You are on page 1of 4

(dT)12_18, and poly(rC) * oligo(dG)12-18.

microliters of virus (which was concentrated Sodiulm-Specific Membrane Channels of Frog Skin Are Pores:
100-fold by centrifugation and then suspended
in 50 mM tris-HCI, pH 7.3, 10 mM dithio- Current Fluctuations Reveal High Turnover
threitol, and bovine serum albumin, I mg/
ml) were incubated with 10 ,ul of 50 mM tris-
HCI, pH 7.3, 10 mM dithiothreitol, 0.06 percent Abstract. The reversible sodium transport blocker amiloride causes current fluc-
Triton X-100 for 10 minutes at 2C. A portion (20 tuations at the apical membrane of the outer stratum granulosum offrog skin. Their
I1A) of assay mixture was added, and the result-
ing 40-,ul reaction mixture was incubated for 60 power density spectra reveal that single transport sites translocate more than 106 So-
minutes at 37C. The reaction was stopped by dium ions per second, which indicates a pore mechanism. The density of open plus
the addition of 0.5 ml of 80 mM sodium pyro-
phosphate, 25 ,u of bovine serum albumin (5 mg/ amiloride-blocked pores is in the order of 108 pores per square centimeter of skin
ml), and 0.5 ml of 25 percent trichloroacetic acid
at 2C. Acid-insoluble material was collected on area with 60 millimolar sodium and 18 micromolar amiloride in the outer solution.
glass fiber filters, washed extensively with 10
percent trichloroacetic acid, dried, and dis-
solved in NCS tissue solubilizer; the radio- Carrier- and pore-mediated transport mined by the apical membrane of the
activity was then counted in toluene-based
scintillation fluid. The reaction mixtures with through biological membranes can be dis- stratum granulosum (4). The current
poly(rA) * oligo(dT)12_18 and poly(dA) * oligo
(dT)12-18 consisted of: 50 mM tris-HCI, pH
tinguished by the turnover number of in- component that did not pass the Na-spe-
7.3, 70 mM KCI, 10 mM dithiothreitol, 0.1 mM dividual transport sites (1). While a car- cific channels was determined as the cur-
each of deoxyriboadenosine triphosphate (dATP), rier molecule, which has to move rent flowing in the presence of 35 pM
deoxyriboguanosine triphosphate (dGTP),
deoxyribocytidine triphosphate (dCTP), 2.3 through the lipid phase of the membrane, amiloride, and was subtracted from the
pM 3H-labeled thymidine triphosphate (40,000 is not likely to transport more than 104 total current to obtain the transcellular
count/mnin per picomole), 20 j.g of poly(rA) or
poly(dA) per milliliter, 20 pg of oligo(dT)12_18 ions or molecules per second, the trans- Na current, INa-
per milliliter, and either 5 mM MgCl2 or 0.4
mM MnCI2. The reaction mixture containing port rate of a pore can be several orders Transepithelial voltage was clamped
poly(rC) oligo(dG)12-18 consisted of: 50 mM
tris-HCI, pH 7.3, 50 ,ug of actinomycin D
of magnitude larger (2). Applying this to 0 mv by a voltage clamp circuit with

Downloaded from on April 22, 2016

per milliliter, 15 mM KCI, 10 mM dithiothreitol, idea to the Na-selective membrane of continuous feedback. The input stage of
10.7 PuM 3H-labeled dGTP (6500 count/min per frog skin, we have attempted to deter- the voltage-sensing amplifier was de-
picomole), 40 ,ug of poly(rC) per milliliter, 40
jAg of oligo(dG)12_8 per milliliter, and either mine the Na turnover of individual trans- signed around a matched pair of low-
25 mM MgCI2 or 0.4 mM MnCI2. port sites by an evaluation of current noise transistors (National Semiconduc-
11. R. C. Nowinski, E. Edynak, N. H. Sarkar,
Proc. Natl. Acad. Sci. U.S.A. 68, 1608 (1971); fluctuations.. The huctuations were in- tor 2N4250) to minimize feedback cur-
H. C. Chopra, I. ZelUadt, E. M. Jensen, M. M. troduced artificially by addition of the rent fluctuations arising from this stage.
Mason, N. J. Woodside, J. Natl. Cancer Inst.
46, 127 (1971). drug amiloride, a pyrazine diuretic The open-loop clamp gain was 25,000.
12. N. C. Goodmati and S. Spiegelman, Proc. Natl.
Acad. Sci. U.S.A. 68, 2203 (1971). known to block Na transport reversibly The short-circuit current was amplified
13. J. W. Abrell and R. C, Gallo, J. Virol. 12, 431 from the outside (3). A reversible with a gain of 50 ,uv/na, fed through a
(1973). blocker can be expected to randomly in- high-pass RC-filter with a characteristic
14. J. Schlom and S. Spiegelman, Science 174, 840
(1971). terrupt the Na turnover of individual frequency of 0.007 hertz, amplified 400
15. Serum was neutralized by titrating the virus on
FCf2Th cells before and after reacting serial ten- transport sites. A site will then either times, and recorded on magnetic tape.
fold dilutions of virus with an equal volume of conduct fully or, when blocked, not con- Recording periods were 10 to 30 minutes
antiserum diluted 1 : 5 at room temperature for
1 hour. Inoculated cultures were split 1: 10 by duct at all. The continuous current i pass- for each amiloride concentration.
trypsinization at weekly intervals. After two ing one site is thus chopped up into small The amplified a-c current signal was
passages, RDDP activity in the culture fluids
was determined. The titer of the virus was 104.5 current pulses of varying duration but sampled from the magnetic tape at fre-
tissue culture infective doses (50 percent effec-
tive) per 0.1 ml with and without serum. The equal amplitude (i). The pulses add up to quencies of 50 and 5000 hertz by use of an
antiserum used was prepared in goats against a mean current which on close in- anti-aliasing filter of the Butterworth
intact'virus and was obtained from Dr. R. Wil-
snack, Huntingdon Research Center, Brookland- spection will show random fluctuations. type (72 db per octave, characteristic fre-
ville, Md. Mouse mammary tumor virus, ba- The mean current per square centimeter quency set at 80 percent of the maximal
boon type C virus (M7), and woolly monkey
sarcoma virus antiserums from the same source will be analyzed frequency). The digitized signal
also failed to neutralize the squirrel monkey
virus when similarly tested. was divided into 20 records of 4096
16. P. K. Nakane and G. B. Pierce, J. Histochem. INa = iMPo (1) words each. A power density spectrum
Cytochem. 14, 929 (1966); E. H. Leduc, R.
Wicker, S. Aurameus, W. Bernhard, J. Gen. where M (cm-2) is the mean density of was computed from each record on an
Virol. 4, 609 (1969).
17. The rabbit antiserum to M-PMV polymerase unblocked plus amiloride-blocked trans- IBM 370/58 by use of a fast Fourier trans-
was supplied by Dr. M. Ahmed, Pfizer Inc.,
Maywood, N.J. port sites. The steady-state probability form program. The spectra of 20 records
18. K. H. Rand and C. Long, Nature New Biol. 240, P0 represents the fraction of 'M not were averaged.
187 (1972). blocked by amiloride and'MPO the mean Figure IA shows five power density
19. M. Ahmed, G. Schidlovsky, D. L. Larson, M.
Manousos, S. A. Mayyasi, J. Gen. Virol. 30, 11 density of open sites. Statistical evalua- spectra in the range 0.3 to 100 hertz.
20. A. F. Bykovsky, I. S. Irlin, V. M. Zhdanov, tion of the fluctuations permits computa- Curves a to d were obtained with a so-
Arch. Gesamte Virusforsch. 45, 144 (1974). tion of i and thus the Na turnover of indi- dium activity of (Na)o = 60 mM in the
21. R. S. Howk, L. A. Rye, L. A. Killeen, E. M.
Scolnick, W. P. Parks, Proc. Natl. Acad. Sci. vidual transport sites in the open state. outer solution and amiloride concentra-
U.S.A. 70, 2117 (1973); D. H. Moore, Cancer
Res. 34, 2322 (1974); A. S. Dion, A. B. Vaidya,
Isolated abdominal skin of Rana escu- tions of (A)O = 1.4, 4.3, 12.7, and 35 ,uM.
G. S. Fout, ibid., p. 3509; D. L. Fine and L. 0. lenta was used at room temperature. It Spectrum e was obtained with (Na)o =
Arthur, Abstr. Annu. Mtg. Am. Soc. Microbiol.
(1974), p. 246 (V273); R. Michalides, J. Schlom, was mounted in a Lucite chamber which 0. Spectra a to d are of the Lorentz
J. Dahlberg, K. Perk, J. Virol. 16, 1039 (1975). left 3 cm2 exposed to the bathing solu- type expected for exponential re-
22. A. Hellman et al., J. Virol. 14, 133 (1974); K. H.
Rand, C. W. Long, T. T. Wei, R. V. Gilden, J. tions. The outer solution was K or Na laxatidn phenomena. They obey the
Natl. Cancer Inst. 53, 449 (1974).
23. J. Schlom, G. Schochetman, D. Colcher, W.
sulfate Ringer solution containing vary- relationship
Drohan, Abstracts, VII International Sympo- ing concentrations of amiloride. The in-
sium on Comparative Leukemia Research No. ner one was K sulfate Ringer solution, S= SO/[l+ (f/fc)] (2)
C-l 1 (1975); D. Colcher, W. Drohan, J. Schlom,
J. Virol. 17, 705 (1976). which can be expected to depolarize the where S is the power density (amp2 sec),
24. Supported by contract NOI-CP-43214 within the K-selective inward-facing membranes of SO the plateau value, f the frequency
Virus Cancer Program, National Cancer Insti-
tute. We thank A. Bodenman, C. Thomas, and the epithelium and to increase their con- (hertz), andf, the corner frequency. The
G. Peacher for technical assistance. ductance. Thus, transepithelial resist- exponent a was found to be in the range
18 June 1976; revised 15 July 1976 ance and potential were largely deter- 1.75 to 2 (1.8 in Fig. IA). It is evident
292 SCIENCE, VOL. 195
Fig. 1. (A) Power density spectra obtained from 3 cm2 of epithelium.
Only the parts of the recorded spectra that are relevant for the argu-
ments of this report are shown. (Curves a to d) Increasing
amiloride concentrations of (A). = 1.4, 4.3, 12.7, and 35 UM at 60 mM
(Na)o. (Curve e) Sodium-free outer solution (Na replaced by K). The
corresponding values of transcellular Na current are (from curve a to
curve e) 41, 24, 9, 0, and 0 ua per 3 cm2. The transcellular Na current
at (A)o = 0 was 78 ,ua per 3 cm2 (spectrum not shown). Plateau values
(SO) and corner frequencies (fe) are indicated with horizontal and
vertical arrows; f, was read off where the power density equaled half
of the plateau value. The f, of curve d (for which INa was zero) was
not evaluated because the a value of this curve fell clearly below the
expected range. (B) Relationship between corner frequency and
amiloride concentration (see Eq. 4a). Arrows point to the intercepts
where k2 and KA' are read off. The linear dependence of l/r on (A)0
shows, in retrospect, that neglecting the term kl(R) in Eq. 4 was (A)o ,M
.1 1 10 hertz100

that the plateau value, S0, of the power where ,8 = (A)m/(A)o is the unknown par- For (Na)0 = 60 mM, values in the range
spectrum decreases with increasing tition coefficient of amiloride between 0.3 to 0.5 pa were computed, which cor-
amiloride concentration while the corner the outer bulk solution and the space at respond to turnover numbers of 1 to 3 x

Downloaded from on April 22, 2016

frequency f, increases. the outer surface of the membrane. After 106 Na ions per second for individual
Following a derivation in Verveen and substituting into Eq. 3 and combining transport sites. These turnover numbers
DeFelice (5), we may set 27Tf, equal to Eq. 3 with Eq. 4a, we find the "on" and are much larger than the values of 104
the chemical rate (1/T) of the current- "off" probabilities to be ion/sec which are expected (1) and were
modulating process. In the simplest pos- observed (8) for shuttle-type mobile car-
sible case this process is due to the ran- Po = 1/[l + (A)o/KA'] = k2r riers, like valinomycin, which have to dif-
dom formation and disintegration of a PA= Po(A)O/KA' k1'r(A)0 (3a)= fuse through the lipid phase of the
transport-blocking complex between The linear relationship between fc and membrane. Turnover rates of nonshuttle-
amiloride (A) and the transport site (R) (A)o predicted by Eq. 4a was found in most type carriers, where the carrier molecule
experiments (for example, see Fig. I B). rotates or where only a part of the carrier
This shows that channel blocking by molecule moves, have, to our knowl-
A + R z AR edge, not yet been obtained experimen-
k2 amiloride can be described in terms of
simple, bimolecular kinetics, and that tally. However, since such carriers will
where AR is the blocked site; k, and k2 kj(R) in Eq. 4 is numerically negligible. also have to overcome the viscosity and
are rate constants, and their ratio The rate constant k2 and the "apparent" the electrostatic energy barrier of the
KA = k2/k, is the dissociation constant constants k1' = k1,8 and KA' = KA/13 can lipid phase, low turnover rates will be
of AR. With M = R + AR, where R and be estimated from the plot. We found expected in this case too. We feel justi-
AR are the densities of unblocked and 1/k2, the mean lifetime of the complex fied, therefore, in concluding that in the
blocked sites, we find from the law of AR, to be in the order of 100 msec at Na-selective membrane of frog skin, trans-
mass action the steady-state probabili- room temperature. The apparent dis- port occurs through pores, which are the
ties that a site is unblocked or blocked sociation constant KA' was close to 1 ,uM only high-rate translocators presently
at 60 mM (Na)0. It decreased with de- known (9). This result is at variance with
Po = RIM = 1/[1 + (A)m/KA] creasing (Na)o, as expected if Na and a previous conclusion by Biber and
PA = ARIM = Po(A)m/KA (3) amiloride compete for transport sites (7). Sanders (10), who maintained that the Na
The apparent constantk1' was found to be transport is carrier mediated.
where (A)m is the effective amiloride con- in the order of 107 liter mole-' sec-I at 60 The mean density of open plus amil-
centration at the outer surface of the mM (Na)0. oride-blocked Na pores can be calculated
membrane. This blocking mechanism im- For randomly blocked transport sites from
plies the following relationship between
with unequal on and off probabilities (5)
chemical rate and equilibrium concentra- the plateau value of the power density
M = lNaI(iPo) =
tions of A and R (6)
spectrum (Eq. 2) is, in our notation, giv- It was found to be in the range 0.7 to
1IT = kl(A)m + kl(R) + k2 (4) en by 2 x 108 pores per square centimeter of
membrane area at 60 mM (Na)O. When
in which (R) is the equilibrium concentra- S0 = 4Mi2POPATa (5) (Na)0 was lowered to 15 mM by sub-
tion of unblocked transport sites, ex- in which a is the experimental membrane stitution with K, larger pore densities
pressed in the same units as (A)m In area. Substituting Eq. 1 into Eq. 5, we were computed, but did not exceed
cases where (R), which decreases with obtain 3 x 109 pores per square centimeter.
increasing (A)m, is numerically negligible Extrapolation to (Na)o = 0 shows that
compared to (A)m, a linear relationship So = 4INaPAria (5a)
the total number of pores (N) will be
between amiloride concentration and Further combination with Eqs. 3a and 4a below 5 x 109 cm-2. This density cor-
chemical rate will be expected. We thus yields an expression for the amplitude of responds to less than 50 pores per square
obtain the prediction individual current pulses micrometer of membrane area if a homo-
geneous distribution of pores over the
217Tf = 1IT = ki(A)m + k2 =
apical membranes of all cells of the outer
kj3(A)o + k2 (4a)
4a1Na (A).
[(A)O + KA']2 (6)
stratum granulosum may be assumed.
21 JANUARY 1977 293
If, in the absence of amiloride, all pores 5. A. A. Verveen and L. J. DeFelice, Prog.
Biophys. Mol. Biol. 28, 189 (1974).
(LRB) (5). The conjugation reactions
were permanently open, a Na current 6. G. H. Czerlinski, Chemical Relaxation (Dekker, were carried out at room temperature in
INa = iN = 2000 p.a/cm2 would result at New York, 1966); G. G. Hammes, Adv. Protein an aqueous-acetone solvent made alka-
Chem. 23, 1 (1968).
60 mM (Na)0. This current is 75 times 7. A. W. Cuthbert and W. K. Shum, Naunyn line with KHCO3. The products were pu-
larger than the INa value of 26 ,ua/cm2 Schmiedebergs Arch. Pharmakol. 281, 261 (1974); rified by ion-exchange and gel-per-
Proc. R. Soc. London Ser. B 189, 543 (1975).
actually observed (see legend to Fig. IA). 8. P. Lauger, "Carrier-mediated ion transport," meation chromatography. The criterion
We deduce from this observation that Science 178, 24 (1972). for purity was the formation of a single
9. Theoretically, translocators that combine prop-
even in the absence of amiloride the pores erties of carriers and pores are conceivable. For fluorescent spot in paper electro-
instance, at a transport site the membrane may
are not permanently open (11). effectively be thinned down to a narrow protein phoresis. Amino acid analysis and end-
B. LINDEMANN structure, part of which can bind an ion selec- group analysis were performed on all
tively and transfer it by a rotational movement
Second Department of Physiology, in low-viscosity surroundings ["translocase"; peptides, except microperoxidase. The
Universitat des Saarlandes, see P. Mitchell, Nature (London) 180, 134 purified compounds were dissolved in
(1957)]. Transfer rates may be high, although
665 Homburg, West Germany this is not very likely in view of the low turnover water and the pH was adjusted to about 7
W. VAN DRIESSCHE numbers of most enzymes. H. Passow (personal with KOH or HCI.
communication) has estimated turnover num-
Laboratorium voor Fysiologie, bers of 2 x 104 sec-' for the anion "carrier" of The solutions of the fluorescent probes
erythrocyte membranes, which might fulfill the
3000 Leuven, Belgium structural requirements mentioned above. It is were injected into cells of isolated Chiro-
an open question of nomenclature how such nomus salivary glands (mid-fourth in-
References and Notes translocators are to be classified. They resemble
pores because the larger part of the membrane's star) with the aid of a micropipette and a
1. C. M. Armstrong, Biophys. J. 15, 932 (1975); Q. diameter is passed by diffusion through a hydro- pneumatic pressure system (6); the
Rev. Biophys. 7, 179 (1975). philic channel, and they resemble carriers be-
2. We suggest that (i) the term carrier be applied cause movement of a membrane component is spread of the fluorescence inside the
only to translocators where the transfer-mediat- essential for the transfer event. In the spirit of cells (excited with wavelengths of 460

Downloaded from on April 22, 2016

ing moiety moves through the lipid phase; (ii) (2) we would not classify such structures, if they
the term pore be applied to water-filled path- exist, as carriers. They may be viewed as pores
in which the selectivity filter [B. Hille, J. Gen.
nm for FITC, 340 nm for DANS, and 540
ways that permit diffusional transport, which
can, however, involve binding of ions to parts of Physiol. S, 599 (1971)] constitutes a peculiar nm for LRB) was observed and photo-
the pore structure; and (iii) different terms be energy barrier. graphed in a microscope darkfield, or,
used for translocators that do not meet these 10. T. U. L. Biber and M. L. Sanders, J. Gen.
specifications [see also (9)]. Physiol. 61, 529 (1973). for velocity determinations, the spread
3. E. N. Ehrlich and J. Crabbe, Pfluegers Arch. 11. A. W. Cuthbert [.J Physiol. (London) 228, 681 was viewed and videotaped with the aid
302, 79 (1968). (1973)] computed, from amiloride binding data,
4. Under short-circuit conditions, the voltage a channel density of 400 i.m-2 [lower limit at 2.5 of an image intensifier-television system
across the apical membrane will come even mM (Na)0], and from this calculated turnover
closer to the transepithelial voltage when the rates for individual Na channels of 2000 sec-1 at (7). The fluorescent emissions of FITC
resistance of the apical membrane is increased 2.5 mM (Na)0 and 8000 sec-1 at 115 mM (Na)0 and DANS peak at 520 and 525 nm (yel-
by amiloride. Also, the decrease of INa due to (upper limits). These turnovers are integrated
amiloride will decrease the K current of the K- over open and closed times of a channel and low-green) and that of LRB, at 590 nm
selective, inward-facing membrane, and thus de- therefore cannot be directly compared with our (red). Thus, in experiments where two
crease the possible effect of this membrane on results.
the recorded fluctuations. Finally, the intra- 12. This work was supported by Deutsch For- tracers were injected together, the LRB
cellular Na concentration can be expected to schungegemeinschaft as project Cl within SFB was easily distinguished from either of
become small in the presence of amiloride. This 38, and by the Humboldt Foundation.
condition simplifies the computation of channel the other two by the use of appropriate
conductance. 14 June 1976; revised 19 August 1976
filters. In some cases, the tracer studies
were combined with measurement of
electrical coupling. Electrical current
was then passed between the exterior
Size Limit of Molecules Permeating the and the interior of the cell injected with
Junctional Membrane Channels the tracer, and the resulting changes in
membrane potential were measured in
Abstract. The permeability of the cell-to-cell membrane channels in salivary gland this cell and the immediate neighbor with
cell junction (Chironomus thummi) was probed with fluorescent-labeled amino acids the use of three microelectrodes (1).
and synthetic or natural peptides. Molecules up to 1200 daltons pass through the The results obtained with the various
channels with velocities depending on molecular size. Molecules of 1900 daltons or tracers are summarized in Table 1. The
greater do not pass. This passage failure seems to reflect the normal size limit for amino acids and peptides with sizes less
junctional channel permeation; the channels continue to be permeated by the mole- than or equal to 1158 daltons passed
cules up to 1200 daltons when these are mixed with the nonpermeant molecules. through the junction. Their fluorescence
From this size limit a channel diameter of 10 to 14 angstroms is estimated. spread from the site of injection through-
out the injected cell and into the cell
Since the finding that fluorescein, a conjugates which incorporate some of neighbors at rates inversely related to
molecule of 330 daltons, passes through the desirable features of the popular trac- molecular size. The tracers generally
a cell junction of low electrical resistance er fluorescein, such as water solubility, crossed several cell junctions on either
(1), other fluorescent and colorant sub- nontoxicity, low cytoplasmic binding, side. The arrival of a tracer at the junc-
stances have been used to probe junc- and high fluorescent yield. To obtain con- tion was marked by an abrupt change in
tional permeability properties (2, 3). jugates of well-defined structure, we the velocity of the fluorescence spread.
However, the range of questions that sought, for the nonfluorescent backbone, With molecules less than or equal to 380
could be studied has been restricted by not only a molecule of known structure daltons, passage through the first junc-
the small number of useful probes avail- but one with few reactive sites, prefera- tion became detectable within a few sec-
able. We have now enlarged the reper- bly only one. Thus, the primary amine onds of the tracer's arrival at the junc-
toire of probes with the aim of determin- group of the synthetic and natural pep- tion; and the fluorescence on the two
ing the size limit of molecules per- tides listed in Table 1 was coupled with sides of the junction appeared to equal-
meating the junctional membrane the fluorescent dyes fluorescein iso- ize within 1 to 10 minutes. With mole-
channels (4). thiocyanate (FITC), dansyl chloride cules between 593 and 1158 daltons, the
We set out to construct fluorescent (DANS), or lissamine rhodamine B transit through junction was slower (tak-
294 SCIENCE, VOL. 195
Sodium-specific membrane channels of frog skin are pores:
current fluctuations reveal high turnover
B Lindemann and W Van Driessche (January 21, 1977)
Science Translational Medicine 195 (4275), 292-294. [doi:

Editor's Summary

This copy is for your personal, non-commercial use only.

Downloaded from on April 22, 2016

Article Tools Visit the online version of this article to access the personalization and
article tools:

Permissions Obtain information about reproducing this article:

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last
week in December, by the American Association for the Advancement of Science, 1200 New
York Avenue NW, Washington, DC 20005. Copyright 2016 by the American Association for
the Advancement of Science; all rights reserved. The title Science is a registered trademark of