Cell, Vol. 5, 1-9,
May 1975,
Copyright01975
by MIT
Morphological Changes in Cultured Mammalian Cells: Prevention by the Calcium lonophore A231 87 R. C. Henneberry, P. H. Fishman, and E. Freese National Institute of Neurological and Communicative Diseases and Stroke National Institutes of Health Department of Health, Education, and Welfare Bethesda, Maryland 20014
Summary The morphological changes induced by butyrate In HeLa cells and by monobutyryl or dibutyryl CAMP in CHO cells are prevented by mlcromolar concentrations of the dlvalent cation ionophore A23187. The lonophore is unable to prevent such changes in medium from which calcium Is omitted. At sllghtly higher (but nontoxic) concentrations, the ionophore lnhlblts the butyrate-mediated induction of the ganglloside biosynthetic enzyme, sialyltransferase, in HeLa. In CHO, sialyltransferase activity is normally high and not altered by any of the compounds tested. Introduction Many methods for inducing changes in the morphology of mammalian cells in culture have been described. Several systems have been studied in detail because regulation of morphology is thought to be important for the understanding of cell growth, differentiation, and malignant transformation. However, the mechanisms controlling these morphological changes are still obscure. Although cyclic nucleotides appear to play an important role in the control of cell shape, unrelated compounds have been found to elicit similar results in some systems. For example, Chinese hamster ovary (CHO) cells respond to monobutyryl or dibutyryl CAMP by elongation and lengthwise association with other cells (Hsie and Puck, 1971). In fibroblasts, phosphodiesterase inhibitors cause elevation of intracellular CAMP and promote a similar elongation; however, in the same cell lines the NQubstituted derivatives of adenine produce the same effects without altering intracellular CAMP levels (Johnson, D’Armiento, and Carchman, 1974). Viral transformation of fibroblasts also causes changes in cell shape that can be reversed by agents which cause increases in CAMP levels in the cell (Johnson, Friedman, and Pastan, 1973). In contrast, in BHK fibroblasts insulin is thought to cause morphological changes by inhibiting adenylate cyclase and thus lowering intracellular CAMP levels (Jimenez de Asua et al., 1973). Short chain fatty acids such as butyrate also cause striking changes in certain cells (Ginsburg et al., 1973; Wright, 1973). Prior to the appearance of neurite-like processes in HeLa cells, the content
of a particular glycosphingolipid, N-acetylneuraminylgalactosylglucosylceramide (Gus), and the corresponding biosynthetic enzyme, CMP-sialic acid:lactosylceramide sialyltransferase (“sialyltransferase”) increase after exposure to butyrate (Fishman et al., 1974; Simmons et al., 1975). The butyrate effects in HeLa are not produced by CAMP or its derivatives and appear to be independent of changes in intracellular CAMP levels (Ginsburg et al., 1973; Simmons et al., 1975; Fishman, Simmons, and Henneberry, manuscript in preparation). Some of the morphological effects described above require RNA and protein synthesis; others do not. However, all are sensitive to colchicine and/or vinca alkaloids, thus implicating assembly of microtubules as a requirement for cell elongation and process formation. Since calcium has an apparent role in the regulation of microtubule assembly (Borisy et al., 1974; Gallin and Rosenthal, 1974) we examined the effects on cellular morphology of the ionophore A23187, which is known to alter intracellular calcium levels (Reed and Lardy, 1972a). In this paper we report that the morphological changes induced in HeLa cells by butyrate and in CHO cells by dibutyryl CAMP are prevented by low concentrations of the ionophore in a calciumdependent manner. At higher concentrations of the ionophore the butyrate-mediated increase in sialyltransferase in HeLa is also inhibited. Results Morphological Changes The epithelioid morphology of HeLa cells incubated in Eagle’s minimal essential medium containing 10% fetal calf serum (“medium”) and the appearance of flattened cells with extensive processes after 24 hr incubation in the presence of 2 mM butyrate can be compared in Figures 1 a and 1 b. Prevention of butyrate-induced process formation by the ionophore A23187 (3.6 PM) is shown in Figure 1 c. Cells incubated in the presence of both butyrate and ionophore assume a slightly extended but more uniform shape than do the untreated cells shown in Figure 1 a; cells incubated in the presence of ionophore alone appear identical to untreated cells. Cells exposed to both butyrate and ionophore undergo no more than one cell division, as has been reported for the effect of butyrate alone (Ginsburg et al., 1973; Simmons et al., 1975). However, these cells remain viable: removal of both butyrate and ionophore after 24 hr exposure permits resumption of exponential growth after a short lag, while removal of only the ionophore is followed by the typical butyrate-induced morphological changes within 24 hr. Cytotoxic effects of the ionophore, indicated
Cell 2
Figure
1. Photomicrographs
Cells were grown (C) 2 mM butyrate
Showing
the Effects
of Butyrate
and lonophore
24 hr in normal medium and incubated an additional plus 3.6 PM A23187. and (D) 3.6 PM A23187 alone.
on HeLa
Cell Morphology
24 hr in medium containing: Bar represents 50 pm.
(A) no additions,
(8) 2 mM butyrate,
Figure 2. Photomicrographs of HeLa Cells Showing the Dependence of the lonophore Effect on Calcium
the Reversal (C and D)
of Butyrate-Induced
Morphological
Ch’anges
by A23187
(A and
B) and
Cells were incubated: (A) 24 hr in normal medium then 46 hr in medium containing 2 mM butyrate; (B) 24 hr in normal medium then 24 hr in medium with 2 mM butyrate, followed by an additional 22 hr in medium with 2 mM butyrate and 3.6 pM A23187; (C) 24 hr in medium lacking calcium then 24 hr in medium lacking calcium but containing 2 mM butyrate and 3.6 PM A23187: and (D) 24 hr in medium lacking calcium, then 24 hr in medium containing 2 mM butyrate, 3.6 pM A231 87, and 1.8 mM calcium. Bar represents 50 flrn.
Cell 4
by the appearance of crenated and/or floating cells, are seen at concentrations higher than about 10 PM. As previously reported (Simmons and Breuer, 1975) butyrate-induced morphological changes in HeLa cells are reversible; removal of butyrate permits resumption of normal morphology within 24 hr. Addition of ionophore to cells exposed to butyrate for 24 hr has a similar effect: the fully induced shape changes are reversed even though butyrate remains in the medium, as can be seen in Figures 2a and 2b. From reported effects of the ionophore A23187 (Reed and Lardy, 1972a), one would assume that it prevents process formation by causing an increase in the intracellular calcium concentration. This assumption is supported by the finding that the ionophore is unable to prevent process formation in medium from which calcium salts have been omitted, as shown in Figure 2c. An increase in extracellular phosphate has been shown to increase intracellular calcium in mamma-
Figure 3. Photomicrographs Showing the Effect of Increased Extracellular Phosphate Concentration on Butyrate-Induced Morphological Changes in HeLa Cells were grown 24 hr in normal medium and incubated an additional 24 hr in medium containing (A) 2 mM butyrate and (8) 2 mM butyrate plus 20 mM sodium phosphate. Bar represents 50 pm.
lian cells (Borle, 1973). If the ionophore prevents process formation in HeLa by causing an influx of calcium, a high phosphate concentration in the medium might be expected to have a similar effect. Figure 3 demonstrates that butyrate-induced process formation in HeLa is prevented when the phosphate concentration is increased from the 1 .l mM in regular medium to 20 mM. The effect of phosphate is less striking than that of the ionophore; further increases in phosphate concentration are not meaningful due to calcium phosphate precipitation in the medium. (Some calcium phosphate granules can be seen in Figure 3b.) The effect of the ionophore on the morphological alterations in CHO cells caused by the derivatives of CAMP is shown in Figure 4. 1 mM CAMP does not cause noticeable shape changes in the 24 hr incubation period used in these studies, as shown in Figure 4b. 1 mM monobutyryl or dibutyryl CAMP induce similar shape changes that are not significantly affected by the addition of testosterone propionate, as can be seen in Figures 4c through 4e. The induction of morphological changes in CHO cells is prevented by the ionophore A23187 (3.6 PM), as shown in Figure 4f. This ionophore concentration prevents both the elongation and the lengthwise association of CHO cells caused by monobutyryl or dibutyryl CAMP with or without testosterone propionate. CHO cells exposed to the ionophore remain viable: removal of both ionophore and dibutyryl CAMP permits resumption of growth within 24 hr. In agreement with earlier results (Wright, 1973), butyrate is also capable of causing morphological changes in CHO cells, but longer incubation times are required for similar results; these butyrate-induced shape changes in CHO cells are also prevented by the ionophore (3.6 PM). Levels of Enzymes and Gangliosides Table 1, which was compiled from several separate experiments, shows the effects of exposure to various concentrations of butyrate and ionophore on sialyltransferase activity and morphology in HeLa. The activity of a second enzyme involved in ganglioside biosynthesis, UDP-galactose:glucosylceramide galactosyltransferase (“galactosyltransferase”) is shown to illustrate the relative specificity of sialyltransferase induction by butyrate. Sialyltransferase activity in cells exposed to butyrate for 18 hr is about 20 fold higher than in untreated cells, whereas galactosyltransferase activity is slightly reduced. The ionophore alone has no effect on the morphology or the sialyltransferase activity of HeLa, while it causes a slight increase in galactosyltransferase activity. The remainder of Table 1 shows that the ionophore completely prevents butyrate-induced process formation even at the lowest con-
lonophore 5
Prevents
Shape
Changes
in HeLa
and CHO
centration used (1.4 ,uM), whereas the sialyltransferase activity is depressed only at higher concentrations, apparently in a dose-dependent manner. The prevention of the butyrate effect by the ionophore is not caused by a nonspecific inhibition of protein synthesis, as can be seen from Table 2. Both the incorporation of amino acids and the spe-
cific activity of alkaline phosphatase are not significantly affected by the ionophore. The inhibition of sialyltransferase induction is dependent on the concentration of ionophore. Both Tables 1 and 2 show that concentrations of A23187 which completely block process formation are insufficient to inhibit sialyltransferase induction. The small increase in alkaline phosphatase activity caused by incubation
Figure
and
4. Photomicrographs
Showing
the Effects
of Cyclic
Nucleotides
lonophore
A23187
on CHO Cell Morphology
Cells were grown 24 hr in normal medium and incubated an additional 24 hr in medium containing: (A) no additions, (B) 1 mM CAMP, (C) 1 mM monobutyryl CAMP, (D) 1 mM dibutyryl CAMP. (E) 1 mM dibutyryl CAMP plus 1.5 x 1O-5 M testosterone propionate. and (F) 1 mM dibutyryl CAMP plus 1.5 x 10-S M testosterone propionate, plus 3.6 PM A23187. Cells incubated in medium with 3.6 pM A23167 alone appear identical to the cells shown in (A). Bar represents 50 pm.
Cell 6
in the presence of butyrate appears to be prevented by the ionophore. Induction of alkaline phosphatase by butyrate and dibutyryl CAMP has recently been reported in HeLa, but longer incubation times (72 hr) than those used in our studies were required (Griffin et al., 1974). These investigators suggested that the butyrate cleaved from dibutyryl CAMP is the mediator of induction of alkaline phosphatase in HeLa. We have not observed any effect of 1 mM dibutyryl CAMP on morphology or sialyltransferase activity in HeLa during 24 hr of exposure. In contrast to the effect on HeLa, in CHO cells neither butyrate nor dibutyryl CAMP significantly influence the levels of GU3 or its biosynthetic enzyme within 18 hr, as can be seen in Table 3. However, the specific activity of sialyltransferase in untreated CHO cells is as high as that seen in butyrate-treated HeLa cells. The relative enzyme levels are reflected in a several fold higher content of Gr.,3 in CHO cell membranes compared to membranes from untreated HeLa cells. The Gu3 content of untreated CHO cells is comparable to that of butyrate-treated HeLa, and the glycolipid composition of CHO membranes is similar to that reported for HeLa (Simmons et al., 1975), containing several neutral glycosphingolipids in addition to GU3. Absence of more complex gangliosides in CHO and HeLa appears to be due to negligible activity of the enzyme UDPN-acetylgalactosamine:GH3 N-acetylgalactosaminyl transferase which is required for their synthesis.
by butyrate in HeLa cells and by monobutyryl or dibutyryl CAMP in CHO cells. Cells treated in this manner continue to synthesize protein and remain viable. The simplest interpretation of these results is that the ionophore prevents shape changes by causing an increase in intracellular calcium. This interpretation, based on reported effects of A23187 (Reed and Lardy, 1972a), is supported by the inability of the ionophore to prevent morphological changes in the absence of calcium. It is also supported by the corresponding effect of increased extracellular phosphate, which is reported to increase calcium levels within the cell (Borle, 1973). Micromolar concentrations of A23187 have diverse biological effects which can all be related to its functioning as a calcium ionophore. A231 87 uncouples oxidative phosphorylation in rat liver mitochondria (Reed and Lardy, 1972b) and activates sea urchin eggs (Steinhardt and Epel, 1974) apparently by releasing calcium from mitochondria. In rat mast cells it stimulates release of histamine, primarily by facilitating calcium transport into the cell (Foreman, Mongar, and Gomperts, 1973). Our results do not permit us to determine whether the ionophore functions primarily by increasing calcium transport into the cell or by stimulating release of calcium from mitochondria. Similarly, we cannot distinguish between direct and indirect effects of calcium. The relationship between calcium, phosphate, and hydrogen ions within the cell is complex (Rasmussen, Goodman, and Tenenhouse, 1972); it may be that the intracellular pH, for example, is the important physiological parameter and is disturbed by an influx of calcium.
Discussion We have shown that the ionophore A231 87, a monocarboxylic acid antibiotic specific for divalent cations, prevents the changes in cell shape induced
Table
1. Effect
Additions
of Butyrate
and lonophore
on Glycosyltransferase
Activities
in HeLa
Cells*
to Medium
Butyrate
A231 87
Cm@
(PM)
2.5 5.4
Glycosyltransferases
(pmoles/mg
protein/hr)
Sialyltransferase
Galactosyltransferase
2.6
284
50.0
181
2.6
443
2.5
1.4
75.1
ND
2.5
2.7
32.9
ND
2.5
5.4
22.8
233
52.3
ND
5.0
5.4
25.0
ND
5.0
9.0
6.1
ND
5.0
*HeLa cells seeded at 20,000 cells/cm2 in 150mm dishes were grown 24 hr in standard medium medium containing the indicated additions. After 18 hr additional incubation, photomicrographs changes, cells harvested, and enzymes assayed as described in Experimental Procedures. ND = Not determined
and the medium replaced were taken for scoring
Process Formation
+
+
with identical morphological
lonophore 7
Prevents
Shape
Changes
in HeLa
and CHO
Prevention of the dibutyryl CAMP-induced morphological changes in CHO cells by the ionophore suggests a relationship between CAMP and calcium. The role of CAMP as the “second messenger” in many systems is well known (Sutherland, Oye, and Butcher, 1965), as is the role of calcium in muscle contraction (Martonosi, 1971) activation of exocytosis (Cochrane and Douglas, 1974), and control of cell proliferation (Boynton et al., 1974; Whitfield et al., 1973). The complex relationship between calcium and cyclic nucleotides in many other systems has become apparent in recent years (Rasmussen, 1970; Rasmussen et al., 1972; Durham, 1974). Many cyclic nucleotide-mediated phenomena are calcium dependent; cytoplasmic calcium levels are known to influence adenylate cyclase activity, and cyclic nucleotides have been found to affect intracellular calcium levels (Rasmussen, 1970; Durham, 1974). While it is difficult to distinguish primary and secondary effects in these systems, some investigators have proposed that calcium may actually be the “second messenger” in some cases (Rasmussen et al., 1972).
Table
2. Effects
Additions Butyrate
A23187
3.5 mM
of Butyrate
and lonophore
on Sialyltransferase,
Calcium has been shown to prevent the assembly of tubulin into microtubules in vitro (Borisy et al., 1974); recent in vivo evidence indicates that cytoplasmic calcium levels have a role in regulating microtubule assembly in human granulocytes (Gallin and Rosenthal, 1974). The morphological changes in the different cell lines used in our studies apparently require microtubule assembly; assembly and alignment of microtubules have been correlated with dibutyryl CAMP-induced shape changes in CHO cells (Porter et al., 1974), and in HeLa process formation is sensitive to colchicine (Ginsburg et al., 1973; Simmons et al., 1975). However, the intracellular content of tubulin is not significantly affected by exposure to butyrate; a comparison of butyratetreated and untreated HeLa cells reveals no significant difference in tubulin content per cell based on binding of radioactive colchicine (R. C. Henneberry, unpublished data). The relationship between the alteration in membrane ganglioside levels and the induction of shape changes by butyrate in HeLa is not yet clear. We have no direct evidence that the increase in Gul
Alkaline
Phosphatase,
Amino Acid Incorporated (pmoles/l06 cells/hr)
Cells/Dish (X 10-b)
Protein
Synthesis,
Alkaline Phosphataseb %
and Morphology
Sialyltransferaseb %
10.8
957
70
May 1975,
Copyright01975
by MIT
Morphological Changes in Cultured Mammalian Cells: Prevention by the Calcium lonophore A231 87 R. C. Henneberry, P. H. Fishman, and E. Freese National Institute of Neurological and Communicative Diseases and Stroke National Institutes of Health Department of Health, Education, and Welfare Bethesda, Maryland 20014
Summary The morphological changes induced by butyrate In HeLa cells and by monobutyryl or dibutyryl CAMP in CHO cells are prevented by mlcromolar concentrations of the dlvalent cation ionophore A23187. The lonophore is unable to prevent such changes in medium from which calcium Is omitted. At sllghtly higher (but nontoxic) concentrations, the ionophore lnhlblts the butyrate-mediated induction of the ganglloside biosynthetic enzyme, sialyltransferase, in HeLa. In CHO, sialyltransferase activity is normally high and not altered by any of the compounds tested. Introduction Many methods for inducing changes in the morphology of mammalian cells in culture have been described. Several systems have been studied in detail because regulation of morphology is thought to be important for the understanding of cell growth, differentiation, and malignant transformation. However, the mechanisms controlling these morphological changes are still obscure. Although cyclic nucleotides appear to play an important role in the control of cell shape, unrelated compounds have been found to elicit similar results in some systems. For example, Chinese hamster ovary (CHO) cells respond to monobutyryl or dibutyryl CAMP by elongation and lengthwise association with other cells (Hsie and Puck, 1971). In fibroblasts, phosphodiesterase inhibitors cause elevation of intracellular CAMP and promote a similar elongation; however, in the same cell lines the NQubstituted derivatives of adenine produce the same effects without altering intracellular CAMP levels (Johnson, D’Armiento, and Carchman, 1974). Viral transformation of fibroblasts also causes changes in cell shape that can be reversed by agents which cause increases in CAMP levels in the cell (Johnson, Friedman, and Pastan, 1973). In contrast, in BHK fibroblasts insulin is thought to cause morphological changes by inhibiting adenylate cyclase and thus lowering intracellular CAMP levels (Jimenez de Asua et al., 1973). Short chain fatty acids such as butyrate also cause striking changes in certain cells (Ginsburg et al., 1973; Wright, 1973). Prior to the appearance of neurite-like processes in HeLa cells, the content
of a particular glycosphingolipid, N-acetylneuraminylgalactosylglucosylceramide (Gus), and the corresponding biosynthetic enzyme, CMP-sialic acid:lactosylceramide sialyltransferase (“sialyltransferase”) increase after exposure to butyrate (Fishman et al., 1974; Simmons et al., 1975). The butyrate effects in HeLa are not produced by CAMP or its derivatives and appear to be independent of changes in intracellular CAMP levels (Ginsburg et al., 1973; Simmons et al., 1975; Fishman, Simmons, and Henneberry, manuscript in preparation). Some of the morphological effects described above require RNA and protein synthesis; others do not. However, all are sensitive to colchicine and/or vinca alkaloids, thus implicating assembly of microtubules as a requirement for cell elongation and process formation. Since calcium has an apparent role in the regulation of microtubule assembly (Borisy et al., 1974; Gallin and Rosenthal, 1974) we examined the effects on cellular morphology of the ionophore A23187, which is known to alter intracellular calcium levels (Reed and Lardy, 1972a). In this paper we report that the morphological changes induced in HeLa cells by butyrate and in CHO cells by dibutyryl CAMP are prevented by low concentrations of the ionophore in a calciumdependent manner. At higher concentrations of the ionophore the butyrate-mediated increase in sialyltransferase in HeLa is also inhibited. Results Morphological Changes The epithelioid morphology of HeLa cells incubated in Eagle’s minimal essential medium containing 10% fetal calf serum (“medium”) and the appearance of flattened cells with extensive processes after 24 hr incubation in the presence of 2 mM butyrate can be compared in Figures 1 a and 1 b. Prevention of butyrate-induced process formation by the ionophore A23187 (3.6 PM) is shown in Figure 1 c. Cells incubated in the presence of both butyrate and ionophore assume a slightly extended but more uniform shape than do the untreated cells shown in Figure 1 a; cells incubated in the presence of ionophore alone appear identical to untreated cells. Cells exposed to both butyrate and ionophore undergo no more than one cell division, as has been reported for the effect of butyrate alone (Ginsburg et al., 1973; Simmons et al., 1975). However, these cells remain viable: removal of both butyrate and ionophore after 24 hr exposure permits resumption of exponential growth after a short lag, while removal of only the ionophore is followed by the typical butyrate-induced morphological changes within 24 hr. Cytotoxic effects of the ionophore, indicated
Cell 2
Figure
1. Photomicrographs
Cells were grown (C) 2 mM butyrate
Showing
the Effects
of Butyrate
and lonophore
24 hr in normal medium and incubated an additional plus 3.6 PM A23187. and (D) 3.6 PM A23187 alone.
on HeLa
Cell Morphology
24 hr in medium containing: Bar represents 50 pm.
(A) no additions,
(8) 2 mM butyrate,
Figure 2. Photomicrographs of HeLa Cells Showing the Dependence of the lonophore Effect on Calcium
the Reversal (C and D)
of Butyrate-Induced
Morphological
Ch’anges
by A23187
(A and
B) and
Cells were incubated: (A) 24 hr in normal medium then 46 hr in medium containing 2 mM butyrate; (B) 24 hr in normal medium then 24 hr in medium with 2 mM butyrate, followed by an additional 22 hr in medium with 2 mM butyrate and 3.6 pM A23187; (C) 24 hr in medium lacking calcium then 24 hr in medium lacking calcium but containing 2 mM butyrate and 3.6 PM A23187: and (D) 24 hr in medium lacking calcium, then 24 hr in medium containing 2 mM butyrate, 3.6 pM A231 87, and 1.8 mM calcium. Bar represents 50 flrn.
Cell 4
by the appearance of crenated and/or floating cells, are seen at concentrations higher than about 10 PM. As previously reported (Simmons and Breuer, 1975) butyrate-induced morphological changes in HeLa cells are reversible; removal of butyrate permits resumption of normal morphology within 24 hr. Addition of ionophore to cells exposed to butyrate for 24 hr has a similar effect: the fully induced shape changes are reversed even though butyrate remains in the medium, as can be seen in Figures 2a and 2b. From reported effects of the ionophore A23187 (Reed and Lardy, 1972a), one would assume that it prevents process formation by causing an increase in the intracellular calcium concentration. This assumption is supported by the finding that the ionophore is unable to prevent process formation in medium from which calcium salts have been omitted, as shown in Figure 2c. An increase in extracellular phosphate has been shown to increase intracellular calcium in mamma-
Figure 3. Photomicrographs Showing the Effect of Increased Extracellular Phosphate Concentration on Butyrate-Induced Morphological Changes in HeLa Cells were grown 24 hr in normal medium and incubated an additional 24 hr in medium containing (A) 2 mM butyrate and (8) 2 mM butyrate plus 20 mM sodium phosphate. Bar represents 50 pm.
lian cells (Borle, 1973). If the ionophore prevents process formation in HeLa by causing an influx of calcium, a high phosphate concentration in the medium might be expected to have a similar effect. Figure 3 demonstrates that butyrate-induced process formation in HeLa is prevented when the phosphate concentration is increased from the 1 .l mM in regular medium to 20 mM. The effect of phosphate is less striking than that of the ionophore; further increases in phosphate concentration are not meaningful due to calcium phosphate precipitation in the medium. (Some calcium phosphate granules can be seen in Figure 3b.) The effect of the ionophore on the morphological alterations in CHO cells caused by the derivatives of CAMP is shown in Figure 4. 1 mM CAMP does not cause noticeable shape changes in the 24 hr incubation period used in these studies, as shown in Figure 4b. 1 mM monobutyryl or dibutyryl CAMP induce similar shape changes that are not significantly affected by the addition of testosterone propionate, as can be seen in Figures 4c through 4e. The induction of morphological changes in CHO cells is prevented by the ionophore A23187 (3.6 PM), as shown in Figure 4f. This ionophore concentration prevents both the elongation and the lengthwise association of CHO cells caused by monobutyryl or dibutyryl CAMP with or without testosterone propionate. CHO cells exposed to the ionophore remain viable: removal of both ionophore and dibutyryl CAMP permits resumption of growth within 24 hr. In agreement with earlier results (Wright, 1973), butyrate is also capable of causing morphological changes in CHO cells, but longer incubation times are required for similar results; these butyrate-induced shape changes in CHO cells are also prevented by the ionophore (3.6 PM). Levels of Enzymes and Gangliosides Table 1, which was compiled from several separate experiments, shows the effects of exposure to various concentrations of butyrate and ionophore on sialyltransferase activity and morphology in HeLa. The activity of a second enzyme involved in ganglioside biosynthesis, UDP-galactose:glucosylceramide galactosyltransferase (“galactosyltransferase”) is shown to illustrate the relative specificity of sialyltransferase induction by butyrate. Sialyltransferase activity in cells exposed to butyrate for 18 hr is about 20 fold higher than in untreated cells, whereas galactosyltransferase activity is slightly reduced. The ionophore alone has no effect on the morphology or the sialyltransferase activity of HeLa, while it causes a slight increase in galactosyltransferase activity. The remainder of Table 1 shows that the ionophore completely prevents butyrate-induced process formation even at the lowest con-
lonophore 5
Prevents
Shape
Changes
in HeLa
and CHO
centration used (1.4 ,uM), whereas the sialyltransferase activity is depressed only at higher concentrations, apparently in a dose-dependent manner. The prevention of the butyrate effect by the ionophore is not caused by a nonspecific inhibition of protein synthesis, as can be seen from Table 2. Both the incorporation of amino acids and the spe-
cific activity of alkaline phosphatase are not significantly affected by the ionophore. The inhibition of sialyltransferase induction is dependent on the concentration of ionophore. Both Tables 1 and 2 show that concentrations of A23187 which completely block process formation are insufficient to inhibit sialyltransferase induction. The small increase in alkaline phosphatase activity caused by incubation
Figure
and
4. Photomicrographs
Showing
the Effects
of Cyclic
Nucleotides
lonophore
A23187
on CHO Cell Morphology
Cells were grown 24 hr in normal medium and incubated an additional 24 hr in medium containing: (A) no additions, (B) 1 mM CAMP, (C) 1 mM monobutyryl CAMP, (D) 1 mM dibutyryl CAMP. (E) 1 mM dibutyryl CAMP plus 1.5 x 1O-5 M testosterone propionate. and (F) 1 mM dibutyryl CAMP plus 1.5 x 10-S M testosterone propionate, plus 3.6 PM A23187. Cells incubated in medium with 3.6 pM A23167 alone appear identical to the cells shown in (A). Bar represents 50 pm.
Cell 6
in the presence of butyrate appears to be prevented by the ionophore. Induction of alkaline phosphatase by butyrate and dibutyryl CAMP has recently been reported in HeLa, but longer incubation times (72 hr) than those used in our studies were required (Griffin et al., 1974). These investigators suggested that the butyrate cleaved from dibutyryl CAMP is the mediator of induction of alkaline phosphatase in HeLa. We have not observed any effect of 1 mM dibutyryl CAMP on morphology or sialyltransferase activity in HeLa during 24 hr of exposure. In contrast to the effect on HeLa, in CHO cells neither butyrate nor dibutyryl CAMP significantly influence the levels of GU3 or its biosynthetic enzyme within 18 hr, as can be seen in Table 3. However, the specific activity of sialyltransferase in untreated CHO cells is as high as that seen in butyrate-treated HeLa cells. The relative enzyme levels are reflected in a several fold higher content of Gr.,3 in CHO cell membranes compared to membranes from untreated HeLa cells. The Gu3 content of untreated CHO cells is comparable to that of butyrate-treated HeLa, and the glycolipid composition of CHO membranes is similar to that reported for HeLa (Simmons et al., 1975), containing several neutral glycosphingolipids in addition to GU3. Absence of more complex gangliosides in CHO and HeLa appears to be due to negligible activity of the enzyme UDPN-acetylgalactosamine:GH3 N-acetylgalactosaminyl transferase which is required for their synthesis.
by butyrate in HeLa cells and by monobutyryl or dibutyryl CAMP in CHO cells. Cells treated in this manner continue to synthesize protein and remain viable. The simplest interpretation of these results is that the ionophore prevents shape changes by causing an increase in intracellular calcium. This interpretation, based on reported effects of A23187 (Reed and Lardy, 1972a), is supported by the inability of the ionophore to prevent morphological changes in the absence of calcium. It is also supported by the corresponding effect of increased extracellular phosphate, which is reported to increase calcium levels within the cell (Borle, 1973). Micromolar concentrations of A23187 have diverse biological effects which can all be related to its functioning as a calcium ionophore. A231 87 uncouples oxidative phosphorylation in rat liver mitochondria (Reed and Lardy, 1972b) and activates sea urchin eggs (Steinhardt and Epel, 1974) apparently by releasing calcium from mitochondria. In rat mast cells it stimulates release of histamine, primarily by facilitating calcium transport into the cell (Foreman, Mongar, and Gomperts, 1973). Our results do not permit us to determine whether the ionophore functions primarily by increasing calcium transport into the cell or by stimulating release of calcium from mitochondria. Similarly, we cannot distinguish between direct and indirect effects of calcium. The relationship between calcium, phosphate, and hydrogen ions within the cell is complex (Rasmussen, Goodman, and Tenenhouse, 1972); it may be that the intracellular pH, for example, is the important physiological parameter and is disturbed by an influx of calcium.
Discussion We have shown that the ionophore A231 87, a monocarboxylic acid antibiotic specific for divalent cations, prevents the changes in cell shape induced
Table
1. Effect
Additions
of Butyrate
and lonophore
on Glycosyltransferase
Activities
in HeLa
Cells*
to Medium
Butyrate
A231 87
Cm@
(PM)
2.5 5.4
Glycosyltransferases
(pmoles/mg
protein/hr)
Sialyltransferase
Galactosyltransferase
2.6
284
50.0
181
2.6
443
2.5
1.4
75.1
ND
2.5
2.7
32.9
ND
2.5
5.4
22.8
233
52.3
ND
5.0
5.4
25.0
ND
5.0
9.0
6.1
ND
5.0
*HeLa cells seeded at 20,000 cells/cm2 in 150mm dishes were grown 24 hr in standard medium medium containing the indicated additions. After 18 hr additional incubation, photomicrographs changes, cells harvested, and enzymes assayed as described in Experimental Procedures. ND = Not determined
and the medium replaced were taken for scoring
Process Formation
+
+
with identical morphological
lonophore 7
Prevents
Shape
Changes
in HeLa
and CHO
Prevention of the dibutyryl CAMP-induced morphological changes in CHO cells by the ionophore suggests a relationship between CAMP and calcium. The role of CAMP as the “second messenger” in many systems is well known (Sutherland, Oye, and Butcher, 1965), as is the role of calcium in muscle contraction (Martonosi, 1971) activation of exocytosis (Cochrane and Douglas, 1974), and control of cell proliferation (Boynton et al., 1974; Whitfield et al., 1973). The complex relationship between calcium and cyclic nucleotides in many other systems has become apparent in recent years (Rasmussen, 1970; Rasmussen et al., 1972; Durham, 1974). Many cyclic nucleotide-mediated phenomena are calcium dependent; cytoplasmic calcium levels are known to influence adenylate cyclase activity, and cyclic nucleotides have been found to affect intracellular calcium levels (Rasmussen, 1970; Durham, 1974). While it is difficult to distinguish primary and secondary effects in these systems, some investigators have proposed that calcium may actually be the “second messenger” in some cases (Rasmussen et al., 1972).
Table
2. Effects
Additions Butyrate
A23187
3.5 mM
of Butyrate
and lonophore
on Sialyltransferase,
Calcium has been shown to prevent the assembly of tubulin into microtubules in vitro (Borisy et al., 1974); recent in vivo evidence indicates that cytoplasmic calcium levels have a role in regulating microtubule assembly in human granulocytes (Gallin and Rosenthal, 1974). The morphological changes in the different cell lines used in our studies apparently require microtubule assembly; assembly and alignment of microtubules have been correlated with dibutyryl CAMP-induced shape changes in CHO cells (Porter et al., 1974), and in HeLa process formation is sensitive to colchicine (Ginsburg et al., 1973; Simmons et al., 1975). However, the intracellular content of tubulin is not significantly affected by exposure to butyrate; a comparison of butyratetreated and untreated HeLa cells reveals no significant difference in tubulin content per cell based on binding of radioactive colchicine (R. C. Henneberry, unpublished data). The relationship between the alteration in membrane ganglioside levels and the induction of shape changes by butyrate in HeLa is not yet clear. We have no direct evidence that the increase in Gul
Alkaline
Phosphatase,
Amino Acid Incorporated (pmoles/l06 cells/hr)
Cells/Dish (X 10-b)
Protein
Synthesis,
Alkaline Phosphataseb %
and Morphology
Sialyltransferaseb %
10.8
957
70
