|
| |
|
|
|
|
||||
| Home Help Feedback Subscriptions Archive Search Table of Contents | ||||
First published online January 18, 2008
Journal of Experimental Biology 211, 409-422 (2008)
Published by The Company of Biologists 2008
doi: 10.1242/jeb.011213
Gap junctions in Malpighian tubules of Aedes aegypti
1 Department of Biomedical Sciences, Cornell University, Ithaca, NY 14853,
USA
2 National Heart, Lung, and Blood Institute, National Institutes of Health,
Bethesda, MD 20892, USA
3 Department of Biological and Environmental Engineering, Cornell University,
Ithaca, NY 14853, USA
* Author for correspondence (e-mail: kwb1{at}cornell.edu)
Accepted 19 November 2007
 |
Summary |
|---|
 TOP Summary INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
between two
cells. This resistance would stem from 6190 open gap-junctional channels,
assuming the high single gap-junction conductance of 375 pS found in
vertebrate tissues. The addition of the calcium ionophore A23187 (2 µmol
l–1) to the peritubular Ringer bath containing 1.7 mmol
l–1 Ca2+ did not affect the gap-junction
resistance, but metabolic inhibition of the tubule with dinitrophenol (0.5
mmol l–1) increased the gap-junction resistance 66-fold,
suggesting the regulation of gap junctions by ATP. Lucifer Yellow injected
into a principal cell did not appear in neighboring principal cells. Thus, gap
junctions allow the passage of current but not Lucifer Yellow. Using RT-PCR we
found evidence for the expression of innexins 1, 2, 3 and 7 (named after their
homologues in Drosophila) in Malpighian tubules. The physiological
demonstration of gap junctions and the molecular evidence for innexin in
Malpighian tubules of Aedes aegypti call for the double cable model
of the tubule, which will improve the measurement and the interpretation of
electrophysiological data collected from Malpighian tubules.
Key words: Malpighian tubule, yellow fever mosquito, electrical coupling, gap-junction resistance, innexin, circuit analysis, cable analysis, ATP
|
INTRODUCTION |
|---|
|
TOPSummaryINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
;
Yu and Beyenbach, 2004).
Nevertheless, principal cells make up more than 90% of the tubule mass owing
to their large size (Beyenbach,
2001; Wu and Beyenbach,
2003; Yu and Beyenbach,
2004). Principal cells mediate transepithelial secretion of
K+ and Na+ from the hemolymph to the tubule lumen by
active transport (Beyenbach,
1995; Beyenbach,
2001; Beyenbach,
2003b). Current–voltage plots of a single principal cell in
the intact Malpighian tubule indicate rather low cell input resistances which,
compared to the transepithelial resistance, suggest the electrical coupling of
5–6 principal cells (Masia et al.,
2000). Moreover, voltage recordings from two or more adjacent
principal cells of the same tubule are strikingly similar in amplitude and
frequency under control and experimental conditions, which is consistent with
electrical coupling.
There are two pathways for the electrical coupling of principal cells. One
pathway leads through gap junctions; the other passes through the apical
membrane of one cell into the tubule lumen and then across the apical membrane
into the neighboring cell. Thus, gap junctions couple neighboring cells
directly, whereas the tubule lumen couples them indirectly. Although insects
do have gap junctions (Phelan,
2005; Phelan et al.,
1998), their presence in Malpighian tubules has not been
established to this date.
In the present study we ascertain the functional presence of gap junctions
between principal cells of Malpighian tubules of the yellow fever mosquito.
The gap junctions permit the passage of current from one cell to the next, but
they prohibit the passage of the dye Lucifer Yellow. Metabolic inhibition of
the tubule, which is known to reduce intracellular ATP concentrations
(Wu and Beyenbach, 2003) and
to halt transepithelial electrolyte and fluid secretion
(Beyenbach and Masia, 2002;
Beyenbach et al., 2000b;
Pannabecker et al., 1992),
increases the gap-junction resistance 66-fold, consistent with gap-junction
gating by ATP. Furthermore, PCR studies on cDNA derived from Aedes
Malpighian tubules reveal the expression of four innexin-like transcripts.
|
MATERIALS AND METHODS |
|---|
|
TOPSummaryINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
) except for feeding larvae with TetraMinTM Tropical
Flakes ground manually with mortar and piston. On the day of the experiment a
female mosquito (3–7 days post-eclosion) was cold-anesthetized and
decapitated. A Malpighian tubule was removed under Ringer solution from its
attachment to the gut and transferred to a LuciteTM perfusion chamber
containing 1.0 ml Ringer solution. The bottom of the chamber was covered with
thinly stretched Parafilm`M' (American National Can, Greenwich, CT, USA), to
which Malpighian tubules adhere and stabilize for the impalement of principal
cells with microelectrodes. The tubules were viewed from above at x50
magnification using a stereomicroscope (Wild, Heerbrugg, Switzerland).
Ringer solution and drugs
Ringer solution contained the following in mmol l–1: 150.0
NaCl, 25.0 Hepes, 3.4 KCl, 1.8 NaHCO3, 1.0 MgCl2, 1.7
CaCl2 and 5.0 glucose. The pH was adjusted to 7.1 with NaOH. The
osmolality of the Ringer solution was approximately 320 mosmol
kg–1 H2O. Synthetic leucokinin-VIII was a gift
from Ron Nachman (USDA, Texas A&M University).
Electrophysiological studies
As shown in Fig. 1, three
adjacent principal cells of a Malpighian tubule were selected for impalement
with conventional microelectrodes. Principal cell 1 typically was 5–10
cells away from the blind (distal) end of the tubule, and principal cells 2
and 3 were downstream towards the open-end of the tubule. Principal cell 1 was
impaled with both current and voltage microelectrodes. The latter measured the
basolateral membrane voltage (Vbl1) as well as
Vbl1 when cell 1 was voltage clamped to a
hyperpolarizing voltage of 40 mV for 50 ms. Principal cells 2 and 3 were each
impaled with a voltage microelectrode for the measurement of
Vbl2 and Vbl3, respectively, and their
Vbl when cell 1 was voltage clamped.
|
Microelectrodes (Omega dot borosilicate glass capillaries, 30-30-1;
Frederick Haer & Co., St Bowdoinham, ME, USA) were pulled on a
programmable puller (Model P-97; Sutter Instruments, Novato, CA, USA) to yield
resistances of 20–30 M when filled with 3 mol
l–1 KCl. The microelectrodes were bridged to the measuring
hardware using Ag/AgCl junctions that were prepared by first degreasing the
silver wire with alcohol, and then by Cl-plating it in 0.1 mol
l–1 HCl for 20 min at a current of 50 µA. The bath was
grounded with the Ag/AgCl junction lodged in a 4% agar bridge of Ringer
solution.
The electronic hardware consisted of (1) the Gene Clamp model 500B voltage
and patch clamp amplifier, (2) head stage HS-2A gain 10MGU for current
injection, and (3) head stage HS-2A gain 1LU for voltage recording (all from
Axon Instruments, Sunnyvale, CA, USA). Voltage deflections in principal cells
2 and 3 were recorded using custom-made high-impedance amplifiers (Burr-Brown,
1011 ). We used Clampfit (pClamp 9) for data analysis (Axon
Instruments).
All current and voltage data were digitized with the aid of a computer and the A/D converter DigiData1332x (Axon Instruments). Current and voltage data from principal cell 1 were also displayed on an oscilloscope (Iwatsu, Tokyo, Japan) and on a strip chart recorder (model BD 41; Kipp and Zonen, Crown Graphic, Totnes, Devon, UK).
Circuit analysis
The transepithelial secretion of electrolytes in Malpighian tubules of
Aedes aegypti can be modeled using an electrical circuit consisting
of two major transepithelial transport pathways; one is active, and the other
is passive (Fig. 2).
Na+ and K+ must take the active transport pathway
through principal cells that provide the energy for moving the two cations
against their electrochemical potentials into the tubule lumen. The active
transport pathway consists of the electromotive forces (E) and the
membrane resistances (R) at apical (a) and basolateral (bl)
membranes. The passive pathway is located outside principal cells and is
represented by the single shunt resistance Rsh
(Fig. 2B).
|
; Beyenbach and Masia,
2002; Pannabecker et al.,
1992; Pannabecker et al.,
1993; Scott et al.,
2004; Wu and Beyenbach,
2003; Yu and Beyenbach,
2002). In the present study, we model the Malpighian tubule as a
cable with two parallel axial resistances, the resistance of the tubule lumen
(Rlu) and the gap-junction resistance
(Rgj) as shown in Fig.
3B.
|
cm2 to 9.9
cm2, thereby decreasing the transepithelial voltage
(Vt) from 59 mV to 6 mV
(Pannabecker et al., 1993). If
we assume that both Rsh and Vt
collapse to zero in the presence of leucokinin-VIII, then the `short-circuit
assumption' eliminates Rsh from the circuit, thereby
placing the tubule lumen at the same ground potential as the peritubular bath.
The tubule lumen is thus eliminated as a current path to adjacent principal
cells, leaving the gap-junction resistance (Rgj) as the
only axial resistance. The value of Rgj is determined as
follows.
We begin with the tubule modeled as the double cable in Fig. 4A, where gap junctions and the tubule lumen present two parallel axial resistances along the length of the tubule. The transepithelial short circuit induced by leucokinin eliminates the paracellular shunt resistance (Rsh) and places the resistances of the apical membrane (Ra) and the tubule lumen (Rlu) in parallel to the resistance of the basolateral membrane Rbl (Fig. 4B). Representing these two parallel resistances as the single non-junctional resistance (Rnj) yields a circuit consisting of Rnj and the gap-junction resistance (Rgj) alone. The two resistances Rnj and Rgj are determined as follows.
|
As shown in Fig. 4C, the
current injected into principal cell 1 (to voltage-clamp
Vbl1) can take three routes. One route is through
Rnj1 to ground; the second route is from cell 1 to the
upstream principal cell Rgj1'; and the third route
is downstream to principal cells through Rgj1. If the
characteristics of all the cells are assumed to be the same, then the
gap-junction current passing into upstream and downstream cells will be the
same (Fig. 4C), and the current
injected into principal cell 1 (Iinject) is the sum of
three currents (Eqn 1):
 | (1) |
 | (2) |
 | (3) |
 | (4) |
 | (5) |
 | (6) |
 | (7) |
 | (8) |
 | (9) |
 | (10) |
 | (11) |
 | (12) |
Vbl) in cells 1, 2 and 3
(Fig. 4).
The input resistance
Voltage-clamping cell 1 yields the input resistance
(Rinput) as the ratio of
Vbl1/Iinject (Figs
1,
4). The input resistance
includes cell 1 and all other cells coupled to it, as illustrated in
Fig. 4B. The input resistance
can also be predicted from the values of Rgj and
Rnj determined in the circuit analysis above
(Eqn 11,
Eqn 12). Good agreement between
values of Rinput measured directly and values predicted
from the circuit analysis would validate the short-circuit assumption needed
for reducing the double cable in Fig.
4A to a manageable level for circuit analysis. To predict values
of Rinput from values of Rgj and
Rnj in circuit Fig.
4C, the experimental preparation illustrated in
Fig. 4A is redrawn in
Fig. 5 from the perspective of
current injected into cell 1 flowing symmetrically into cells upstream and
downstream from cell 1. The parallel symmetry of all resistances in
Fig. 5 allows the reduction of
all resistances to a single resistance, which is the predicted input
resistance. We have limited the number of coupled cells to 7 in this data
reduction because previous studies suggested that 5–6 principal cells
are coupled electrically (Masia et al.,
2000).
|
V) decay exponentially from cell to cell along the length of
the tubule according to Eqn 13
where x is the distance from cell 1 and 
is the cell length
constant:
 | (13) |
(Vx at 37% of Vbl1) is:
 | (14) |
 | (15) |
 | (16) |
Iinject is the current injected into cell 1
(Fig. 4B,
Fig. 5), but one half of the
injected current passes upstream and the other downstream the tubule.
Since rr and ra are normalized to
tubule length, the gap-junction resistance Rgj and the
non-junctional resistance Rnj can be estimated. Since two
sets of gap junctions are expected to couple cell 1 to cell 3,
Rgj is:
 | (17) |
 | (18) |
Lucifer Yellow injections of principal cells
To visualize the coupling of principal cells via gap junctions,
one principal cell was injected with Lucifer Yellow to observe whether the dye
appears in neighboring principal cells. Microinjecting pipettes were made from
borosilicate glass (TW100F-4, WPI, Sarasota, FL, USA) using a horizontal
puller (P87, Sutter Instrument Co., Novato, CA, USA) and a two-step pulling
protocol. Injection pipettes had an average resistance of 4.33±0.37
M (N=9) when filled with 3 mol l–1 KCl. For
injections of Lucifer Yellow, the pipette was back-filled by capillary action
with a 2.5 mmol l–1 Lucifer Yellow CH dilithium salt (Sigma,
St Louis, MO, USA) dissolved in water.
After isolation, the Malpighian tubule was transferred to a perfusion bath pre-coated with 0.125 mg ml–1 poly-L-lysine (Sigma), which prevents the movement of tubules when a single principal cell is injected with dye. After impaling a principal cell, a hydrostatic pressure of approximately 5200 mmHg was applied to the pipette for 250 ms with the aid of a pneumatic picopump (PV830, WPI). The pressure pulse injected a volume of approximately 0.7 pl, as determined in pre-experiment pipette calibrations. The cytoplasmic volume of a principal cell is about 200 pl. The intracellular Lucifer Yellow was immediately visible when viewed with an inverted microscope (Diaphot, Nikon, Kawasaki, Japan) equipped with a B-3A filter (Chroma Technology, Brattleboro, VT, USA) and a mercury lamp light source (Chiu Technical Corp., Kings Park, NY, USA). Typically, we observed the tubule for more than 1 h for signs of Lucifer Yellow diffusing from the injected principal cells to the adjacent principal cells or stellate cells. To prevent photo bleaching of the dye, we turned on the light source briefly every 10 min.
Computational simulation of the electrical properties of a Malpighian tubule
After finding the evidence for gap junctions in Aedes Malpighian
tubules and determining the gap-junction resistance, it was clear that the
tubule should be modeled as a double cable. In previous studies we have
treated the tubule as a single cable in order to determine electromotive
forces (E) and the resistances (R) of transcellular and
paracellular transport pathways shown in
Fig. 2B
(Beyenbach and Masia, 2002;
Pannabecker et al., 1992;
Pannabecker et al., 1993). It
was therefore important to determine values of E and R for
the double cable model of the tubule. In brief, we modeled the tubule as a
linear series of ten principal cells where each cell was represented by the
electrical circuit shown in Fig.
3B. Analysis of this tubule model with the software of Electronics
Workbench® 5.12 (National Instruments, Austin, TX, USA) allows current to
be injected at any position in the circuit to observe voltage deflections
across any two points of the circuit. By simulating our previous in
vitro microperfusion experiments, where a known current was injected into
the tubule lumen at one end of the tubule
(Pannabecker et al., 1992), we
fitted data collected from the single cable to the double cable, to obtain new
values of E and R (Fig.
3).
Generation of Malpighian tubule cDNA
To prepare Malpighian tubule cDNA, 175 Malpighian tubules were isolated
from 35 female mosquitoes. The tubules were immersed in ice-cold Trizol
reagent (Invitrogen, Carlsbad, CA, USA) and then stored at –80°C.
Total RNA from the tubules was isolated by homogenization in Trizol reagent
(Invitrogen), followed by a phenol:chloroform phase separation and an
isopropyl alcohol precipitation
(Chomczynski and Sacchi,
1987). The resulting RNA was used as a template to synthesize a
pool of single-stranded cDNA using Superscript III Reverse Transcriptase
(Invitrogen) and a GeneRacer oligo dT primer (Invitrogen). Before use in PCR,
the cDNA was diluted tenfold with nuclease-free H2O (Integrated DNA
Technologies, Coralville, IA, USA).
PCR of innexin-like transcripts
The genome of Drosophila lists eight genes encoding innexins
(Stebbings et al., 2002).
Using the predicted amino-acid sequences encoded by these genes, we searched
the genomic database of Aedes
(http://aaegypti.vectorbase.org)
with a Basic Local Alignment Search Tool (BLAST). The search yielded six
innexin-like genes in Aedes (see Results). We designed
oligonucleotide-primer pairs (Table
1) to amplify fragments of the open-reading frames (ORF) for each
Aedes innexin using PCR. All of the PCRs were conducted on 0.5 µl
of Malpighian tubule cDNA in Platinum PCR Supermix HF (Invitrogen) using the
following cycling parameters: one cycle of 94°C for 2 min; 35 cycles of
94°C for 30 s, 50°C for 30 s, and 68°C for 1 min; one cycle of
68°C for 10 min. As a negative control, each PCR was conducted on 0.5
µl of Malpighian tubule RNA that was not reverse transcribed, but was
diluted to the same degree as the cDNA. To verify PCR results, we also
performed all PCRs on 3 µl of a double-stranded cDNA library derived from
adult Aedes Malpighian tubules (generously provided by Dr Dimitri
Boudko, The Whitney Laboratory, St Augustine, FL, USA). All PCR products were
separated via electrophoresis on a 1%-agarose gel and stained with
ethidium bromide.
|
DNA sequencing of PCR products
PCR products were either ligated into a pCR 4-TOPO vector (Invitrogen) or
purified using a QIAquick PCR purification kit (Qiagen, Valencia, CA, USA).
The ligated PCR products were transformed into TOP 10 Chemically Competent
E. coli (Invitrogen) according to the manufacturer's protocol.
Plasmid DNA from four resulting colonies was purified using a QIAprep Spin
Miniprep kit (Qiagen). DNA sequencing of plasmid DNA and purified PCR products
was performed in both the 5' and 3' directions by the Cornell DNA
Sequencing Center (Cornell University, Ithaca, NY, USA).
Statistical evaluation of data
Each tubule/cell was used as its own control so that the data could be
analyzed for the difference between paired samples, control vs
experimental (paired Student's t-test).
|
RESULTS |
|---|
|
TOPSummaryINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
; Sawyer and Beyenbach,
1985; Williams and Beyenbach,
1984; Yu and Beyenbach,
2002). The oscillations stem from spontaneous changes in the
paracellular shunt Cl– conductance
(Beyenbach et al., 2000a) and
have a frequency not unlike those of spontaneous cyclical changes in
cytoplasmic Ca2+ concentration
(Fig. 6). Leucokinin-VIII
increases the paracellular Cl– conductance via the
activation of a Ca2+ channel in the basolateral membrane of
principal cells, thereby increasing the cytoplasmic Ca2+
concentration (Yu and Beyenbach,
2002). Elevated cytoplasmic Ca2+ levels obliterate the
spontaneous changes in cytoplasmic [Ca2+], eliminate spontaneous
voltage oscillations, and maintain an increased paracellular
Cl– conductance as long as leucokinin-VIII is present
(Beyenbach, 2003a;
Yu and Beyenbach, 2001).
|
Estimate of the gap-junction resistance by circuit analysis
Principal cells 1, 2 and 3 were impaled with current and voltage electrodes
as shown in Figs 1 and
4. Principal cell 1 was then
voltage-clamped at a hyperpolarizing voltage of 40 mV for 50 ms and the
voltage deflections were recorded in cells 2 and 3
(Fig. 7). After obtaining
control data, the tubule was treated with 1 µmol l–1
leucokinin-VIII. In the presence of this diuretic peptide, the voltage
deflections in principal cells 2 and 3 were less than those under control
conditions, in view of (1) the increase in the Cl–
conductance of the paracellular shunt pathway
(Yu and Beyenbach, 2001), and
(2) the increase in the Ca2+ conductance of the basolateral
membrane of principal cells (Yu and
Beyenbach, 2002).
|
Vbl in principal cell 2
(Vbl2) hyperpolarized by 20.7 mV, and the
Vbl in principal cell 3
(Vbl3) hyperpolarized by 11.9 mV
(Table 2).
|
In the presence of leucokinin-VIII, Vbl hyperpolarized
significantly from –82.2 mV to –100.4 mV
(Table 2). The
hyperpolarization is due to the increased coupling of Vbl
to the apical membrane voltage as the shunt resistance drops to 12% of control
values, nearly short-circuiting the transepithelial voltage
(Pannabecker et al., 1993). In
the presence of leucokinin-VIII, cells 1, 2 and 3 are electrically coupled by
primarily gap junctions, which allows measurement of the gap-junction
resistance (Rgj, Table
2) and the non-junctional resistance (Rnj,
Table 2). Importantly,
Rgj was derived from the circuit analysis of voltage
deflections described in Eqn 1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12.
Measured and predicted input resistances
The input resistance of principal cell 1 (Rinput) can
be measured directly as the ratio of
Vbl1/Iinject.
Rinput was 246.5±17.8 k under control
conditions and fell significantly (P<0.05) to 194.8±14.3
k in the presence of leucokinin-VIII in the 14 experiments summarized
in Table 2. The decrease
reflects in part the increased transcellular secretion of Na+ and
K+ in the presence of leucokinin-VIII
(Beyenbach, 2003a;
Hayes et al., 1989;
Pannabecker et al., 1993), and
also the activation of Ca2+ channels in the basolateral membrane
(Yu and Beyenbach, 2002).
The input resistance Rinput can be predicted from
values of Rgj and Rnj determined by
circuit analysis (Eqn 11,
Eqn 12;
Fig. 5). The predicted
Rinput is 209.6±15.4 k, which comes close to
the measured Rinput of 194.8 k in the same tubules.
Good agreement between the Rinput measured directly and
the Rinput predicted from the circuit analysis validates
the `short-circuit assumption' needed to obtain measures of
Rgj.
Cable analysis of Rgj
Keeping track of the distances that separated the microelectrodes impaling
cells 1, 2 and 3 (see Fig. 7)
allowed us to re-evaluate Rgj, but by cable analysis. A
two-parameter exponential decay curve was fitted to the mean voltage
deflections of Fig. 7. The
equation of this curve for tubules treated with leucokinin was:
 | (19) |
(regression correlation coefficient of 0.99). Using
Eqn 15 and
Eqn 16 yields 26.7 M
cm–1 as ra and 6.5 kcm as
rr, both normalized to tubule length in the presence of
leucokinin-VIII (Table 3). Use
of Eqn 17 and
Eqn 18 yields 327.3 k as
Rgj and 524.9 k as Rnj in the
same 14 tubule experiments shown in Fig.
7 and Table 2.
|
Coupling of principal cells tested with Lucifer Yellow
One principal cell of the Malpighian tubule was microinjected with Lucifer
Yellow at time zero (Fig. 8).
At the time of injection, the fluorescence of Lucifer Yellow was confined to
the injected principal cell. About 5 s after the injection, the fluorescence
of Lucifer Yellow started to appear in the tubule lumen and 1 min later the
dye exited the tubule lumen at its open end, indicating the secretion of
Lucifer Yellow across the apical membrane of the injected principal cell.
|
Physiological regulation of gap junctions
Since it is known that intracellular calcium and metabolic inhibition can
close gap junctions, we tested the effects of the calcium ionophore A23187 and
dinitrophenol on the gap-junction resistance (Rgj) and the
non-junctional resistance (Rnj). In these experiments we
first treated the isolated Malpighian tubule with 1 µmol
l–1 leucokinin-VIII to measure Rgj and
Rnj by the circuit analysis shown in Figs
1,
4 and
Table 2. After 5 min in the
presence of leucokinin-VIII we added 2 µmol l–1 A23187 to
the peritubular medium and took measurements of Rgj and
Rnj 5 min later. We then added 0.5 mmol
l–1 dinitrophenol (DNP) to the peritubular Ringer bath (in
the presence of leucokinin-VIII and A23187) and measured
Rgj and Rnj 2 min later.
Table 4 summarizes the
results.
|
The addition of leucokinin-VIII to the peritubular bath significantly
(P<0.05) hyperpolarized Vbl1 from –83.5
to –103.3 mV (Table 4).
The circuit analysis of voltage deflections in neighboring cells yielded a
mean Rgj of 468.0 k and a mean
Rnj of 578.5 k in this series of experiments
(Table 4). After 5 min in the
presence of A23187, neither Vbl1 nor
Rgj changed significantly. In contrast,
Rnj decreased significantly which is further reflected by
the significant decrease in Rinput
(Table 4).
Upon the addition of 0.5 mmol l–1 DNP to the peritubular
medium, Vbl1 significantly depolarized from –99 mV
to –13 mV while Rgj significantly increased from 526
k to 30 762 k (Table
4). In the presence of DNP, Rnj significantly
increased from 504 k to 1483 k, and Rinput
significantly increased from 194 k to 1474 k
(Table 4). In summary,
significant effects of Ca2+ ionophore A23187 were limited to a
small reduction in the non-junctional resistance that may reflect the
pore-forming action of the ionophore at the basolateral membrane. The effect
of metabolic inhibition was profound on every measured variable. It increased
Rnj threefold, it dropped Vbl
sevenfold, increased Rinput eightfold, and increased
Rgj 60-fold (Table
4).
Identification of putative innexins in Aedes
A BLAST-search for homologues of the known eight Drosophila
innexins in the genome of Aedes identified six innexin-like genes
(Table 5). We refer to the
innexins in Aedes after their most similar homologues in
Drosophila. Among the Aedes innexins, AeInx1,
AePassover, AeInx2, and AeInx3 shared over 50%
amino-acid identity to their Drosophila homologues, whereas
AeInx4 and AeInx7 were less than 50% identical to their
Drosophila homologues (Table
5). Homologues of Inx5 and Inx6 from Drosophila were not
apparent from BLAST-searches in the Aedes genome.
|
To better resolve the relationships between Aedes and Drosophila innexins, we constructed a neighbor-joining phylogenetic tree (Fig. 9) using the predicted amino-acid sequences encoded by the genes. For Inx1, Passover, Inx2 and Inx3, the tree demonstrates that: (1) the Aedes genes cluster into discrete branches with their respective Drosophila homologues and (2) the branch length from a node (e.g. arrow in Fig. 9) to each homologue is short (see legend of Fig. 9 for explanation). For Inx7, the tree demonstrates that the Aedes gene clusters with its Drosophila homologue, but that the length between the node of the branch and the two homologues is long. For Inx4, the tree demonstrates that the Aedes gene does not cluster discretely with its Drosophila homologue, but it appears within a larger branch that includes Inx4, Inx5 and Inx6 of Drosophila. Among these three Drosophila innexins, the cumulative branch length between AeInx4 and DrInx4 is the shortest. The insights gained in the phylogenetic tree are consistent with the results of the BLAST search (Table 5). The implications of these results are analyzed further in the Discussion.
|
Expression of innexin transcripts in Aedes Malpighian tubules
In Aedes Malpighian tubule cDNA we detected partial PCR products
near the expected sizes for AeInx1, AeInx2, AeInx3
and AeInx7 cDNAs (Fig.
10, lanes a). In contrast, we did not detect a PCR product for
AePassover, and the PCR product for AeInx4 was extremely
weak (Fig. 10, lanes a). DNA
sequencing verified that (1) the identities of the amplified transcripts were
as expected and (2) the PCR products from primers that bracketed introns were
not derived from genomic DNA contamination. When total RNA was used as a
template for PCR in lieu of cDNA, no products were amplified for any
of the innexins (Fig. 10,
lanes b). Both the positive and negative results for each gene were verified
using (1) a combination of at least two additional, independent primer pairs
(data not shown) and (2) cDNA from adult Aedes Malpighian tubules
generated by the Boudko laboratory (data not shown). The above results
indicate that Aedes Malpighian tubules primarily express mRNA
transcripts for four innexins.
|
|
DISCUSSION |
|---|
|
TOPSummaryINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
(Table 2), and the cable
analysis yielded a gap-junction resistance of 327 k
(Table 3). Although the two
measurements are significantly different, they both document electrical
coupling between principal cells.
A short-coming of the cable analysis is the reliance on the length
constant, which is very sensitive to curve-fitting the data. Accordingly, we
consider the circuit analysis to deliver the more accurate measurement of the
gap-junction resistance, because this analysis yields values of both
gap-junctional and non-junctional resistances from which the input resistance
can be predicted, i.e. 210 k. The input resistance measured directly as
the ratio of Vbl1/Iinject, is
195 k (Table 2). The
good agreement between the predicted and measured values confirms the validity
of the circuit analysis.
Critical assumptions
Our measurement of the gap-junction resistance by circuit analysis depends
on two assumptions: the current distribution assumption and the short-circuit
assumption.
The current distribution assumption states that the current injected into
principal cell 1 for voltage clamping splits equally in both directions along
the tubule (Figs 4,
5). Indeed, in the course of
our study we have found that the decay of Vbl from
cell to cell was quantitatively similar in both directions (data not shown).
Furthermore, the good agreement between the measured and predicted input
resistances supports the symmetrical distribution of injected current, since
the predicted value depends on the symmetrical circuit shown in
Fig. 5, where no more than
seven principal cells are needed for predicting the input resistance. In our
experiments we have at least ten cells with at least five cells on either side
of principal cell 1.
The short-circuit assumption states that the Malpighian tubule is analyzed
at true short circuit. Specifically, we assume that leucokinin completely
reduces the paracellular shunt resistance to zero, thereby short-circuiting
the epithelium and leaving gap junctions as the major if not exclusive pathway
for significant electrical coupling between cells. However, in reality,
leucokinin reduces the paracellular shunt resistance nearly sixfold and causes
the transepithelial voltage to drop from 59 mV to 6 mV
(Pannabecker et al., 1993),
which falls short of ideal short-circuit conditions. Thus it is relevant to
evaluate the effect of the paracellular shunt resistance on the gap-junctional
resistance using circuit analysis (Table
6).
|
The analysis shows that as the shunt resistance changes 500-fold, the gap-junction resistance changes by only 2% (Table 6). Thus, the effect of the shunt resistance on the gap-junction resistance is negligible, and our assumption, that leucokinin completely short circuits the epithelium, introduces negligible error to our gap-junction resistance measurements.
Revised electrical circuit model of the Aedes Malpighian tubule
In previous studies we have modeled the Malpighian tubule as a single cable
that consists of a single axial resistance, i.e. the core resistance
(Rco) and a radial resistance, the transepithelial
resistance (Rt) as shown in
Fig. 11A. Our finding of gap
junctions in the present study calls for a double cable model of the tubule
with two axial resistances, the lumen resistance and the gap-junction
resistance (Fig. 11B).
Accordingly, the electrical parameters determined in the single cable model
should be corrected for the double cable model. The new set of electrical
parameters were determined by fitting previous data of voltage and resistances
(Pannabecker et al., 1992) to
the new equivalent circuit model that includes the gap-junction
resistance.
|
cm in the single cable to 22.0
kcm in the double cable, and it increases the resistance of the apical
membrane (Ra) from 11.4 kcm to 18.0 kcm
(Fig. 11). Since the
fractional membrane resistance of the basolateral membrane
(fRbl) is the ratio of the basolateral membrane resistance
and the transcellular resistance,
Rbl/(Rbl+Ra), we
now find that 55% of the transcellular resistance resides at the basolateral
membrane, instead of 68% as determined previously in the single cable model.
It follows that 45% of the transcellular resistance resides at the apical
membrane, instead of 32% as determined by the previous model.
The second revision concerns the value of the paracellular shunt resistance
Rsh, which increases from 16.8 kcm in the single
cable model to 23.2 kcm in the double cable model
(Fig. 11). Among other
revisions of the equivalent circuit are the change of the electromotive force
of the basolateral membrane (Ebl) from 17.5 mV to
–8.1 mV, and the increase of the electromotive force of the apical
membrane (Ea) from 146.1 mV to 151.4 mV
(Fig. 11).
Ebl and Ea summarize all electromotive
forces operating, respectively, at the basolateral and apical membrane of
principal cells. Ebl includes primarily the
electrochemical potentials of ions, Na+, K+,
Mg2+, Ca2+, Cl– and
HCO3– across the basolateral membrane. Therefore,
the reversal of Ebl from 17.5 mV to –8.1 mV is an
improved estimate of Ebl without identifying the more
dominant electrochemical potential. In contrast to the basolateral membrane,
Ea derives largely from the V-type H+ ATPase
inhabiting the apical membrane. Here the revised value of
Ea (151.4 mV) suggests a stronger electromotive force of
the V-type H+ ATPase than previously assumed.
Splitting the core resistance
The double cable model also resolves the discrepancy between the diameter
of the tubule lumen measured optically and the electrical diameter of the core
when the tubule is modeled as a single cable. Measured through a microscope,
the lumen of the Aedes Malpighian tubule has a diameter between 10
and 15 µm. However, the diameter calculated from the core resistance (22.9
Mcm–1) of the single cable model is considerably
larger, 22 µm (Fig. 11A).
In the presence of dinitrophenol, which shuts down ionic traffic through
apical and basolateral membranes (Wu and
Beyenbach, 2003), the core resistance increases to 32.8
Mcm–1, reducing the core diameter to 17 µm. Thus,
blocking ion transport across the apical membrane reduces the core diameter to
values near the optical diameter of the tubule lumen. The double cable
confirms this conclusion as follows.
The product of the single cell-to-cell gap-junction resistance (431
k; Table 2) and the
average number of principal cells per cm tubule (125 cells
cm–1) yields 53.9 Mcm–1 as the
gap-junction resistance normalized to a Malpighian tubule of 1 cm length
(Fig. 11B). The core
resistance in the presence of dinitrophenol yields the resistance of the
tubule lumen, 32.8 Mcm–1. Combining these two axial
resistances yields a core resistance of 20.4 Mcm–1
that approximates the core resistance of 22.9 Mcm–1
measured in the single cable model. The good agreement of these axial
resistances not only confirms the validity of the short-circuit assumption,
but it also confirms the electrical diameter of the core calculated in the
presence of dinitrophenol as the actual lumen diameter. Moreover, the
electrical diameter represents an average diameter of the tubule lumen, which
is more accurate than the optical diameter, especially in view of the
elaborate luminal brush border.
The good agreement between the calculated (20.4
Mcm–1) and measured (22.9
Mcm–1) core resistances suggests that gap junctions of
stellate cells contribute little to 53.9 Mcm–1, the
gap-junction resistance of a tubule 1 cm long. Unfortunately, stellate cells
are too thin and spongy for intracellular electrical recordings. The
sponginess derives from deep invaginations of the basolateral membrane that
leaves only 2–3 µm of cytoplasm before the microelectrode penetrates
the apical membrane to arrive in the tubule lumen.
Magnitude of the lumen and gap-junction resistances
Comparison of the gap-junction resistance of 53.9
Mcm–1 and the lumen resistance of 32.8
Mcm–1 is instructive. The lumen resistance stems from
the saline occupying a 1 cm length of tubule lumen, without any barrier
(Fig. 11B). Since the
gap-junction resistance is only 1.6 times greater than the lumen resistance,
there must be a substantial number of gap-junction channels coupling one cell
to the next.
The gap-junction resistance between one principal cell and the next is 431
k, which is equivalent to a gap-junction conductance of 2.3 µS
(Table 2). The single channel
conductances of invertebrate gap junctions range from 100 pS in earthworm
giant axons (Brink and Fan,
1989) to 248 pS in epidermal cells of the flour beetle
(Churchill and Caveney, 1993)
and to 375 pS in a mosquito cell line derived from Aedes albopictus
(Bukauskas and Weingart,
1994). If the single channel conductance in the mosquito cell line
is similar to that in Malpighian tubules of Aedes aegypti, then there
must be approximately 6190 gap junctions present in a single principal cell.
The estimate is not unreasonable in view of as many as 10 000, the number of
gap junctions per µm2 in a single gap junction plaque of the
heart or liver (Unwin and Zampighi,
1980).
Permeable to current but not to Lucifer Yellow
The failure of Lucifer Yellow to move from the injected principal cell into
neighboring cells (Fig. 8)
calls for an explanation since these cells display substantial electrical
coupling (Tables 2,
3,
4). Lucifer Yellow has a
molecular mass of 457 Da. Since insect gap junctions are thought to permit the
passage of hydrophilic molecules up to 1200 Da
(Simpson et al., 1977), the
dye would be expected to pass freely from one principal cell to another.
However, permeation through gap junctions is also influenced by charge.
Lucifer Yellow is a divalent anion. Hence, negative fixed charges in the gap
junction pore may thwart its passage through the gap junction in
Aedes Malpighian tubules (Brink
and Dewey, 1980; Veenstra,
1996).
In both vertebrate and invertebrate cells, reports of electrical coupling,
but not dye coupling, are not uncommon. Micromeres of the starfish
(Tupper and Saunders, 1972),
early embryonic cells of the amphibian
(Slack and Palmer, 1969),
embryonic cells of the killifish (Bennett
et al., 1972), cells in the developing insect epidermis
(Warner and Lawrence, 1982)
and cells of mouse blastocysts (Lo and
Gilula, 1979), all exhibit electrical coupling but do not allow
the passage of Lucifer Yellow. The permeability of gap junctions to Lucifer
Yellow depends upon the proteins that compose the gap junctions. Vertebrate
gap junctions composed of connexin 43 (Cx43) permit the passage of both
Lucifer Yellow and current, whereas those composed of Cx45 only allow the
passage of current (Martinez et al.,
2002; Steinberg et al.,
1994). Developmental changes involving the downregulation of Cx43
and Cx26 and the upregulation of Cx31 and Cx31.1 result in a dramatic
reduction of the transfer of Lucifer Yellow between cells without affecting
their electrical coupling (Brissette et
al., 1994).
The permeability of gap junctions to Lucifer Yellow may also be regulated
by heteromerization and/or post-translational modification of connexins.
Coexpression of Cx45 with Cx43 in vertebrate HeLa cells reduces the
permeability of the gap junctions to Lucifer Yellow and results in a unique
electrical conductance not seen before the coexpression
(Martinez et al., 2002). When
gap junctions composed of Cx40 are exposed to cAMP their permeability to
Lucifer Yellow is enhanced (van Rijen et
al., 2000). In contrast, the permeability of gap junctions made of
Cx43 to Lucifer Yellow is decreased by exposure to protein kinase C
(Bao et al., 2004).
Although the above examples are from vertebrate gap junctions composed of
connexins, similar functional and regulatory properties may also apply to
invertebrate gap junctions made of innexins. For example, disruption of the
gene encoding the innexin `passover' prevents dye coupling in the giant fiber
system of Drosophila (Phelan et
al., 1996). Since the transcript for `passover' is not expressed
in Aedes Malpighian tubules (Fig.
10), the absence of this innexin may explain the failure of
Lucifer Yellow to pass between principal cells in our study.
In addition to the above explanations, previous investigators have
suggested that a failure to observe dye coupling between insect cells can be
attributed to (1) the use of an unphysiological saline
(Bohrmann and Haas-Assenbaum,
1993), and (2) impalement damage that allows Ca2+ to
leak into the injected cell and close dye-permeable gap junctions
(Lang and Walz, 1999). The
first artifact can be ruled out in the present study, because isolated
Malpighian tubules of Aedes aegypti bathed in the saline that we used
in the present study secrete fluid for hours
(Beyenbach and Dantzler, 1990),
which would not be expected with use of an unphysiological solution. The
second artifact can also be ruled out, because after impaling a principal cell
for the injection of Lucifer Yellow, the dye arrives first in the tubule lumen
and later in the fluid exiting the open end of the tubule, which confirms that
the cell was not damaged to the point of stopping secretory transport
(Fig. 8).
A glimpse at the regulation of gap junctions in Aedes Malpighian tubules
Although it is known that cytoplasmic Ca2+ can close gap
junctions, the Ca2+ concentration required to do so ranges from
nanomolar to millimolar (Peracchia,
2004). If the gap junctions in Aedes Malpighian tubules
are closed by a rise in intracellular Ca2+, then a cytoplasmic
Ca2+ concentration higher than that achieved in the presence of the
Ca2+ ionophore A23187 and a peritubular Ca2+
concentration of 1.7 mmol l–1 is necessary. A23187
significantly decreased the input resistance from 215 k to 194 k
(Table 4), which importantly
stems from the decrease in the non-junctional resistance but not from the
gap-junctional resistance. Accordingly, the ionophore induced changes at the
level of basolateral and/or apical cell membranes of principal cells but not
gap junctions. These obse
