Pfliigers Arch (1985) 405 (Suppl 1): S 17- S22
EuropeanJournal
Pathways for volume flow and volume regulation in leaky epithelia
G. Whittembury, A. Paz-Aliaga*, Angela Biondi**, Paola Carpi-Medina, E. Gonzfilez***, and H. Linares
Centro de Biofisica y Bioquimica, Instituto Venezolano de Investigaciones Cientificas, IVIC, P.O. Box 1827, Caracas 1010A, Venezuela
Abstract. Continuous pathways must pierce the cell mem-
brane to be used by water during osmotic equilibration
between proximal straight tubular cells and the external
medium, because a) the water osmotic permeability
coefficient of the basolateral plasma membrane, P~, is high;
b) its activation energy, E,, is as that of free water movement
and c) pCMBS inhibits markedly (but reversibly) Wo~ and
increases E, to values similar to those observed in lipid
bilayers without pores, d) Preliminary measurements of Pd
the water diffusive permeability coefficient using NMR indi-
cate that P~ob/pdis near 4-- 5. The following two observations
indicate that a significant paracellular water flow must exist
in leaky epithelia. Namely, a) large extracellular solutes are
dragged by water in four leaky epithelia: gall bladder,
Necturus proximal tubule, rat proximal tubule and Rhodnius
malpighian tubule, b) The transcellular water osmotic
permeability coefficient is smaller than the transepithelial
values available in the rabbit proximal straight tubule. This
requires a significant paracellular permeability.
Key words: Water osmotic permeability - Leaky
epithelia - Paracellular water flow - Osmotic equilibra-
tion - Cell membrane pores - Tubular permeability -
Epithelial water transport - Kidney proximal tubule
Introduction
Leaky epithelia absorb or secrete nearly isosmotic solutions
[15]. Although cells are kept together by junctional
complexes, a paracellular pathway for ion flow has been
shown to exist in these epithelia, based in three independent
observations. First the transepithelial resistance is much
lower than the cell membrane resistance indicating that there
exists a low paracellular resistance which effectively
constitutes a pathway for current and therefore for ion flow
across these epithelia. Then the demonstration that the
electrondense ion La 3 + crosses these epithelia at the level of
the junctional complexes. Hydrated La 3 + is at least as big
Present addresses:
* Departamento de Fisiologia, Universidad Nacional de San
Agustin, de Arequipa, Arequipa, Peril
** Instituto de Quimica-Fisica, Facultad de Bioquimica, Qulmica
y Farmacia, Universidad Nacional de Tucumgm, San Miguel
de Tucumfin, Argentina
*** Escuela de Medicina, Jos6 Maria Vargas, Caracas, Venezuela
Offprint requests to: G. Whittembury at the above address
Pflfigers Archiv
of Physiology
9 Springer-Verlag1985
as Na + [22]. Finally the direct demonstration that the highest
current flow occurs indeed where cells are joined together
([7 a], see also [4] and [24] for references).
In these tissues it has been shown that water moves
passively, secondary to ion flow [24]. But the pathway that
water may follow is still under controversy [4]. This second-
ary movement of water may only occur across the cells, or
part of the water may also move between cells [4, 23]. We
will review here first the nature of the transcellular path-
ways; and the evidence showing that water flows between
cells via the junctional complexes in addition to the known
transcellular water flow.
I. Pathways for water flow between cell and peritubular space
in the rabbit proximal straight tubule (PST)
Water may cross the cell membrane by diffusion through
the lipid matrix, and/or by flow through continuous path-
ways that may be opened at a given moment piercing the
cell membrane from cytoplasm to extracellular space. Infor-
mation concerning the nature of these pathways in kidney
cells may be obtained by taking advantage of previous ex-
perience with artificial lipid bilayers and with red blood cells
[7, 13, 17]. (A) In artificial bilayers doped with pore-forming
antibiotics and in untreated red blood cells (under control
conditions), (1) the water osmotic permeability coefficient,
Pos, is high; (2) its Arrhenius activation energy, Ea, is as low
as that of free water movement, i.e. around 4 kcal/mole; (3)
the water diffusive permeability, Pa, is much lower than Pos,
so that the ratio of PoJPd is clearly greater than 1. (B) In
non-doped artificial bilayers and in red blood cells which
have been treated with the suifhydryl reagent para-
chloromercuribenzene sulfonate, pCMBS, (1) Po~ is low; (2)
its ER is higher than 10 kcal/mole; (3) Pd has values similar
to Pos so that Po~/Pd is about 1. These results are taken to
indicate that in condition A there are continuous pathways
piercing the artificial membranes and the red cell
membranes. Thus Pos is high, water flow is relatively free of
great interactions with the membrane fabric, therefore E, is
low. Current views hold that the ratio of Pos/Pd indicates the
number of water molecules occupying the pore in a single
file (rather than the pore radius, [11]). In the case of the non-
doped membranes and of the red cells treated with pCMBS,
water flows across these membranes essentially by diffusion
through the membrane fabric; thus the Po~ observed is
mainly due to diffusion secondary to the gradient in water
activity created by the osmotic difference. The degree of
interaction between water and the lipid membrane fabric is
high, thus E, is high.
S18
pcb
OS Eo
C pCMBS C pCMBS
l~ig. 1. Water osmotic permeability coefficients of the basolateral
membranes of proximal straight tubular cells, Pg~, in units of
l 0 - 4 cm3/cmz of epithelium - s 9 osmolar. Left: values obtained
under control conditions (C) and after treatment with 2 mmole/1 of
the sulfhydryl reagent pCMBS (pCMBS). Right: the corresponding
Arrhenius energies of activation, E,, (kcal/mole). The effects of
pCMBS onPo~s" and on its E, are fully reversible upon addition of
dithiothreitol
To explore whether the basolateral cell membranes of
the proximal tubule adhere to any of these alternatives, we
have measured Po~b across the contraluminal cell membrane
and their temperature dependence in order to obtain the
Arrhenius E~, under control conditions. Then the tubules
were treated with pCMBS and measurements of P~ob and of
E, were repeated. For this purpose PST were isolated from
the rabbit kidney, they were mounted, between crimping
pipettes, in a chamber that could be thermostated from 4
to 37~ The tubules were observed with an inverted
microscope and a TV camera. Since the lumen is occluded,
changes in diameter reflect volume changes, which were
recorded with a special processor which provides a continu-
ous record of diameter vs. time with precision better than
0.03 gm and 17 ms [5]. Po~b were estimated from the rate of
diameter change with time per unit osmotic concentration
difference. Osmotic steps were set up across the basolateral
membrane in less than 100 ms. After control measurements
of Po~ at two temperatures, 2 mmole/1 pCMBS was added
and P ~ was determined again at the same temperatures. The
results are summarized in Fig. 1. It may be noticed that in
the control experiments p~b is high and E~ is not different
from a value of about 4 expected from free movement.
Treatment with pCMBS lowers markedly Po~b and increases
E,. It must be pointed out that this effect of pCMBS readily
reverses upon addition of dithiothreitol (DTT) to the prepa-
ration [191.
Experiments are in progress to determine Pa using pro-
ton N M R signals and Mn 2 § in the external solution as has
been used in red cells [13]. Preliminary results indicate that
Po,/Pa = 4 - 5 [6]. All these experimental results indicate, by
analogy with the artificial membranes and red cells that
continuous pathways must be piercing the membrane at
a given moment under control conditions. The action of
pCMBS to change the permeability characteristics of the
channel and the reversibility of the action of pCMBS with
D T T indicates that sulfhydryl groups (i.e. proteins) are in-
volved in the stability and structure of these water channels,
which must be upset by addition of pCMBS so that E,
increases to high values (not different from 10 kcal/mole).
Band 3 proteins are being associated with the water channel
in the red cell [3, 17]. Whether this is also the case of the
contraluminal cell membrane of the proximal tubule remains
an open question.
2. Pathways for water flow across leaky epithelia
There is little doubt that the lateral intercellular spaces
facilitate water flow during absorption or secretion in leaky
epithelia, because they have actually been seen to dilate
during this process [18]. The question is whether the water
that flows through them comes from transcellular flow of
from transjunctional flow, i.e. from paracellular flow. To
establish whether in addition to the transcellular flow, water
also flows between cells via the junctional complexes, the
ideal experiment would consist in tightly locating one
micropipette in front of the cell, and a second one just
where cells meet. Comparison of the rate of movement of a
physiological solution within the pipettes would indicate
their relative importance. Since the experiment is technically
difficult at present, two indirect approaches have been
sought to look for an answer to this problem. The first
studies the solvent drag of large extracellnlar molecules. The
second compares the transepithelial and transcellular water
osmotic permeability coefficients.
A. Paracellular flow in leaky epithelia as explored by studying
solvent drag of large extraeellular noneleetrolytes
Large extracellular solutes - which are known not to enter
the cells and therefore to stay exclusively as extracellular
markers - are added to the side of the preparation from
where the flow originates and their net flow (Js) towards the
other side is explored as a function of the net volume flow
across the preparation (Jr) which is allowed to vary.
Js is made up of two terms: a diffusive, ~ and a
convective term, ~. J~ = Is" ACs, and ~ = (1 - o-) C*Jv,
where Ps is the solute permeability of the membrane, a the
reflection coefficient of the membrane to s, C* the average
solute concentration in the membrane.
9Is = Ps" AC~ + (1 -- a)C*Jv. (l)
Therefore a plot of Js as a function of & will have a
diffusive term as intercept and the slope (1 - a)C* will be
different from zero if the convective term is important.
At first sight A Cs seems independent of Jr. However,
because of unstirred layer effects the volume flow piles up
solutes on the side of the membrane from which Jv origi-
nates, and washes away them from the other side. There-
fore ACs across the membrane proper is indeed a func-
tion of Jv and is larger than the A Cs originally set up between
the two solutions. In consequence, before concluding that
Js is a function of J~ one must subtract the diffusive terms
taking into account whatever relationship with Jv may
exist due to the unstirred layer effects (Eq. (12), [1] and
Eq. (83), [2]). The observation of a significant relationship
between Y~and .Iv, after correction for unstirred layer effects,
explores J~ (conv) alone, and therefore it shows the presence
of "true" solvent drag which indicates that there exists in-
teraction between solutes and water within the pathway and,
therefore, that water and solutes share a common pathway
which would be extracellular if the solutes are true extra-
cellular markers.
Such experiments have been performed in the gall
bladder [10, 23]. Flows of sucrose [10, 23], inulin, dextran
(MW 15,000-18,000), hemoglobin and albumin were stud-
ied as a function of Yv [23]. It was found that sucrose and
inulin did show a true solvent drag relation with J~ while the
small relationship observed between Ys for dextran,
hemoglobin and albumin and Jv was due to unstirred layer

IC
b,
%= ............. _-_. . . . . . #EF_USt_V!E. . . .
-5 ; ; ,ro ,; 2O 2'5
Jv[nl crn-2.s-t )
Fig. 2. Net sucrose flux in the absorptive direction, i.e. out of the
rat proximal tubule (Ys)per unit average lumen concentration (C1)
examined as a function of the net absorptive volume flow (Jr). (O)
represent spontaneous variations in J~ obtained by varying either
the length of tubule under observation or the rate of tubular lumen
perfusion. (x) corresponds to isosmotic perfusion solutions
containing 70 mM raffinose which reduce the rate of net absorption.
Thefull line is the regression line best fitting the experimental points.
The dashed line corresponds to the diffusive flow calculated con-
sidering unstirred layer effects (Eq. (12), [1], Eq. (83), [2]. It may be
noticed that the diffusive flow is slightly dependent on Jv. The
difference between the full line (total JJC~) and the line representing
the fraction of Js that is calculated to move exclusively by diffusion
corresponds to the flow of sucrose that moves exclusively by convec-
tion. (From [21] and G. Whittembury, G. Malnic, M. Mello-Aires
and C. Amorena, unpublished)
effects. Similar explorations have been performed in
Necturus proximal tubule ([23], and references therein).
Flows of mannitol and sucrose have been shown to be truly
related to Jv (see Fig. 7 of [23]), and not due to unstirred
layer effects, which fi'om the lumen side little affect the
results, since the tubular lumen due to its small diameter is
indeed a self-stirred compartment.
Figure 2 shows unpublished experiments performed in
the rat proximal tubule [21]. Rat proximal tubules were
microperfused using known techniques, with fluids con-
taining labelled sucrose. Variations in Jv were achieved either
by perfusing shorter segments at a higher perfusion rate
or by adding 70 mM raffinose to the isosmotic perfusion
solution. Notice that at Jv > 5 nl/cm2, s the estimated
diffusive flow is only a fraction of the total flow of sucrose
observed. The convective flow is important, leading to the
conclusion that sucrose is dragged by solvent during
spontaneous absorption in the rat proximal tubular lumen.
Figure 3 shows results of similar experiments performed
during secretion in the isosmotic transporting portion (distal
segment) of the malpighian tubules from the insect Rhodnius
prolixus. This preparation was chosen because it secretes,
(rather than absorbs) an "isosmotic" fluid; the rate of secre-
tion can be varied at will by addition of given doses of
serotonin. The junctional complexes are 1 0 - 2 0 gm long,
formed by 7 gm long septate junctions (rather than by
zonulae occhidentes). The linear extent of the junctions
frontal to the epithelium is much lower than in the
mammalian preparation since the cells in this preparation
are 100 gm in diameter (as compared to 10 ~tm in the
mammalian proximal tubule). Therefore not only the direc-
$19
~
Mannitol
5.0
,x 2,0
rose..
20 [aJ
>
8Z i.o
O f
dv (lO-9k'crn'IZ"s-l)
Fig. 3. Net solute flux in the secretory direction, i.e. into tile tubular
lumen (as) per unit bath concentration (C1), exclusively due to
convection, as a function of the net secretory volume flow (Jr) in the
isosmotic transporting (distal) segment of the Malpighian tubules of
Rhodnius prolixus. The net diffusive flow, which takes into account
unstirred layer effects has already been subtracted using Eq. (12)
and (83) of [1] and [2], respectively. For the sake of clarity
the experimental points have been suppressed. ([14] and G.
Whittembury, A. Paz-Aliaga, A. Biondi and H. Linares, un-
published)
tion of the flow but the geometrical characteristics of the
paracellular pathway are different from the other three
epithelia observed. In addition, since the tubules can be
extracted from the animal without extracellular material,
the spaces of distribution of the non-electrolytes under study
can be readily evaluated [12]. Figure 3 summarizes the re-
sults of unpublished experiments [14]. It shows only the
convective flow of several non-electrolytes (the diffusive flow
has already been subtracted from the total flow. Mannitol
shows the highest dependence with J~. This cannot be judged
as due to solute entrainment in the paracellular pathway,
since mannitol distributes itself in 90% of the cell water. In
the case of sucrose and PEG, polyethylene glycol 800, the
slope relating Js with Jv is clearly significant (P < 0.001, [14,
21 a]). These solutes distribute in about 5% of the tubule
water, as is the case of raffinose, inulin, and dextran MW
15-18,000; however the last three substances are not
dragged by water during secretion in this preparation. In-
terestingly sucrose has a molecular radius of 4.5 * (being
quasi-spherical), while PEG is a rod-like molecule with the
small diameter of 3.4 ?t and a length of 18.5 A. Raffinose is
again quasi-spherical with a diameter 5.9 A and inulin and
dextran are larger. One could therefore deduce that the
pathway through which water entrains sucrose and PEG
must have dimensions near or smaller than the raffinose
diameter.
B. Pathways for water flow as deduced from Po~ values
We assume a simple model in which water osmotic
permeabilities are treated as (their reciprocals) resistances in
$20
pg p~b
p~ - _ _ +~
PST
k__l MEASURED PERMEABILITIES
1 ~I 300 mOsm REQUIRED
TE __(__29_5_} ! j OSMOTIC DIFFERENCES
r lowest value of 40. The other possibility is that our values
] 290 FOR dv ~ 50 nl/cmZ.s
27O /
TC
Fig. 4. Simple model of the proximal tubular wall. Top: Water
osmotic permeabilities are depicted as arrows. Middle: from the
available measured values of apical, basolateral and total Po~, a
value for the paracellular Posis calculated using a model in which the
permeabilities are treated as (their reciprocals) simple resistances.
Bottom: assuming simple osmosis, .Iv = Pos AC. At the bottom, the
osmotic differences that would account for a net absorptive flow of
50 nl/cmz 9s have been calculated. TE, the transepithelial osmotic
difference, for the transepithelial Po~ values of 94 is 5 mosmolar.
For the apical P~ the osmotic difference across the apical cell
membrane would be 20 mosmolar and of 10 mosmolar for the
basolateral Po% The value for Jv represents an upper limit that
encompasses values obtained in the rat proximal tubule and in the
rabbit PCT. Clearly for smaller Jv~ the magnitude of AC must be
proportionally smaller.
a circuit (see Fig. 4). Apical and basolateral cell membrane
water osmotic permeabilities (Po~ and Pop'b, respectively)
would be in series forming P oC s--- Posca cb ca cb
9 Pos/(Po~ + Pos), the
transcellular permeability. The paracellular permeability
(P~s) would be in parallel to the transcellular one to form
the transepithelial permeability, P~s = PoPs+ P~s, therefore if
this simplified model holds measurements of the transepithe-
lial permeability should equal the transcellular permeability
if the paracellular permeability (PoPs)is negligible. This would
clearly not require any paracellular water flow. On the other
hand, if Po%is larger than PoPs,one is bound to conclude that
a significant paracellular water permeability and therefore
a paracellular water flow must exist.
The central part of Fig. 4 makes such a comparison. As
already illustrated in Fig. 1 the value for P ~ is 50 in units
of 10 -4 cm3/s osmolar per cm a of basement membrane [5].
To measure Po~ across the apical cell membrane, the tubules
were cannulated with double barreled pipettes. Through
one barrel an isosmotic solution and through the other an
anisosmotic solution could be perfused. Changes in cell
volume were followed as already described [8]. The tubules
were externally bathed with oil during the performance of
the measurement to avoid undesirable osmotic equilibration
of the cell which would have lead to under estimations of
PXL The value obtained is 23 (same units) which combined
with the value for Po~ of 50 yields a transcellular P~s of 14
(in the same units) which is smaller than any of the transepi-
thelial values available. All these values are adjusted to 25 ~C.
In the rabbit PST the measured published value [16] is 94
(Fig. 4) (also at 25 ~C). Range 40--138 (in the same units).
Doubts can be casted on the transepithelial values because
they were obtained using steady state conditions in which
certainly the cells have adjusted to a different osmolality
from the one originally set up as an osmotic step [4]. In
addition extrapolation to infinite stirring was used to
approximate the data. The fact remains, however, that any
of the other transepithelial calculated values (and values
obtained in the rat are near 40, see [5]) are significantly
higher than the transeellular value given above of 14. Clearly
fast methods to record transepithelial volume changes are
needed to reassess the transepithelial values in order to pro-
vide more secure grounds for comparison, but in the mean
time the conclusion must be drawn that a paracellular PoPsof
80 is required for the transepithelial value of 94 (Fig. 4), or
at least a paracellular value of 26 for the transepithelial
are falsely low. This does not seem to be case at least for the
contraluminal pxb values, since figures in agreement with
ours have been published. As yet there are no other values
for the apical P~. But even if they were to be as high as the
contraluminal values of 50 available, the total transcellular
Pgs would be of at least 25 which still calls for significant
paracellular permeabilities.
Simple osmosis' and transepithelial and transcellular osmotic
differences. The above mentioned Pos values may be used to
calculate, assuming simple osmosis (i.e. Jv = Po~ " A C), the
concentration differences (osmotic pressure differences) re-
quired to account for the measured transepithelial volume
flows (Jr). An example is illustrated in Fig. 4 (bottom). If Jv
is 50 nl/cm 2 9 s, the transepithelial concentration difference
would be 5 mosmolar, if one uses the transepithelial Po% of
94. Actual values for Jv are 10 nl/cm 2 - s in PST and range
from 2 0 - 4 2 in PCT. In Fig. 4 a value of 50 has been given
in order to emphasize that under hypoosmotic conditions .Iv
values higher than those obtaining under isosmotic circum-
stances may be obtained, which call for higher osmotic
concentration differences, and which therefore represent
more stringent tests to osmotic theories [20]. Therefore even
lower osmotic differences are required which are just now
being measured [9, 15]. The problem returns when the avail-
able transcellular values are taken into consideration. Thus
Fig. 4 illustrates that for a Jv of 50, transcellular differences
of 20 and 10 for the apical and basolateral plasma
membranes would be required. If a Jv of 10 is used the
differences would be 5 and 2 mosmolar, respectively.
Areas available for transcellular and f o r paracelIular volume
flows. The following calculations illustrate that the pathways
for transepithelial water flow occupy only a small fraction
of the epithelial area. This holds both for the possibility that
all the water flow is transcellular and for the extreme that it
is all paracellular. Pos can be defined in terms of the geometry
of the pathways across which the osmotic gradients act to
produce the observed flow. For the transcellular calculations
we assume Pos = 50 x 10-4 cm3/cm z s osmolar. We assume
for simplicity that the pathways are cylindrical pores
piercing the cell membranes with a radius of 3 A, which is
a reasonable value. Thus Pos = nnr4/8tlAx where t/is the
viscosity of water (7 centipoise = 2.7 x 10-i0 osmolar s at
37~ A x is the membrane thickness, which we take as
100 A. With these values we calculate that n, the number
of pores piercing the cell, will be 6.8 x 10 - l z per cm z of

epithelium, which will occupy a total area nr~r z o f 0.01 cm z
per cm 2 o f epithelium. Alternatively we can assume that Po~
reflects only paracellular water flow t h r o u g h rectangular
s/its. Thus, Po~ = 8 9 w 3 . L / 1 2 A x . rl, where L is the linear
extension o f the slits o f 4,460oCm/cm 2 o f epithelium, with
depth of a p p r o x i m a t e l y 1,000A and width w. These values
using PX~ o f 94 x 10 - 4 crn3/cm 2 9 s 9 o s m o l a r lead to the re-
asonable value o f w = 19 A. The junctional area used by the
flow would be 0.001 cm2/cm z o f epithelium. These calcu-
lations are only a first a p p r o x i m a t i o n and m a y be an
oversimplification because neither the resistance to water
flow o f the cytoplasm nor that o f the LIS is taken into
account, but they nonetheless indicate that only a small
fraction o f the whole epithelial area, be it transcellular or
paracellular must be used for water transport. As a conse-
quence, the speed o f water flow through each o f these indi-
vidual pathways, be it each o f the pores piercing the cell
membranes or the junctions keeping cells together must be
quite high indeed. Thus, with a Jv value o f say 50 nl/cm z
epilhelium 9 s represents a speed o f 0.5 gm/s when the whole
epithelium is considered, but if all this flow through 0.01 or
0.00l o f this area, the speed o f fluid flow will then be 1 0 0 -
1,000 times as high.
Conclusion
Continuous pathways piercing the cell m e m b r a n e allow for
osmotic equilibration across the contraluminal plasma
m e m b r a n e o f the p r o x i m a l tubule. A t present no definite
decision can be m a d e as to whether in addition to transcellu-
lar flow there is flow between the cells across the junctional
complexes. However, the observations a) that large extra-
cellular non electrolytes are dragged by solvent in four leaky
epithelia; and b) that the present available figures for the
transepithelial water permeability are clearly larger than
those recently measured for the transcellular water
permeability are m o r e easily explained if there is paracellular
water flow in these tissues.
Acknowledgements. Supported in part by grants from CONICIT,
PNUD/UNESCO (RLA 78/024), Dr. Gonzfilez work was supported
in part by CDCH Project M-09.5182. GW is also a member of
Instituto Internacional de Estudios Avanzados, IDEA, Caracas. It
is a pleasure to thank the kind help of Drs. T. J. Pedley and G. K.
Aldis (Cambridge) in approaching the unstirred layer problems.
Mrs. Rebeca Godoy kindly typed the manuscript. Mr. Jos6 Mora,
F. Murillo, A. Cazorla, J. Bigorra and M. Diaz helped in several
stages of this research.
References
1. Andreoli TE, Schafer JA, Troutman SL (1981) Coupling of
solute and solvent flows in porous bilayer membranes. J Gen
Physiol 57: 479 - 493
2. Barry PH, Diamond JM (1984) Effects of unstirred layers on
membrane phenomena. Physiol Rev 64: 7 6 3 - 873
3. Benz R, Tosteson MT, Schubert D (1984) Formation and
properties of tetramers of band 3 protein form human erytliro-
cyte membranes in planar lipid bilayers. Bioehlm Biophys Acta
775:347-355
4. Berry CA (1983) Water permeability and pathways in the prox-
imal tubule. Am J Physiol 245: F279 -- F294
$21
5. Carpi-Medina P, Lindemann B, Gonz~lez E, Whittembury G
(1984) The continuous measurements of tubular volume
changes in response to step changes in contraluminal osmolality.
Pfliigers Arch 400: 343 - 348
6. Carpi-Medina, P, Le6n V, Linares H, Whittembury G,
Gonz~lez E (1984) Permeabifidad acuosa difusional (Pa) Y
osm6tica (Pos) de la membrana basolateral de tfibulos pro-
ximales de conejo. Primeras Jornadas Venezol. Nefrol Caracas,
p31
7. Finkelstein A (1984) Water movement through membrane
channels. Curr Top Membr Transport 21:295--308
7a. Froemter E (1972) The route of passive ion movement through
the epithelium of Necturus gall bladder. J Membr Biol 8: 2 5 9 -
301
8. Gonzfilez E, Carpi-Medina P, Linares H, Whittembury G (1984)
Water osmotic permeability of the apical membrane of proximal
straight tubular (PST) cells. Pflfigers Arch 402:337-339
9. Green R, Giebisch G (1984) Luminal hypotonicity as a driving
force for fluid absorption from the proximal convoluted tubule
of the rat. Am J Physiol 246:F167-F174
10. Hill AE, Hill BS (1978) Sucrose fluxes and junctional water
flow across Necturus galI bladder epithelium. Proc Roy Soc B
200:151-162
11. Levitt DG (1984) Kinetic of movement in narrow channels.
Curr Top Membr Transport 21 : 181 - 197
12. Maddrell SHP (1969) Secretion by the Malpighian tubules of
Rhodnius. The movement &ions and water. J Exp Bio151 : 71 --
97
13. Moura T, Macey RI, Chien DY, Karan D, Santos H (1984)
Thermodynamics of all or none water channel closure in red
cells. J Membr Biol 8:105-111
14. Paz-Aliaga A, Linares H, Whittembury G (1983) Flujo de agua
por la via paracellular en tubo de Malpighi. Acta Cientif
Venezol 34: suppl 1,215
15. Schafer JA (1984) Mechanisms coupling the absorption of
solutes and water in the proximal nephron. Kidney Int 25: 708 -
716
16. Schafer JA, Patlak CS, Troutman SL, Andreoli TE (1978)
Voltune regulation in the pars recta. Am J Physiol 234:F340--
F348
17. Solomon AK, Chasan G, Dix JA, Lukacovic MF, Toon MR,
Verkman AS (1983) Possible relation between anion transport
and water flow in red cells. Ann NY Acad Sci 414: 97-104
18. Spring KR, Hope A (1979) Size and shape of the lateral in-
terspaces in a living epithelium. Science 200:54-58
19. Whittembury G, Carpi-Medina P, Gonz/dez E, Linares H (1984)
Effect of para-chloromercuribenzenesulfonic acid and tempera-
ture on cell water osmotic permeability of proximal straight
tubules. Biochim Biophys Acta 775:365--373
20. Whittembury G, Hill BS (1982) Fluid reabsorption by Necturus
proximal tubule perfused with solutions of normal and reduced
osmolarity. Proc Roy Soc B 215:41l -431
21. Whittembury G, Malnic G, Mello-Aires M, Amorena C (1981)
Flujo paracelular de agua en tfibulo renal proximal de rata.
Acta Cientif Venezol 32: (suppl 1)40
21a. Whittembury G, Biondi AC, Paz-Aliaga A, Linares H, Linares
N (1985) Transcellular and paracellular flow of water during
secretion in the upper segment of the Malpighian tubule of
Rhodnius prolixus: Solvent drag of grades sized molecules. J
Exp Biol (submitted)
22. Whittembury G, Rawlins FA (1971) Evidence o f a paracellular
pathway for ion flow in the kidney proximal tubule: electron-
microscopic demonstration of lanthanum precipitate in the tight
junction. Pflfigers Arch 330: 3 0 2 - 309
23. Whittembury G, Verde-Martinez C, Linares H, Paz-Aliaga A
(1980) Solvent drag of large solutes indicates paracellular water
flow in leaky epithelia. Proc Roy Soc B 211:63- 81
24. Windhager EE (t979) Sodium chloride transport. In: Giebisch
G, Tosteson DC, Ussing HH (eds) Membrane transport in
biology, vol4a. Springer, Berlin Heidelberg New York,
pp 143-231
$22
Audience discussion
Spring: I have two questions that relate to the distribution
of the water permeabilities. You said, that when there is no
oil present, changes in the luminal composition result in a
half osmometric response. One can easily calculate that a
half osmometric response occurs if both membranes have
the same water permeability. Would this not suggest that in
the oil experiments the lumen permeability might be
understimated?
Whittembury: Although I said "half' this is only a rough
estimate. The real point is that we have Ringer's solution
outside and inside initially and when we change either one
of them, the cell adjusts and we get a different value from
the one we ought to have obtained.
Spring: My second question regarding the water perme-
abilities is that Berry has argued that the water permeabi-
lities from Schafer and Andreoli are too high by about a
factor of 2. If that is the case, then there is rough agreement
between the cellular water pathway and the transepithelial
pathway.
Whittembury: The water permeability mentioned from
Berry is about 40. The apical plus basolateral permeabilities
we measured add up to 14 or 15. Even if we call it 20, since
the other is 40, there is a real discrepancy. I think what needs
to be done is to measure the transepithelial permeabilities
with a fast method, instead of a steady-state method, in
which the cells adjust to the new osmolalities and are in an
entirely different state after 10 or 15 rain. Although I find
the possibility that water moves between cells attractive, my
mind is open. The solvent drag of extracellular markers
which we found does however support the idea of paracellu-
lar water flow.
Spring: If you use the numbers that you have calculated for
the water permeability of the paracellular pathway and you
take into account the very small area of the paracellular
pathway, do you not get impossibly high required
permeability values for each tight junction?
Whittembury: Yes, one finds a high value for the tight junc-
tion permeability per square centimeter of tight junction.
But the same argument applies to the cells. If water moves
through the cells and if there are pores for this movement,
only a tiny fraction of the cell membrane will be occupied
by these pores and there will be a huge water flow going
through each of these pathways.
Ullrich: Dr. FrSmter and I found a considerable solvent
drag transport component for Na, C1 and urea in the rat
proximal convoluted tubule, and also a high water permeabi-
lity. We have difficulty with the interpretation of other in-
vestigators, who claim that water flows through the cells,
because we cannot see how water flowing through the cell
can provide the force for solvent drag, which occurs
paracellularly. Thus, I think that Dr. Whittembury's data
fits with the finding of high solvent drag in the proximal
tubule much better than the hypothesis that all the water
flows through the cell, rather than paracellularly.