- Email: [email protected]
Expression of RNA isolated from the water-shunting complex of a sap-sucking insect increases the membrane permeability for water in Xenopus oocytes
EXPERIMENTAL
CELL
RESEARCH
200,301-305
(l%&?)
Expression of RNA Isolated from the Water-Shunting Complex of a SapSucking Insect Increases the Membrane Permeability for Water in Xenopus Oocytes MARIE-TH~~~SEGUILLAM,FABIENNEBEURON,NATHALIEGRANDIN, JEAN-FFUNCOISHUBERT, CLAUDE BOISSEAU, ANNIE CAVALIER, ANNE COUTURIER, JEAN GOURANTON, AND DANIEL THOMAS’ Laboratoire de Biologic Cellulnire URA CNRS No. 256, Universit& de Rennes 1, Campus de Beaulieu, 35000 Rennes, France
The highly specialized membranes of the filter chamber found in the digestive tract of some homopteran insects could represent a favorable material for characterizing water channels. In order to demonstrate that membrane proteins of this epithelial complex serve as water channels, we have investigated the membrane permeability for water in Xenopus oocytes injected with RNA isolated from the filter chamber. Volumes of oocytes injected with filter chamber RNA were increased by 16% following a 16-min osmotic shock, while volumes of oocytes injected with RNA from midgut not of Alter chamber or with water were increased only by 8.6 and lo%, respectively. This significant difference in oocyte swelling leads us to conclude that RNA isolated from the Alter chamber contains mRNA coding for water channel proteins. 0 1992 Academic Prese, IIIC.
INTRODUCTION There is much evidence that specialized channels are facilitated pathways for water through many plasma membranes [S]. Nevertheless water channels have not yet been characterized. In amphibian urinary bladder [3] and in mammalian kidney collecting ducts [ 121 they have been associated with particle aggregates which appear inserted into the luminal membrane of the epithelial cells under ADH treatment. In these ADH-sensitive cells, water permeability is regulated by the cycling of functional water channels between the apical membranes and endosomes [lo, 16,191. The fact that some membrane proteins serve as water channels can be demonstrated by the functional expression of these proteins after introduction of their corresponding mRNA into a foreign cell. The most frequently used cell for the expression of plasma membrane proteins is the Xenopus oocyte. In many cases, injection of crude RNA is sufficient to effect functioning 1 To whom reprint requests should be addressed.
of the expressed proteins [17]. Recently, mRNA isolated from red cells, kidney proximal tubules, kidney collecting tubules, and toad urinary bladder has been injected into Xenopus oocytes and expressed, increasing dramatically the water permeability of the oocyte membrane, thus suggesting that water channels are proteins [21, 221. At present, there is no specific marker for water channel proteins. Moreover, in the most extensively investigated model, the amphibian urinary bladder, the particle aggregates cover only l-2% of the apical surface of the epithelial cells. Thus, water channel proteins probably represent only a tiny fraction of the total membrane proteins. Moreover in the amphibian bladder it is difficult to isolate the apical membrane of epithelial cells [7, 181. We have decided to look for a more favorable model for isolating water channel proteins. A material with highly specialized membranes should have the advantage that water channel proteins would appear as major components and therefore the number of candidates for water-conducting proteins should be reduced. Thus, we investigated an epithelial complex found in the digestive tract of some homopteran insects feeding on plant sap. In this complex, called the “filter chamber,” a significant water transfer is believed to occur down a transepithelial osmotic gradient [9,14]. In a previous study on the filter chamber of CicadcUu uiridis, we observed that epithelial cell membranes exhibit a pattern of intramembrane particles, which cover the whole surface of the membranes, and resemble those of the amphibian urinary bladder, which are associated with water permeability modifications. SDS-PAGE performed on purified membranes of the filter chamber revealed the existence of two major protein components, 25 and 75 kDa. These components appear to be specific of the filter chamber since they were not detected in membranes isolated from other parts of the midgut [ll]. To test the hypothesis that some proteins of these membranes serve as water channels, we have achieved 301 Copyright All
rights
0 1992 of reproduction
0014-4827/92 $5.00 by Academic Press, Inc. in any form reserved.
302
GUILLAM
ET
AL.
in the present study the functional expression of RNA isolated from C. viridis filter chambers in Xenopus oocytes. The results reported here indicate that membrane permeability for water is significantly increased in oocytes injected with filter chamber RNA, when compared to oocytes injected with RNA of midgut distinct from the filter chamber or with water and to punctured noninjected oocytes.
5 MATERIAL
AND
METHODS
0 5
Insects. meadows
C. uiridis L. (Homoptera, Jassidae) were harvested from in summer and autumn. RNA. Total RNA was prepared from approximately 200 freshly collected filter chambers or from midguts, using the LiCl/urea procedure [2]. RNA concentrations were evaluated by 260-nm absorption and its integrity was controlled by agarose gel electrophoresis. Expression of mRNA was assayed in a reticulocyte lysate system. Synthesis of the 25-kDa protein, the major membrane protein of the filter chamber, was detected by Western blot, using a rabbit antiserum raised against purified 25-kDa protein. Oocytes. Ovary fragments were surgically removed from previously anesthetized females of Xerwpus Zuevis and stored in OR, physiological solution modified from Wallace et al. [20], which contains (mhf): 82.5 NaCl, 2.5 KCl, 1.0 CaCl,, 1.0 MgCl,, 2.0 NaHCO,, 3.8 NaOH, buffered at pH 7.4-7.5 with 10.0 m&f Hepes. The follicular cell layers were mechanically removed, using fine forceps, from immature full-grown (stage VI of Dumont [4]) oocytes. Defolliculated oocytes, stored at 16°C in OR, until use, were pressure microinjected with RNA solution (see below), using micropipettes drawn from glass capillaries GC 150 (Clark Electromedical Instruments) which had their tips broken to lo-30 pm using a microforge and were calibrated to deliver around 50 nl. Total RNA, from filter chambers (RNAfc) or from parts of midguts distinct from filter chamber (RNAmi), was dissolved in H,O (1 pglpl). Control experiments consisted of microinjetting oocytes with 50 nl of Hz0 or puncturing the oocyte cortex with a glass micropipette. oocyte water permeabdity. Oocytes water permeability was assayed by measuring their volume variation by videomicroscopy. Forty eight hours after injection, oocytes were removed from OR, buffer (176 mosmol) and transferred to a 20-fold diluted OR2 (9 mosmol). Oocytes were monitored using a low magnification objective (~2) on an Olympus microscope equipped with a video camera linked to a Sony image printer. Focus was done on the oocyte equatorial plane and pictures were recorded every 2 min for 16 min. Oocyte area was measured with a digitizing pad linked to a microcomputer. Results were expressed in relative volume oocyte (V/V,) computed from the relative oocyte area (A/A,,) in the focal plane, V/V, = (A/A,J3n, where V, is the initial oocyte volume when transferred into diluted OR, 1221. Statistical analysis. Osmotic water permeability was assayed on 41 oocytes. Results were quoted in a data table of 41 X 8 elements, where each row corresponds to the ratio V/V,, for a given oocyte and each column to the experimental time. In order to extract essential characteristics from each experiment through each time, we carried out principal component analysis (PCA) on this table and results were expressed in a score plot [l, 6, 151. Before applying PCA, data were preprocessed either (1) by normalizing them to zero mean and unit variance, or (2) by only centering the data. Since these two methods gave similar results, we present here results obtained using the first one. Experimental groups were then reexamined according to their scores along the first component: the confidence interval was calculated for each group and outliers were excluded for a confidence level of 0.001. Then they were analyzed by a one-factor analysis of variance, followed by a Dunnett’s test, and a Student t test, using the
-0
10
OJ
I P
I W
0
00
1 Mi
I Fc
FIG. 1. Oocyte response to osmotic shock. Results in relative volume increase per minute for all individual Higher values are recorded for RNAfc-injected oocytes. W, water; Mi, midgut RNA; Fc, filter chamber RNA.
RNAfc-injected time course ment, using
are expressed experiments. P, punctured;
group as control. Finally, an averaged curve of the of oocyte swelling was computed for each kind of experisignificant values.
RESULTS In OR2 medium just after injection, oocytes have a volume of 1.14 +- 0.02 mm3 (mean + SD) (n = 52), and after 48 h in the same medium the volume found was 1.10 5 0.02 mm3 (n = 41), reflecting a slight shrinking, but no significant difference was found between the volume of the different injected or punctured oocyte groups. In response to an osmotic shock there is quite good homogeneity in values computed for the relative volume increase per minute (d( VIV,,)ldt) between oocytes within groups injected with water or punctured only. Results obtained for oocytes injected with RNAmi and RNAfc are apparently more variable, yet with higher values for RNAfc-injected oocytes (Fig. 1). Since a simple calculation of mean and standard deviation will be biased by anomalous values and is not the appropriate method to analyze our data, we have used principal component analysis, a multivariate method able to take into account each V/V, value for each time and for each experiment. The results of such analysis are shown on a score plot in which each dot represents
FUNCTIONAL
EXPRESSION
OF
WATER
CHANNEL
mRNA
IN
303
OOCYTES
n
m
m
m
FI n m
n
0A
*
I
-4
m
-2
I
I
I
I
0
2
4
FIG. 2. Principal component analysis score plot. (A) The analysis was conducted on each oocyte for each time; each dot represents one oocyte. The first component, which accounts for 75% of the total interexperiment variance, corresponds to the predominant differences between oocytes. (B) Correlation circle between the experimental times demonstrating that the classification obtained in A along the first component is highly related to the swelling of the oocytes during the experiment. *, punctured; *, water injected, q , RNAmi, n, RNAfc-injected oocytes.
one oocyte (or experiment) (Fig. ZA). Experiments having a close resemblance are positioned closely on the score plot. In our study the first component accounts for 75% of the total interexperiment variance, while components 2, 3, and 4 account for only 11, 8.5, and 3.6%, respectively. Figure 2B displays the correlation circle between experimental times in the factorial plane of components 1 and 2 and demonstrates that the classification for the experiments along component 1 is strongly related to the experimental values at times 4,6, 8,10,12,14 min, while values for times 2 and 16 min are of less importance. Component 2 reflects variations within the experiments for time 2 and 16 min. Thus, the first component corresponds to a classification of the oocyte population, according to the variations of their relative volume increase. Along the first component there is a continuous dispersion of the points, with a cluster of oocytes from the various experiments in the middle, but there is an obvious location of points corre-
sponding to RNAfc-injected oocytes on the right side of the graph. Groups of reduced variance for each experimental group were determined by computing a confidence interval (P = 0.001) from the group score on component 1. For RNAfc-injected oocytes seven points were excluded, while for RNAmi-injected oocytes one point was excluded. The analysis of variance conducted between groups led us to reject the hypothesis of equality between the means of experimental groups. A Dunnett’s test using the RNAfc group as “control” (taking n = 5 for the number of repeats corresponding to the minimum sample size) revealed two groups for which the mean differs significantly from the RNAfc group: RNAmi-injected oocytes (P < 0.01) and punctured oocytes (P < 0.05). For the water-injected oocytes group, the test did not permit us to reject any hypothesis of equality with the control group. However, a Student t test performed on the same data and relative to the
304
GUILLAM 1.20
I
*o i
l.lO-
LOO2
8 Tim.
16 mln
FIG. 3.
Time course oocyte swelling in response to an osmotic shock. For each kind of experiment, an averaged curve was computed using oocytes belonging to a reduced variance group. The RNAfc-injetted oocytes group is significantly different from the other groups and displays a higher V/V, increase in response to an osmotic shock. Punctured or injected oocytes are fragile and very often ruptured during the experiment, therefore we limited the experimental time to 16 min after transfer to hypotonic medium. *, punctured (n = 6); *, water (n = 5); 0, RNAmi (n = 8); n , RNAfc (n = 14).
RNAfc-injected group revealed that the means for the RNAmi group, the water-injected group, and the punctured group differ significantly from the RNAfc group for P < 0.01, P < 0.05, and P < 0.01, respectively. Using values from each reduced variance group, an average curve was obtained for each experimental group (Fig. 3). Such curves show clearly that there is a higher increase in the ratio V/V, for RNAfc-injected oocytes than for the other experiments. Oocytes injected with RNAmi display a similar volume increase to an osmotic shock, as do water-injected or punctured oocytes. The response to an osmotic shock for 16 min corresponds to a volume increase of 15% for RNAfc-injected oocytes, 8.5% for RNAmi-injected and punctured oocytes, and 10% for water-injected oocytes.
ET AL.
eliminated. Our previous observations on the filter chamber of C. viridis revealed membranes extraordinarily infolded [9] and highly specialized in both structure and composition [ll]. The volume of water transferred through the filter chamber and the rapidity of transfer strongly suggested that the membranes of filter chambers are specialized for water transfer. The goal of this study was to demonstrate that these membranes possess proteic water channels. The present study shows that total RNA, isolated from filter chambers, and injected into Xenopus oocytes, significantly increases oocyte water permeability, while RNA isolated from parts of midgut not of the filter chamber has no significant effect. This leads us to conclude that total RNA, isolated from the filter chamber, contains mRNA coding for water-conducting proteins. Moreover, since the percentage of mRNA in total RNA is very small, the quantity of mRNA entering the oocyte is limited. The significant response observed in water permeability for Xenopus oocytes could reflect that RNA from filter chambers is highly enriched in mRNA coding for water channel proteins or is better expressed. These results do not preclude that such proteins are solely water channels and could not act as ionic channels. However there is a large body of evidence which indicates that the major permeation routes across the membrane for water and anions are independent [8]. Since the major components of the membrane fraction of the filter chamber are two polypeptides of 25 and 75 kDa, it is tempting to associate water channels with one or both of these components. A high resolution electron microscopy study is in progress to characterize the fine structure of the membrane particles. Using monoclonal and polyclonal antibodies raised against the 25and 75-kDa components we expect to be able to map the polypeptides within the native membrane particle. Molecular cloning of these two components will be a necessary step to establish their possible relationship with water channels. We thank Dr. Michel Charbonneau for an informative discussion on the experimental procedures and a careful reading of the manuscript. This work was supported by a grant from INSERM(CRE 897003).
REFERENCES 1.
DISCUSSION
Since the early studies of Licent [13], the filter chamber found in the digestive tract of homopteran insects is considered as a water-shunting complex which permits the transfer of the excess dietary water from the initial midgut to the terminal midgut or the Malpighian tubules, in order to rapidly reach the hindgut and be
Anderson, T. W. (1958) An Introduction tical Analysis, Wiley, New York.
2. Auffray,
C., and Rougeon,
F. (1980)
to Multivariate
Statis-
Eur. J. Biochem. 107,303-
314.
3. Chevalier, J., Bourguet, J., and Hugon, J. S. (1974) Res. 152,129-140. 4. Dumont, J. N. (1972) J. Morphd. 136,153-180. 5. Dunnett, C. W. (1964) Biometrics 20, 482-491. 6. Escofier, B., and PagBs, J. (1988) Analyses factorielles multiples,
p. 7-24.
Dunod,
Paris.
Cell Tissue
simples
et
FUNCTIONAL 7.
8. 9. 10. 11. 12. 13. 14.
EXPRESSION
OF WATER
Favard, P., Favard, N., Zhu, Q. L., Bourguet, J., and Lechaire, J. P. (1989) Biol. Cell 66,99-106. Finkelstein, A. (1987) Water Movement Through Lipid Bilayers, Pores and Plasma Membranes: Theory and Reality, p. 153-222. Wiley, New York. Gouranton, J. (1968) J. Microscopic (Paris) 7, 559-574. Harris, H. W., and Handler, J. S. (1988) J. Membr. Biol. 103, 207-216. Hubert, J. F., Thomas, D., Cavalier, A., and Gouranton, J. (1989) Biol. Cell 66, 155-163. Kubat, B., Lorenzen, M., and Reale, E. (1989) Bid. Cell 66,5963. Licent, E. (1912) La Cell& (Louvain) 28,1-161. Marshall, A. T. (1983) Cell Tissue Res. 231, 215-217.
Received November 11, 1991 Revised version received January 27, 1992
CHANNEL
mRNA IN OOCYTES
305
15. Pearson, K. (1901) Philos. Mag. 2,559-572. 16. Shi, L. B., Brown, D., and Verkman, A. S. (1990) J. Gen. Physiol. 95,941-960. 17. Sigel, E. (1990) Membr. Biol. 117, 201-221. 18. Verbavatz, J. M., Calamita, G., Hugon, J. S., and Bourguet, J. (1989)
Biol.
Cell 66,91-97.
19. Verkman, A. S., Lencer, W. I., Brown, D., and Ausiello, D. A. (1988) Nature 333, 268-269. 20. Wallace, R. A., Jared, D. W., Dumont, J. N., and Sega, M. W. (1973) J. Exp. 2001. 184,321-334. 21. Zhang, R., and Verkman, A. S. (1991) Am. J. Physiol. 30, C26c34. 22. Zhang, R., Logee, K. A., and Verkman, A. S. (1990) J. Biol. Chem. 266, 15,375-l&378.
CELL
RESEARCH
200,301-305
(l%&?)
Expression of RNA Isolated from the Water-Shunting Complex of a SapSucking Insect Increases the Membrane Permeability for Water in Xenopus Oocytes MARIE-TH~~~SEGUILLAM,FABIENNEBEURON,NATHALIEGRANDIN, JEAN-FFUNCOISHUBERT, CLAUDE BOISSEAU, ANNIE CAVALIER, ANNE COUTURIER, JEAN GOURANTON, AND DANIEL THOMAS’ Laboratoire de Biologic Cellulnire URA CNRS No. 256, Universit& de Rennes 1, Campus de Beaulieu, 35000 Rennes, France
The highly specialized membranes of the filter chamber found in the digestive tract of some homopteran insects could represent a favorable material for characterizing water channels. In order to demonstrate that membrane proteins of this epithelial complex serve as water channels, we have investigated the membrane permeability for water in Xenopus oocytes injected with RNA isolated from the filter chamber. Volumes of oocytes injected with filter chamber RNA were increased by 16% following a 16-min osmotic shock, while volumes of oocytes injected with RNA from midgut not of Alter chamber or with water were increased only by 8.6 and lo%, respectively. This significant difference in oocyte swelling leads us to conclude that RNA isolated from the Alter chamber contains mRNA coding for water channel proteins. 0 1992 Academic Prese, IIIC.
INTRODUCTION There is much evidence that specialized channels are facilitated pathways for water through many plasma membranes [S]. Nevertheless water channels have not yet been characterized. In amphibian urinary bladder [3] and in mammalian kidney collecting ducts [ 121 they have been associated with particle aggregates which appear inserted into the luminal membrane of the epithelial cells under ADH treatment. In these ADH-sensitive cells, water permeability is regulated by the cycling of functional water channels between the apical membranes and endosomes [lo, 16,191. The fact that some membrane proteins serve as water channels can be demonstrated by the functional expression of these proteins after introduction of their corresponding mRNA into a foreign cell. The most frequently used cell for the expression of plasma membrane proteins is the Xenopus oocyte. In many cases, injection of crude RNA is sufficient to effect functioning 1 To whom reprint requests should be addressed.
of the expressed proteins [17]. Recently, mRNA isolated from red cells, kidney proximal tubules, kidney collecting tubules, and toad urinary bladder has been injected into Xenopus oocytes and expressed, increasing dramatically the water permeability of the oocyte membrane, thus suggesting that water channels are proteins [21, 221. At present, there is no specific marker for water channel proteins. Moreover, in the most extensively investigated model, the amphibian urinary bladder, the particle aggregates cover only l-2% of the apical surface of the epithelial cells. Thus, water channel proteins probably represent only a tiny fraction of the total membrane proteins. Moreover in the amphibian bladder it is difficult to isolate the apical membrane of epithelial cells [7, 181. We have decided to look for a more favorable model for isolating water channel proteins. A material with highly specialized membranes should have the advantage that water channel proteins would appear as major components and therefore the number of candidates for water-conducting proteins should be reduced. Thus, we investigated an epithelial complex found in the digestive tract of some homopteran insects feeding on plant sap. In this complex, called the “filter chamber,” a significant water transfer is believed to occur down a transepithelial osmotic gradient [9,14]. In a previous study on the filter chamber of CicadcUu uiridis, we observed that epithelial cell membranes exhibit a pattern of intramembrane particles, which cover the whole surface of the membranes, and resemble those of the amphibian urinary bladder, which are associated with water permeability modifications. SDS-PAGE performed on purified membranes of the filter chamber revealed the existence of two major protein components, 25 and 75 kDa. These components appear to be specific of the filter chamber since they were not detected in membranes isolated from other parts of the midgut [ll]. To test the hypothesis that some proteins of these membranes serve as water channels, we have achieved 301 Copyright All
rights
0 1992 of reproduction
0014-4827/92 $5.00 by Academic Press, Inc. in any form reserved.
302
GUILLAM
ET
AL.
in the present study the functional expression of RNA isolated from C. viridis filter chambers in Xenopus oocytes. The results reported here indicate that membrane permeability for water is significantly increased in oocytes injected with filter chamber RNA, when compared to oocytes injected with RNA of midgut distinct from the filter chamber or with water and to punctured noninjected oocytes.
5 MATERIAL
AND
METHODS
0 5
Insects. meadows
C. uiridis L. (Homoptera, Jassidae) were harvested from in summer and autumn. RNA. Total RNA was prepared from approximately 200 freshly collected filter chambers or from midguts, using the LiCl/urea procedure [2]. RNA concentrations were evaluated by 260-nm absorption and its integrity was controlled by agarose gel electrophoresis. Expression of mRNA was assayed in a reticulocyte lysate system. Synthesis of the 25-kDa protein, the major membrane protein of the filter chamber, was detected by Western blot, using a rabbit antiserum raised against purified 25-kDa protein. Oocytes. Ovary fragments were surgically removed from previously anesthetized females of Xerwpus Zuevis and stored in OR, physiological solution modified from Wallace et al. [20], which contains (mhf): 82.5 NaCl, 2.5 KCl, 1.0 CaCl,, 1.0 MgCl,, 2.0 NaHCO,, 3.8 NaOH, buffered at pH 7.4-7.5 with 10.0 m&f Hepes. The follicular cell layers were mechanically removed, using fine forceps, from immature full-grown (stage VI of Dumont [4]) oocytes. Defolliculated oocytes, stored at 16°C in OR, until use, were pressure microinjected with RNA solution (see below), using micropipettes drawn from glass capillaries GC 150 (Clark Electromedical Instruments) which had their tips broken to lo-30 pm using a microforge and were calibrated to deliver around 50 nl. Total RNA, from filter chambers (RNAfc) or from parts of midguts distinct from filter chamber (RNAmi), was dissolved in H,O (1 pglpl). Control experiments consisted of microinjetting oocytes with 50 nl of Hz0 or puncturing the oocyte cortex with a glass micropipette. oocyte water permeabdity. Oocytes water permeability was assayed by measuring their volume variation by videomicroscopy. Forty eight hours after injection, oocytes were removed from OR, buffer (176 mosmol) and transferred to a 20-fold diluted OR2 (9 mosmol). Oocytes were monitored using a low magnification objective (~2) on an Olympus microscope equipped with a video camera linked to a Sony image printer. Focus was done on the oocyte equatorial plane and pictures were recorded every 2 min for 16 min. Oocyte area was measured with a digitizing pad linked to a microcomputer. Results were expressed in relative volume oocyte (V/V,) computed from the relative oocyte area (A/A,,) in the focal plane, V/V, = (A/A,J3n, where V, is the initial oocyte volume when transferred into diluted OR, 1221. Statistical analysis. Osmotic water permeability was assayed on 41 oocytes. Results were quoted in a data table of 41 X 8 elements, where each row corresponds to the ratio V/V,, for a given oocyte and each column to the experimental time. In order to extract essential characteristics from each experiment through each time, we carried out principal component analysis (PCA) on this table and results were expressed in a score plot [l, 6, 151. Before applying PCA, data were preprocessed either (1) by normalizing them to zero mean and unit variance, or (2) by only centering the data. Since these two methods gave similar results, we present here results obtained using the first one. Experimental groups were then reexamined according to their scores along the first component: the confidence interval was calculated for each group and outliers were excluded for a confidence level of 0.001. Then they were analyzed by a one-factor analysis of variance, followed by a Dunnett’s test, and a Student t test, using the
-0
10
OJ
I P
I W
0
00
1 Mi
I Fc
FIG. 1. Oocyte response to osmotic shock. Results in relative volume increase per minute for all individual Higher values are recorded for RNAfc-injected oocytes. W, water; Mi, midgut RNA; Fc, filter chamber RNA.
RNAfc-injected time course ment, using
are expressed experiments. P, punctured;
group as control. Finally, an averaged curve of the of oocyte swelling was computed for each kind of experisignificant values.
RESULTS In OR2 medium just after injection, oocytes have a volume of 1.14 +- 0.02 mm3 (mean + SD) (n = 52), and after 48 h in the same medium the volume found was 1.10 5 0.02 mm3 (n = 41), reflecting a slight shrinking, but no significant difference was found between the volume of the different injected or punctured oocyte groups. In response to an osmotic shock there is quite good homogeneity in values computed for the relative volume increase per minute (d( VIV,,)ldt) between oocytes within groups injected with water or punctured only. Results obtained for oocytes injected with RNAmi and RNAfc are apparently more variable, yet with higher values for RNAfc-injected oocytes (Fig. 1). Since a simple calculation of mean and standard deviation will be biased by anomalous values and is not the appropriate method to analyze our data, we have used principal component analysis, a multivariate method able to take into account each V/V, value for each time and for each experiment. The results of such analysis are shown on a score plot in which each dot represents
FUNCTIONAL
EXPRESSION
OF
WATER
CHANNEL
mRNA
IN
303
OOCYTES
n
m
m
m
FI n m
n
0A
*
I
-4
m
-2
I
I
I
I
0
2
4
FIG. 2. Principal component analysis score plot. (A) The analysis was conducted on each oocyte for each time; each dot represents one oocyte. The first component, which accounts for 75% of the total interexperiment variance, corresponds to the predominant differences between oocytes. (B) Correlation circle between the experimental times demonstrating that the classification obtained in A along the first component is highly related to the swelling of the oocytes during the experiment. *, punctured; *, water injected, q , RNAmi, n, RNAfc-injected oocytes.
one oocyte (or experiment) (Fig. ZA). Experiments having a close resemblance are positioned closely on the score plot. In our study the first component accounts for 75% of the total interexperiment variance, while components 2, 3, and 4 account for only 11, 8.5, and 3.6%, respectively. Figure 2B displays the correlation circle between experimental times in the factorial plane of components 1 and 2 and demonstrates that the classification for the experiments along component 1 is strongly related to the experimental values at times 4,6, 8,10,12,14 min, while values for times 2 and 16 min are of less importance. Component 2 reflects variations within the experiments for time 2 and 16 min. Thus, the first component corresponds to a classification of the oocyte population, according to the variations of their relative volume increase. Along the first component there is a continuous dispersion of the points, with a cluster of oocytes from the various experiments in the middle, but there is an obvious location of points corre-
sponding to RNAfc-injected oocytes on the right side of the graph. Groups of reduced variance for each experimental group were determined by computing a confidence interval (P = 0.001) from the group score on component 1. For RNAfc-injected oocytes seven points were excluded, while for RNAmi-injected oocytes one point was excluded. The analysis of variance conducted between groups led us to reject the hypothesis of equality between the means of experimental groups. A Dunnett’s test using the RNAfc group as “control” (taking n = 5 for the number of repeats corresponding to the minimum sample size) revealed two groups for which the mean differs significantly from the RNAfc group: RNAmi-injected oocytes (P < 0.01) and punctured oocytes (P < 0.05). For the water-injected oocytes group, the test did not permit us to reject any hypothesis of equality with the control group. However, a Student t test performed on the same data and relative to the
304
GUILLAM 1.20
I
*o i
l.lO-
LOO2
8 Tim.
16 mln
FIG. 3.
Time course oocyte swelling in response to an osmotic shock. For each kind of experiment, an averaged curve was computed using oocytes belonging to a reduced variance group. The RNAfc-injetted oocytes group is significantly different from the other groups and displays a higher V/V, increase in response to an osmotic shock. Punctured or injected oocytes are fragile and very often ruptured during the experiment, therefore we limited the experimental time to 16 min after transfer to hypotonic medium. *, punctured (n = 6); *, water (n = 5); 0, RNAmi (n = 8); n , RNAfc (n = 14).
RNAfc-injected group revealed that the means for the RNAmi group, the water-injected group, and the punctured group differ significantly from the RNAfc group for P < 0.01, P < 0.05, and P < 0.01, respectively. Using values from each reduced variance group, an average curve was obtained for each experimental group (Fig. 3). Such curves show clearly that there is a higher increase in the ratio V/V, for RNAfc-injected oocytes than for the other experiments. Oocytes injected with RNAmi display a similar volume increase to an osmotic shock, as do water-injected or punctured oocytes. The response to an osmotic shock for 16 min corresponds to a volume increase of 15% for RNAfc-injected oocytes, 8.5% for RNAmi-injected and punctured oocytes, and 10% for water-injected oocytes.
ET AL.
eliminated. Our previous observations on the filter chamber of C. viridis revealed membranes extraordinarily infolded [9] and highly specialized in both structure and composition [ll]. The volume of water transferred through the filter chamber and the rapidity of transfer strongly suggested that the membranes of filter chambers are specialized for water transfer. The goal of this study was to demonstrate that these membranes possess proteic water channels. The present study shows that total RNA, isolated from filter chambers, and injected into Xenopus oocytes, significantly increases oocyte water permeability, while RNA isolated from parts of midgut not of the filter chamber has no significant effect. This leads us to conclude that total RNA, isolated from the filter chamber, contains mRNA coding for water-conducting proteins. Moreover, since the percentage of mRNA in total RNA is very small, the quantity of mRNA entering the oocyte is limited. The significant response observed in water permeability for Xenopus oocytes could reflect that RNA from filter chambers is highly enriched in mRNA coding for water channel proteins or is better expressed. These results do not preclude that such proteins are solely water channels and could not act as ionic channels. However there is a large body of evidence which indicates that the major permeation routes across the membrane for water and anions are independent [8]. Since the major components of the membrane fraction of the filter chamber are two polypeptides of 25 and 75 kDa, it is tempting to associate water channels with one or both of these components. A high resolution electron microscopy study is in progress to characterize the fine structure of the membrane particles. Using monoclonal and polyclonal antibodies raised against the 25and 75-kDa components we expect to be able to map the polypeptides within the native membrane particle. Molecular cloning of these two components will be a necessary step to establish their possible relationship with water channels. We thank Dr. Michel Charbonneau for an informative discussion on the experimental procedures and a careful reading of the manuscript. This work was supported by a grant from INSERM(CRE 897003).
REFERENCES 1.
DISCUSSION
Since the early studies of Licent [13], the filter chamber found in the digestive tract of homopteran insects is considered as a water-shunting complex which permits the transfer of the excess dietary water from the initial midgut to the terminal midgut or the Malpighian tubules, in order to rapidly reach the hindgut and be
Anderson, T. W. (1958) An Introduction tical Analysis, Wiley, New York.
2. Auffray,
C., and Rougeon,
F. (1980)
to Multivariate
Statis-
Eur. J. Biochem. 107,303-
314.
3. Chevalier, J., Bourguet, J., and Hugon, J. S. (1974) Res. 152,129-140. 4. Dumont, J. N. (1972) J. Morphd. 136,153-180. 5. Dunnett, C. W. (1964) Biometrics 20, 482-491. 6. Escofier, B., and PagBs, J. (1988) Analyses factorielles multiples,
p. 7-24.
Dunod,
Paris.
Cell Tissue
simples
et
FUNCTIONAL 7.
8. 9. 10. 11. 12. 13. 14.
EXPRESSION
OF WATER
Favard, P., Favard, N., Zhu, Q. L., Bourguet, J., and Lechaire, J. P. (1989) Biol. Cell 66,99-106. Finkelstein, A. (1987) Water Movement Through Lipid Bilayers, Pores and Plasma Membranes: Theory and Reality, p. 153-222. Wiley, New York. Gouranton, J. (1968) J. Microscopic (Paris) 7, 559-574. Harris, H. W., and Handler, J. S. (1988) J. Membr. Biol. 103, 207-216. Hubert, J. F., Thomas, D., Cavalier, A., and Gouranton, J. (1989) Biol. Cell 66, 155-163. Kubat, B., Lorenzen, M., and Reale, E. (1989) Bid. Cell 66,5963. Licent, E. (1912) La Cell& (Louvain) 28,1-161. Marshall, A. T. (1983) Cell Tissue Res. 231, 215-217.
Received November 11, 1991 Revised version received January 27, 1992
CHANNEL
mRNA IN OOCYTES
305
15. Pearson, K. (1901) Philos. Mag. 2,559-572. 16. Shi, L. B., Brown, D., and Verkman, A. S. (1990) J. Gen. Physiol. 95,941-960. 17. Sigel, E. (1990) Membr. Biol. 117, 201-221. 18. Verbavatz, J. M., Calamita, G., Hugon, J. S., and Bourguet, J. (1989)
Biol.
Cell 66,91-97.
19. Verkman, A. S., Lencer, W. I., Brown, D., and Ausiello, D. A. (1988) Nature 333, 268-269. 20. Wallace, R. A., Jared, D. W., Dumont, J. N., and Sega, M. W. (1973) J. Exp. 2001. 184,321-334. 21. Zhang, R., and Verkman, A. S. (1991) Am. J. Physiol. 30, C26c34. 22. Zhang, R., Logee, K. A., and Verkman, A. S. (1990) J. Biol. Chem. 266, 15,375-l&378.