K-ATPase ot (catalytic) subunit
NaKct1 Na,K-ATPase x, subunit
NaKct2 Na,K-ATPase o2 subunit
NaKcX3 Na,K-ATPase (x3 subunit
NaKc4 Na, K-ATPase cx4 subunit
NaKi31 Na,K-ATPase 3i subunit
PAGE polyacrylamide gel electrophoresis
PCR polymerase chain reaction
pH1 intracellular pH
PMSF phenylmnethylsulfonyl fluoride
RACE Rapid Amplification ofcDNA Ends
RNA ribonucleic acid
RT reverse transcription
RT-PCR reverse transcriptase-polymerase chain reaction
SDS sodium dodecyl sulfate
UTR untranslated region
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
CHARACTERIZATION OF THE RABBIT RENAL H,K-ATPASES
W. Grady Campbell
Chairperson: Dr. Brian D. Cain
Major Department: Biochemistry and Molecular Biology
H,K-ATPases are located on the apical membrane of epithelial cells lining the
collecting duct in the kidney. ATP hydrolysis drives luminal acidification and potassium
reabsorption by the enzyme. H,K-ATPases consist of two subunits. The catalytic sites are
located on the ac subunits, and the P3 subunits play a role in intracellular trafficking. The
goal of this work was to elucidate the molecular identity ofH,K-ATPase a and 3 subunits
in kidney. Northern analysis demonstrated two H,K-ATPase (3 subunit species in rabbit
renal medulla; only the smaller of these was present in renal cortex and gastric tissues. A
search for H,K-ATPase a subunit isoforms in rabbit renal cortex was conducted using
reverse transcriptase-polymerase chain reaction with degenerate primers. A full-length
cDNA of the H,K-ATPase cX2 subunit was obtained. Two 5' ends of this transcript were
observed by 5' Rapid Amplification of cDNA Ends. One (HKcx2a) had high homology to
previously known H,K-ATPases, and the second was a novel variant (a2c). Phylogenetic
analysis showed that the rabbit cax,, subunit clusters near H,K-ATPases from human axillary
skin and rat distal colon, but is more distant from gastric H,K-ATPase and
Na,K-ATPases. Gene sequence analysis showed that the first HKo_,c exon was located
within the first HKX2a intron. Western analysis using antipeptide polyclonal antibodies
demonstrated the expression ofHKc(x, HKac2,, and HKcx2c subunits in rabbit renal cortex.
The mRNAs encoding the HKO3, HKc,, HKIx2,, and HKc 2, subunits were detected in the
rabbit cortical collecting tubule intercalated cell line RCCT-28A. These results indicated
that acid-secreting intercalated collecting duct cells were a cell type containing
H,K-ATPase, and these cells appear capable of expressing multiple H,K-ATPase isoforms.
BACKGROUND AND SIGNIFICANCE
H,K-ATPase was first identified in the stomach, where it is expressed at an extremely
high level to mediate the extrusion of acid that aids digestion. A great deal has been
learned about regulation of H,K-ATPase in the stomach, enough to feel as though its
regulation is understood in a meaningful way. At the molecular level, gastric-specific
transcription factors have been found that may be responsible for the marked
tissue-specific regulation of expression of H,K-ATPase (Maeda et al., 1991; Oshiman et
al., 1991; Tamura el al., 1992). The principal means of induction of gastric H,K-ATPase
from its basal state is by unwinding a tightly coiled tubular network, exposing pumps
localized in the network and thereby increasing the size of the secretary surface (Pettit et
al., 1995). This morphological transformation is triggered by various stimuli including
activation of stretch receptors in the stomach by entering food, the sight of food, or even
the thought of food (reviewed by Wolfe and Soil, 1988; Hersey and Sachs, 1995). Thus,
there is a framework of understanding of the regulation of H,K-ATPase in the stomach at
the molecular and cellular levels, and the root stimuli that result in the H,K-ATPase
response are known.
H,K-ATPase is expressed at much lower level in the kidney, but its role in the kidney
is even more critical to life than its role in the stomach. The kidney plays the major role in
K' homeostasis, maintaining extracellular (including plasma) K' concentration within the
relatively narrow normal range of 3.8 to 5.0 mEq/L (Merck, 1992; Guyton and Hall,
1997) despite a wide variation in K' intake. If plasma K' levels depart significantly from
these values, then cardiac arrhythmias lead to life-threatening conditions.
K' transport in the kidney is also important to blood pressure regulation. It has been
observed that urinary K+ levels and urinary Na'/K' ratio are related to systolic and
diastolic blood pressures (INTERSALT, 1988; Whelton, 1993). Excretion rates show a
higher degree of correlation with elevated blood pressure than serum levels, implicating
the kidney as an important organ in regulating the effect of K' on blood pressure. Dietary
K+ supplementation has been shown to lower blood pressure since early this century
(Ambard and Beaujard, 1904; Addison, 1928).
The substantial role of renal ion transport in blood pressure regulation is emphasized
by studies conducted by Lifton (1996), who determined the molecular bases underlying
certain types of monogenic inherited extreme hypo- and hypertension disorders. Each of
the genetic defects found affected renal ion transport. Such well-defined inherited diseases
of blood pressure regulation are not common. In 95% of people experiencing hypertension
their diagnosis is essential hypertension, the root cause being essentially unknown. To
understand this majority of hypertensive disorders, a more complete understanding of
renal ion transport is needed. Because of the role of H,K-ATPase in the final regulation of
K' excretion in the kidney, understanding its contribution to renal ion transport could well
be important in understanding the problem of essential hypertension.
In order to understand the role of H,K-ATPase in the kidney, the exact H,K-ATPase
pumps of the kidney must be characterized at a molecular level. This characterization must
be carried out to a high level of detail, to explore variations in the ion pumps conferred by
differences in changes in transcriptional start or adenylation sites, or by alternative
splicing. With the identity of the H,K-ATPase subunits known at a molecular level,
various reagents can then be developed to characterize H,K-ATPase regulation. These
include cDNA probes, antibodies, and activity assays. With these tools available, it will be
possible to develop a more complete framework of understanding of renal H,K-ATPase.
The studies in this dissertation contribute to our knowledge of the H,K-ATPase
subunits at a molecular level. At the time these studies began, it was not known what
H,K-ATPase subunits were responsible for the active H' and K' exchange that had been
observed in renal collecting duct. Further evidence is presented that all the currently
known H,K-ATPase subunits are present in kidney, and that the primary cell type in which
all are expressed is the collecting duct acid-secreting intercalated cell. When these studies
were begun, there was no known alternative splicing of P-type ATPases. Here alternative
splicing of an HKca subunit mRNA is described, and expression of a protein product is
shown in kidney. In addition, a variant renal medulla HK3 transcript is described that has
tissue-specific expression even within the kidney. In summary, this work has generated a
new appreciation for the complexity of H,K-ATPase molecules that underlie renal
collecting duct H,K-ATPase activity.
H,K-ATPase in the Kidney
The kidney is the principal organ responsible for potassium homeostasis and plays a
major role in maintenance of the acid-base balance of the body. The renal collecting duct
(CD) is the primary site of regulation of the excretion of potassium and the acidification of
urine. A number of studies implicate an apical H,K-ATPase as an important mediator of
these functions. The enzyme (pump) actively transports HF into the lumen of the nephron
in a nonelectrogenic exchange for K' (for review, see Wingo and Cain, 1993, Wingo and
A K' activated ATPase was observed in frog gastric microsomes by Ganser and Forte
(1973). Lee et al. (1974) found that gastric microsomes isolated from dog mucosa were
able to accumulate H' in the presence of ATP and K'. These studies were the initial
observations of the now well-known gastric H,K-ATPase. The designation of
H,K-ATPase was given the enzyme by Sachs el al. (1976).
Gustin and Goodman (1981) isolated apical brush-border membrane of the rabbit
descending colon by isolation of epithelial cells, homogenization, and centrifugation on a
Percoll gradient. They found a membrane-bound, K'-activated ATPase, which had a
KIt=2xlO4M for K', was competitively inhibited by Na:, but had no activation by Na. It
was vanadate sensitive, but oligomycin and ouabain (I mM) insensitive. This represented
the first observations of enzymatic activity for the colonic H,K-ATPase.
Smolka and Sachs (in Sachs et al., 1982), employing monoclonal antibodies, detected
a protein at least similar to the gastric proton-potassium-translocating ATPase protein in
renal distal tubule and colon. Sachs et al. (1982) advanced the idea that this protein might
be involved in the K' reabsorption or the acidification known to take place in kidney or
By quantitating the hydrolysis of [7y-32p]ATP, Doucet and Marsy (1987) observed a
K+-stimulated ATPase activity in rabbit kidney. The level of activity was related to the
density of intercalated cells in microdissected segments of rabbit connecting segment
(highest activity), cortical collecting duct (intermediate), and outer medullary collecting
duct (lowest), and not detectable in any other nephron segments. The ATPase affinity for
K' was high (Km=0.2-0.4mM). The pharmacological properties of the renal H,K-ATPase
was examined as a preliminary indication of the pumps present in the kidney. Omeprazole
is an inhibitor of the gastric isoform ofH,K-ATPase, vanadate inhibits all P-type ATPases,
and ouabain is a Na,K-ATPase inhibitor that also inhibits non-gastric H,K-ATPase
isoforms. The renal ATPase was inhibited by omeprazole and vanadate, but not by
ouabain. They also observed a K'-ATPase activity with potassium restriction in rat renal
outer medullary collecting duct that roughly doubled, changing little after 0.5 weeks
low-K+ diet. In cortical collecting duct activity doubled, rising steadily during a five week
low-K+ diet (Doucet and Marsy, 1987).
Wingo (1987) did a more complete study of dietary K' influence on K' transport in the
kidney. He first established that rabbits on a K'-replete diet, or on a K1-depleted diet, or
on a K+-deplete Na+-supplemented diet all consumed similar quantities. Wingo (1987)
used restricted diets containing 0.55% K', marginally less than the 0.6% generally thought
to be required for normal rabbit growth. In time-course studies he found that after an
initial two-week period of K'-replete meals, 72 hr was sufficient for the response to
equilibrate as judged by urinary Na' and K' excretion. Muscular K' levels had not altered
significantly in that time, showing that renal response preceded effects deleterious to the
animal. Serum K' levels were slightly higher in K'-replete animals, and the same in
K+-depleted Na-supplemented animals compared to K'-depleted animals. The sodium
supplementation evidently maintained the K' serum level. Serum aldosterone was found to
be 2.88.57 ng/dl in K'-depleted Na'-supplemented rabbits, 15.6+5.3 for Kt-depleted
rabbits, and 44.714.0 for potassium-replete animals, and thus any changes in transport
observed cannot be correlated with aldosterone level.
In his experiments, Wingo (1987) found that perfused collecting ducts from outer
medullary inner stripe had similar rates of fluid reabsorption, gauged by 3H-inulin flux, and
similar transepithelial voltages. K' reabsorption, however, roughly tripled among rabbits
on K+-restricted diets, measured as a flux of 42K. K'-replete rabbits had
5.9O.8%,K+-depleted rabbits had 1722.0%, and K+-depleted Na'-supplemented rabbits
had 20.84.2% reabsorption. The data were not significantly different in either of the two
K'-restricted cases. He concluded that K' reabsorption in the medullary collecting tubule
is comparable to K' secretion in the cortical segments at normal fluid flux rates, and could
have important contributions to the amount of K excreted in urine. In a later report by
Wingo (1989), the name H,K-ATPase was first explicitly given to the K*-stimulated
ATPase that had been studied by others in the kidney.
Utilizing a fluorometric microassay in which ATP hydrolysis is coupled to the
oxidation ofNADH, Garg and Narang (1989) detected the presence of a Ky-dependent,
ouabain-insensitive ATPase activity in rabbit kidney that was also omeprazole-,
SCH28080-, and vanadate-sensitive. Sch-28080, like omeprazole, is a potent gastric
H,K-ATPase inhibitor. Earlier work by this group (Garg and Narang, 1988) noted H'
secretion coupled to an ATPase that was not NEM-inhibited, and could not account for
the vacuolar HF-ATPase they were studying. Significant K'-ATPase activity was seen in
microdissected connecting segment (17.03.3 pmol min' mm'), cortical collecting duct
(6.6+0.7), and outer medullary collecting duct (8.81.7), but not in proximal straight
tubule, convoluted tubules, nor the thick ascending limbs. They found the K+-ATPase
activity to be affected by diet. Activity was found in the connecting segment (13.04.0
pmol min1 mm-'), cortical collecting duct (10.13.0), and outer medullary collecting duct
(10.8+2.2) in animals on a low K' diet. Cortical collecting duct activity varied with pH;
optimal pH of 7.4 gave activity of 10 pmol min1' mm-', falling at lower pH to 3 at pH of
7.0, and falling at higher pH to 5 at pH 7.8. Thus, a high K" diet completely suppressed
the K-ATPase activity, while rabbits fed a normal diet had similar activity to the low K*
fed rabbits of the second study.
Studies by Cheval and co-workers (Cheval el al., 1991) examined the 86Rb flux (86Rb is
a K' analog used as a radioactive tracer for transporter studies) and K-ATPase activities.
These activities were blocked by Sch-28080 in cortical and medullary collecting duct.
Activities they attributed to Na,K-ATPase and H-ATPase were unaffected by Sch-28080.
In this study, ouabain was used at a concentration of 2.5 mM to inhibit Na,K-ATPase.
Although there are inconsistencies in the exact concentration at which ouabain inhibits
H,K-ATPases containing the HKa_ subunit, all expression studies carried out to date
(Modyanov et al., 1995; Codina e al., 1996; Cougnon el al., 1996; Grishin et al., 1996)
except for one (Lee et al., 1995) indicate that 2.5 mM ouabain is sufficient to block the
HKa2 enzyme. Lee et al. (1995) detected no ouabain sensitivity at 1 mM, and it is not
known what effect 2.5 mM would have had in their system. In this work by Cheval et al.
(1991) the activities of Na,K-ATPase and an HKa, H,K-ATPase would be
These early studies do not distinguish between different isoforms of H,K-ATPase that
may together account for the H,K-ATPase activity observed in the kidney. Given
historical perspective, this makes sense because at the time these studies were conducted,
only HKxci and HKP3 subunit isoforms were known, and the ouabain-sensitive
H,K-ATPase was not yet known. Later studies (Younes-lbrahim el cal., 1995;
Buffin-Meyer el al., 1997) addressed the issue of multiple isoforms of H,K-ATPase in the
kidney and their distribution along the nephron and collecting duct. These investigators
found three distinct H,K-ATPase activities. Two were found in the collecting duct; one of
these was sensitive to Sch-28080; the other was insensitive. The third was found in
proximal tubules and thick ascending limbs, and was inhibited by Sch-28080. In normal
rats, the sole H,K-ATPase activity present in collecting duct was the Sch-28080-sensitive
activity. In K'-depleted rats, the overall H,K-ATPase activity in collecting duct increased,
while the activity in proximal tubule decreased. The increase was abolished by ouabain,
but not by Sch-28080, implying that the increase is due to an H,K-ATPase isoform that is
pharmacologically dissimilar to HKcx. The proximal tubule/thick ascending limb
H,K-ATPase may have a basolateral polarity, rather than the apical localization of the
collecting duct H,K-ATPase. If the H,K-ATPase role in proximal tubule and thick
ascending limb involves K+ homeostasis, then its down regulation by dietary K' depletion
would imply a basolateral location. Together, these results argue for three different
H,K-ATPase isoforms in kidney. One was constitutively expressed in the collecting duct,
and was Sch-28080-sensitive and ouabain-insensitive, like the HKcai isoform of
H,K-ATPase. The second was presumably located basolaterally in proximal tubule and
thick ascending limb, reduced by low K', and was ouabain-sensitive. The third was located
apically in the collecting duct, stimulated by low K', and Sch-28080-insensitive but
Earlier studies naturally concentrated on establishing the existence of H,K-ATPase in
the kidney. The early studies demonstrated H,K-ATPase activity in connecting segment,
cortical collecting duct, and outer medullary collecting duct. Although measurements were
made for HK-ATPase activity in other nephron segments, no H,K-ATPase activity was
found outside the connecting segment and collecting duct. More recently, proximal tubule
and thick ascending limb have been added as regions having H,K-ATPase activity. These
studies also began to address the differences among H,K-ATPases that are present in the
kidney, defining the characteristics of the H,K-ATPase activities and their localization.
The identity of the H,K-ATPase molecules responsible for the two collecting duct
activities have been found. The identity of the H,K-ATPase molecule responsible for the
activity in the more proximal nephron is not yet known.
H,K-ATPase subunit isoforms in the kidney
Until the last five years, the presence of H,K-ATPase in the kidney has been defined
only on the basis of activity. We have now begun to determine the molecular identities of
the pumps responsible for this activity. And only very recently it has become appreciated
that the complexity of H,K-ATPase expression also includes alternative splicing of HKIx
and HK3 isoforms.
An H,K-ATPase has long been known to acidify the lumen of the stomach. This pump
is a member of the P-type ATPase family, which shares similarity of sequence, structure
and mechanism. The gastric H,K-ATPase is comprised of a catalytic cta subunit and a P
subunit; the active form of the enzyme is an (ao3)2 oligomer (for review see Hershey and
Sachs, 1995; Van Driel and Callaghan, 1995). Full-length cDNAs for rabbit HKa1
(Bamberg etal., 1992) and HKp3 (Reuben el a., 1990) have been cloned and sequenced.
RT-PCR followed by sequencing of the amplified products has been used to demonstrate
that mRNA (Ahn and Kone, 1995) and protein (Callaghan el al., 1995) for HKoct and that
mRNA and protein for HK3 (Callaghan el al., 1995) isoforms are present in kidney. This
author was included in the latter study.
The non-gastric H,K-ATPase a subunit isoforms (HKa2) have been cloned from
several tissues. A partial cDNA for an HKa2 isoform was obtained from human axilla skin,
and mRNA observed in brain and kidney (Modyanov el al., 1991). A full-length cDNA
was found subsequently (Grishin et a/., 1994). An HKa2 cDNA was cloned from rat distal
colon by Crowson and Shull (1992); the corresponding amino acid sequence has 86%
amino acid identity to the cDNA derived from human skin. This rat HKa,_ mRNA was
detected in kidney, uterus, and heart using two separate cDNA probes from HKa2 3' UTR
and C-terminal transmembrane domains. The 3' UTR derived probe detected HKo- 2 in
forestomach as well. These human and rat cDNAs were shown to encode H+/K' exchange
activity when expressed in Xenopus /laevis oocytes (Modyanov et a/., 1995; Cougnon et
al., 1996), although recent work questions the stoichiometry of the exchange in the human
isoform (Grishin et al., 1996). In two of these studies rabbit HKp3 subunits were
cotransfected (Modyanov el cal., 1995; Grishin etl a/., 1996), and in the other Bufo marinus
HK3 subunits were cotransfected (Cougnon et al., 1996). A partial HKcx2 cDNA was
cloned from a rabbit cortical CD (CCD) library having 84% amino acid identity to the
human HKot2, and mRNA was detected in CCD and colon (Fejes-Toth et a[., 1995).
Watanabe et al. (1992) cloned and sequenced a similar cDNA fromrn distal colon of guinea
pig. The degree of identity at the amino acid level among human, rat, guinea pig, and
rabbit HKoc2 clones is less than the amino acid identity of HKa1 (>97%) between the three
species, but much greater than the typical amino acid identity between HKa1 and other
P-type ATPases (<64%). Therefore, although controversy on this point exists, this author
considers these to be orthologous and refers to them collectively as HKKa2. As part of this
dissertation, further evidence will be presented that the HKca genes are indeed orthologs.
A related cDNA (75% identity at the amino acid level) was cloned from Bufo marinus
bladder, an analog of mammalian collecting duct (Jaisser et al., 1993). While mRNA was
detected in toad bladder, none was observed in either stomach or colon. The evolutionary
distance between toad and mammal coupled with the tissue distribution dissimilarity of the
toad HKc make it difficult to evaluate the relationship between the toad and mammalian
isoforms. In addition to uncertainty in the number of H,K-ATPase isoforms in the kidney,
it has been recently been discovered that there are two alternatively spliced transcripts of
the rat HKc2 in kidney (Kone, 1996, Higham and Kone, 1998). Thus, several
H,K-ATPase c subunit isoforms have been reported in kidney, but it is at present
uncertain whether these account for all the renal H,K-ATPases. Also uncertain is their
relative contributions to renal H,K-ATPase activity.
Experiments have been conducted in attempts to find more members of the gene
family that includes the x subunits ofNa, K-ATPases and H,K-ATPases. Shull and Lingrel
(1987) probed a human genomic library at low stringency with probes made from
Na,K-ATPase sheep ai, rat cl, rat ao2, and rat U3. Five different sequences were obtained,
with two to ten clones representing each sequence. Three of these were known to be
human NaKcxi, NaKoa2, and HKocx genes. The fourth was later identified as HKa2
(Modyanov el al., 1991). The fifth, which is physically linked to the NaKat2, was later
identified as NaKa4 by Shamraj and Lingrel (1994). In a similar experiment, Sverdlov et
al., (1987) screened a human genomic library with a probe made from porcine kidney
NaKa1. The probe was constructed to contain the well-conserved region surrounding the
active site aspartate residue. They obtained five distinct clones. They were recognized as
the three known NaKa subunits and the two known HKa subunits. In sum, the results of
these two experiments imply that all members of the gene family of X,K-ATPases are
At present, there are no known specific HKP3 isoforms in addition to the one originally
discovered in gastric tissues. However, it is thought that the NaKPI subunit isoform is the
partner to HKaC2 in active H,K-ATPase pumps in colon and kidney (DuBose et al., 1998;
Kraut et al., 1998). These observations are the first suggesting that an individual P-type
ATPase 3 subunit has more than one primary P-type ATPase ac subunit partner in vivo.
This was the determination reached by two separate groups using different antibodies to
specifically immunoprecipitate a(3 pairs, lending strength to their independent and identical
conclusions. However, it is possible that during tissue processing some of the o3/P pairs
may mix partners, leading to an erroneous conclusion. Pairs that are not thought to
associate in vivo have been seen in expression systems to give rise to functional activity
(Horisberger et al., 1991; Codina et al., 1996), showing that the in vivo P-type ATPase
pairs are not exclusive when expressed in expression systems. Expressed in their proper
cell types, there may be compartmentation that controls HKc and HK3 selectivity. This
could segregate the subunits from their incorrect partners on the basis of translation of the
various H,K-ATPase and Na, K-ATPase isoforms at different times or in different
Alternative transcriptional start sites for HK3 in the stomach were detected by
Newman and Shull (1991). The primer extension method used by Newman and Shull
(1991) would not have necessarily differentiated between alternative transcriptional start
sites and alternative splicing. Primer extension would only give information about the
distance from the primer to the beginning of any or all of the transcripts that contain the
primer site. So it is possible that they were detecting alternative splicing as well as
alternate transcription start sites. Thus far, however, there have been no reports of
multiple HKO3 subunit transcripts in any tissue seen by RNA analysis techniques such as
northern analysis or ribonuclease protection assay. In present work, I present evidence
that the transcriptional start site of HK3 subunit in the medulla has slight differences
relative to stomach, and that in renal medulla there are two transcripts expressed at
comparable level, one of which appears to be the product of alternative splicing.
In summary, the first H,K-ATPase subunit detected in kidney was the non-gastric
isoform HKa2, shown by RT-PCR to be present in human kidney (Modyanov et al.,
1991). Shortly before this author joined his lab, Dr. Brian Cain found that the rabbit 3
subunit isoform of stomach was present in kidney (Callaghan et al., 1995;
Campbell-Thompson el al., 1995). After the work described here was begun, work by
other investigators (Ahn and Kone, 1995) established that the catalytic subunit isoform
found in stomach, HKc(i, is also found in kidney. HKcax protein was shown by immunoblot
in rabbit kidney (Callaghan el al., 1995). To further complicate the picture, an
alternatively spliced HKcL2 (dubbed HKo2,, with the canonical HKcC2 being renamed
IHKO2a) was discovered in kidney (Kone and Higham, 1998). A search for the 3 subunit
partner to HKxc2a established that NaKP3 couples to HKct2i, in vivo (Dubose et al., 1998;
Kraut et al., 1998). Presumably, HKcx, pairs with HK3 in kidney as it does in stomach
(Hall etal., 1991; Shin and Sachs, 1994, Mathews el al.; 1995). All the known
H,K-ATPase subunit isoforms are known in kidney. HKoi/HKf3 and HKKa2/NaKf3
H,K-ATPases contribute to the activities observed in the collecting duct. Experiments
have been done to find other candidate H,K-ATPase genes, and to date none have been
detected. So far no molecular identity can be associated with the H,K-ATPase activity in
the proximal tubule and thick ascending limb.
Deficiencies of H.K-ATPase
There are two lines of evidence to indicate that H,K-ATPases are critical to
maintenance ofK+ homeostasis, and that disturbances in this balance are a serious health
problem. First, transgenic mice with deranged H,K-ATPase function show substantial
disturbances ofK+ balance (Meneton el al., 1998). Second, an environmentally high level
of vanadium in northeast Thailand effectively inhibits H,K-ATPase function among
humans and water buffalo living in the area (Dafnis el al., 1993).
Experiments have been carried out to show the critical nature of H,K-ATPase
expression and its relevance in the kidney (Meneton el al., 1998). Meneton showed that
transgenic mice homozygous with respect to an HKci: subunit deletion were normal when
fed diets that had normal levels of K' (1% K). However, when fed K'-free diets
(<0.004% Kt) these mice experienced loss in body weight, plasma K', and muscle K.
HKa2 is not the major mechanism of K' conservation in the kidney; the urinary K'
excretion rate in both wild-type and HKo2-deficient mice declined 100-fold compared to
HKc2-deficient mice on normal diets. However, mean urinary K' excretion per day was
consistently higher (typically 120% of wild-type) following a week of K*-free feeding.
This suggests that there may be a role played by renal HKcx2 in K' conservation. The role
played by HKa2 in the colon was more clear. Fecal excretion rate increased four-fold in
HKa2-deficient compared to normal mice that had been fed the same K+-free diet,
indicating an inability to reabsorb K' by the digestive tract. Cardiac arrhythmias were
observed in some of the HKa2-deficient mice, presumably due to low plasma K'.
Anecdotal evidence concerning H,K-ATPase activity in humans comes to us from
Thailand, where environmental vanadium levels are high. Vanadate is a transition state
inhibitor for all P-type ATPases, and its presence in drinking water and soil in northeast
Thailand was associated with an epidemic of renal distal tubular acidosis. It is interesting
that although vanadate would be expected to inhibit Na,K-ATPases and Ca-ATPases,
along with H,K-ATPase in stomach and distal colon, the predominant symptoms were
renal. The disease is characterized by an inability to lower urine pH to below 5.5 pH units.
Affected persons had generalized paralysis, hypokalemia, metabolic acidosis, muscle and
bone pain, and nocturia (Nilwarangkar el al., 1990). The connection between acid
secretion in the kidney and stomach had been made, and some patients tested positive for
gastric hypoacidity (Sitprija el al., 1988). However, the exact cause of the acid secretion
defect was unknown. As recently as 1990, a genetic predisposition had not been ruled out
(Nilwarangkar et al., 1990). Patients are treated by K' and alkali supplements, and some
deaths occurred with non-compliance.
It was shown in rats that intraperitoneal injections ofvanadate (5 mg/kg) led to
hypokalemic distal renal tubular acidosis and loss in muscle K' (Dafilis et al., 1993).
Cortical collecting duct K-ATPase activity that was sensitive to Sch-28080 (200 uM)
declined 75% in vanadate-treated rats. Medullary collecting duct activity declined less than
50% by the same assay. However, it should be noted that this is a sufficiently high
Sch-28080 concentration to inhibit HKa2 enzyme activity according to some investigators
(Modyanov et al., 1995; Grishin et al., 1996), but other studies have not seen sensitivity
to Sch-28080 even at that relatively high concentration (Cougnon el al., 1996; Codina et
al., 1996). Therefore, a vanadate effect on l-IKc(2 enzyme activity in this study might have
remained undetected. Na,K-ATPase activity was measured also, and showed a decline in
vanadate-treated animals as well. This assay also may or may not have detected a
contribution due to HKcx2 activity, which is expected to be ouabain-sensitive at
concentrations of between >10 pM (Modyanov et al., 1995) and <1 mM (Codina et al.,
1996) depending on the study. The evidence suggests that vanadate could be a factor in
the endemic hypokalemic distal renal tubular acidosis in northeast Thailand. More precise
measurements need to be made to quantitate the effect of Sch-28080 and ouabain on
In summary, these studies illustrate the essential nature of H,K-ATPases in the renal
collecting duct (and in distal colon) for the maintenance of K" balance. In the extreme
case, deranged H,K-ATPase function can result in disease or even death. These studies
isolating the activity of H,K-ATPase do not address the role H,K-ATPase might play in
blood pressure regulation. The other more acute affects of impairments of H,K-ATPase
activity probably obscured the affects on blood pressure regulation. Although in some
cases a defect in a single gene leads to a dramatic loss in blood pressure control,
hypertension is most often a multifactorial disease. H,K-ATPase is likely to be one of a
combination of activities that fail to control blood pressure in the hypertensive patient.
H.K-ATPase Structure and Function
HK-ATPase catalytic subunit
The HKax subunit contains the active sites relating to its catalytic action (Figure 1-1).
A number of functionally important sites in the HKax have been defined. The location of
the phosphorylated aspartyl residue of the catalytic intermediate is known and the region
immediately surrounding it is extremely well conserved within the P-type ATPase family
(Walderhaug et al., 1985). The location of a residue that binds the fluorescent molecule
FITC is also known; FITC competes with ATP for binding to H,K-ATPase and therefore
is thought to be at or near the HKo ATP binding site (Jackson et cil., 1983; Farley and
Faller, 1985). The site of the K'-competitive inhibitor Sch-28080 which acts at an
Figure 1-1. Schematic diagram of H,K-ATPase. Dashed line indicates the
amino-terminal extension of HKac protein.
Kx \ H +
extracellular site (Munson and Sachs, 1988) is known by studies with a photoaffinity
analog to lie between the first two transmembrane domains (Munson et al., 1991). The
Na,K-ATPase inhibitor ouabain binds to the corresponding site in that enzyme, but it is
not known whether this is the binding site conferring ouabain's less sensitive inhibition of
HKcL2. The binding site of the medically important inhibitor omeprazole is also known, it
may bind to any of three extracellular cysteines located in the C-terminal quarter of the
subunit (Besancon et al., 1993). This inhibitor has achieved some popular acclaim,
television advertisements may be seen for it under its brand name Prilosec.
Using hydropathy analysis, the existence of four transmembrane helices in the
amino-terminal half ofHKci is clear. The number of transmembrane domains of the
carboxyl-terminal half of HKo is not as apparent from the hydropathy plot, but may be
placed somewhere between three and five by that analysis. However, the presence of an
odd number of membrane-spanning regions would be inconsistent with the placement of
amino- and carboxy-termini in the cytoplasm by antibody reactivity (Smolka el ali., 1992;
Mercier el al., 1993). Also, the localization of the active sites discussed in the preceding
paragraph, predicts a large intracellular cytosolic loop following the first four
transmembrane domains. The picture is further complicated because recent results
(Raussens et al., 1998) imply that some of the transmembrane domains may be comprised
of 3-strand secondary structure. Using limited tryptic digestion followed by fluorescent
labelling of cysteines, the four transmembrane segments of the amino-terminal half could
be confirmed (Besancon el al., 1993; Shin et al., 1993). In the carboxyl-terminal half of
the protein, only three transmembrane domains were observed. The two most
carboxyl-terminal membrane spans strongly implied by hydropathy analysis were not
observed by fluorescent labelling despite the presence of multiple cysteines. In vitro
translation experiments have also been preformed to elucidate H,K-ATPase topology
(Bamberg and Sachs, 1994). These experiments involve synthesis of putative
transmembrane segments in the presence and absence of microsomes, then detection of
glycosylation of the resulting polypeptides by electrophoresis. Positive glycosylation of a
fusion protein containing glycosylation sites would imply the insertion of a single
transmembrane domain. When a pair of membrane spanning domains were translated, no
glycosylation would result. Both single transmembrane domains and pairs of
transmembranes were tested. In this system, the predicted final two transmembrane
segments were inserted into the membrane. Taken together these experiments predict a
secondary structure for HKxc1 that has amino- and carboxy-termini located intracellularly.
There are four transmembrane domains in the amino-terminal half of the protein, and
between four and six transmembrane domains in the carboxyl-terminal half If
H,K-ATPase membrane topology is conserved with that of Ca-ATPase, recently imaged
by cryoelectron microscopy (Zhang el al., 1998), then HKai subunits probably have ten
The significance of H,K-ATPase enzyme quaternary structure is presently under active
consideration. Studies suggest that there are specific and stable associations between
catalytic subunits. Radiation inactivation studies showed the minimum functional unit to be
an (c(3)2 heterotetramer (Rabon et act., 1988) The electron microscope diffraction pattern
in images of two-dimensional H,K-ATPase crystals suggested a tetrameric arrangement of
HKci subunits (Hebert el al., 1992). The apparent contradiction of the radiation
inactivation and electron microscopy experiments can be explained. An ax/a interaction has
been shown to be necessary for optimal activity of H,K-ATPase (Morii et al., 1996).
Dimeric Wca interactions were necessary for H,K-ATPase activity, and tetrameric ao/a
interactions confer higher affinity binding of ATP. When coexpressed in insect cells,
NaKai, NaKa2, and NaKcX3 were found to coimmunoprecipitate (Blanco et al., 1994).
Using chimeric constructions combining NaKca and HKcio alternating regions, it was
found that the large intracellular loop was necessary for this association (Koster et al.,
1995). Using a yeast two-hybrid system and fusions of Gal4 to aX subunit cytosolic loops
Colonna et al. (1997) further explored this interaction. In two-hybrid assays, there was no
apparent direct interaction between pairs of the large cytosolic loop (Figure 1-1). Likewise
there was no interaction observed between the smaller loop between the second and third
transmembrane domains. However, the two-hybrid assay gave a positive interaction
between the smaller and larger loops. The interaction is apparently between the loop
between transmembrane domains two and three of one member of the al/ca pair and the
larger loop between transmembrane domains four and five of the other. These results
imply that interactions between H,K-ATPase molecules play a role in their function. There
may be interactions with other molecules as well, such as those governing cell polarity.
H.K-ATPase B subunit
Antibodies to the HK3 subunit protein (Chow and Forte, 1993) and reduction of any
of the three disulfide bonds in the HK3 subunit (Chow et al., 1992) have been observed to
affect catalytic subunit function. In addition to a role in modulating enzymatic activity, the
HKp subunit is also thought to participate in shepherding the H,K-ATPase through the
Golgi and to the plasma membrane (Renaud et al., 1991). The HKI3 subunit contains a
tyrosine-based signal required for internalization of the H,K-ATPase pump
(Courtois-Coutry et al., 1997) and thus required for proper regulation of pump activity.
The HK3 subunit protein contains 7 consensus N-glycosylation sites and all are
glycosylated (Chow and Forte, 1993). Thus, the HK3 subunit is required for H,K-ATPase
pumps reaching the cell surface and proper function of the enzyme. Its role in
internalization means that it is required to down-regulate gastric acid secretion.
Hydropathy analysis predicts a single membrane domain for the HK3 subunit protein.
Interactions between HKa and HKp3 subunit proteins have been studied by limited tryptic
digestion. After digestion, detergent solubilization, and lectin binding of the HK3 subunit,
an HKa subunit fragment corresponding to the putative loop between transmembrane
domains seven and eight was recovered (Shin and Sachs, 1994). The sequence of this
fragment is well-conserved, implying some important function. Yeast two-hybrid analysis
has confirmed this area of the HKca subunit protein (Arg-898 to Arg-922) as a region of
association with the HKp3 subunit (Melle-Milovanovic el al., 1998). The yeast two-hybrid
analysis has also shown that two extracellular domains of the HK3 protein to be regions of
association with the HK(a subunit. The HKf3 subunit amino acids involved are Gln-64
through Asn-130 (adjacent to the membrane) and AkIa-156 to Arg-188.
The HK3 subunit is quite important to H,K-ATPase activity due to its role in
intracellular trafficking of the holoenzyme. It apparently also affects the conformation of
the intact complex, because HK3-specific effects modify enzymatic activity. Just as there
are a/a interactions that are now known, the regions of the subunit important to a/3
interactions are also becoming known. These may be important in conferring the effects on
enzyme activity conferred by oligomeric structure. They may also be important in
controlling the interaction between the various ca and P3 subunit proteins of the
H,K-ATPases and Na, K-ATPases.
Renal Tissue Culture Cells
A number of renal continuous cell lines exist, but for the purpose of studying
H,K-ATPase relevance to the cortical collecting duct the selection narrows considerably.
In cortical collecting duct H,K-ATPase activity is relatively high, and based on
immunohistochemical (Wingo et al., 1990) and in situ hybridization evidence
(Campbell-Thompson el al., 1995; Ahn and Kone, 1995). H,K-ATPase is found in
intercalated cells in cortical collecting duct. There are two cell lines with characteristics of
these cell types, MDCK cells and RCCT-28A cell.
The Madin-Darby canine kidney (MDCK) cell line is an epithelial cell line that
appears to have originated from distal tubule (Herzlinger et al., 1982) or cortical
collecting tubule (Valentich, 1981), and has properties of intercalated cells (Pfaller et al.,
1989). MDCK cells are aldosterone-responsive (Simmons, 1978), and immunostaining of
HWiK' ATPase has been observed (Adam Smolka, personal communication). Oberleithner
et al. (1990) detected an aldosterone-stimulated, omeprazole-inhibited transport activity
that was blocked by increased apical extracellular [H'], or decreased apical extracellular
[K+], consistent with the presence of an apical H'/K' ATPase in MDCK cells. MDCK cells
are available for purchase from the American Type Culture Collection (ATCC, Rockville,
MD). Because of the uncertainty of their origin, and because we wanted to take full
advantage of the knowledge of rabbit renal physiology accumulated over the years by our
collaborator Dr. Charles Wingo, we selected the RCCT-28A cell line for use in our
The RCCT-28A transformed cell line was derived from immunodissected rabbit CCD
by Arend el al. 1989. The RCCT-28A line was created by microdissecting cortical
collecting tubule and dispersing cells on plates coated with monoclonal antibody specific
to collecting tubule cells. The antibody coating the plates, known as IgG3(rct-30), had
been made against rabbit renal cortical cells injected into mice. The resultant monoclonal
antibody stained only collecting tubules on cryotome sections. Staining was primarily
basolateral in intercalated and principal cells. Cells from the dissected rabbit collecting
duct that had bound in the antibody-coated dish were immortalized with an adenovirus
2-SV40 hybrid, and then cloned by limited diffusion. A population of cells that continued
to proliferate while retaining epithelial morphology was obtained.
The antigenic and hormone response of these cells is specifically consistent with their
origin in the cortical collecting tubule. Immunocytochemistry showed reactivity in 100%
of cells to an antibody (mr-mct) against mitochondria-rich cells of the medullary collecting
duct (Schweibert el aL., 1992). The mr-mct antibody was seen to be specific for
acid-secreting, or type A cells (Burnatowska-Hledin and Spielman, 1988).
Immunoreactivity was also observed to an antibody to band 3 protein (1VF12), another
marker for acid-secreting intercalated cells (Schweibert el al., 1992). The presence of
carbonic anhydrase in >95% of cells was indicated by the binding of a fluorescent
acetazolamide analog (Dietl el al., 1992). As expected for an intercalated cell,
conductance of Cl, but not Na' or K', was indicated by patch clamp measurements (Dietl
etal., 1992). Schweibert el al. (1992) saw no antigenicity of the cells toward two
antibodies specific for base-secreting (or type B) cells or toward four antibodies specific
for principal cells. The origin, collection, and characterization of these cells indicates that
they are a good model of the acid-secreting intercalated collecting duct cell.
The first study undertaken using these cells showed that adenosine analogs increase
intracellular calcium by stimulating phosphoinositide turnover (Arend et al., 1989).
Inositol turnover was measured by labeling cells with myo-[3H]inositol and detecting
[3H]inositol phosphate formation. A 1-receptor agonists increased phosphoinositide
turnover, and the increase was blocked by an A l-receptor antagonist. Adenosine also
regulated a 305 pS chloride channel in RCCT-28A cells via protein kinase C and a G
protein (Schwiebert et al., 1992). Chloride channels in these cells have been characterized
by the patch clamp method, showing Cl conductance to be stimulated by isoproterenol
(Dietl et al., 1992; Dietl and Stanton, 1992). Cell swelling of RCCT-28A cells activated a
Cl conductance by altering the organization of actin filaments (Mills el al., 1994;
Schwiebert el al., 1994). Activation of the channel was mimicked by stretching the
membrane and disruption of F-actin by dihydrocytochalasins. Stabilizing F-actin with
phalloidin blocked activation of the Cl channel. Bello-Reuss (1993) reported H,K-ATPase
activity in RCCT-28A cells, observing an apical acidification mechanism that had a
component sensitive to withdrawal of K' and two H,K- ATPase inhibitors, Sch-28080 and
omeprazole. She also detected activities suggestive of an apical H-ATPase and a
basolateral Cl /HCO3 exchanger. Acid secretion by these cells was diminished in cells
grown at low pCO2, evidence of regulation of the acidification mechanism by alkaline
conditions. These cells have proven useful in studying various processes normally
associated with acid-secreting intercalated cells. The observation of H,K-ATPase activity
in these cells made them particularly appealing for our studies.
H,K-ATPase activity in the collecting duct has very real implications for maintenance
of health and well-being. Studies have shown a myriad of derangements in the absence of
functional H,K-ATPase activity, including a real possibility of death. The H,K-ATPase
provides an excellent example of how critical some of the enzymes are that fine tune the
environment of the body. It is interesting that an enzyme that has such an immense level of
activity in one organ, the stomach, actually is more important in another, the kidney, in
which its level of activity might be called small.
Several challenges exist in studying the H,K-ATPase enzyme and its function. One of
these is the lack of knowledge of the quantitative effects of ouabain and Sch-28080 on the
HKaC2 isoform of the enzyme. Another its low level of expression, making assays of
mRNA, protein, and activity difficult. There was also a dearth of knowledge about the
molecular forms of H,K-ATPase in the kidney when these studies were undertaken. This
situation has changed, and the contributions described herein have been part of that
MATERIALS AND METHODS
The RCCT-28A cell line was derived from immunodissected renal cortical collecting
duct (Arend et al., 1989). These cells were the kind gift of Dr. William Spielman, and
experiments were performed using cells between passages 11 and 31. Cells were grown in
DMEM media supplemented to 10% with FBS and to 1% with Penicillin-Streptomycin.
All media was bubbled overnight with a mixture of 5% CO2, 21% 02, 74% N2, then filter
sterilized by use of a 0.20 pm cellulose acetate syringe-mounted filter (Corning, Corning,
NY). Cells were maintained in culture in tissue culture flasks at 37C in a 5% CO2
atmosphere. Media was changed on alternate days, and cells were split 4:1 at confluency.
For experiments, cells were passage to Corning Costar Transwell Collagen-coated
semipermeable inserts. Cells were plated to a density of 2x 105/cm2 on the inserts, grown
for 2 days in media containing 10% FBS, and then shifted to 0.1% FBS for a period of 24
hr prior to the experiment.
Isolation of total RNA
Total RNA was isolated from cells or tissues employing the method described by
Chomczynski and Sacchi (1987). After aspiration of media, tissue culture cells were rinsed
in ice cold sterile PBS (10 mM sodium phosphate, pH 7.4, 150 mM NaCI). Next, 0.5 mL
of GTC solution (4 M guanidine thiocyanate, 25 mM sodium citrate, 0.5%
n-lauroylsarcosine, 100 mM P3-mercaptoethanol) was added to promote cell lysis. The
resulting viscous fluid was scraped from the insert and emptied into polypropylene
centrifuge tubes on ice. New Zealand White Rabbits were sacrificed by decapitation and
the kidneys removed immediately. Kidneys were dissected under a microscope, slicing
coronally to separate cortex and medulla. Distal colon was prepared by clipping the distal
25 mm of colon, cutting longitudinally, then rinsing away contents by a thorough spray of
ice cold PBS. Gastric mucosa was collected by slicing the stomach transversely, rinsing in
ice cold PBS, then scraping the rugae with a Scoopula (Fisher, Pittsburgh, PA). Tissues
were dounce homogenized in 4 mL GTC solution until the suspension appeared to be
homogneous. The resulting viscous fluid was poured into polypropylene centrifuge tubes
While being held on ice, 1 volume phenol was added to the homogenized tissue, then
1/10 volume 2 M sodium acetate (pH 4.0), and 0.22 volume chloroform-isoamyl mixture
(24:1). After each addition, the samples were briefly vortexed. The samples were
subjected to centrifugation at I 1000G X 25 min at 4'C. The upper phase was retained, and
RNA was precipitated twice in isopropanol and twice in ethanol, then resuspended in 100
gL DEPC-treated water. Storage of the RNA was at -20C. RNA concentration was
determined by OD26o reading on a spectrophotometer. For absorbance of 1.0 at 260 nm,
the concentration was taken to be 40 ptg/ml.
Northern blots were done following the procedures of Sambrook el al. (1989).
Samples of total RNA (20 [tg per lane) or mRNA (2 ug per lane) underwent
electrophoresis in a 1% agarose, 0.22 M formaldehyde denaturing gel. Capillary transfer
to a nylon membrane (Hybond N, Amershain Corp., Arlington Heights, IL) was conducted
overnight in 20X SSC (3M NaCI, 300 mM Na citrate). Absorbent paper towels were
changed twice during transfer. RNA is immobilized on the membrane by baking 2 h at
80"C in a vacuum oven. 32P-labelled probes were prepared by random primer extension of
75-150 ng DNA according to the protocol of the Megaprime Kit (Amersham).
Membranes were prehybridized a minimum of 15 mrin and hybridized with labelled probe
for 24 h in hybridization solution at 65C. When heterologous probes were employed, such
as cross-species probing, temperatures as low as 42C were used to reduce stringency.
Washing was done first at room temperature for 20 min in IX SSC, 0.1% SDS, then 3
times at hybridization temperature for 20 mini in 0.2X SSC, 0.1% SDS. After washing,
membranes were exposed to Kodak BioMax MS film at -80C with Kodak BioMax
intensifying screens. Exposures between three and six days were often required to detect
H,K-ATPase subunit mRNA. An mRNA probe for glyceraldehyde-3-phosphate
dehydrogenase, a glycolytic enzyme, was used as a control to ensure even loading
amounts of RNA between lanes. An exposure of eight hr was typically sufficient to
visualize this control.
RT-PCR using degenerate primers
RT- PCR was carried out as described by Davis ei a!., 1994. The cDNA template for
the PCR reaction was produced by incubating 1 lag total RNA from microdissected rabbit
renal cortex and 0.1 uig random hexamers in a volume of 11 IPL at 70*C X 2 min. The
reaction mixture was quenched by placing the tube on ice. The reverse transcription
reactions were carried out at 37C X 1 hr in a volume of 25 uL. The mixture contained
RNA-random hexamers mix in the final concentrations indicated: First Strand Buffer (50
mM Tris-HCl (pH 8.3), 75 mM KCI, 3 mM MgCl), dithiothreitol (10mrnM), dNTPs
(2mM), RNAsin (30 U, Promega, Madison, WI), and Superscript 11 (200 U, Gibco BRL,
Gaithersburg, MD). PCR reactions were primed with pairs ofoligonucleotides shown in
Table 2-1 synthesized by the University of Florida Interdisciplinary Center for
Biotechnology Research (UF ICBR) DNA Synthesis Core. PCR reactions were
performed in a volume of 100 uaL at the final concentrations indicated: dNTPs (0.22 mM),
PCR Buffer (20 mM Tris-HCl, pH 8.4, 50 mM KCI), MgCl, (50 mM), primers (1 Pug
each), and Taq DNA Polymerase (5 U). The reactions were overlaid with 50 JAL mineral
oil. Thermal parameters of the reactions included a 5 mrin X 94C presoak followed by
94C X 40 sec denaturation, 55C X 1 min anneal, and 72C X 2 min extension for a total
of 30 cycles followed by a final extension of 5 min. Reactions were held at 4C overnight
before PCR product cloning into the pCR 11 vector and transformed into OneShot E. coli
cells utilizing the TA Cloning Kit (Invitrogen, San Diego, CA). E. coli cells containing the
plasmid of interest were grown overnight in a 5 mL culture and plasmids isolated using the
QiaPrep Spin Mini Kit (Qiagen, Santa Clarita, CA). Sequencing of plasmids was carried
out by the UF ICBR DNA Sequencing Core.
RT-PCR using standard primers
A reverse transcription reaction was carried out as in the RT-PCR method described
above for degenerate primers. PCR reactions were primed by pairs of oligonucleotides
synthesized by the UF ICBR DNA DNA Synthesis Core and summarized in Table 2-1.
Reactions were carried out using I pg total RNA in a 50 pL volume in KlenTaq PCR
Reaction Buffer (40 nmM Tricine-KOH, pH 9.2, 15 mM KOAc, 3.5 mM Mg(OAc)2, and
75 gg/mL Bovine Serum Albumin), primers as listed above (0.5 PM), and dNTPs (0.8
mM), and Advantage KlenTaq Polymerase Mix (Clontech, Palo Alto, CA). Thermal
parameters of the reactions included a 2 min X 94C presoak followed by 94C X 30 sec
denaturation and 68C X 1 min extension with a final extension of 5 min. The numbers of
cycles were specific to each experiment and are indicated in the 1 legends. PCR reactions
were purified for sequencing by phenol/chloroformn (1:1) extraction and two 3000 G X 5
min centrifugation steps in Ultrafri-ee-MC 30000 NMWL regenerated cellulose columns
(Millipore, Bedford, MA). Due to low yield of the HKal PCR reaction, a 45 cycle
reaction was carried out to generate product for sequencing. Sequencing of PCR products
was carried out by the UF ICBR DNA Sequencing Core.
To amplify genomic DNA for sequencing, template used was 1 Pug Clontech Rabbit
Genomic DNA. Clontech KlenTaq Advantage PCR Mix was used with components and
I 0 r-
co 4/- cc n c
*l Cl Cl Cl Cl -
U U U U U U
o 2 .2
< 0 U
concentrations specified in the description of RT-PCR above. Primers were
ACCCGCGGCGCCTCCAGCGCGACAT (nucleotides 16-40, BC386) located in the
first exon of HKc2,. and TATCTGTAGCTGCATGGTGCTCCAC (nucleotides 69-93,
BC334) located in the second exon of HKcx2,. Thermal parameters of the reactions
included a 1.5 min X 94C presoak followed by 5 cycles of 94C X 15 sec denaturation
and 72*C X 2 min extension, 5 cycles of 94'C X 15 sec denaturation and 70C X 2 min
extension, and 25 cycles of 94C X 15 sec denaturation and 68C X 2 min extension with a
final extension of 8 min. Reactions were held at 15'C overnight before ligation of the
products into the pCR 2.1 vector. The ligation mixture was transformed into OneShot E.
coli cells utilizing the TOPO-TA Cloning Kit (lnvitrogen, San Diego, CA). E. coli cells
containing the plasmid of interest were grown overnight in a 5 mL culture and plasmids
isolated using the QiaPrep Spin Mini Kit. Sequencing of plasmids was carried out by the
UF ICBR DNA Sequencing Core.
A 3' RACE reaction was carried out as described by Davis et cil., 1994. mRNA was
prepared from rabbit renal cortex total RNA using the PolyATtract (Promega, Madison,
WI) system. A reverse transcription reaction was carried out as in the RT-PCR method
described above, using 1 pg mRNA as template and 0.1 pg primer
(GACTCGAGTCGACATCGA[T]17, BC229). The complementary strand was next
synthesized using 5 pL of RT reaction along with 0.1 utg of sense primer
(TGCGGAAACTCTTCATCAGG, nucleotides 3088-3107, BC262) in a reaction volume
of 98 pL that contained the following components: PCR Buffer (20 mM Tris-HCl, pH
8.4, 50 mM KCI), MgCl2 (50 mM), dNTPs (0.22 mrM), overlaid with 50 pL mineral oil.
After a 95C incubation for 5 min, the temperature was lowered to 70C and 5 U Taq
DNA Polymerase was added. A 2 min annealing phase followed at 55C, then the
complementary strand was extended at 72C for 10 min. Antisense primer (0.1 pg,
GACTCGAGTCGACATCG, BC230) was added, and a PCR reaction was initiated with a
94"C X 40 sec denaturation, followed by a 55C X 1 min anneal, and a 72C X 2 min
extension for a total of 40 cycles. Final extension was for a duration of 5 min. Reactions
were held at 4C overnight before cloning PCR products into the pCR II vector utilizing
the TA Cloning Kit (Invitrogen, San Diego, CA) per manufacturer instructions. The
library of clones produced in this manner was screened using an oligo
(CTCTACCCTGGCAGCTGGTG, nucleotides 3108-3127, BC261) 5' labeled using the
ECL kit (Amersham, Arlington Heights, 1L). Sequencing was carried out by the UF ICBR
DNA sequencing core.
A 5' RACE reaction was performed using the Marathon cDNA Amplification Kit
(Clontech, Palo Alto, CA) following manufacturer's instructions with the following
modifications. RT reactions were carried out using 5.5 pg total RNA of rabbit renal
cortex, incubated with the gene-specific primer TTGCCATCTCGCCCCTCCTT
(nucleotides 121-102, BC331) for 30 min at 50C, then at 550C for 15 min. PCR reactions
were carried out using the anchor primer (CCATCCTAATACGACTCACTATAGGGC,
API) included in the Marathon kit paired with gene-specific primer
TATCTGTAGCTGCATGGTGCTCCAC (nucleotides 93-69, BC334). Concentrations
and components of the reactions were detailed above for use of Clontech KlenTaq
Advantage Polymerase Mix. Thermal parameters of the reactions included a 1 min X 94C
presoak followed by 5 cycles of 94C X 15 sec denaturation and 72C X 1 min extension,
5 cycles of 94C X 15 sec denaturation and 70'C X 1 min extension, and 25 cycles of 94C
X 15 sec denaturation and 68C X 1 min extension with a final extension of 8 min.
Evaluation of PCR products
PCR products were analyzed by agarose gel electrophoresis. 20 U1L of reactions were
run along with 1 mL DNA loading dye (50% glycerol, 1% xylene cyanol, 1%
bromophenol blue) on a 1.2% agarose gel in TAE buffer (Tris mM, acetic acid mM,
EDTA mM). Low DNA Mass Ladder (Gibco BRL) was electrophoresed alongside PCR
products to quantitate DNA concentration, and 100 bp ladder (Gibco BRL) to evaluate
the size of the products.
In some cases, products were not visible after ethidium bromide staining, so Southern
blotting was used to visualize the products. Procedures followed were as described by
Davis etal., (1994). After running products on gels as described above, gels were soaked
30 min in denaturation solution (1.5 M NaCI, .5 M NaOH), and 30 min in neutralization
buffer (1 M ammonium acetate, .02 M NaOH). Capillary transfer in neutralization solution
and baking 80* X 1 hr was utilized to adhere DNA to nylon membrane (Hybond N).
Probes to hybridize specifically to the expected products (Table 2-1) were created using
RT-PCR with rabbit renal cortex as template, cloning the inserts using the TA Cloning Kit
as described above. Inserts were sequenced to confirm their identity. Probes were labelled
using the ECL Direct System(Amersham) following manufacturers instructions. Exposure
to Hyperfilm ECL (Amersham) for a period of two hr was required to visualize the results
of PCR reactions.
Preparation of memrnbrane protein from tissue culture cells
Two clusters of six wells each were used for each experimental condition; this
typically yielded 50 pg of protein. All solutions and glass douncers are cooled on ice, and
centrifuges are either refrigerated or located in cold rooms. Cells are held on ice at all
times. Cells were rinsed in PBS containing 0.5 mM PMSF, 1.5 mL in top well of
Transwell, 2.5 mL in bottom well. Following aspiration of the rinse solution, to each
single well was added 0.5 mL PBS, 0.5 mM PMSF, 1 mM EDTA. A cell scraper was used
to dislodge cells from the insert. This solution containing cells was pipetted from one well
to the next dislodging and gathering all the cells from one cluster. This was repeated with
a fresh 0.5 mL solution on the same cluster to remove any remaining cells. The two passes
combined for a total of 1 nmL, and the other clusters were processed in the same manner.
Cells were spun in a centrifuge for 5 min at 500 X g to remove cellular debris.
Supernatant was discarded, and the pellet was resuspended in swelling buffer (tris 10 mM
pH 7.8, 1 mM EDTA, 1 mM PMSF, 2 .tM aprotinin, 2 i.M leupeptin, 2 paM pepstatin) for
15 min. Cells were homogenized by 50 strokes in a glass douncer. After moving cells back
to a microfuge tube, 0.1 I mL of 10X salts (300 mM NaCI, 20 mM MgCI2, 10 mM Tris
pH 7.8) was added before vortexing. Mixture was spun in a centrifuge for 1 minute at
1000 X g to remove nuclei. Supernatant was retained and spun in a centrifuge for 5 min at
1500 X g. Supernatant was retained and spun in a centrifuge for 30 min at 23000 X g.
Cells were resuspended in 20 itL resuspension buffer (1 volume swelling buffer, 1/10
volume o10X salts) and stored at -20"C.
Preparation of membrane protein from rabbit tissues
Kidney, distal colon, and stomach tissues were obtained in the same manner as for
RNA isolation. Care was taken to maintain solutions and apparatus ice cold. Two rabbit
distal colons were required per preparation to yield useful concentrations of protein.
Tissues were homogenized for 15 seconds at 12500 rpm (setting 8 on Omni-Sorvall tissue
homogenizer) in Buffer A (50 mM sucrose, 10 mM Tris pH 7.4, 1 mM EDTA, 1 mM
PMSF). After allowing 15 seconds for settling, homogenization was repeated. Three
volumes buffer B (250 mM sucrose, 10 mM Tris pH 7.4, 1 mM EDTA, 1 mM PMSF)
were added, then homgenate spun in a centrifuge for 10 min at 1000 X g. The supernatant
was subjected to centrifugation for 20 min at 10000 X g three times. Final centrifugation
was for 1 hr at 100000 X g. After discarding supernatant, pellet was resuspended in 500
piL loading solution (1 mM Tris, 10 mM MgCI2, 150 mM NaCI). After transferring
resuspended pellet to glass douncer, dounce is dropped into the douncer, turned three
times, then raised and dropped again. This was done ten times. The resultant preparation
was stored at -20C.
Peptides used as immunogens were designed for maximum antigenicity and minimum
homology to other proteins. Avoiding homology to other P-type ATPases was particularly
important. Rabbit H,K-ATPase catalytic subunit sequences were scanned using the
computer program PEPTIDESTRUCTURE (Genetics Computer Group, 1997) to find
regions relatively high in charged, hydrophilic residues (Jameson and Wolf, 1988). Such
regions were then searched by BLAST (Altschul el tl., 1990) to eliminate those that could
be predicted to cross-react with other proteins. Once these constraints were met the
candidate sequences were examined using the program MOTIFS (Genetics Computer
Group, 1997; Bairoch and Apweiler, 1996) to ensure that there were no potential sites for
protein modification that might affect reactivity. In the case of HKcL1 and HKo2a,, there
was only one region that satisfactorily met all of these criteria. In the case of HK2c,, there
were two, one being the region in common with HKX2,, the other in the extended amino
terminal region of HKot2c not contained in HKx:,.
Peptides were synthesized with an N-terminal cysteine and conjugated to keyhole
limpet hemocyanin (UF ICBR Protein Core) using the Pierce (Rockford, IL) Imject
system. Three peptides were used as immunogens, the first was designed to react with a
portion of HKcL1, the second was designed to recognize a portion of HKoC2 found in both
HKac2 and HKctx2,c, the third to a portion unique to HKax. The peptide chosen for HKo.i
corresponded to amino acids 569-582 (CLYLSEKDYPPGYAF). The peptide chosen
within the common region contained amino acids 18-37 in HK 2., and 79-88 in HKa2c
(CDIKKKEGRDGKKDNDLELKR). The peptide chosen within the HKazc-specific
region corresponded to amino acids 13-25 (CGEERKEGGGRWRA). Antipeptide
antibodies were raised in chickens by Lofstrand Laboratories (Bethesda, MD). Chickens
received boost innoculations at 21 day intervals, and were exsanguinated at day 73 to
produce antisera. Preimmune sera was collected prior to initial innoculation. Yolks of eggs
collected over the two week period prior to final bleed were pooled, and immunoglubulins
purified from yolk material by the Promega EGGstract method. Concentration of the
EGGstracted yolks was determined by the modified Lowry procedure ofMarkwell et al.
(1978) and the concentrations adjusted to 2 mg/mrnL by addition of lgY buffer solution
(Promega). Purity and concentration of the IgY obtained was confirmed by non-reducing
SDS-PAGE and staining with Coomassie blue.
Protein concentrations in tissue and cell samples were determined by modified Lowry
(Markwell el al., 1978). Proteins (10 ag/lane) were separated on 4-20% reducing
SDS-polyacrylamide gels (BioRad, Hercules, CA ), 10 utg per lane. Vesicle preparations
were suspended in buffer (62.5 mM Tris-HCI pH 6.8, 10% glycerol, 5%
P-mercaptoethanol, 3% SDS) and incubated 2 min X 90C prior to electrophoresis. Gels
were rinsed 10 min in TBS (10 mM Tris- HCI pH 7.2, 150 mM NaCI), and 10 min in
transfer buffer (20 mM Tris-HCL, 150 mM glycine, 20% methanol, pH 8.3). Proteins
were electrotransferred at 104 V, .25 A, 4'C to Hybond ECL nitrocellulose membranes
(Amersham) in transfer buffer. Blocking of membranes was done in TBS-T (TBS, 0.1%
TWEEN-20) containing sodium azide and 5% non-fat dry milk at 40C overnight or room
temperature for 1 hr. Antibody incubations were one hour each, carried out in TBS-T
containing 5% non-fat dry milk. Following blocking and following primary and secondary
antibody incubations, immunoblots were rinsed in TBS-T with continuous agitation. This
was done three times for one minute each, then twice for 5 min each. Primary antibodies
purified from egg yolk were used at a dilution of 1:200, the anti-chicken IgY-horseradish
peroxidase conjugated secondary antibody (Promega) was diluted to 1:10000. When
antisera or preimmune sera were used as primary antibodies, dilution was 1:2000. A final
wash was carried out in TBS for 10 min, and then antibody reactivity was detected using
chemiluminescence (Pierce). Apparent molecular masses were established using the High
Mass Range Molecular Weight Markers (BioRad).
Measurement of pH,
The fluorescent, pH-sensitive dye BCECF-AM was used to directly measure pH, in
RCCT-28A tissue culture cells. BCECF-AM was stored as a stock at -20C in a 30 mM
solution in DMSO. Cells were incubated at room temperature for a period of 30 min in
solution 1 (Table 2-2) with BCECF-AM at a final concentration of 5 M. A minimum of
5 min perfusion with solution I delivered at 37C was allowed to rinse BCECF away at
the beginning of each experiment.
Cells were imaged by epifluorescence at 530 nm emission on an inverted microscope
(Boyarsky el al., 1988, Weiner and Hamm, 1989) using excitation wavelengths of 440 nm
and 490 nm. These wavelengths correspond to the isosbestic point and to a highly
pH-sensitive wavelength of BCECF, respectively. The ratio of emission intensities at the
two wavelengths is directly proportional to pH over the pH range being studied, and is
constant with respect to such variables as cell-to-cell variations in dye uptake and leakage.
A field containing approximately 50 cells could be visualized using a Nikon X40, 0.55
Table 2-2. Solutions for determination of pH,
1 2 3 4 5 6 7
102.2 122.2 119.2
I I I
1 1 1
1.2 1.2 1.2 1.2 1.2 1.2 1.2
1 1 1 1 1 1 1
5 5 5 5 5 5 5
8.3 8.3 8.3 8.3 8.3 8.3 8.3
25 25 25 25 25 25 25
Note: Concentration units are in miM. Osmolality was adjusted to 2905 mosmol/kg H20
by addition of the principal salt. pH was adjusted to 7.40.05 by addition of
tetramethylammonium-OH. Solutions were bubbled with 100% 0.
LWD lens on a Nikon Diaphot-TMD inverted microscope. Excitation light was provided
by a 100 W mercury lamp. The light was split into two beams by a 470 nm low-pass
dichroic mirror. The two split light beams then passed through filters to yield beams of the
desired wavelengths. The transmitted light path contained a 440 nm filter; the reflected
light path contained a 490 nm filter. Computer-actuated shutters on each light path
alternated the incident light wavelengths and minimized the time the cells were subjected
to the high intensity light. The two light beams were recombined by a second 470 nm
low-pass dichroic mirror. Light was directed to the microscope stage by a 510 nm
high-pass mirror, and the emitted light was directed to a Videoscope KS-1381 image
intensifier coupled to a Dage 72 CCD camera. Because the optics necessary to image cells
grown on inserts required relatively intense incident light, measurement frequency was
limited to avoid phototoxicity. Therefore, measurements were made at 30 second
intervals. Images were digitized and stored by computer allowing subsequent analysis of
single cells using the Image 1/FL software package (Universal Imaging Corp.,
During measurements, cells were constantly perfused at a rate of-10 mL/min by
HEPES-buffered solutions that were continuously bubbled by 100% 02 and heated to be
delivered at a temperature of 37C. Switches for the various solutions were located
physically near the input to the cell chamber to minimize the time required to switch
solutions, and fluid was continuously removed from the opposite side of the chamber by
vacuum suction. For some experiments, the apical side of the cells was perfused by
different solutions than the basolateral side of the cells by utilizing separate input tubes in
the upper and lower chambers of the Transwell inserts.
Cells were acid-loaded using the NH4Cl prepulse technique. Briefly, cells were
incubated with 20 mM N1-H4CI (equimolar substitution for NaCI) for 5 mrin, then
ammonium chloride was removed from the perfusing solution. Addition of inhibitors or
K1 removal began at the start of the ammonium pre-pulse. The ethylisoproplamiloride
(EIPA) stock solution was 1 mM in dimethyl sulfoxide (DMSO), and Sch-28080 was 10
mM in DMSO. EIPA was obtained from Research Biochemicals, International (Natick,
MA) and Sch-28080 was the kind gift of Dr. James Kaminski at Schering Corporation
(Bloomfield, NJ). Inhibitors were stored as stock solutions at -20C and diluted into the
solutions indicated immediately preceding their use.
Calibration of the pH, measured by BCECF fluorescence was carried out by the
nigericin calibration technique of Thomas el cil. (1979). Calibration solutions (120 mM
KCI, 1.2 mM CaCI2, 1 mM MgCI2, 25 mM HEPES, 14 pM nigericin) were adjusted to
6.6, 6.8, 7.0, 7.2, and 7.4 pH units using 1 M NaOH and HCI. Cells were incubated with
each solution for a minimum of three min, and three fluorescence ratio measurements were
taken at each pH.
pH, recovery rates for individual cells were calculated using least-squares linear
regression. Rates were calculated for the period beginning two min after NFLCI
withdrawal, allowing time for cells to equilibrate after acidification, and ending with Na+
addition or EIPA withdrawal. Data were collected for independent experiments involving
separate passages of cells. pH, recovery rates for each experimental condition were
presented as the mean of the rates determined for the individual experiments SE. Cells
without NaVH' exchanger activity, defined as an increase in pH, recovery with EIPA
removal, were classified as non-viable and excluded from further analysis. P<0.05 by
Student's t-test was taken as significant.
H,K-ATPASE P3 SUBUNITS IN THE RABBIT RENAL MEDULLARY COLLECTING
There are multiple H,K-ATPase subunit isoforms that are involved in coupled H* for
K+ active exchange in the kidney. These now include the NaKP1, subunit that is the newly
recognized partner for the HKx,, subunit, in addition to its long-known role as partner to
NaKaxi. In addition to the complexity involved in having multiple H,K-ATPase isoforms
present, there is the possibility of variant transcripts of each H,K-ATPase subunit.
Differences in polyadenylation sites, transcription start sites, and alternative splicing in a
tissue-specific manner might further complicate the picture.
When rat HK3 subunit was originally sequenced, multiple transcriptional start sites
were observed by primer extension experiments (Newman and Shull, 1991). It was not
known what role these various transcriptional start sites play. When the Cain, Wingo and
Nick laboratories did preliminary studies concerning the regulation by K' status of the
HK3 subunit in rabbits, it was noted that on some northern blots of renal tissues HKP3
subunit mRNA appeared as a doublet. This was evidence for a second H,K-ATPase
subunit transcript present at a level comparable to the primary transcript.
Here we show that there is a variant HKp3 subunit mRNA that is expressed in renal
medulla, and not in renal cortex or stomach. We have designated the variant mRNA
HK3'. The HKJ3' transcript is found in medulla at a level comparable to the quantity of
the HK3 subunit, and the translation start site is unchanged, so both HK3 and HK3'
encode the same protein. HK3' may be the product of alternative splicing.
Renal Medulla HKI3 mnRNA Variant
In order to investigate the doublet band in northern probed with Hk3 cDNA probes,
RNA was prepared from stomach, renal cortex, and stomach. In northern analyses (Figure
3-1) involving twenty different rabbits in experiments spaced over several years, only
single transcripts were observed in either gastric or renal cortical tissues. Gastric and renal
cortical transcripts appeared to be of identical size. However, in renal medullary tissues, a
doublet was universally visible. Northern analysis showed that the smaller of the two
medullary HK3 transcripts was the same size as in the other two tissues, and the second
renal medullary transcript was larger. The quantity of the second mRNA, HK3', was
generally comparable to the quantity of HKP3 mRNA.
In order to determine the molecular nature of the different transcripts of the medulla, 3'
and 5' RACE experiments were conducted using mRNA isolated from rabbit renal
medulla. The rabbit RNA was selected from those that showed prominent HKO subunit
upper bands by northern analysis because these would have a higher abundance of the
novel transcript. Representative PCR products generated by the 3' and 5' RACE
experiment are shown in Figure 3-2. PCR often 3' RACE products sequenced, all were
the same, and all extended to within 5 bp of the published HK3 cDNA sequence (Reuben
et al., 1990). Ten 5' RACE products were also sequenced. Nine of these were similar to
- 1.4 kbp
- 1.3 kbp
Figure 3-1. Northern analysis showing presence of HKP mRNA in renal cortex, renal
medulla, and stomach. The existence of two mRNA species was clearly visible in renal
medulla, whereas a single mRNA species was detected in stomach and renal cortex.
GAP3DH was used to show the condition of the RNA samples.
Figure 3-2. 3' and 5' RACE reactions to amplify HKP cDNAs using rabbit renal
medulla RNA as template.
the published HK3 cDNAsequence. However, all ended approximately 20 bp from the 5'
end of the published sequence making it impossible to confirm the extreme 5' end of the
One of the sequences had a very different 5' end than the others. In that sequence
(Figure 3-3A), the first 11 bp of the published HK3 sequence was not found, and in its
place there was a different 118 bp region (Figure 3-3B). The sequence was confirmed by
5' RACE using a primer specific to the variant extension
(CCCTGCACCCCGACTGAGG, nucleotides 104-121, BC388).
To examine the possibility that this renal medulla-specific transcript might be regulated
by K' status in rabbits, northern analysis was carried out using total RNA from animals fed
a low K' diet. Renal medulla total RNAs from four rabbits fed a Harlan (Indianapolis, IN)
control diet and four rabbits fed a Harlan K'-restricted diet for two weeks were probed to
visualize HK3 subunits using a probe made from 570 bp of coding region (Table 2-1).
Neither of the two HK3 transcripts of renal medulla had systematic variation in HIK3
mRNA level due to K' restriction in these animals. Northern analysis of renal cortex and
stomach RNA from these same rabbits showed uniform level of HKp3 mRNA between
rabbits. No regulation of HK3 was observed, although in renal medulla there was a great
deal of individual variation between rabbits.
A rabbit renal medulla-specific HK3 variant mRNA was cloned by 5' RACE and
sequenced. It was not known why only one out of the ten sequences proved to be a
---------+--------- --..---...__- + ....--.-- -+.......-+ --....+-- -+
.---------+-----------.----+--+----------+-.......-------+ --.........--- +
S+--------------. .. +--------------+ ..-----+-- -+-.- --... --....- ....- +
1 M A A L Q E K K S C S Q R M E E F R H Y C W N P D T G
28 ... Remainder of sequence omitted for clarity
HKP' L118bp /
Figure 3-3. HKp and P' subunits. A) The 5' end of the HKp' cDNA and its deduced
amino acid sequence. Numbering begins at the start of the HKp' transcript. HKp'
nucleotide 119 corresponds to nucleotide 12 of published gastric HKp3 sequence
(Reuben, et al., 1990), and the point of divergence is marked by B) Comparison of
HKp3 and HKp' mRNAs. There is no difference in the deduced amino acid sequence.
variant, since both transcripts are observed by northern analysis at comparable levels.
Perhaps the extra length or RNA secondary structure reduced amplification efficiency.
Because the variant 5' end differed in sequence from the 5' end reported for rabbit
(Reuben etal., 1990), the variant is presumed to arise from alternative splicing. This was
not pursued because our focus was on the H,K-ATPase ca isoforms. HK3 mRNA and
HK3' mRNA produced identical HK3 subunit proteins. The splice sites of the related rat
and mouse HK3 genes (Newman and Shull, 1991; Canfield and Levenson, 1991) are not
conserved, and working out this issue might have been a considerable target for our
The deduced amino acid sequence is unchanged by the variant 5' end. Differential
regulation by alternate promoters may be the functional reason for the two transcripts. K+
restriction was studied as a possible stimulus for regulating the promoters, but K' status
alone did not correlate with differential regulation of the two transcripts. Other stimuli,
such as aldosterone levels, may regulate these transcripts in renal medulla. It is interesting
to note that although HK3 subunit is expressed at extremely high levels in gastric tissues,
this variant transcript was not observed in stomach. The tissue-specificity of the variant is
striking. Study of its promoter region may be fruitful in terms of finding the cis acting
factors that turn the gene on specifically in renal medulla.
H,K-ATPASE X SUBUNITS IN THE RABBIT RENAL CORTICAL COLLECTING
H,K-ATPase has been shown to play a role in acid/base and K' transport in kidney
collecting duct (for review, see Wingo and Cain, 1993; Wingo and Smolka, 1995).
However, many questions remained to be addressed, such as finding which H,K-ATPase
isoforms are present in kidney, and determining their distribution along the nephron and
collecting duct. This included defining the cell type specificity of each isoform. For many
years, the rabbit has been the archetypal experimental animal for use in experiments
involving microperfusion of renal tubules, To take advantage of the extensive knowledge
of rabbit renal physiology in exploring H,K-ATPase function at the molecular level,
primary structure information was needed for H,K-ATPase subunits in that species.
Because of the discovery of H,K-ATPase catalytic subunit isoforms in addition to the
gastric isoform in several species, the relationship between these isoforms needed to be
examined. Perhaps most importantly, the role that the multiplicity of H,K-ATPase subunit
isoforms plays in kidney function should be examined.
At the time the experiments described here were undertaken, little was known of the
identity ofH,K-ATPase subunits in the kidney. mRNAs encoding catalytic subunit
isoforms cloned from human axillary skin (Modyanov et acil., 1991) and rat distal colon
(Crowson and Shull, 1992) were the first to be detected in kidney. In collaborations
including the Cain and Wingo laboratories, HK3 (Callaghan et aC., 1995,
Campbell-Thompson el a!., 1995) was observed in rabbit kidney. The key experiment
identifying HKP3 in the kidney dated from 1992, prior to the arrival of this author, so that
information was available in designing this dissertation project. As this work progressed, it
was found that the HKax1 subunit isoforms are present in renal tissues (Ahn and Kone,
1995). In addition, an alternatively spliced isoform of rat HKcx was recently reported in
kidney (Kone and Higham, 1998). The rat HKl;2 alternatively spliced variant was
designated HKoth, with the original renamed HK 2.,. The previous experiments were not
designed to discriminate between HKa;2, and HKo-2b.
In these studies a full-length cDNA sequence was determined for the HKoa2 subunit in
rabbit. An alternatively spliced variant of this HKc2 subunit isoform was found in rabbit.
The pattern of splicing was identical to that found in rat (Kone and Higham, 1998).
However, the translation start site was not conserved. For this reason, the alternatively
spliced variant in rabbit was designated HKcx2,. Antipeptide polyclonal antibodies were
raised and used to show expression of both HKax2, and HKc(x, proteins in rabbit renal
cortex. These results allude to a potentially complex pattern of regulation of H,K-ATPase
activity in renal cortex.
Multiple H.K-ATPase cx Subunits in the Kidney
We set out to find the HKac subunits that mediate the H,K-ATPase activities that had
been observed in kidney. Because HKa;,, mRNA had been observed in kidney, it was
necessary to find the rabbit sequence of HKa,,, in rabbit for use in designing tools for
studying H,K-ATPase function in the kidney. Rabbit HKa(. sequence was known
(Bamberg et al., 1992). In designing an experimental approach to this goal, we wanted to
consider 1) the possibility that HKccX was expressed in kidney (unknown at the time) in
addition to HKca2, 2) the possibility that the HKa2 mRNA might be different in kidney than
in the other tissues for which full length sequence was known, and 3) that novel HKa
isoforms might be present in renal tissues. A good technique to address all these
possibilities was to do RT-PCR using degenerate primers designed to anneal to regions of
sequence that were highly conserved among P-type ATPases. This approach was expected
to amplify any P-type H,K-ATPase that might be present in rabbit kidney.
The aspartyl residue that is phosphorylated as an intermediate in the catalytic cycle of
H,K-ATPase lies within a motif that is highly conserved in the entire P-type ATPase
family. Another well-conserved region among P-type ATPases surrounds a lysine residue
that can be modified by FITC. FITC competes with ATP for binding (Jackson et al., 1983;
Farley and Faller, 1985), so this region is thought to make up part of the ATP binding site.
Advantage was taken of these two regions to design degenerate primers to amplify any
P-type ATPase using renal cortex RNA as template Fifteen P-type ATPase sequences
were aligned using the computer program PILEUP (Genetics Computer Group, 1997).
These sequences were selected to give a wide range of mammalian Na,K-ATPases and
H,K-ATPases. Chicken Na,K-ATPases were included because all three ct subunit isoforms
were known for that species so they comprised a good example of P-type ATPases from
an organism less related to mammals. Toad sequences were included because toad
H,K-ATPase was one of only three non-gastric H,K-ATPases known at the time.
Degenerate primers were then designed to amplify any of these ATPases (Figure 4-1).
The products of the first two RT-PCR reactions using these primers, when cloned and
sequenced, had high homology to the NaKcxi subunit in other species. The third reaction
product (Figure 4-2) was more related to the non-gastric H,K-ATPase a subunits than to
the HK(X| subunit or to any of the Na, K-ATPase x subunits. It was a 419 bp fragment
that corresponded to nucleotides 1182-1600 of the full-length rabbit HKa2, cDNA
sequence. A BLAST search listed three non-gastric H,K-ATPases as the most highly
aligned sequences to the 419 bp sequence, with other H,K-ATPases and Na,K-ATPases
being less well aligned (Figure 4-3). The sequence of the fragment shared
89% nucleic acid identity with the rat distal colonic H,K-ATPase oa subunit, 88% with a
guinea pig distal colonic H,K-ATPase cx subunit, and 86% with the H,K-ATPase X subunit
cloned from human axillary skin. Based on these similarities, this fragment was tentatively
identified as belonging to an H,K-ATPase.
Using degenerate primers designed to anneal to conserved sequences we identified
farther 5' and 3' in the sequence alignment paired with gene-specific primers designed
based on the new sequence (Table 2-1), fragments containing more of the transcript were
cloned (Figure 4-4). Outside the coding sequence, homology between cDNA sequences
declines. With much of the coding region cloned and sequenced, 5' and 3' RACE were
used to clone the full extent of the cDNA including the two ends. The 3' end was easily
obtained; a single 3' RACE product was observed the first time the procedure was
performed. The 5' end proved far more interesting, but more challenging to find. Initial
G G G
G G G
Figure 4-1. Design of degenerate primers for RT-PCR of novel P-type ATPases. A)
Upstream primer was designed to anneal to well-conserved sequence at the enzyme active
site phosphorylated aspartyl residue. GenBank loci of 15 aligned sequences are shown at
left. Oligonucleotide is shown beneath aligned sequences. B) Downstream primer was
designed to anneal to well-conserved sequence at the putative ATP binding site.
Oligonucleotide is reverse complement of consensus sequence shown.
Figure 4-2. RT-PCR product amplified from rebbit renal cortex RNA using degenerate
primers. Gel isolated product was cloned and sequenced.
Sequences producing High-scoring Segment Pairs:
gb U02076 HSU02076 Human ATP-driven ion pump (ATPlALl...
gb M90398 RATATPASEZ Rat H+,K+-ATPase mRNA, complete cds.
gb U94912 RNU94912 Rattus norvegicus H-K-ATPase alpha...
dbjID21854IGPIHKAAS Guinea pig mRNA for distal colon H...
gbjU94913jRNU94913 Rattus norvegicus H-K-ATPase alpha...
embi Z25809IBMHKATPAS B.marinus mRNA for H,K-ATPase
gbIM599601CHKNAKAT3 Chicken Na,K-ATPase alpha-3-subuni...
embIX05883IRNATPAHO Rat mRNA homologous to alpha subun...
gb M14513 RATATPA3 Rat Na+, K+-ATPase alpha(III) isof...
gbI M28648 RATNALPH2 Rattus norvegicus Na,K-ATPase alph...
embIZ11798 IBMNKAAl B.marinus mRNA for Na, K-ATPase al...
gb U10108 XLU10108 Xenopus laevis Na+-K+-ATPase alpha...
gb U49238 XLU49238 Xenopus laevis adenosine triphosph...
gb J02649 RATATPAST Rat stomach (H+,K+)-ATPase mRNA, c...
gb U17249 XLU17249 Xenopus laevis gastric H(+)-K(+)-A...
embIX64694 oCATPRNA O.cuniculus mRNA for ATPase (alpha...
gb J03230 CHKATPAS Chicken (Na+ + K+)-ATPase mRNA, co...
gb S66043 S66043 sodium pump alpha subunit (Ctenoce...
gb U17282 MMU17282 Mus musculus gastric H(+)-K(+)-ATP...
gb L42565 HUMATP1G04 Homo sapiens (clone 1SW34) non-gas...
gb L11568 DOGHKATP Dog H+,K+-ATPase mRNA, complete cds.
embIX02813 OAATPMR Sheep mRNA for (Na+ and K+) ATPase...
gb M22724 PIGATPHK Pig (H+ + K+)-ATPase mRNA, complet...
gb M28647 RATNALPHl Rattus norvegicus Na,K-ATPase alph...
gb S74801 S74801 H(+)-K(+)-ATPase alpha-subunit [ra...
gb M14511 RATATPA1 Rat Na+,K+-ATPase alpha isoform ca...
gb M74494 RATNAKATP Rat sodium/potassium ATPase alpha-...
emb X76108 AASPAA A.anguilla mRNA for sodium/potassi...
emb X05882 RNATPAR Rat mRNA for alpha subunit kidney-...
emb X02810 TCATPMR Torpedo californica mRNA for (Na+ ...
gb M59959 CHKNAKAT2 Chicken NA,K-ATPase alpha-2-subuni...
gb U16798 HSU16798 Human Na,K-ATPase alpha-i subunit ...
embjX04297|HSATPAR Human mRNA for Na,K-ATPase alpha-s...
gb J03007 HUMATPAS Human Na+,K+ ATPase alpha-subunit ...
gb L42173 DOGNKAA Canis familiaris Na, K-ATPase alph...
gb M38445 PIGATPBSEN Pig NA+, K+-ATPase alpha subunit m...
embIX03938|SSATPAR Pig mRNA for (Na+, K+)-ATPase alph...
gbIM14512|RATATPA2 Rat Na+,K+-ATPase alpha(+) isoform...
emb X58629 CCNAKATP C.commersoni mRNA for Na(+)/K(+) A...
emb X56650 AFNAKATP A.franciscana mRNA for Na/K ATPase...
gb M75140 HYDATPASE H.vulgaris Na,K-ATPase alpha subun...
gb L42566 HUMATP1G05 Homo sapiens (clone 1SW11-1) non-g...
gb S76581 S76581 Na,K-ATPase alpha-i subunit [dogs,...
gb J05451 HUMATPGG Human gastric (H+ + K+)-ATPase gen...
gb M63962 HUMHKATPC Human gastric H,K-ATPase catalytic...
Figure 4-3. BLAST search using sequence of 419 bp fragment of HKa2.
pWGC 11 RACE
Figure 4-4. Cloning of HKao. and HKa, cDNAs. Individually cloned cDNAs are
indicated by bars and the plasmid numbers are indicated. AUG and UGA indicate start
and stop codons of the mRNA, respectively. Symbols: *, degenerated primers; RACE,
RACE primer. All other primers were gene-specific.
attempts employed the classical technique of Frohmian el al. (1988), which is analogous
to the 3'RACE process. In several attempts, this technique produced only short cDNA
sequences which provided little new sequence information. Taking a new approach, the
Marathon cDNA Amplification Kit by Clontech (Palo Alto, CA) was tried. This kit uses a
ligation technology to add a known upstream sequence for PCR amplification, rather than
the less efficient terminal transferase technology employed by Frohman et al. (1988). Also,
the reverse transcription step was changed to a higher temperature to lessen the possibility
of secondary structure blocking full extension by the RNA polymerase. With the new
protocol, after some optimization, two different 5' ends were obtained when 5' RACE
products were sequenced.
Two rabbit renal HK_2 cDNA sequences were found in this manner, having 4035
bases in common at the 3' end but different 5' ends. The GenBank records of these two
cDNAs are shown in Figure 4-5. A segment of the shared 3' portion of the sequences was
identical to the 1456 bp sequence obtained by Fejes-Toth (1995) except for two single
base mismatches. One mismatch was a transition of C-,T at nucleotide 2927, which does
not change the primary protein sequence and the other a transversion of G-T at
nucleotide 3259, in the 3' untranslated region. The full-length rabbit HKoaa nucleotide
sequence was 86% identical to human HKc2 and 83% to rat, whereas identity of HKoCz, to
rabbit HKc, was only 67%. The deduced amino acid sequence shared 87% identity with
human HKaX2, 87% identity with rat HKca2, but merely 64% with rabbit HKai. Northern
analysis using an HKc2,,-specific probe (Table 2-1) is shown in Figure 4-6.
AF023128 4079 bp mRNA MAM 13-OCT-1997
Oryctolagus cuniculus H+,K+-ATPase alpha 2a subunit mRNA,
Eukaryotae; Metazoa; Chordata; Vertebrata; Mammalia;
Eutheria; Lagomorpha; Leporidae; Oryctolagus.
1 (bases 1 to 4079)
Campbell,W.G., Weiner,I.D., Wingo,C.S. and Cain,B.D.
H,K-ATPase in the RCCT-28A rabbit cortical collecting duct
2 (bases 1 to 4079)
Campbell,W.G., Wingo,C.S. and Cain,B.D.
Submitted (08-SEP-1997) Biochemistry, University of Florida,
JHMHC 100245, Gainesville, FL 32610, USA
/strain="New Zealand White"
/product="H+,K+-ATPase alpha 2a subunit"
Figure 4-5. GenBank accession records for rabbit HKca2 sequences. A) Genbank record
for rabbit HKoc2,,. B) Genbank record for rabbit lHK,.>
1037 a 1073 c 1024 g 945 t
2761 aagacaacta tggacaggaa
4422 bp mRNA
Oryctolagus cuniculus H+,K+-ATPase alpha 2c subunit mRNA,
Eukaryotae; Metazoa; Chordata; Vertebrata; Mammalia;
Lagomorpha; Leporidae; Oryctolagus.
1 (bases 1 to 4422)
Campbell,W.G., Weiner,I.D., Wingo,C.S. and Cain,B.D.
H,K-ATPase in the RCCT-28A rabbit cortical collecting duct
2 (bases 1 to 4422)
Campbell,W.G., Wingo,C.S. and Cain,B.D.
Submitted (08-SEP-1997) Biochemistry, University of Florida,
100245, Gainesville, FL 32610, USA
/strain="New Zealand White"
/product="H+,K+-ATPase alpha 2c subunit"
1174 c 1136 g 1009 t
Figure 4-6. Northern analysis showing presence of HKza, in distal colon and renal
cortex. GAP3DH was used to show the condition of the RNA samples.
To determine the relationship between the rabbit HKa2 mRNAs and other P-type
ATPases, phylogenetic analysis was necessary. Several programs and algorithms were
sampled, all giving the same general pattern of sequence relationships. Programs employed
include CLUSTALW (Thompson et al., 1994), DISTANCES (Genetics Computer Group,
1997), and PAUP (Genetics Computer Group, 1997). All three programs use the distance
algorithm, and the maximum parsimony algorithm of PAUP was also used. In addition to
distance and maximum parsimony analyses, a third popular algorithm to assay relatedness
of sequences is used, the maximum likelihood analysis. Distance analysis generates a tree
showing relatedness of sequences by simply counting
dissimilarities in aligned sequences as a test of their homology. Maximum parsimony
considers only positive information in making a comparison. For instance, in an alignment
of four sequences, a position that had one each of A, C, G, and T would not be included in
the analysis because that position would not give information about any pair of the
sequences being related. A position that had two As and two Cs would be counted
because it would show the grouping of two pairs. Distance analysis would include both
positions in the analysis. Maximum likelihood takes into account the tendency for a given
type of mutation to occur. In two sequences that have a point mutation, that point
mutation was more probably created by a "likely" change than a change that has less
tendency to occur. Maximum likelihood would measure two sequences as more related if
their alignment shows a "likely" mutation. If the mutation had less tendency, it may
indicate a more distant relationship between sequences arrived at only by multiple changes
or higher pressure of selection. Thile maximnlum likelihood analysis requires information that
is not readily available for rabbit, and was not attempted. A representative phylogram
arrived at by distance analysis by the CLUSTALW program is shown in Figure 4-7.
The 5' end of the second rabbit sequence (HKa2c) differed dramatically from the
published rat HKat_,, (Crowson and Shull, 1992) and human HKa_,, (Grishin el cil., 1994)
sequences. The HKcx2c 5' untranslated region bore no homology to the comparable
segments from the HKc2,, cDNAs from human or rat, or to any GenBank sequence. It
was clear from the beginning that this second 5' end might represent an alternative splicing
product; the sequence homology diverges from the human HKa2 sequence at a point
known in the human ATP1ALI gene to be a splice junction (nucleotide 177 ofATP1AL1)
(Sverdlov et cial., 1996).
Alternative Splicing of H.K-ATPase cx Subunits in the Kidney
When a sequence for the 5' end of the rat HKKc2 gene was published (Kone and
Higham, 1998), we wanted to compare the sequence at the 5' end of the rabbit gene. In
order to determine the exon arrangement at the 5' end of the HKa2 gene, rabbit genomic
DNA was amplified by PCR. The antisense primer was selected within the region of
sequence that is common to HKa;2, and HKa-,. Because it was unknown which of the two
cDNA 5' ends was located more 5' in the gene, sense primers that anneal to sequence
within each were selected. The oligonucleotide primer specific to HKca2., gave the larger
amplified product, and its sequence is shown in Figure 4-8A. The exon specific to HKc2,a
splices to the common core sequence at the splice site shown. The exon specific to HKKa2,
rat HKa2 I
.. human NaKa2
--- rat NaKal
Figure 4-7. Distance analysis of selected HlKa and NaKa. subunit coding.
T HKa_,. transcription start 71
A ) 1 GCCCCCTGCC CGCCGACCCG CGGCGCCTCC AGCGCGACAT GCGCCAGgtg
M R Q
51 tgtgaggaag tgacgcggtg cggactggcg agaagtgcgg gaaagggtga
101 agggctccgt ccgggggtct ttactctgca accctgttcc agccgccgag
151 cacccgtgtg tcactcggga actggctggg vaagaggtc aatccagaca
201 cgcggggaag gagttccagg ggtcctgggc cagCTCCGCC CTCGCACCTG
251 CGGGCTCGGA TTCGGAGAAA AGTGCTAGAC TGGAGCTACA CGTATGCGTA
301 GCGGTCTGGA AAATGCCCCA GGCTCGGGTC TGAGGGGCCC AAGTCTATGC
351 ACCGCTGGTG TGACCCCGCA GGGCAACCCC GCGGTTAACT TCTCTCCTGC
401 CCACCCCTAG AGGTGTCTTC CTGGGAAGAC GATGGCAGGC GGTGCCCACC
M A G G A H R
451 GAGCCGACCG TGCAACAGGG GAAGAGAGGA AGGAGGGAGG TGGGAGGTGG
A D R A T C G E E R K E G G G W
501 CGCGCTCCCC ACAGCCCTTC CCCTCCTGGC CCTCGAGGGT GTCCGGTCCC
R A P H S P S P P G P R G C P V P
551 ACTCAAGGCA GCTGCGCAGA GBCTGTGCAG AAAACCCACC TGGGGCCGGT
L K A A A Q S L C R K P T W G R Y
HKa,. splice site
601 ATTGCACTCT GCTTCTCTTT CAGAGAAAGC TGGAAATTTA CTCCGTGAG
C T L L L F Q R K L I I Y S V E
B) 651 ... remainder of sequence omitted for clarity
...ATG, .ATG ATG... COMMON
Figure 4-8. Rabbit HKa, gene sequence at the 5' end. A) HKa2 gene sequence and
deduced amino acid sequences are shown. Intron sequence is in lower case type. The
amino acid sequence that is shared by HKaa and HKacx is in boldface type. B) Pattern of
alternative splicing at the 5' ends of the HKa2 transcripts.
is continuous with the core sequence and lies within the first intron of the HKo2a
pre-mRNA. This general intron/exon structure is the same as that reported for rat HKa2a
and HKoC2b (Kone and Higham, 1998)
Expression of HK-ATPase x Subunits in the Kidney
The start codon of HKa2a,, was omitted in the HKC(2c sequence. Instead, a probable
start codon was located upstream in frame with the HKo2,, open reading frame. Thus, the
deduced amino acid sequence encoded a protein 61 amino acids longer at the amino
terminal end than the rabbit HKc2,, sequence. The deduced amino acid sequence of the
HKot2za and HKc(X2 proteins are shown in Figure 4-8A, indicating translational start sites.
Although the HKx2c. cDNA indicated a continuous open reading frame including the
amino-terminal extension, the possibility remained that translation might be initiated at the
ATG codon homologous to that reported for the rat (Kone and Higham, 1998).
Therefore, to determine whether the upstream ATG served as a translational start site,
antipeptide antibodies were generated for peptides corresponding to amino acids 13-25
and 79-98 of the HKx2,c subunit. The former (antibody LLC27) was HKa:2c-specific, while
the latter (antibody LLC25) recognized a segment common to both HKoa2a, and HKax2,
subunits. Western analyses of rabbit kidney tissues and RCCT-28A cells using antibody
LLC25 (anti-HKa2, common) revealed a doublet migrating at an apparent molecular mass
of approximately 90KDa (figure 4-9A). Experiments using antibody LLC27 (anti-HKa(2)
indicated a single band with a migration comparable to the upper band of the doublet
(figure 4-9B). Both antibodies appear to recognize the same protein providing strong
OL2c N j ,
_, '- 66.2
Figure 4-9. Western analysis showing presence of HKa2, and HKa2c protein in renal
cortex. A) Detected with the anti-HKa2 common antibody LLC25. B) Detected with the
anti-HKac2-specific antibody LLC27. A and B show separate lanes from the same
SDS-PAGE gel. Membrane was cut after electrotransfer and re-aligned after
evidence that the HKcX2, subunit containing thile amino-terminal extension was indeed
present in both rabbit renal tissue and RCCT-28A cells.
To test the hypothesis that enzyme localization in the kidney or in the cell is changed
as a result of the alternative splicing of the mRNA, immunohistochemnistry experiments
were performed. Using our antibodies, Dr. Jilt Verlander and Ms. Robin Moudy
conducted immunohistochemistry experiments. Neither antibody we had made to the
anti-HKa2 common region (LLC24 or LLC25) worked well for immunohistochemistry.
However LLC22 (anti-HKc1) and LLC26 (anti-HKac2) gave satisfactory results. A
representative section photographed in Dr. Verlander's laboratory is shown in Figure
4-10. There is no visible reactivity except in the apical membranes of some cells of
connecting segment and collecting duct. A majority of the cells in which labelling is
observed bulge out into the lumen, a hallmark of the intercalated cell. Fewer cells are
labelled in connecting segment compared to more distally in the collecting duct. This is in
accordance with the normal distribution of acid-secreting intercalated cells, fewer in
connecting segment than farther down in cortical collecting duct. There appears to be
some labelling of cell types other than intercalated cells, including principal cells. Labelling
continues to be visible in more distal sections of collecting duct than one would expect
based on distribution of intercalated cells, but individual rabbits may vary widely in this
distribution. Therefore, the few observations made do not permit a certainty that HKot2c
subunit protein is expressed more distally in cell types not normally thought to be
associated with acid secretion. The lack of basolateral labelling implies that there is not a
cell polarity change in HKxc protein compared to HKaxa.
Figure 4-10. Immunohistochemistry by Dr. Jill Verlander and Ms. Robin Moudy.
Anti-HKcta, antibody reacts with the apical surfaces in some cells in rabbit renal
collecting duct. The section shown is from the outer stripe of outer medulla.
We have generated complete cDNAs for both the rabbit renal HKac2, and the novel
HKoX2c subunits. They were more closely related to other HKaX2 nucleotide sequences than
to HKoa, or to Na,K-ATPases. The exon structure at the 5' end was found and compared
to the exon structure found at the 5' end of rat (Kone and Higham, 1998). Proteins
corresponding to both HKac2 isoforms were detected by immunoblot analysis, indicating
that the novel HKct2, has the predicted amino-terminal extension.
The HKca2 cDNAs reported here generated using rabbit renal cortex RNA as template
have high homology to previously known HKa_ sequences from human skin axilla and the
rat distal colon (Grishin el al., 1994; Crowson and Shull, 1992). The level of homology is
relatively low compared to the homology typically found when comparing HKa, or NaKa
subunits across mammalian species. However, it is much higher than the homology
between the rat HKo., and HKoa: cDNAs (Crowson and Shull, 1992) or among the
different Na,K-ATPase isofotbrms within a species The HKax2 isoforms appear to have
undergone somewhat greater evolutionary divergence than HKoi or the Na,K-ATPase
catalytic isoforms. Phylogenetic analysis provides a clearer picture of the relationships
between these P-type ATPases In a phylogram (figure 4-7), the HKa;' subunit isoforms
cluster together compared to the other P-type ATPases. The HKoa2 subunit nucleotide
sequences all represent H,K-ATPases that are expressed at high level in colon, at lower
level in other tissues including the kidney, and not at all in stomach. At least in rat and
rabbit, an alternatively spliced variant has been shown to be transcriptionally competent,
and the pattern of introns and exons that give rise to this alternative splicing are strikingly
similar. It would be highly interesting to attempt to detect a similar transcript in human.
Based on tissue distribution, phylogenetic analysis, and the similarity of alternatively
spliced transcripts, it continues to be a valid premise that the HKc X2 subunits known for
rat, human, and rabbit form an orthologous group.
Like the rat, the rabbit appears to have alternatively spliced transcripts of HKo2 in the
kidney. The organization of the rat alternatively spliced cDNA (Kone and Higham, 1998)
omitted the exon containing the start codon of the sequence previously reported for rat
distal colon (Crowson and Shull, 1992), giving rise to a protein truncated by 108 amino
acids at the amino-terminal end. The rabbit HKa2, sequence also omits the start codon
present in the HKx2,, sequence. However, an upstream 5' ATG codon lies in the same
uninterrupted reading frame as the coding sequence for HKU2. Initiation of translation at
this position yields a subunit having an extended amino-terminus 61 amino acids longer
than the canonical HKo2, protein. Western analysis demonstrated that a protein having
this extension is present in rabbit kidney. An antibody designed to detect the common
core region of HKct, was reactive to two proteins very close in apparent mass, whereas an
antibody designed to detect the amino-terminal extension was reactive to a single species.
The extended portion of the HKox,_-encoded protein contains a casein kinase II
phosphorylation motif at thr-12, and a cAMP-dependent protein kinase phosphorylation
motif at thr-53. These sites impart a potential for regulation ofHKcx, distinct from HK2,a.
There are no apparent glycosylation sites or signal sequences. The HKac(, extension is
hydrophilic in nature and lacks any conspicuous membrane-spanning domains. Because the
N-terminus of HKct1 was found to be cytosolic (Smolka et al., 1992), the elongation can
be expected to have a cytosolic location. Figure 1-1 shows the putative location of the
amino-terminal extension. Chou-Fasman calculations predict this segment to be
predominantly alpha- helical with a turn structure in a region containing seven prolines
between amino acids 26 through 40. A similarly proline-rich hinge region is found in band
3 protein, and an ankyrin-binding site has been localized to that area of band 3 (Willardson
etal., 1989). Products of alternative splicing of the band 3 protein AEI gene have been
characterized in chicken kidney, where it is thought that the variation in transcripts serves
to determine the membrane domain to which the polypeptides are targeted (Cox et al.,
1995). A similar situation may exist for the various HKa_ transcripts; alternative splicing
may mediate the polarity of expression. In cortical thick ascending limb, an H,K-ATPase
activity has been described that diminished in rats fed a low K' diet (Younes-Ibrahim et
al., 1995). Basolateral polarity of expression would be consistent with potassium
homeostasis in that case. However, no basolateral staining using the anti-HKc(x2 antibody
was seen in this segment of the nephron (discussed below). If the HKc2b protein is the
molecule responsible for that activity, its absence in RCCT-28A could be explained by the
normal expression being in a region of the cortex other than CCD.
A number of possibilities exist to answer the question of why there are multiple
H,K-ATPases in the kidney. One possibility is differential regulation. In rat (Kone and
Higham, 1998) and rabbit (this work) the alternatively spliced HKa, subunit mRNA
contains multiple upstream open reading flames. This may be associated with an inhibition
of translation, leading to a decrease in HKc, subunit protein in the cell. Covalent
modification is another possible means of differential regulation. In this case,
phosphorylation is an unlikely candidate for differential regulation because although motifs
recognized by kinases are present, they are not conserved between rat (Kone and Higham,
1998) and rabbit. Having a different promoter region immediately 5' to the HKa2e cDNA
gives rise to potential differences in regulation relative to HKox2a. Of course, different
promoter regions giving different responses to conditions such as aldosterone status and
low K' may account for why there are the three catalytic subunit isoforms HKcc, HKa2a,
and HKa2c. There may also be differences in the kinetics between HKa2 ., and HKCa2c.
Another possibility is localization, both within the kidney and within the cell. Using
our antibodies, Dr. Jill Verlander and Ms. Robin Moudy conducted immnunohistochemistry
experiments. They found a similar pattern of expression of expression of HKa2, compared
to other H,K-ATPase proteins for which localization is known (Wingo el al., 1990;
Campbell-Thompson el al., 1995; Aihn and Kone, 1995; Ahn eli al., 1996; Haragsim and
Bastani, 1996). These results imply that the alternative splicing does not confer on the
protein a different polarity of expression or localization within the kidney. They do imply
that the HKca2c protein joins other acid and ion transporters in cell types specialized for
high transport activity.
Immunohistochemistry showed that differences in HKa2, subunit protein cellular
distribution had no major departures from the distribution of other H,K-ATPase subunits.
Because the phosphorylation motifs in rat and rabbit alternatively spliced isoforms are not
conserved, regulation by phosphorylation is a less likely candidate to lead to differential
regulation. There are myriad possibilities for why there is such a heterogeneity of
H,K-ATPases in the kidney, such as differential regulation at the gene level by different
promoter regions, differential regulation of protein synthesis by the short upstream open
reading frames, or differences in enzyme kinetics.
H,K-ATPASES IN A RABBIT KIDNEY CORTICAL COLLECTING TUBULE A-TYPE
INTERCALATED CELL LINE
The kidney is a well-organized and complicated organ, with many different cell types
contributing to its morphology and function. Any understanding of the kidney must
include an appreciation of the processes mediated by each cell type. Expression of
H,K-ATPase has been shown to be complex in rabbit kidney, so one issue of interest is the
identification of the cell types possessing the various isoforms of the pump.
Immunohistochemical evidence was presented in the previous chapter indicating that the
collecting duct A-type intercalated cell is one of the primary cell types that expresses
H,K-ATPase catalytic subunit proteins. An independent approach showing that A-type
intercalated cells express these pumps is to detect H,K-ATPase in a well-characterized cell
line of intercalated cell origin. A cell line in which H,K-ATPase was present would also
serve as a model cell system offering the potential for future experiments pertaining to
The rabbit cortical collecting tubule cell line RCCT-28A was selected by
immunodissection (Arend el il., 1989). Cortical collecting duct was dissected under a
microscope and the collagenase-dispersed cells incubated in culture dishes to which the
antibody rct-30 was attached (Spielman ei ul., 1986). The rct-30 antibody binds
specifically to collecting duct cells. The resultant cell line was shown to have
characteristics of the cortical collecting duct A-type intercalated cell (Arend el al.,
1989;Schwiebert el al., 1992; Dietl e al., 1989; Bello-Reuss, 1993) The existence of
H,K-ATPase mRNA, protein, and activity in this cell line was studied to demonstrate that
this is a cell type that expresses H,K-ATPase. Here we show by RT-PCR that mRNA for
HKai, HKoa2,, HKac2,, and HK3 are present in RCCT-28A cells. By western analysis it is
shown that protein corresponding to the novel alternatively spliced isoform HKcx2c is
present. Lastly, fluorescence microscopy measurements are used to demonstrate that
RCCT-28A cells have a mechanism for pH, regulation that is K'-dependent and sensitive
to the HK-ATPase inhibitor Sch-28080. Together the data represent very strong evidence
for H,K-ATPase expression in a cell line derived from the type-A intercalated cell in the
Detection of H,K-ATPase in RCCT-28A Cells
Detection ofH.K-ATPase mRNAs in RCCT-28A cells
To determine the presence of H,K-ATPase transcripts, we examined whether the
RCCT-28A cells possessed mRNA for the known H,K-ATPase 3 or a subunits. In the
initial experiments, there were no H,K-ATPase subunits detectable by northern analysis,
nor were there RT-PCR products amplified from RCCT-28A cells that could be directly
visualized by agarose gel electrophoresis. A more sensitive technique, Southern blotting of
the RT-PCR products using cDNA probes designed to hybridize to the expected products
was employed to demonstrate the presence of H,K-ATPase subunit mRNA in these cells.
RT-PCR products were amplified using RCCT-28A cell total RNA as a template. The
presence ofHK3 subunit mRNA was observed as a product of approximately 309 bp
hybridizing to a probe containing nucleotides 304-873 of the rabbit HKO3 sequence
obtained by Reuben el cal. (1990) in gastric tissues (Figure 5-1 A). Sizes were estimated by
measurements from the original agarose gel on which a 100 bp ladder was visualized by
ethidium bromide staining. The "-RT" lanes in this and all subsequent figures depict
negative controls in which RT was omitted from reactions to show that the RNA template
was free of contaminants. Products amplified from mRNA isolated from rabbit renal
cortex tissue were also included in these experiments as a positive control for the RT-PCR
reaction. The hybridization of a probe specific for HKf3 to an RT-PCR product of the
expected size implied the presence of HK3 in these cells.
At the time these experiments were carried out, the existence of alternatively spliced
HKcX2 variants had not yet been discovered Therefore, these experiments were designed
only to show the presence of HKc_2 transcript in general, and do not differentiate between
the HKcx2a and HKcx2c species. For HKca2 mRNA, a primer pair yielded the anticipated
product size of 305 bp. This product was hybridized with a probe of nucleotides
1264-1569 of the HK_;., sequence, contained in the region now known to be common to
HKa2a and HKa2, (Figure 5- 1 B).
A similar strategy was employed to show that HK.Oj was present in the RCCT-28A
cells. Primers were designed to amplify a 611 bp region ofHKca mRNA (nucleotides
2537-3147), and again a product of the expected size was observed to hybridize to a
probe containing the same 611 bp region (Figure 5- IC). At the time these experiments
600 bp -
300 bp -
100 bp -
600 bp -
300 bp -
100 bp -
600 bp -
300 bp -
100 bp -
800 bp -
400 bp -
200 bp -
Figure 5-1. Southern blots ofH,K-ATPase subunit mRNA in RCCT-28A cells. A)
Products amplified by RT-PCR were shown by Southern blot analysis. PCR primers
were designed to amplify a 309 bp region within the 3' UTR ofH,K-ATPase P subunit.
B) PCR primers designed to amplify a 305 bp region within the coding sequence of
H,K-ATPase a2 subunit produced the products shown. C) PCR primers designed to
amplify a 611 bp region within the 3' UTR of H,K-ATPase ao subunit produced the
products shown. D) Restriction digests of the 611 bp fragment ofH,K-ATPase a,.
Barn HI digestion was predicted to yield fragments of 423 and 187 bp. Pst I digestion
was predicted to yield fragments of 353 and 257 bp.
were ongoing, there were no reports in the literature of HKca in kidney, so a further step
was taken to confirm the identity of these products. Restriction digests were carried out
on these products, and the anticipated size fragments were obtained (Figure 5-ID), further
evidence that these products corresponded to HKa, mRNA in RCCT-28A cells.
With further optimization of RT-PCR protocols, including such parameters as
annealing temperatures and number of amplification cycles, it was found that the RT-PCR
technique was capable of generating products visible by ethidium bromide staining of
agarose gels. This was a substantial improvement over the preceding experiments, because
it offered the opportunity to obtain nucleotide sequences of the PCR products to confirm
their identity. RT-PCR products were generated and sequenced using RCCT-28A cell
total RNA as a template. The presence of HKP subunit mRNA was observed by
amplification of a 570 bp cDNA corresponding to nucleotides 304-873 of the sequence
obtained by Reuben el al. (1990) in gastric tissues (Figure 5-2A). Nucleotide sequencing
of the RCCT-28A product confirmed that the amplified product was identical with the
rabbit HK3 subunit mRNA.
HKai was also shown to be present in the RCCT-28A cells by the same technique.
Primers were designed to amplify a 611 bp region of HKci mRNA (nucleotides
2537-3147), and again a product of the expected size was observed (Figure 5-2B). The
nucleotide sequence was identical to that reported by Bamberg et al. (1992) except for
two single base mismatches. One mismatch was a transition of G-,A at nucleotide 2567
and the other a G-*C transversion at nucleotide 3089; neither affected the deduced amino
S + 1 +
Figure 5-2. HK3 and HKa, subunit mRNA in RCCT-28A cells. PCR products were
generated using RCCT-28A cell and rabbit renal cortex total RNA. Amplification was
40 cycles for RCCT-28A cell PCR products and 30 cycles for renal cortex PCR
products. 100 bp ladder is shown for size reference.