Main

The human fetus accumulates some 160 mmol of chloride throughout gestation(1), and it is likely that most of this is transferred across the placenta. The syncytiotrophoblast is the main barrier to the transfer of small solutes across the placenta, and data from both in vivo and in vitro studies of placenta at term suggest that there are potentially two pathways for solute transfer across this barrier in either direction. They are a paracellular route(2,3) and a transcellular route utilizing membrane transport proteins to cross both the maternal facing plasma MVM and the fetal facing plasma BM(4). The diffusional permeability of the human placenta to inert hydrophilic solutes at term, estimated from in vivo studies(2,3), suggests that there may be large bidirectional paracellular fluxes of chloride across the placenta. An in vitro study of transplacental chloride clearance, using the isolated, dually perfused, human placental cotyledon, supports this by showing that a large component of unidirectional maternofetal chloride clearance occurs via routes uninhibitable by the chloride transport blockers DIDS and DPC(5). However, this study also provided evidence that there is a small inhibitable transcellular component to maternofetal chloride clearance. This was in accord with studies on chloride uptake by MVM vesicles prepared from term human placenta, which had demonstrated the presence in these membranes of an anion exchanger, sensitive to both DIDS and DPC(6–8), together with a DIDS-sensitive voltage-dependent chloride conductance and a separate DPC-sensitive chloride conductance(9). More recently, preliminary uptake studies using BM vesicles prepared from term placenta have also suggested the presence of a chloride conductance that is sensitive to both DIDS and DPC(10).

Studies of human transplacental chloride transfer have been confined to work on placentas at term, by which time fetal growth has slowed compared with earlier in gestation and when net accretion of chloride by the fetus may be relatively small. We have previously shown that the activity of the system A amino acid transporter and the Na+/H+ exchanger is lower in MVM vesicles formed from first trimester placentas than those from term placentas(11). However, there have been no studies of chloride transporter activity in either MVM or BM from placentas other than at term.

Therefore, the aims of this study were to examine and characterize the chloride-transporting moieties present in first trimester human syncytiotrophoblast MVM and to compare these with those of term MVM. To do this we have measured chloride uptake into vesicles prepared from MVM of both first trimester and term placentas in the presence and absence of the inhibitors DIDS and DPC and in the presence and absence of a membrane potential. We also performed Western blotting to compare the expression of the AE-1 protein in the MVM of first trimester and term placentas.

METHODS

Preparation of MVM vesicles. These studies were approved by the local Ethical Committee and were carried out in Manchester, UK, except for Western blotting, which was done in Goteborg. MVM vesicles were prepared from first trimester and term human placentas by methods previously described(11,12). Briefly, first trimester tissue was obtained from surgical therapeutic terminations of pregnancy, by the method of suction and curettage under general anesthesia and were performed for psychosocial reasons on healthy women at a gestation of 7-13 wk. Placentas were pooled until adequate amounts of tissue (>50 g) were available for processing. Due to the rapid increase in placental weight during this period of gestation, the majority of tissue used in each preparation came from placentas from pregnancies of 10-13-wk gestation. The umbilical cord and its insertion were removed, and the tissue was stored in cold (4°C) homogenization buffer (300 mM mannitol, 10 mM HEPES, 1 mM MgSO4, pH 7.4, with Tris) for up to 4 h before homogenization. Preliminary experiments using term placentas suggested that this form of storage did not affect the properties of the MVM. Term tissue came from single placentas after normal vaginal delivery or elective cesarean section at term (37-42 wk) and after normal pregnancy. The umbilical cord and chorion were completely removed, and 100-200 g of diced tissue were used in each preparation. First trimester and term MVM vesicles were prepared by homogenization in homogenization buffer followed by magnesium precipitation (10 mM MgCl2) and differential centrifugation, all at 4°C. The process was repeated, and the final pellet was suspended in IVB (160 mM sucrose, 6.6 mM HEPES, 10 mM HCl, pH adjusted to 7.5 with 10 M KOH) by repeatedly passing the suspension through a 25-gauge needle. As in the previous studies, the protein content of the vesicle suspension was assayed as was alkaline phosphatase activity(12); alkaline phosphatase enrichment was expressed as the ratio of the activity per mg of protein in the vesicle suspension to that in placental homogenate. Any preparation with an alkaline phosphatase enrichment below 10-fold was excluded from the study. Each MVM vesicle suspension was stored at either 4°C for less than 48 h for use in uptake experiments, or stored frozen at -80°C before use in Western blotting studies.

36Cl- uptake experiments. The method used has been previously described(9) and was adapted from the method of Shennan et al.(6). Experiments were carried out at room temperature and in duplicate when quantities of MVM allowed. A MVM vesicle suspension was diluted with IVB to a uniform concentration of 10 mg of protein/mL, incubated for 1 h with valinomycin at a concentration of 20 nmol/mg of protein, and then 90 µL of this vesicle suspension was mixed with 210 µL of EVB (10 mM H36Cl (Amersham International), 6.6 mM HEPES, 160 mM sucrose, pH 7.5, with KOH) for a set time (see "Results"). An aliquot of 200 µL of this suspension was then passed through an anion exchange column formed from a disposable Pasteur pipette plugged with aquarium floss and filled with strongly basic anion-exchange resin (chloride form, 50-100 mesh, 8% cross-linked, Sigma Chemical Co. Co, Dorset, UK). The 36Cl- content of the eluent was estimated by scintillation counting with correction for quenching(Packard 2000CA), after addition of scintillation fluid (Optiphase II, Wallac). Recovery of vesicles from the anion exchange resin (measured in terms of recovery of alkaline phosphatase activity), efficiency of the anion exchange resin, and background radiation were all taken into consideration in calculating the chloride uptake by the vesicles (nmol/mg of protein).

DIDS and DPC (from concentrated solution in DMSO, i.e. 10 and 50 mM, respectively) were added to the EVB in some experiments, to concentrations of 0.1 and 0.5 mM, respectively (previously shown to be maximally effective concentrations in blocking 36Cl uptake by term vesicles(6,7,9). An equivalent volume of DMSO was added to control experiments. In some experiments, 27-30 mM potassium gluconate was added to the EVB, substituting iso-osmotically for sucrose, to create a 25-mV (inside positive) electrical potential difference across the membrane, as calculated according to the Nernst equation. To test that chloride uptake was occurring into an osmotically active space, experiments were also conducted with a range of osmolarities of EVB, effected by varying the sucrose concentration(9).

Western blotting. Membrane proteins were separated by PAGE in the presence of SDS/lauryl sulfate PAGE as follows. Vesicle suspension was thawed on ice and diluted with a sucrose buffer (250 mM sucrose, 10 mM HEPES, pH 7.4, with Tris) to a final concentration of 0.5 mg/mL. One volume of sample buffer (8 M urea, 5% SDS, 0.04% bromphenol blue, 455 mM DTT, 50 mM Tris, pH 6.8, with HCl) was then added to 2 volumes of this diluted vesicle suspension and mixed thoroughly. The Bio-Rad Mini-Protean II electrophoresis system (Bio-Rad Laboratories, Hemel Hempstead, Herts, UK) was used for electrophoresis; 5 µg of MVM protein were loaded into wells 2-9 of a 10-well, 7.5% polyacrylamide, 0.375% Tris/HCl precast ready gel (Bio-Rad Laboratories) which was mounted into a holding cassette surrounded and filled by electrophoresis buffer (0.3% Tris, 1.44% glycine, 0.1% SDS). SDS molecular weight markers (Sigma Chemical Co.) were loaded in lane 1, and equal protein loadings of either term placental homogenate or blood cell ghost membrane were loaded in lane 10 as positive controls. Electrophoresis, at 200 mV, was continued until the smallest molecular weight marker ran off the end of the gel (approximately 30 min).

The proteins were transferred from the gel to a nitrocellulose sheet(Hybond-ECL, Amersham International) using the Bio-Rad Mini Trans-Blot electrophoresis transfer cell (Bio-Rad Laboratories). Gels were initially washed in transfer buffer (0.30% Tris, 1.44% glycine, 20% vol/vol methanol in H2O) for 30 min. A Bio-Ice cooling unit (Bio-Rad) was incorporated. Transfer was carried out at 30 mV for 12-18 h, and then the nitrocellulose sheet was immersed in PBS (228 mM Na, 27 mM H2PO4, 90 mM HPO4, pH adjusted to 7.4 with 3 N NaOH) containing 0.1% Tween 20 before blotting and probing.

The nitrocellulose sheet was immersed for 60 min in a fresh, filtered solution of 5% fat-free cow's milk (Semper, Stockholm, Sweden) made up in 0.1% Tween in PBS and then washed in 0.1% Tween in PBS. To probe for AE-1 protein, the sheet was then covered with a 1:2000 vol/vol solution of mouse monoclonal anti-human Band 3 antibody (Sigma Chemical Co.) in 0.1% Tween, 0.05% thimerosal in PBS and gently shaken for 1 h. The negative control gel was covered with 0.1% Tween in PBS for 1 h. The sheet was again washed in 0.1% Tween in PBS and then covered with secondary antibody, a rabbit peroxidase-conjugated anti-mouse IgG (Sigma Chemical Co.) at a dilution of 1:1000 vol/vol in 0.1% Tween, 0.05% thimerosal in PBS. The sheet was again gently shaken for 1 h and then washed with 0.1% Tween in PBS.

Detection of labeled secondary antibody was by an enhanced chemiluminescence system (ECL, Amersham International) and using Hyperfilm ECL photographic film (Amersham International). The film was developed using Ilford Multigrade Developer (Ilford, UK), dilute acetic acid as stop and Ilfospeed 2 Fixer (Ilford, UK). A plot of the distance from the end of the blot to the position of the band for each molecular weight marker, versus the molecular weight of the marker was fitted to a best-fit curve (Calcurve, WG Bardsley, University of Manchester, UK). The distance of a detected protein in the sample lanes from the edge of the blot was then calculated, and the molecular weight of that protein was then estimated by using the standard curve. Blots were photographed and stored onto a disk using a computer program (Neotech Image Grabber 2.1, Neotech, Hampshire, UK), and the photograph then used for analysis by densitometry by using a computer program (IPLab Gel, Vienna, VA). The mean volume density of bands from first trimester and term MVM was then calculated.

Statistical analysis. Data are expressed as mean ± SEM where n is the number of different vesicle preparations for both first trimester and term experiments. Statistical analyses used were repeated measures ANOVA followed by t tests with Bonferroni correction, and paired or unpaired t tests as appropriate. The figures show SEM error bars unless they fall within the symbol. Statistical analysis was performed by a computer software package (Instat2, Graphpad, San Diego, CA).

RESULTS

Preparation of MVM Vesicles

Fifteen first trimester and eight term MVM vesicle preparations were made and used in this study; one preparation of first trimester MVM vesicles was excluded from the study because alkaline phosphatase enrichment (9-fold) was below our exclusion criterion. The alkaline phosphatase enrichment of the first trimester and term MVM vesicle preparations used was 19.3 ± 1.1-fold (n = 15) and 17.8 ± 2.1-fold (n = 8), respectively, and these were not significantly different. The protein recovery from first trimester tissue was 0.13 ± 0.01 mg/g of placenta(n = 15), significantly lower (p < 0.05) than that from term tissue 0.27 ± 0.03 mg/g of placenta (n = 8).

36Cl- Uptake Experiments

Protein recovery from the anion exchange columns. The fraction of total vesicle protein recovered from the columns of anion exchange resin was 34 ± 2% (n = 12) for first trimester vesicles and 36 ± 4% (n = 7) for term vesicles.

Time course of 36Cl- uptake.Figure 1 shows the time course of36 Cl- uptake in the absence of inhibitors and with zero imposed membrane potential. The data show clearly that equilibrium uptake is reached by 1 h for both first trimester and term vesicles. However, it was not clear from the data whether uptake at 30 s approximated the initial rate because uptake at 60 s appeared to deviate from linearity. Therefore two further experiments, one with first trimester and one with term MVM vesicles, with more time points were performed as illustrated in the inset to Figure 1. In these experiments there was linearity of uptake between 15 and 45 s (first trimester, r = 0.98; term, r = 1.00) suggesting that 36Cl- uptake at 30 s does approximate to initial rate for the purpose of this study.

Figure 1
figure 1

36Cl- uptake plotted against time for first trimester and term vesicles formed from the microvillous membrane of placental syncytiotrophoblast. The main graph shows mean ± SEM with n = 3 preparations each of first trimester and term MVM vesicles. Inset, individual experiments with each of first trimester and term vesicles showing 36Cl- uptake plotted against time with linear regression analysis applied to uptakes from 15 to 45 s.

Full size image

Equilibrium uptakes and membrane binding. The equilibrium uptake at 1 h for first trimester MVM vesicles was significantly higher than that for term vesicles (7.44 ± 0.83 nmol/mg of protein, n = 10 versus 4.16 ± 0.16 nmol/mg of protein, n = 7, respectively; p < 0.01, t test). Assuming that, at equilibrium, 36Cl- concentration is the same inside as outside the vesicle and is 6.8 mM (10 mM 36Cl- in EVB, diluted 210/300 by IVB/vesicle suspension), then the intravesicular volume can be calculated. Expressed per mg of membrane protein, it is significantly higher for first trimester than for term vesicles (1.09 ± 0.12µL/mg protein, n = 10 versus 0.61 ± 0.02µL/mg protein, n = 7, respectively; p < 0.01, t test).

Figure 2 shows the equilibrium36 Cl- uptake plotted against the reciprocal function of the extravesicular osmolarity. The linear relationship between uptake and the reciprocal function of osmolarity suggests that chloride is entering an intravesicular space. The y intercept provides an estimate of the binding of 36Cl- to membranes at equilibrium (0.31 nmol/mg of protein for first trimester vesicles, 0.24 nmol/mg of protein for term vesicles). The binding was therefore 4 and 5% of total equilibrium uptake in isosmolar EVB, for first trimester and term vesicles, respectively.

Figure 2
figure 2

Equilibrium (1 h)36 Cl- uptake plotted against the reciprocal of the osmolarity of EVB for first trimester and term vesicles formed from the microvillous membrane of syncytiotrophoblast (mean ± SEM). Linear regression analysis was fitted to each set of data. For first trimester data, r = 1.00, p < 0.01, y intercept = 0.31 nmol/mg of protein. For term data, r = 0.99, p < 0.01, y intercept = 0.24 nmol/mg of protein.

Full size image

The effects of Cl- transport inhibitors. The effects of inhibitors on 36Cl- uptake by vesicles are displayed in Table 1. It can be seen that, in first trimester vesicles, the presence of DIDS or DPC significantly reduced36 Cl- uptake at 0 mV. DIDS and DPC together had a similar effect to DPC alone. In the presence of valinomycin and a K+ gradient to impose a 25-mV, inside-positive transmembrane pd, control36 Cl- uptake at 30 s by the first trimester vesicles was significantly higher than with a 0 mV pd (p < 0.01). At 25 mV pd, there was no statisically significant change in uptake when DIDS was present. However, first trimester uptake in the presence of DPC at this membrane potential was significantly lower than that with no inhibitor or that in the presence of DIDS.

Table 1 36Cl-uptake (mean ± SEM, nmol/mg protein) at 30 s for MVM vesicles formed from first trimester (n = 8) or term(n = 6) placentas
Full size table

In term vesicles at 0 mV, DIDS, DPC, or both DIDS and DPC together significantly reduced 36Cl- uptake by similar amounts. In the presence of a 25 mV pd, the 30-s control 36Cl- uptake at 30 s was higher than at 0 mV (p < 0.01), and DIDS, DPC, and their combination significantly reduced 36Cl- uptake compared with control. At 25 mV, DPC reduced uptake significantly more than did DIDS, in contrast to the experiments conducted at 0 mV. Again, the mean uptake by term vesicles in the presence of both inhibitors was similar to that with DPC alone.

At 0 mV, control (total) 36Cl- uptake at 30 s into first trimester vesicles was higher (p < 0.05) than that into term vesicles, but there was no difference between the two groups at 25 mV(Table 1). To directly compare DIDS-sensitive and DPC-sensitive uptakes in the first trimester vesicles with those uptakes at term, the inhibition of 36Cl- uptake was calculated as(control uptake) - (uptake in the presence of inhibitor); the results are illustrated in Figure 3. In agreement with statistical analysis of the mean uptakes displayed in Table 1, DPC-sensitive uptake was significantly greater than DIDS-sensitive uptake in first trimester vesicles at 25 mV and in term vesicles at 25 mV, but there was no difference between DIDS-sensitive and DPC-sensitive uptakes in either set of vesicles at 0 mV. The inhibitor-sensitive uptakes into first trimester vesicles are similar to those at term, statistical analysis finding no significant differences between them (unpaired t test).

Figure 3
figure 3

Inhibitor-sensitive 36Cl uptake at 30 s into microvillous membrane vesicles formed from syncytiotrophoblast of first trimester and term placentas. *p < 0.05, a, vs DIDS-sensitive uptake at 25 mV into first trimester vesicles; b, vs DIDS-sensitive uptake at 25 mV into term vesicles (paired t tests).

Full size image

To compare the activity of the DPC-sensitive conductance in first trimester MVM with that in term MVM, we also calculated, from the raw data, the DPC-sensitive, DIDS-insensitive uptake at the two pds (as the difference between DPC-sensitive and DIDS-sensitive uptake) and found them to be similar for first trimester and term MVM vesicles (0.28 ± 0.06 nmol/mg at 0 mV, 0.38 ± 0.09 nmol/mg at 25 mV for first trimester; 0.14± 0.07 nmol/mg at 0 mV, 0.24 ± 0.09 nmol/mg at 25 mV for term).

An alternative method of analysis of the effects of inhibitors, rather than comparing absolute inhibitor-sensitive uptakes, is to estimate the percentage contribution of the inhibitor-sensitive pathway to total control uptake. With this form of analysis, the percentage contributions of the inhibitor-sensitive were, as above, similar for first trimester and term vesicles (examples: the DIDS-sensitive contribution at 0 mV is 25 ± 3%versus 30 ± 7%, and the DPC-sensitive, DIDS-insensitive component at 25 mV is 18 ± 5% versus 15 ± 3%, first trimester versus term, respectively).

Western Blotting

Figure 4 shows the three blots probed with anti-human band 3 antibody, together with the negative control, which shows no signal. The positive control (blood cell ghosts) shows an AE-1 band of estimated MW 93-113 kD. This band was clearly present in lanes loaded with both first trimester and term MVM vesicles; it was also present, faintly, in the placental homogenate lane.

Figure 4
figure 4

Three Western blots (A, B, and C) probed with a MAb raised against the AE-1 protein, together with a negative control probed with PBS (D).f1-f8, 5 µg of protein loadings from eight different first trimester vesicle preparations. t1-t8, 5 µg of protein loadings from eight different term vesicle preparations. h, 5µg of protein loading of term placental homogenate, bc, 5µg of protein loading of blood cell membranes. MW, molecular weight markers, estimated molecular weights of markers as indicated.

Full size image

As well as the AE-1 band, there was also a faint band in some MVM lanes at around 53-55 kD, which might be either specific binding to a fragment of the anion exchanger protein or binding to another protein. Two bands, around 55 and 50 kD, were also present in the blood cell ghost lanes. None of these bands was present in the negative control lane. For the purposes of densitometric analysis, these bands were excluded.

Densitometric analysis strongly suggested that expression of AE-1 protein in MVM vesicles from first trimester tissue was the same as that in term tissue; densitometry values were 0.23 ± 0.02 for first trimester(n = 8), and 0.23 ± 0.02 for term (n = 8), arbitrary units. This was comparable to that of blood cell membranes (0.17, n = 2) but markedly higher than that for homogenate (0.02, n = 1).

DISCUSSION

Preparation of the MVM vesicles. Both the alkaline phosphatase enrichment and the protein recovery of first trimester and term MVM vesicle preparations were similar to those previously reported by this laboratory(11,12). This demonstrates the consistency of this method of producing relatively pure MVM vesicles from placentas at both ends of gestation. Interestingly we found here, as previously(11), that the recovery of MVM protein was significantly lower from first trimester placental tissue compared with that from term placentas. This agrees with data from morphometric studies(13) showing that the surface area of the syncytiotrophoblast per g of placenta is lower in early gestation.

36Cl- uptake-membrane binding and equilibrium uptake. The binding of 36Cl- to membranes at equilibrium, as calculated from the effect of osmolarity, was relatively small for both first trimester and term vesicles and was much smaller than that previously reported(6,9). This may be because the range of osmolarities chosen in the previous studies was wider, a number of hypotonic media being used. Bathing vesicles in hypotonic medium for 1 h might lead to lysis of some vesicles (their osmotic fragility has not been tested) resulting in a spuriously high estimate of binding. In support of this, there does seem to be some deviation from the linear relation between equilibrium uptake and osmolarity with the hypotonic solutions in the studies of Byrne et al.(9) and Shennan et al.(6). Interestingly Cole(14) used the same range of osmolarities as used here to test the effects of osmolarity on equilibrium uptake of sulfate, and estimates of binding for the divalent anion were similar to the results from the present study.

We found that there was a difference in equilibrium36 Cl- uptake, and therefore calculated intravesicular volume/mg of protein between first trimester and term vesicles. There are a number of possible explanations for this, including a difference in protein density on the MVM, a difference in the surface area to volume relationship for the MVM vesicles, and a difference in any fixed intravesicular charges that might be present. Interestingly, with regard to the latter, the estimated intravesicular volume from 22Na+ uptake experiments(11) showed a lower estimated volume in first trimester MVM vesicle compared with term, suggesting, along with the results of this study, that there is a change in intravesicular charge between first trimester and term vesicles. However, given that we do not know the actual reason for the differences, we have presented the initial rates (30 s) of36 Cl- uptake uncorrected for equilibrium uptake and shown effects of inhibitors in absolute terms as well as in proportion to control uptakes.

36Cl- uptake-initial rates and effects of inhibitors. As in a previous study on term vesicles(9), 36Cl- uptake in first trimester and term vesicles was linear up to at least 30 s. We therefore accepted uptake as approximating to initial rate at this time. In first trimester MVM vesicles, in the absence of an imposed pd (0 mV), both DIDS and DPC inhibited36 Cl- uptake, there being no significant difference between the effects of the two inhibitors. Both substances are known to block the anion exchanger(7,9), and these data strongly suggest that the Cl-/HCO3- exchanger is active in first trimester MVM. At 25 mV, total chloride uptake was higher, as would be expected if there was a diffusional uptake into the vesicles. In first trimester vesicles DPC, but not DIDS, significantly reduced chloride uptake at this potential, suggesting that a DPC sensitive conductance was now predominant over the Cl-/HCO3- exchanger.

The data obtained with the term vesicles was generally in agreement with previous studies from this(9) and other laboratories(6–8). There was evidence for the presence of the DIDS-sensitive, DPC-sensitive electroneutral anion exchanger as well as the DIDS-insensitive, DPC-sensitive conductance. The DIDS-sensitive uptake at 0 mV contributed 30 ± 7% to total 30-s uptake. There was no DPC-sensitive, DIDS-insensitive uptake at 0 mV, but at 25 mV there was a significant DPC-sensitive, DIDS-insensitive uptake contributing 15 ± 3% to total 30-s uptake. Total uptake at 30 s was higher than in the study of Byrne et al.(9) and proportionately the effects of the inhibitors were lower, although the absolute values for inhibitor-sensitive uptakes were similar in the two studies. No evidence was found in the present study, with either first trimester or term vesicles, for a DIDS-sensitive, DPC-insensitive voltage-dependent conductance previously reported in term preparations(9). However, this conductance constituted only a very small component of 36Cl- uptake in that study and it is voltage-dependent, so that small differences in membrane potential would affect its open/closed state(9).

There was no difference in the DIDS-sensitive or DPC-sensitive uptakes by first trimester vesicles when compared with term vesicles at either pd, as calculated in absolute terms or as a proportion of control uptakes. This suggests that the activities of the anion exchanger and DPC-sensitive conductance in isolated MVM do not change (per mg of membrane protein) with gestation. Comparing control uptakes at 0 mV for the two gestations, there was, however, a significantly higher uptake at 30 s in the first trimester vesicles than there was at term. This might be explained by a difference in the fraction of unvesiculated membrane protein between first trimester and term vesicles (which could also explain the difference in equilibrium uptakes noted above). Alternatively, it might be due to there being a faster initial rate of uptake in the first trimester vesicles at 0 mV. At 25 mV, control uptake at 30 s was similar in both first trimester and term MVM vesicles. Therefore, the data raise the possibility of an additional, voltage-sensitive36 Cl- uptake pathway in the first trimester MVM compared with that at term.

As in previous studies of 36Cl- uptake into term MVM vesicles(6,7,9), a large proportion of uptake at initial rate into both first trimester and term vesicles was not inhibited by DIDS or DPC. There must therefore be other routes for chloride uptake, such as a specific transport mechanism, which is not blocked by any currently tested inhibitor and/or nonspecific "leakage" through the lipid bilayer(15).

Western blotting. The expression of AE-1 in microvillous membranes has not been reported previously for placentas at any gestation. The MAb used in this study is highly specific to human AE-1(16). It was raised to a specific amino acid sequence in the N terminus of the cytosolic part of the protein. This is the area in which the different AE proteins share the least homology, the amino acid sequences in the rest of the protein being very similar for all the anion exchanger family(17,18). Therefore, the protein detected, which has the same molecular weight as AE-1, in the MVM of both first trimester and term placentas is highly likely to be homologous with the anion exchanger found in the red blood cell membrane. To date, it appears that the MVM and, from preliminary data, basal membranes(19) of placenta are the only membranes, besides that of the red cell, which express AE-1 in quantities detectable by immunoblotting. It seems highly unlikely that contamination of our preparations with red blood cell membranes contributed to the high expression of AE-1. A previous study, in which placental MVM vesicles were prepared by a method similar to our own, showed that red cell contamination was less than 0.5%(20).

The results of the Western blot and densitometric analysis shown here suggest that AE-1 is equally expressed in both term and first trimester MVM. AE-1 is likely to be the protein accounting for most, if not all, of the DIDS-sensitive, DPC-sensitive electroneutral chloride uptake by the syncytiotrophoblast MVM vesicles prepared from both first trimester and term placentas. The implication from both parts of the study is that neither the activity nor the expression of the AE-1 anion exchanger is different between MVM vesicles from first trimester compared with those from term placentas. It is still possible that activity of the anion exchanger may change with gestation in vivo. Anion exchanger activity is known to be sensitive to change in intracellular pH, and may also be sensitive to cAMP and protein kinase C(21,22). Thus, changes in these regulatory mechanisms may dictate activity of the exchanger in vivo, and this needs to be investigated by other techniques.

Our data here showing no change in the activity or expression of the anion exchanger in first trimester compared with term MVM contrasts with previous reports that the activity of the Na+/H+ exchanger and of the system A amino acid transporter are markedly lower in first trimester MVM vesicles compared with those from term(11,23). Furthermore the potential difference across the microvillous membrane is significantly more negative in the first trimester than at term(24). Altogether these data suggest that there are quite selective changes in the activity and expression of transport proteins by the placenta as gestation proceeds. The overall contribution of these alterations in net flux to the fetus in vivo will be dependent on the resultant effect of several variables, including possible controlling influences of hormones. Further studies should address these as well as the underlying cause of the changes and will be important in understanding how net flux of solute across the placenta is matched to the acceleration in the growth rate of the fetus, relative to that of the placenta, which occurs toward the end of the first trimester in normal pregnancy(25).