Differentiation-dependent expression of carbonic anhydrase II and III ...

Differentiation-dependent of carbonic anhydrase

expression II and III in 3T3 adipocytes

CHRISTOPHER J. LYNCH, STACY A. HAZEN, RICK L. HORETSKY, NICHOLAS D. CARTER, AND SUSANNA J. DODGSON Department of Cellular and Molecular Physiology, The Pennsylvania State University College of Medicine, Hershey 17033; Department of Physiology, School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania 19104; and Department of Child Health, St. George’s Hospital Medical School, London S W17 ORE, United Kingdom Lynch, Christopher J., Stacy A. Hazen, Rick L. Horetsky, Nicholas D. Carter, and Susanna J. Dodgson. Differentiation-dependent expression of carbonic anhydrase II and III in 3T3 adipocytes. Am. J. Physiol. 265 (Cell Physiol. 34): C234-C243, 1993.-Carbonic anhydrase (CA) was examined in two adipocyte cell lines, 3T3-Ll and 3T3-F442A. Both CA III and non-CA III activities, measured by I80 mass spectrometry, were present in 3T3-Ll and 3T3-F442A adipocytes; however, no CA activity was detected in 3T3 preadipocytes of either line. These observations were supported by immunoblot experiments employing CA III and CA II isoform-specific antisera. CA III, a major protein in rodent and murine adipocytes, and CA II, another isoform known to be present in adipose tissue, were observed only in the differentiated 3T3 adipocytes. The differentiation-dependent expression of these isozymes may imply an adipocyte-related role for CA. Compared with cultures maintained in the absence of insulin, 3T3 adipocytes maintained in the presence of insulin exhibited 65-90% lower concentrations of CA III. CA II was unaffected. This negative effect of insulin on CA III may explain the metabolic regulation of adipose CA III observed in vivo. After media changes, 3T3 adipocyte cultures rapidly lower media pH, which in turn lowers the bicarbonate/CO, of bicarbonate/CO,-buffered media. Cultures maintained at low pH displayed 50-90% lower concentrations of CA II and CA III. Similarly, cultures maintained in a low bicarbonate/CO, media (GibCO,-I medium containing 1 mM bicarbonate under an atmosphere of 100% humidified air) displayed 30-50% lower CA II and CA III concentrations. Thus CA II and CA III concentrations are influenced by pH and bicarbonate/CO,. Neither effect, the pH or the GibCO,-I media effect, was associated with changes in the concentration of pyruvate carboxylase or ATP citrate lyase (2 markers of adipocyte differentiation). Because the regulation by pH and bicarbonate/CO, may be relatively selective for CA in adipocytes, a simple method for reducing the concentration/activity of CA in 3T3 adipocytes is described that may be a useful tool for studies on the physiological role of the enzyme.

Other interesting properties of CA II and CA III are their tissue-specific expression and abundance. CA II is widely distributed, whereas relatively few tissues contain CA III (41). In the few tissues where CA III is expressed, it is unusually abundant, which compensates for its lower turnover number relative to CA II. For example in red skeletal muscle, CA III comprises ~8% of the cytosolic protein content, where its original appellation was “muscle basic protein” (7, 17, 18, 24). Despite its remarkable abundance in red muscle, there is little CA III in white skeletal muscle or any other muscle type for that matter. Another example of unusual abundance is exemplified by rat adipocytes, where CA III can comprise up to 24% of the cytosolic protein content (32, 40). Interestingly, CA III appears to be metabolically regulated in fat cells. Expression of adipocyte CA III decreases in several obesity models and increases in response to drug-induced diabetes mellitus (31, 32, 42). Adipose tissue CA III is also reciprocally regulated during feeding and fasting (42). The factors responsible for CA III’s metabolic regulation, tissue-specific distribution, and unusual abundance have not been elucidated. Recently, our laboratory reported that rodent adipocytes contain high specific activities of CA (32). Because the physiological role of this enzyme in fat cells is not known, we sought information on the enzyme’s differentiation-dependent expression. An adipocyte model system in which regulation of the enzyme could be easily studied and concentrations of the enzyme could be manipulated was also desirable. We report here that CA II and CA III are present in the two adipocyte cell lines, 3T3-Ll and 3T3-F442A. These cell lines have proven to be very useful in the past for these types of studies (12,

glyceraldehyde-3-phosphate hydrogen ion concentration;

26, 41, 43).

dehydrogenase; cell lines; insulin; bicarbonate; carbon dioxide

(CA) family (EC 4.2.1.1., carbonate dehydratase) comprises at least eight members (13,X, 44). Three of the isoforms are cytosolic, and two of these are present in adipose tissue (40). The adipocyte isoforms, CA II and CA III, display several interesting properties. For instance, CA II has arguably one of the highest turnover numbers of any enzyme in nature (27, 35). CA II and other CA isoforms are inhibited by sulfonamide CA inhibitors, like acetazolamide, in the range of 10 -g to low7 M (13, 34). In contrast, CA III is relatively insensitive to even micromolar concentrations of acetazolamide. This insensitivity to inhibitors has made it possible to selectively quantitate CA III activity, even when other CA isoforms are present (16). THE

c234

CARBONIC

ANHYDRASE

0363-6143/93

$2.00

Copyright

Using IsO mass spectrometry enzyme activity measurements, we show that CA III activity, along with the activity of at least one other non-CA III isoform, presumably CA II, is found in differentiated 3T3-Ll and 3T3-F442A adipocytes, but not undifferentiated preadipocytes. This was confirmed by isoform-specific antibody immunoblotting. The differentiation-dependent expression of CA in adipocytes is different from the expression of CA III in myogenic G-8 cells. In that system, induction of CA III is not seen during the differentiation of myoblasts into myotubes (for review, see Ref. 17). The differentiation-dependent expression of CA in fat cells may imply an adipocyte-related role of the enzyme. Other novel findings in this report are that CA III, but not CA II, is negatively regulated by insulin

0 1993 the American

Physiological

Society

CARBONIC

ANHYDRASE

(a finding that could explain CA III’s metabolic regulation in vivo) and that both CA II and CA III concentration are inversely regulated by their substrate/ products, bicarbonate/C02, and hydrogen ions. A relatively simple method for lowering the concentration of CA in 3T3 adipocytes is described that may prove to be useful for studies on the physiological role of the enzyme. EXPERIMENTAL

PROCEDURES

CeZZculture of37'3 cell lines. 3T3-LI (passage 6) cells obtained from Dr. Daniel M. Lane (Dept. of Physiological Chemistry, The John Hopkins University, School of Medicine) were maintained and differentiated using minor modifications of the method of Student et al. (43) to obtain differentiated adipocytes and undifferentiated preadipocytes.1 Unless otherwise indicated, we used the recommended medium for these cells, which is Dulbecco’s modified Eagle’s medium (DMEM) with 4.5% glucose and 10% calf serum (CS) before the 48-h differentiation program and DMEM with 10% fetal calf serum (FCS) and 10 mg/ml insulin after the differentiation program (43). Undifferentiated preadipocyte 3T3-Ll cells were harvested when cultures were 50-70% confluent. To initiate the differentiation program in 3T3-Ll preadipocytes, cultures were allowed to grow to confluency then incubated for 48 h with 10% FCS, 3-isobutyl-1-methylxanthine (115 pg/ml), insulin (IO hg/ml), and dexamethasone (390 rig/ml). Unless otherwise indicated, cells were maintained under an atmosphere of 95% humidified air-5% CO, (see footnote 1). Experimental media and media manipulations were used only after the 48-h differentiation protocol. 3T3-F442A cells, a gift from Dr. Howard Green (Dept. of Cellular and Molecular Physiology, Harvard Medical School), were cultured with minor modifications according to Djian et al. (12) to obtain differentiated adipocytes and undifferentiated preadipocytes (see footnote I). We used DMEM with 4.5% glucose and 10% CS to maintain undifferentiated 3T3-F442A preadipocytes, which we harvested when cultures were 50-70% confluent. A “differentiation program” analogous to that used for 3T3-Ll cells is not required for 3T3-F442A cells (12). Instead, differentiation is typically initiated in postconfluent 3T3F442A cultures by employing DMEM supplemented with 10% FCS and 5 pg/ml of insulin; thereafter, cultures are maintained in this medium. In the present communication, however, the first 48 h after the switch to FCS and insulin were referred to as the differentiation program for the 3T3-F442A cells. This was only done to allow a convenient framework for comparing the two cell lines. This seemed reasonable, since the morphological changes in the two cell lines during this period were similar. Thus, as with the 3T3-Ll cells, experimental media and media manipulation in 3T3-F442A cultures were not employed until after the completion of this 48-h period. 3T3-F442A cells were l Methods for maintaining these cultures recommend atmospheres of 90-92% humidified air and 8-10% CO, (5% CO, is normally used with a 25 mM bicarbonate buffer to obtain pH levels of %4), which, when coupled with 48-h feeding intervals, further acidifies the media (12, 43). It has been found empirically that these conditions improve the differentiated appearance of cultured adipocytes (i.e., as determined by the size and number of lipid droplets per cell). The mechanism for this phenomenon has not been explored to our knowledge; however, we speculate that this procedure may inhibit hormone-sensitive lipase, which has a very narrow pH optima (9,22, 36,38). Thus, after media changes, the drop in pH may be a reflection of triglyceride breakdown and fatty acid mobilization. The leveling off of the pH may then be a reflection of the pH sensitivity of hormone-sensitive lipase. Some workers have proposed this as a mechanism for feedback inhibition of hormone-sensitive lipase.

IN 3T3

ADIPOCYTES

C235

maintained under an atmosphere of 95% humidified air-5% CO, unless otherwise indicated (see footnote 1). In some experiments differentiated adipocyte cultures were maintained for 5-7 days in GIBCO CO,-independent medium (GibCO,-I) modified to include similar concentrations of additives to the DMEM medium (e.g., FCS, high glucose, insulin, antibiotics, and glutamine) .2 The primary difference between these two media is the bicarbonate concentration [l mM (GibCO,-I) vs. 25 mM (DMEM)] and the proprietary buffering system employed in the GibCO,-I. Cells maintained in GibCOz-I were kept under an atmosphere of 100% humidified air rather than 95% air-5% CO,. In some experiments, medium with a pH of 6.2-6.4 was employed (“low pH”). The pH of the low-pH media was adjusted with NaOH after equilibration to the temperature and atmosphere used for growing the cells and was rechecked just before use. For gel electrophoresis and CA assays, undifferentiated and differentiated 3T3 cells were harvested by mechanical scraping with ice-cold phosphate-buffered saline. Concentrated cells were frozen in IOO-~1 aliquots at -85°C. Gel electrophoresis, immunoblotting, and NH,- terminal microsequencing. One-dimensional sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Coomassie Blue staining were performed as previously described (28, 31, 32) using a Hoefer SE600 apparatus with 21°C cooling. Western blotting of proteins in SDS-PAGE gels to Nitroplus membranes (Micron Separations, Westboro, MA) was performed overnight at 15 V (CA II and CA III) or 30 V (pyruvate carboxylase and ATP citrate lyase) in a Bio-Rad transfer apparatus. The transfer solution contained 192 mM glycine, 25 mM tris(hydroxymethyl)aminomethane base and 20% methanol. Blots were temporarily stained with Ponceau S stain and destained with water. In most experiments the molecular weight region of interest (based on protein standards) was excised from the stained blot and used for further processing to limit the amount of antibody needed per experiment. Immunoblotting was performed with washing and incubation solutions (blotto) containing 5% nonfat dried milk to prevent nonspecific binding. 1251-labeled goat anti-rabbit immunoglobulin G F’ab, fragments (2.5 X lo5 counts. min-l *ml blotto-I; from New England Nuclear, Boston, MA) and autoradiography with Du Pont enhancing screens at -84°C were employed as a means of identifying antibody binding proteins (3 1, 32). The production and characterization of rat CA II (29 kDa) and CA III (28 kDa) antisera are described elsewhere (24). Relative concentrations of 112 kDa immunoreactive ATP citrate lyase were measured using antisera provide by Dr. Charles Rubin (Albert Einstein College of Medicine, New York, NY). Pyruvate carboxylase (116 kDa) was measured by either immunoblotting or avidin binding. For immunoblotting, antisera prepared against the purified rat adipocyte enzyme was used (3 1). In adipocytes, pyruvate carboxylase is the major avidin binding protein and the only 116-kDa protein that binds avidin; therefore, pyruvate carboxylase was also quantitated in Western blots by measuring 116-kDa avidin binding activity (31). We used 1251-avidin and autoradiography for detection of biotin-containing proteins in blots. These experiments are conducted with the same washing times and solutions used for immunoblotting except that the primary antibody step is omitted and 1251-avidin (2.5-5 X lo5 countsmin-l . ml blotto-l; obtained from New 2 On the condition that such information would not be published, a GIBCO representative provided us with information on the nonbuffering constituents of the GibCO,-I, allowing adjustment of salts and amino acids to concentrations found in DMEM. The proprietary buffering system used in GibC02-I is similar to phosphate-based buffering systems.

C236

CARBONIC

ANHYDRASE

England Nuclear) is substituted in what would be the l-h second antibody incubation. Laser densitometry scansof one-dimensionalelectrophoresis gels and autoradiograms were analyzed using “Quantitation one” software (Protein Databases,Huntington Station, NY). Protein staining, immunoreactivity, and activity were measured in three or more experiments using different samples.Where indicated, statistically significant changes(P < 0.05) were determined using Student’s t test. NH,-terminal microsequencing from Coomassie Bluestained polyvinylidene fluoride (PVDF) blots was performed as previously described(31, 32). The DNA-STAR computer program was used to search the National Biomedical Research Foundation Protein Identification Resource for matching amino acid sequences. CA enzyme activity. CA activity of cell lysates was measured at 37°C by the la0 massspectrometric technique of Itada and Forster (23) using the previously describedmodifications (16). Frozen cell lysates were thawed and in some instances (with fatty samples)diluted with water and incubated at 37°C for 4-5 min before assay.The 3-ml reaction chamber contained 25 mM NaHCOs (1% labeled with IsO). The mass46 peak i2Ci80i60 decreasesas IsO exchangeswith the 55 M 160 pool in water in the absenceof CA. CA increasesthe rate of this exchange.CA activity is calculatedfrom the half time of the rate of disappearance of the mass46 peak heights, both before and after the addition of a sample,from the equationsof Mills and Urey (35). Valueswere then corrected for cellular protein measuredby the method of Lowry et al. (30). CA III enzyme activity is calculated from the following equation: CA III activity = total CA activity - activity of other isoforms(inhibited by low concentrations of CA inhibitors), that is, CA III activity is taken asthe portion of the total CA activity that is not inhibited by 7 FM acetazolamide (16, 32).

IN 3T3 ADIPOCYTES

3T3-L UD

1

-CA

3T3-F442A U D

III

UD

’ .

-CA

Ill

UD

.ys ‘s .., ;,

-CA

II

’ &* *

-CA

II

Fig. 1. Immunoblotting of 3T3-Ll and 3T3-F442A cell lysates with carbonic anhydrase (CA) II and CA III specific antisera. Frozen 3T3 cell proteins (100 pg protein/gel lane) were solubilized in Laemmli sample buffer (28) and separated by SDS-polyacrylamide gel electrophoresis (PAGE). Immunoblotting was performed as described in EXPERIMENTAL PRO-CEDURES on a strip of filter including proteins in the 28- to 29-kDa range. Lanes labeled U are undifferentiated preadipocytes, and lanes labeled D are differentiated adipocytes. Antibody step employed a 1:500 dilution of rabbit anti-rat CA III (top) or CA II (bottom) serum, previously described (for review, see Ref. 24). CA II and CA III immunoreactivities are demonstrated in an autoradiograph which is representative of 3 such studies.

RESULTS

expression of CA. Unlike

Differentiation-dependent

many enzymatically catalyzed reactions, the reversible dehydration of bicarbonate proceeds to a significant extent nonenzymatically, which accounts for the decrease in 180/160 exchange in the absence of cell lysate. At least 80% of the CA enzyme activity in differentiated 3T3 adipocytes was sensitive to low concentrations of acetazolamide (Table l), suggesting that other isoforms besides CA III emerge as part of the differentiation process. Because CA III and CA II, but not CA I, have previously been reported in adipocytes, immunoblotting studies employing CA II- and CA III-specific antisera were performed (Fig. 1). Differentiated 3T3 adipocytes, but not Table 1. CA enzyme activities in adipoblasts and adipocytes Cell

Type

Total

Activity

CA III Activity

Activity of Other Isoforms

3T3-Ll adipoblast ND ND ND 3T3-F442A adipoblast ND ND ND 3T3-Ll adipocyte* 10.1+4.0 0.8kO.3 8.1 3T3-F422A adipocyte 4.022.2 0.6kO.3 3.4 Values are means + SE and are expressed as rmol CO, converted.mg protein-‘. min-‘. Activity of other isoforms was determined as follows: total activity - carbonic anhydrase (CA) III activity = activity of other isoforms. CA activity was assayed by ‘“0 mass spectrometry as described in EXPERIMENTAL PROCEDURES. ND, none detected. * A passage-related decrease in CA mass and specific activity was observed between passages 6 and 12.

preadipocytes, displayed 29- and 28-kDa immunoreactivity to CA II- and CA III-specific antisera, respectively, in agreement with the results of the activity measurements. A passage-related decrease in CA mass and specific activity was observed in 3T3-Ll adipocytes between passages 6 and 12. This passage-related phenomenon was not observed in 3T3-F442A adipocytes (data not shown). For that reason some of the later studies were conducted only on the 3T3-F442A adipocytes. A passage-related alteration of adipocyte phenotype was not observed for either 3T3-Ll or 3T3 F442A cells. Effect of insulin on CA II and CA III concentrations.

After the differentiation program, cell lysates were prepared from 3T3 adipocytes maintained for 5 days in media containing the usual insulin concentration, 10 pg/ml for 3T3-Ll adipocytes and 5 pg/ml for 3T3-F442A adipocytes, or in media in which insulin was removed.3 The removal of insulin does not cause the cells to “dedifferentiate,” i.e., to revert to their fibroblast appearance. The lysates were subjected to SDS-PAGE and immunoblotted with CA II and CA III antisera. Figure 2 shows that on a per milligram of protein basis (which is in reference to the amount of protein loaded in the gel) insulin addition resulted in a statistically significant decrease in CA III immunoreactivity in 3T3-Ll adipocytes and 3T3-F442A 3 Immunoreactive insulin in the fetal calf serum used for these studies was measured as the only other source of insulin in the medium. The results indicate that in the absence of added insulin the cell culture media contained 0.012 rg/ml insulin.

CARBONIC

-

+

INSULIN

-CA

3T3-L

1

ANHYDRASE

Ill

3T3-F442i

Fig. 2. Effect of insulin removal on CA III concentrations in 3T3 adipocytes. After the 48-h differentiation program, 3T3 cultures were maintained in presence of fetal calf serum with (usual treatment, + lanes) or without added insulin (- lanes) (see footnote 3). Insulin is normally added to a concentration of 10 pg/ml medium (3T3-Ll) or 5 pg/ml medium (3T3-F442A). Cell lysates were prepared 5 days after differentiation program ended. CA III was analyzed in 100~pg samples of cell lysate by immunoblotting with CA III antisera.

adipocytes. The decrease ranged from 65 to 90% in different experiments with both cell lines. In contrast, no changes were seen in CA II concentrations (Fig. 3). Media acidification by 3T3 adipocytes. A few days after the differentiation program, the rate of media acidification after feeding increases. 3T3 adipocyte cultures can acidify their medium from pH 7.4 to between pH 6.2 and 6.8 in 24 h or less (Fig. 4A; see footnote 1). The change in pH over time decreased asymptotically below pH 6.8-7.0 in either DMEM, containing 25 mM bicarbonate, or Gib+-

+-

INSULIN

8 / * / I

-CA

3T3-L

II

1 3T3-F442A

Fig. 3. Effect of insulin removal on CA II concentrations in 3T3 adipocvtes. 3T3 cultures were maintained in presence of fetal calf serum with (usual treatment, + lanes) or without-added insulin (- lanes) as described in Fig. 2. CA II was analyzed in lOO-pg samples of cell lysate by immunoblotting with CA II antisera.

IN 3T3 ADIPOCYTES

C237

CO,-I, a media which contains only 1 mM bicarbonate (data not shown). Effect of PH. To evaluate the possible influence of pH on steady-state CA II and CA III concentrations, we allowed 3T3-F442A cells to complete the differentiation program in DMEM and then switched them to media with either normal or low starting pH values. Differentiated cells were maintained for 5 days with either GibC02-I (under 100% humidified air) having a starting pH of 6.3-6.4 (low pH) or GibC02-I having a starting pH of 7.4-7.6 (normal pH) and were then harvested. Insulin was not added to either the low- or normal-pH maintenance media. Medium pH was measured several times a day before media changes. 3T3 adipocyte cultures decreased the pH of the normal-pH media but did not appreciably alter the pH of the low-pH media (Fig. 4A). By changing the media several times a day, it was possible to reduce the amount of time that the normal-pH group spent at lower pH levels. 3T3 adipocytes maintained in the low-pH media had lower concentrations of CA II and CA III compared with cells maintained at overall higher pH levels (Fig. 4B). Similar effects were observed in 3T3-Ll cultures (data not shown). The pH-related percent decrease in CA was similar for both isozymes (89 + 4% decrease in CA II, n = 4; and 89 f 3% decrease in CA III, n = 7). The pH effect was also observed in cultures maintained with DMEM or in the presence of insulin (data not shown). These findings support a negative role for hydrogen ion in the expression of CA II and CA III. Despite its effect on CA II and CA III concentrations, pH manipulation did not significantly decrease the concentrations of two adipocyte differentiation markers, ATP citrate lyase and pyruvate carboxylase (Fig. 5), or most other proteins observed in silver-stained SDSPAGE gels (Fig. 4C). However, a pH-related decrease in a major protein of 34.5 kDa was observed (Fig. 4C). Gel lanes containing this protein were transferred to PVDF and were Coomassie Blue stained. The 34.5-kDa protein was subjected to NH,-terminal microsequencing. The resulting amino acid sequence, VKVGVNGFGRIGRLV, had 100% homology with mouse glyceraldehyde-3-phosphate dehydrogenase (GAPDH), a major 3T3 adipocyte protein containing subunits of the approximate size we found (41). GAPDH was the only protein outside the 28to 29-kDa range that consistently demonstrated a pHdependent range. The CA isoforms are located within the 28- to 29-kDa range. Effect of bicarbonate/C02. A pH-related decrease in the bicarbonate concentration can be predicted from the law of mass action as a result of organic acid secretion from the cells and the resultant media acidification; therefore, we investigated the possible role of bicarbonate/COz on steady-state CA II and CA III concentrations. We allowed 3T3 cells to complete the differentiation program in DMEM and then maintained one-half of the cultures in either DMEM (containing 25 mM bicarbonate) or GibC02-I (a media which contains only 1 mM bicarbonate) for the next 5 days. Daily media changes were employed in these experiments, and changes in media pH were similar for either medium (data not shown). Figure 6

C238 A

CARBONIC

ANHYDRASE

IN 3T3 ADIPOCYTES

B

7.6: 7.4-r 7.2:

Yi

7: 6.8; 6.6: 6.4.: 0

1

2

3

4

5

6

Days after Diff. Program

C

3 16, Q 7 - ^_ .x ki,

290

-GAPDH

Fig. 4. Effect of media pH on 3T3 adipocyte CA II and CA III concentrations. 3T3-F442A cultures were differentiated in DMEM (in an atmosphere of humidified air with 5% CO,) and then switched and maintained (with insulin and fetal calf serum) in GIBCO COz-independent medium (GibCO,-I; gassed with humidified air only). Starting pH values were either 6.3-6.4 or 7.4-7.6. After differentiation program, media were changed 1-3 times per 24-h period (as opposed to every 2 days as recommended) and analyzed for hydrogen ion content. Cell lysates were prepared 5 days after differentiation program ended. A: media pH was measured at times indicated after completion of differentiation program. B: CA II and CA III concentrations in 100~pg cell lysate protein samples were analyzed in Western blots with CA IIspecific antisera. C: silver staining of 3T3-F442A proteins separated by SDS-PAGE (10 pg of cell lysate protein per lane). Lane 1, normal pH; lane 2, low pH. Position of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) is indicated. CA II migrates at 29 kDa and CA III at 28 kDa; however, bands at 28-29 kDa contain several other proteins in addition to CA II and CA III as determined by 2-dimensional isoelectric focusing/SDS-PAGE (data not shown).

shows that CA II and CA III concentrations decreased when 3T3 adipocytes were maintained in GibC02-I (under an atmosphere of 100% humidified air) rather than DMEM (under an atmosphere of 5% CO,-90% humidified air). The decreases ranged from 28 to 66% in different experiments. In contrast, no media-dependent changes in the expression of other major proteins in silver-stained gels (Fig. 6B) or in the immunoreactive levels of ATP citrate lyase and pyruvate carboxylase were ob-

served (Fig. 7). Based on the extracellular pH measurements, the change in CA II and CA III seen when GibC02-I is replaced for DMEM is not due to changes in the extracellular pH. These findings support a positive role for bicarbonate in CA II and CA III expression. Table 2 shows that different concentrations of CA activity can be achieved in 3T3 adipocytes by the following method. After the differentiation protocol, the media change interval is altered to three or more times per day

CARBONIC

ANHYDRASE

Fig. 5. Effect of media pH on 3T3 adipocyte ATP citrate lyase and pyruvate carboxylase concentrations. 3T3-F442A adipocytes from Fig. 4 experiment were analyzed by immunoblotting with antisera to ATP citrate lyase (ACL) and pyruvate carboxylase (PC).

to keep the pH near physiological levels. The cultures are maintained in the absence of added insulin for 5 days in either low-pH GibCOs-I or normal-pH DMEM. This method takes advantage of both the pH and the GibCOz-I effects. Silver-stained gels comparing proteins from adipocytes cultured under these two conditions (data not shown) had no significant changes as determined by laser densitometry, with the exception of GAPDH which decreased in the cultures maintained at the low pH as described above. DISCUSSION

Recently, we reported that rodent fat cells contain unusually high specific activities of CA; indeed, the activities are so high that they may be second only to those found in mammalian erythrocytes (32). CA II and CA III, the latter of which comprises -24% of the cytosolic protein content in adipocytes from 5 to 8-wk-old Zucker rats (32), are the two isoforms that appear to account for the majority of this activity (41). Despite the abundance of CA in adipose tissue, very little is known about its function there. It was of considerable interest, therefore, to obtain information on the expression of these isoforms during adipocyte differentiation. Of the several model systems available for such studies, 3T3-Ll and 3T3F442A cells appear to be in widest use and best characterized in terms of the differentiation-dependent expression of specific proteins (12, 26, 41, 43). These cells initially have a fibroblast appearance (adipoblast), thought to be representative of committed stromal stem cells. At confluence they can be induced to differentiate into functional adipocyte-like cells. Differentiation-related expression of CA. In the present study, total CA and CA III specific enzyme activities were measured in adipoblast and adipocyte lysates by IsO mass spectrometry (Table 1). CA II and CA III immunoreactivities were measured by immunoblotting. The results of these studies support the conclusion that CA expression is differentiation dependent in adipocytes, since enzyme was present in the differentiated cells but no enzyme could be measured in the adipoblasts (Fig. 1 and Table 1). Differentiation-dependent CA II expression in these cells was somewhat surprising, since this isoform is relatively ubiquitous compared with CA I or CA III (13, 17, 21, 40, 44). A differentiation-dependent pattern of expression for adipose CA III is also novel, because in skeletal muscle this is not the case. Thus, despite CA III’s

IN 3T3 ADIPOCYTES

c239

highly tissue-specific pattern of expression, in skeletal muscle its expression is not differentiation related (for review, see Ref. 21). For example, in vitro studies of the myogenic cell line G-8 revealed that CA III is present in myoblasts and myotubes and is not induced during differentiation like other classic muscle-specific genes (such as myosin, myoglobin, creatine kinase, etc.). Additionally, in vivo studies on mouse embryos and fetuses detected CA III mRNAs as early as 9.5 days in myotomes, i.e., the part of the somite that develops into skeletal muscle (33). Possiblephysiological roles of CA in adipocytes. In view of the fact that CA III represents -8% of the cytosolic protein content in type 1 slow-twitch muscle, it is not surprising that a number of investigators have worked toward providing a muscle-specific role for the enzyme (for review, see Ref. 18). While several hypotheses have been forwarded from these studies, no muscle-specific role has been definitively demonstrated, and Edwards (17) has argued that a role of the enzyme in contractile activity may not be forthcoming given CA III’s pattern of developmental expression in muscle. In contrast, a role for CA in red blood cells is well established (13, 44), and CA emerges in the late stages of erythropoiesis (3, 10). If we can accept these arguments, our findings would seem to imply a role for CA in one or more adipocyte-related function not present in adipoblasts. Two possible roles emerge from the literature, namely de novo lipogenesis and pH regulation. For instance, roles for CA V and CA II in carbon-fixing intermediary metabolic pathways, such as ureagenesis and de novo lipogenesis, have been proposed from studies on liver and central nervous system (e.g., Refs. 6, 11, 14, 15). Because pyruvate carboxylase and acetyl-CoA carboxylase preferentially utilize bicarbonate as opposed to CO2 as a substrate (for review, see Ref. 37), it has been hypothesized that CA might provide the bicarbonate for these reactions. The important evidence in support of this hypothesis is that a reduction in de novo lipogenesis is observed in the presence of CA inhibitors, such as acetazolamide. In addition, in rat, mouse, and hamster brains, fatty-acid synthase, acetyl-CoA carboxylase, and CA have been colocalized in oligodendrocytes. Problems with the theory that CA plays a role in de novo lipogenesis can be also be cited. An important question yet to be resolved is whether or not intracellular bicarbonate concentrations are actually rate limiting for these processes. An alternative explanation of data in which CA inhibitors are shown to block fatty acid synthesis (e.g., Refs. 4, 11, 15) is that the effects of CA inhibitors may not be specific for CA. The colocalization studies of Cammer (6) are quite interesting; however, although it is true that glial cells do play an important metabolic role in the central nervous system, recent evidence also supports the theory that glial CA regulates pH in interstitial space and spinal fluid (e.g., Refs. 8, 39). These pH shifts are associated with changes in neuronal activity; thus the colocalization of anabolic metabolic events and CA may not necessarily be related. The major unresolved problem with the lipogenesis theory is that CA III and the classic lipogenic enzymes (e.g., fatty-acid synthase, ATP citrate lyase, and acetyl-CoA carboxylase)

C240

CARBONIC

ANHYDRASE

IN 3T3 ADIPOCYTES

A

0

5

116, 970

-CA

F, P

679

in fact are not coordinately regulated either in obesity and other metabolic states or in response to insulin; in fact, the metabolic and hormonal regulation of CA III and the lipogenic enzymes is opposite (1,5, 19,31,33,41-43). CA II regulation in rodent adipocytes has not been studied; however, CA II concentrations were not influenced by insulin in 3T3 adipocytes (Fig. 3). A more reasonable role for adipocyte CA II and CA III would seem to be in pH regulation or buffering associated with fatty acid mobilization and transport. CA has been proposed to play an integral role in pH regulation in several tissues (for review, see Refs. 8, 13, 18,39,44). The activity of hormone-sensitive lipolysis is associated with large changes in extracellular and presumably intracellular pH; this could partly explain the increased acid secretion that occurs following 3T3 adipocyte differentiation (38). It is tempting to speculate that CA II and CA III in adipocytes could play a role in buffering against changes in pH occurring as a result of fatty acid mobilization. If this is the case, the findings of several investigators that adipocyte glucose metabolism and hormone-sensitive lipolysis is highly sensitive to even small changes in extracellular pH (9, 22, 29, 36) could explain, at least teleolog-%CO*0

o

5

5

Fig. 7. Effect of GibCOe-I on 3T3 adipocyte ATP citrate lyase and pyruvate carboxylase concentrations. 3T3-F442A adipocytes from Fig. 6 experiment were analyzed by immunoblotting with antisera to ATP citrate lyase (ACL) and pyruvate carboxylase (PC).

-

Fig. 6. Effect of GibCO,-I on 3T3 adipocyte CA II and CA III concentrations. 3T3-F442A cultures were differentiated and maintained (with insulin and fetal calf serum) in either GIBCO DMEM [their usual medium which contains 20 mM N-2-hydroxyethylpiperazine-W-2ethanesulfonic acid (pH 7.4) and 25 mM bicarbonate (gassed with 5% CO,)] or GibCO,-I (gassed with humidified air only). Starting pH values were the same with both media, and pH declined to the same extent with both media. Cell lysates were prepared 5 days after differentiation program ended. A: CA II and CA III concentrations were analyzed in 100-pg samples of cell lysate by immunoblotting with isoform-specific antisera. B: silver staining of 3T3-F442A proteins separated by SDS-PAGE (10 pg of cell lysate protein per lane). Lane 1, cells maintained in DMEM; he 2, cells maintained in GibCO*-I. CA II migrates at 29 kDa and CA III at 28 kDa; however, bands at 28-29 kDa contain several other proteins in addition to CA II and CA III as determined by 2-dimensional isoelectric focusing/SDS-PAGE (data not shown).

ically, the need to maintain cytosolic CA isozymes at such high specific activities in this tissue. Interestingly, hormone-sensitive lipolysis like CA enzyme activity is a process that emerges during adipocyte differentiation (26) and is impaired in obesity (1,5,19), i.e., when CA activity emerges and declines respectively. Insulin, pH, and bicarbonate/COg effects on CA. Previous in vivo experiments in mice and rats had shown that adipose CA III is metabolically regulated. No metabolic regulation has been reported for the other adipocyte isoforms (32). Because insulin is involved in the metabolic regulation of many genes, we evaluated the influence of insulin on 3T3 adipocyte CA II and CA III concentrations. CA III concentrations rose in 3T3 adipocyte cultures maintained in the absence of insulin, suggesting that insulin inhibits the expression of CA III (Fig. 2). In contrast, CA II was unaffected by changes in media insulin concentrations (Fig. 3). Because only the immunoreactivity of CA III was altered by this manipulation,

Table 2. CA enzyme activities in 3T3-F442A adipocytes Cell Culture Medium

Total Activity

CA III Activity

Activity of Other Isoforms

1.72 DMEM (normal pH) 3.7 1.96 GibCOs-I (low pH) 0.747 ND 0.743 Values are means + SE and are expressed as hmol COe converted’mg protein-‘. min-‘. CA activity was assayed by IsO mass spectrometry as described in EXPERIMENTAL PROCEDURES. Cells were differentiated, maintained, and harvested as described in legend to Fig. 6, except DMEM was used as the “normal pH” medium. ND, none detected.

CARBONIC

ANHYDRASE

the insulin experiments help establish the specificity of the CA II and CA III antisera toward their respective antigens and indicate independent regulation of CA II and CA III by the hormone. The specificity of insulin toward the CA III isoform helps explain the previously observed obesity-related decreases in Zucker rat adipocyte CA (32). Obesity in the Zucker rat is characterized by hyperinsulinemia (1,4, 19) and hyperresponsiveness of adipose tissue to insulin (4, 20) in the young animals that were the subjects of this study. On the basis of our findings, it seems likely that the changes in adipose tissue CA III in obesity and in other metabolic states may be inversely related to the plasma concentration of insulin. In early experiments we noted considerable variability in the specific activity and mass of CA isozymes when data from different experiments were compared (e.g., large standard errors in Table 1). Because we hoped to use 3T3 cells in studies of the regulation and adipocyte-related role of CA III, we explored the underlying basis of this problem. Some inconsistencies were determined to be due to a passage-related decrease in CA II and CA III in the 3T3-Ll line. In the original starter cultures we received, 3T3-Ll adipocytes had more total CA activity than 3T3-F442A cells. In subsequent passages, concentrations of CA in the 3T3-Ll cultures declined and eventually disappeared altogether (data not shown). We did not observe this phenomenon in the 3T3-F442A cells; therefore, these cells were used exclusively for some of the later studies. In both cell lines a passage-independent source of variability in the specific activities and mass of total CA present in 3T3 adipocytes was also noted (data not shown). Feeding intervals of 48 h and CO2 concentrations of 8-10% (with 25 mM bicarbonate) were recommended for optimal differentiation of the cells by the investigators who generously supplied the cells and in the literature (12, 43). We changed the media daily and lowered CO2 concentrations to 5% in some, but not all, of our early experiments; due to a concern about postdifferentiation media acidity, 3T3 adipocytes are normally exposed to an acidic environment for extended periods as a result of organic acid secretion, which we hypothesize is largely the result of lipolysis and fatty acid mobilization. This acidity is accompanied by a decrease in cell culture medium bicarbonate concentrations. The acidity is further exacerbated when cells are exposed to the recommended CO:! concentrations of 8-10%. Another variable in our studies was that cells were initially harvested anywhere between 6 and 48 h after a medium change, which meant they spent variable amounts of time at low pH. The variability in total CA concentrations was discontinued when harvest time and intervals between media changes were controlled (data not shown). Recently, Kaiser and Curthoys (25) have reported effects of bicarbonate/CO, on phosphoenolpyruvate carboxykinase (PEPCK) and glutaminase expression. Hydrogen ion has also been reported to regulate PEPCK and other gene expression (e.g., Refs. 2,25, 27). We wondered if some of the variability in the CA concentrations we had observed in our early studies might be related to fluctu-

IN 3T3

ADIPOCYTES

C241

ations in the bicarbonate/CO2 concentrations and/or the pH. This hypothesis was clearly supported by the data presented in Figs. 4 and 6. These experiments support the conclusion that both changes in pH and bicarbonate/CO, concentration are capable of influencing the concentration of CA II and CA III in the cells. To the best of our knowledge, effects of hydrogen ions and bicarbonate/CO2 on CA gene expression or protein concentrations have not been previously studied. The mechanism or level at which the effects we observed manifest themselves was not investigated, since our main interest was in determining conditions for stabilizing the expression of the enzyme in 3T3 adipocytes. The negative effects of hydrogen ion concentration do not rule out a role of the enzyme in pH regulation, since the mechanism of the effects remains to be determined and may in fact be a cell culture artifact. Thus we found that the recommended culture conditions for optimal differentiation and yield of 3T3-Ll and 3T3-F442A adipocytes were not optimal for the expression of CA isozymes. It seems that the highest concentrations of total CA are observed when insulin is removed from the media (which specifically increases CA III) and when care is taken to avoid acidic media conditions. With these modifications, cells retain their differentiated characteristics (Figs. 4C, 5, 6B, and 7); however, protein content per dish is lower at harvest, as expected, when insulin is removed (data not shown). The removal of insulin from the media does alter the cell number, but it does not alter the phenotype of the cell or cause dedifferentiation. The effects of bicarbonate/CO2 manipulation appeared to be fairly specific for CA isozymes in terms of the concentrations of specific adipocyte proteins detected by silver staining (Fig. 4C), and only one major protein band in silver-stained SDS-PAGE gels was altered by low pH in the absence of added insulin [Fig. 6B; it should be noted that lower pH levels altered the concentration of several specific proteins in media containing insulin (data not shown)]. Two markers of adipocyte differentiation were also not influenced by these manipulations (Figs. 5 and 7). Because the effects of many CA inhibitors are probably not selective at the millimolar concentrations required to inhibit CA III, it may be possible to take advantage of these media effects on CA II and CA III concentrations (Table 2) to explore the role of CA in various adipocyte functions. Such studies, when coupled with data from experiments employing CA inhibitors, might provide more convincing evidence for a role of CA in particular adipocyte processes. Thus 3T3 adipocyte cell lines may be useful not only for studying the tissuespecific and hormonal regulation of CA III expression, but also for evaluating the physiological role of CA in adipocytes. We thank Lizabeth Bohlen and Ken McCall for help with these studies as well as Drs. Robert Forster II, Leonard Jefferson, Kathryn LaNoue, and Patrick Quinn for scientific advice. This work was supported in part by a feasibility grant from the American Diabetes Association, by a grant from the Juvenile Diabetes Foundation, and by National Institue of Diabetes and Digestive and Kidney Diseases Grants DK-38041 and DK-43402. The mass spectrometer facility is supported by R. E. Forster II, Program Director. C. J. Lynch is the recipient of a Ciba-Geigy New Investigator Award. Address for reprint requests: C. J. Lynch, Dept. of Cellular and

C242

CARBONIC

Molecular Medicine, Received

Physiology, The Pennsylvania Hershey, PA 17033. 7 December

1992; accepted

State in final

form

ANHYDRASE

University 26 February

College

IN 3T3 of

1993.

REFERENCES 1. Argiles, J. M. The obese Zucker rat: a choice for fat metabolism 1968-1988: twenty years of research on the insights of the Zucker mutation. Prog. Lipid Res. 28: 53-66, 1989. B., H. Purcell, P. Hemken, P. Trigg, and S. 2. Bastani, Gluck. Expression and distribution of renal vacuolar protontranslocating adenosine triphosphatase in response to chronic acid and alkali loads in the rat. J. Clin. Inuest. 88: 126-136, 1991. 3. Brady, H. J., M. Edwards, D. C. Linch, L. Knott, J. H. Barlow, and P. H. Butterworth. Expression of the human carbonic anhydrase I gene is activated late in fetal erythroid development and regulated by stage-specific trans-acting factors. Br. J. HuematoZ. 76: 135-142, 1990. 4. Bray, G. A. Lipogenesis in human adipose tissue: some effects of nibbling and gorging. J. Clin. Invest. 51: 537-548, 1972. 5. Bray. G. A. The Zucker-fatty rat: a review. Federation Proc. 36: 148-153, 1977. W. Immunostaining of carbamoylphosphate synthase 6. Cammer, II and fatty acid synthase in glial cells of rat, mouse, and hamster brains suggests roles for carbonic anhydrase in biosynthetic processes. Neurosci. Lett. 129: 247-250, 1991. N. D. Hormonal and neuronal control of carbonic anhy7. Carter, drase III gene expression in skeletal muscle. In: The Carbonic Anhydruses: Cellular Physiology and Molecular Genetics, edited by S. J. Dodgson, R. E. Tashian, G. Gros, and N. D. Carter. New York: Plenum, 1991, p. 247-256. 8. Chen, J. C. T., and M. Chesler. pH transients evoked by excitatory synaptic transmission are increased by inhibition of extracellular carbonic anhydrase. Proc. NutZ. Acud. Sci. USA 89: 7786-7790, 1990. 9. Chiappe de Cingolani, G. E., and F. V. Vega. Effect of pH on lipolysis, CAMP and CAMP-dependent protein kinase activity in isolated rat fat cells. Arch. Int. Physiol. Biochem. 88: 497-503, 1980. J. F., and W. Meier. Coordinate expression of 10. Conscience, eurythroid marker enzymes during dimethylsulfoxide-induced differentiation of Friend erythroleukemia cells. Exp. CeZl Res. 125: 111-119, 1980. 11. Coulson, A., and J. D. Herbert. A role for carbonic anhydrase in intermediary metabolism. Ann. NY Acud. Sci. 429: 505-515, 1984. 12. Djian, P., M. Phillips, and H. Green. The activation of specific gene transcription in the adipose conversion of 3T3 cells. J. Cell Physiol. 124: 554-556, 1985. 13. Dodgson, S. J. The carbonic anhydrases: overview of their importance in cellular physiology and in molecular genetics. In: The Carbonic Anhydruses: Cellular Physiology and Molecular Genetics, edited by S. J. Dodgson, R. E. Tashian, G. Gros, and N. D. Carter. New York: Plenum, 1991, p. 3-14. 14. Dodgson, S. J., and R. E. Forster II. Carbonic anhydrase: inhibition results in decreased urea production by hepatocytes. J. Appl. PhysioZ. 60: 646-652, 1986. 15. Dodgson, S. J., R. E. Forster II, and B. T. Storey. The role of carbonic anhydrase in hepatocyte metabolism. Ann. NY Acud. Sci. 429: 516-524, 1984. 16. Dodgson, S. J., and M. Watford. Differential regulation of hepatic carbonic anhydrase isozymes in the streptozotocin-diabetic rat. Arch. Biochem. Biophys. 277: 410-414, 1990. 17. Edwards, Y. H. Structure and expression of the carbonic anhydrase III gene. In: The Carbonic Anhydruses: Cellular Physiology and MolecuZur Genetics, edited by S. J. Dodgson, R. E. Tashian, G. Gros, and N. D. Carter. New York: Plenum, 1991, p. 215-224. 18. Geers, C., and G. Gros. Muscle carbonic anhydrases: function in muscle contraction and in the homeostasis of muscle pH and pC0,. In: The Carbonic Anhydruses: Cellular Physiology and Molecular Genetics, edited by S. J. Dodgson, R. E. Tashian, G. Gros, and N. D. Carter. New York: Plenum, 1991, p. 227-240. 19. Greenwood, M. R. C. Enzymatic alterations in the obese: long

ADIPOCYTES

or short-term regulatory errors. Int. J. Obesity 8: 561-569, 1984. 20. Guerre-Millo, M., M. Lavau, J. S. Horne, and L. J. Wardzada. Proposed mechanisms for increased insulin-mediated glucose transport in adipose cells from young, obese Zucker rats. J. BioZ. Chem. 260: 2197-2210, 1985. 21. Hewett-Emmett, D., and R. E. Tashian. Structure and evolutionary origins of the carbonic anhydrase multigene family. In: The Carbonic Anhydruses: Cellular Physiology and Molecular Genetics, edited by S. J. Dodgson, R. E. Tashian, G. Gros, and N. D. Carter. New York: Plenum, 1991, p. 15-32. P., and B. B. Fredholm. Direct antilipolytic effect 22. Hjemdahl, of acidosis in rat adipocytes. Actu Physiol. Scund. 101: 294-301, 1977. N., and R. E. Forster II. Carbonic anhydrase activity in 23. Itada, intact red blood cells measured with IsO exchange. J. Biol. Chem. 252: 3881-3890, 1977. S., N. D. Carter, and A. Smith. Immunocytochem24. Jeffery, ical localization of carbonic anhydrase isozymes I, II, and III in rat skeletal muscle. J. Histochem. Cytochem. 34: 513-516, 1986. S., and N. P. Curthoys. Effect of pH and bicarbonate 25. Kaiser, on phosphoenolpyruvate carboxykinase and glutaminase mRNA levels in cultured renal epithelial cells. J. Biol. Chem. 266: 93979402, 1991. M., D. F. Jensen, E. V. Wancewicz, L. L. Joy, 26. Kawamura, J. C. Khoo, and D. Steinberg. Hormone-sensitive lipase in differentiated 3T3-Ll cells and its activation by cyclic AMPdependent protein kinase. Proc. Nutl. Acud. Sci. USA 78: 732-736, 1981. E. W. Acidic extracellular environment induces only 27. Khandjian, a subset of heat-shock proteins in primary mouse kidney cell cultures. Biochem. CeZZ. BioZ. 68: 804-807, 1990. U. K. Cleavage of structural proteins during the as28. Laemmli, sembly of the head of bacteriophage T4. Nature Lond. 227: 680685, 1970. W. J., B. R. Landau, E. S. Baker, A. B. 29. Longmore, Hastings, D. M. Lum, and H. R. Williams. Effect of pH and CO, concentration on glucose metabolism by rat adipose tissue in vitro. Am. J. Physiol. 215: 582-586, 1968. 30. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. Protein measurements with the Folin phenol reagent. J. Biol. Chem. 193: 265-275, 1951. 31. Lynch, C. J., K. M. McCall, M. L. Billingsley, L. M. Bohlen, S. P. Hreniuk, L. F. Martin, L. A. Witters, and S. J. Vannucci. Pyruvate carboxylase in genetic obesity. Am. J. Physiol. 262 (Endocrinol. Metub. 25): E608-E618, 1992. 32. Lynch, C. J., S. J. Vannucci, W. A. Brennan, Jr., T. G. Vary, N. Carter, and S. J. Dodgson. Carbonic anhydrase III in obese Zucker rats. Am. J. Physiol. 264 (Endocrinol. Metub. 27): E621-E630, 1993. 33. Lyons, G. E., M. E. Buckingham, S. Tweedie, and Y. H. Edwards. Carbonic anhydrase III, an early mesodermal marker, is repressed in embryonic mouse skeletal muscle and notocord. Development 111: 233-244, 1991. 34. Maren, T. H. Carbonic anhydrase: chemistry, physiolology, and inhibition. Am. J. Physiol. 232 (Renal Fluid Electrolyte Physiol. 1): F291-F297, 1967. C. A., and H. C. Urey. The kinetics of isotopic exchange 35. Mills, between carbon dioxide, bicarbonate ion, cabonate and water. J. Am. Chem. Sot. 62: 1019-1026, 1940. 36. Poyart, C. E., and G. G. Nahas. Inhibition of activated lipolysis by acidosis. Mol. PhurmucoZ. 4: 389-401, 1968. 37. Rubio, V. Enzymatic HCO, fixation: a commom mechanism for all enzymes involved? Biosci. Rep. 6: 335-347, 1986. D., and P. W. Shank. Observations on the produc38. Rudman, tion of hydrogen ions during mobilization of fatty acids from adipose tissue. Endocrinology 79: 1 lOO- 1108, 1966. 39. Schule, W. R., R. Dorner, L. Rempe, and B. Riehl. Glial H+ transport and control of pH. Ann. NY Acud. Sci. 633: 287-305, 1991. 40. Spicer, S. S., Z. H. Ge, R. E. Tashian, D. J. Hazen-Martin, and B. A. Schulte. Compartive distribution of carbonic anhydrase isozymes III and II in rodent tissues. Am. J. Anut. 187: 55-64, 1990. 41. Spiegleman, B. M., M. Frank, and H. Green. Molecular cloning of mRNA from 3T3 adipocytes. Regulation of mRNA content

CARBONIC

ANHYDRASE

for glycerophosphate dehydrogenase and other differentiation-dependent proteins during adipocyte development. J. Biol. Chem. 258: 10083-10089, 1983. 42. Stanton, L. W., P. A. Ponte, R. T. Coleman, and M. A. Snyder. Expression of CA III in rodent models of obesity. MOL. Endocrinol. 5: 860-866, 1991.

IN 3T3

ADIPOCYTES

C243

43. Student, A. K., R. Y. Hsu, and M. D. Lane. Induction of fatty acid synthetase in differentiating 3T3-Ll pre-adipocytes. J. BioZ. Chem. 255: 4745-4750, 1980. 44. Tashian, R. E. The carbonic anhydrases: widening perspectives on their evolution, expression and function. Bioessuys 10: 186192, 1990.