- Email: [email protected]
Ligand-induced conformational changes in cytosolic protein kinase C
Ligand-induced conformational changes in cytosolic protein kinase C Davis S. Lester* and Vlad Brmnfeld Department of Membrane Research, The Weizmann Institute of Science, Rehovot 76100, Israel
(Received 5 September 1989; revised 21 January 1990) The changes in intrinsic spectral properties of protein kinase C were monitored upon association with its divalent cation and lipid activators in a model membrane system. The enzyme demonstrated changes in both its intrinsic fluorescence and far ultraviolet circular dichroism spectra upon association with lipid vesicles in the absence of calcium. The acidic phospholipid, phasphatidylserine, sionificantly quenched the intrinsic tryptophan fluorescence and was also the most potent lipid support for the phosphorylating activity of the enzyme. The enzyme was fully activated by a number of Ca2+-lipid combinations which correlated with maximal fluorescence quenching (40-50 %) of available tryptophan residues in hydrophobic domains. The circular dichroism structure of the associated active-protein Ca 2 +-lipid complexes suooested different active enzyme secondary structures. However, the Ca2+-dependent changes in fluorescence and circular dichroism spectra were observed only after the enzyme associated with the lipid vesicles. These data suggest that protein kinase C has the properties of a complex multidomain protein and provides an additional perspective into the mechanism of protein kinase C activation. Keywords: Protein kinaseC; lipids; diacylglycerol;phorbol ester; conformation;tryptophan;circular dichroism
Introduction The Ca 2 +/phospholipid-dependent protein kinase, protein kinase C (PKC), is considered to play a significant role in numerous cellular processes including transmembrane signal transduction and tumour promotion 1'2. PKC is activated in vitro by Ca 2÷, phosphatidylserine (PS), and lipid activators such as diacylglycerol (DAG) or the exogenous class of tumour promoting agents (TPA) known as the phorbol esters t. In addition, Mg 2+ is required for optimal tritiated phorbol ester binding a'4, while Mg2+'ATP is the preferred substrate for the phosphotransferase activity2,s. Ca z+ ions are generally considered to induce the translocation of the enzyme from the cytosol to the plasma membrane ~,2,6, possibly by binding to the soluble enzyme and eliciting a conformational change promoting translocation6. However, examination of the primary structure deduced from cDNA sequence does not reveal any of the motifs typifying a Ca 2+ binding site 7'8. Additionally, PKC has been shown to associate and penetrate lipid bilayers in the absence of divalent cations 9. The enzyme is a single polypeptide, however, it has been subdivided into two functional domain structures. The N-terminal region is considered to be the regulatory domain as it has specific binding regions for Ca 2 +, PS and the lipid activator, while the C-terminal region is the catalytic domain containing the Mg 2+ "ATP and the substrate binding site 7'a'1 o. Different regions of each domain have been proposed to be involved in binding of the various ligands to PKC ~°. We have recently shown that half of the PKC tryptophan residues * To whomcorrespondenceshouldbe addressedat: NIH, Sectionon Neural Systems,Laboratoryof Molecularand CellularNeurobiology, Park 5 Bldg,Room 435, Bethesda,MD 20892, USA. 0141-8130/90/040251-06 © 1990 Butterworth-HeinemannLimited
are in regions of the protein that undergo conformational change(s) upon addition of the activating lipids 9. In this report, we have examined the contribution of the various activating cofactors by monitoring changes in the intrinsic spectroscopic properties of the protein. The data obtained suggest that PKC has a number of conformational states and the process of PKC activation may occur in a different sequence than that which has been proposed according to the models based on biochemical analyses.
Experimental Cytosolic PKC was purified from male Wistar rat brains (6 weeks) according to the procedure of Huang et a l ) I. Generally, 50 such brains yielded an average of 1.0 mg of pure enzyme (single band on a Coomassie and silver stained an SDS-polyacrylamide gel) containing all three isozymes (Types I, II and III). The specific activity of the enzyme preparation was approximately 1000 units/mg proteins. Lipids were purchased from Lipid Products Inc. or Serdary Research Laboratories. Phorbol myristate acetate (TPA), histone III-S and all other chemicals were purchased from Sigma Chemical Co. [7-32p]ATP (3500 Ci/mmol) was obtained from New England Nuclear. All buffers were prepared from double distilled, deionized water that was further washed against Chelex 100 (Bio-Rad) and Amberlite MB-3 mixed bed monovalent resin. All buffers were corrected to pH 7.5 after divalent cation addition. Lipid vesicle preparations
The grade 1 phospholipid sources were from natural sources; phosphatidylserine (PS) from bovine brain and
Int. J. Biol. Macromol., 1990, Vol. 12, August
251
Structural changes in protein kinase C: D. S. Lester and V. Brumfeld phosphatidylcholine (PC) from egg (lecithin). 1,2Dioleoyl-glycerol (18:1, DAG) was of synthetic origin. The phospholipids were used above their phase transition temperatures as measured by differential scanning calorimetry, thus, they were in the liquid-gel state/2. Lipid mixtures were all prepared according to weight ratios. PS and PC were assumed to have an Mr of 800, while synthetic DAG had an Mr of 620.9 (Serdary Research Laboratories). Lipid mixtures were dried under a stream of nitrogen and further under a high vacuum for at least 4 h to reduce residual organic solvents. Small unilamellar vesicles (SUV) were prepared by suspending the dried mixtures in 20 mM Tris buffer, pH 7.5, 0.5 mM EGTA and exposed to ultrasonic irradiation in a bath sonicator for 30 min. SUV were used for all spectral measurements to minimize problems of light scattering by the relatively high lipid vesicle concentration.
Activity measurements Phosphorylating activity of the enzyme was measured using the exogenous substrate, histone I (Sigma histone III-S) as previously described la. The reaction mixture routinely contained 20mM Tris/HCl, pH 7.5, 0.5mM EGTA, 0.5 mM D T r , histone I (0.05 mg/ml), 20 #M [Va2p]ATP (1000 counts/min/pmol ATP), leupeptin (1 #M), 2.5 mu MgC12, plus or minus 100 #M 'free' Ca 2 +. Free Ca 2+ was defined according to calculations for Ca 2 +-EGTA buffers 14. The various lipid vesicle mixtures (50#M) as indicated in the figures and tables were prepared as previously described. Enzyme samples (40 ng) were added to the reaction mixture and incubated at 30°C for 4 min. The reaction was stopped by the addition of 1 ml 25 % trichloroacetic acid. The mixtures were kept at 4°C for 30 min and then filtered through MFS cellulose nitrate filters (0.45 #m) and washed with 4 x 2 ml of 5 % trichloroacetic acid. Scintillant (10 ml) was added to the dry filters and the radioactivity estimated on a Kontron fl-scintillation counter. Intrinsic fluorescence Changes in intrinsic tryptophan fluorescence were monitored by recording the fluorescence emission spectra between 300 and 400nm on a Perkin Elmer L5 Luminescence Spectrophotometer (spectral pathwidth of 5 nm for excitation and emission) at an excitation wavelength of 280 nm. Changes in the environment of tryptophans were monitored as changes in the peak emission wavelength and the relative fluorescence, i.e. quenching of or increased emission fluorescence1s. These changes were considered to be a result of the binding of the activators 9. A protein concentration of 5 #g/ml was used. All measurements were done at 25 + I°C. The effect of divalent cations upon intrinsic fluorescence were measured by addition of Ca 2 + (200 #M free) or Mg 2+ (100/~M free) to the protein in 20 mM Tris buffer (pH 7.5) containing divalent cation chelators (0.5 mM EGTA and 0.5 mM EDTA). These samples were briefly stirred and then left for 30 min before recording of the fluorescence spectra. This equilibration period had no significant effect on the structure of the native enzyme (data not shown). For lipid effects, the initial protein spectrum was measured in the presence of 1 mM free Mg 2+ and 100 mM KCI in the same buffer used for the divalent cation effects. Lipid vesicle mixtures (compositions and concentrations as indicated in the figure and table legends) were added
252
Int. J. Biol. Macromol., 1990, Vol. 12, August
and the spectra recorded. Free C a 2 +, at concentrations indicated in Figure 2c, was mixed with the PKC-lipid vesicle mixture and changes in intrinsic fluorescence monitored. The lipid vesicles and Ca 2 + were allowed to interact with the protein before recording the fluorescence spectrum. The peak fluorescence was between 338340 nm for all measurements. Subsequently, for Figure 2b and 2c the emission was monitored at 338nm. For comparative analyses of lipid-dependent fluorescence changes and activity measurements a fixed lipid vesicle concentration of 6.25#M was used, i.e. a protein:lipid ratio of 1:100. The pH of the samples did not significantly change (___0.1 pH unit) upon addition of the divalent cations.
Circular dichroism Circular dichroism (c.d.) spectra were measured on a Jasco 500A. Spectropolarimeter. Protein concentrations were 50-100 #g/ml (,-, 0.6-1.2 #M). The samples showed approximate u.v. absorption values of 0.3-0.6 at 200 nm measured on a Hewlett Packard Diode Array Spectrophotometer at room temperature (25_ 1°C) in a 1 - m m pathlength cuvette. This absorption value enabled an optimal c.d. signal to be measured between 200250 nm. All the c.d. measurements were done in a 1-mm pathlength cylindrical cuvette. C.d. spectra presented are an average of 8 scans (20nm/min) between 200 and 250 nm. Digitally recorded curves were fed through the Jasco J-DPY data processor for signal averaging and base-line subtraction. The estimated experimental error was less than _+600 deg.cm2/dM between different protein preparations, however, the relative response to the addition of the ligand was equivalent. No significant protein concentration changes due to ion or lipid addition were recorded, thus, correction for dilution was not necessary. The addition of divalent cations and lipids was as described in the previous section on intrinsic fluorescence. Only one lipid vesicle concentration (50 #M) was used so as to limit problems of light scattering Titrations were not performed due to the extensive time required for recording each spectrum. One concentration of Ca 2+ (100#M free) was used for addition for the Mg 2 +-protein-lipid complex to demonstrate the Ca 2 + specific effect. Molar elipticity [0], was calculated from the observed elipticity, 0, according to the following formula: OM E°] = lOlc where M is the molecular weight (80kD), 1 is the pathlength (1 mm), and c is the concentration (mg/ml). Estimations of secondary structure were not made due to lipid effects (flattening) on the spectra. Protein concentrations were measured using the Coomassie dye protein binding procedure of Bradford 16. Results
The potential role of established activators in changes in PKC conformation was monitored by measuring changes in spectral properties of the native enzyme using c.d. and intrinsic tryptophan fluorescence. We have previously shown that half of the PKC tryptophan residues (4) are in hydrophobic regions of the protein that upon association with ligands which optimally activate the
Structural changes in protein kinase C: D. S. Lester and V. Brumfeld deg. cm2/dM. There was considerable reduction of ~helicity as seen by the decrease at wavelengths of 208 and 220 nm. These changes were irreversible for both divalent cations. Subsequent addition of Ca 2÷ upon Mg 2÷ induced changes had no further effects on either the c.d. or fluorescence spectra (data not shown). Thus, K P C undergoes strong conformational changes upon association of Ca 2 ÷ and Mg 2 +. No Ca 2 +-specific change was detectable using these techniques. In a previous report from our laboratory, it was shown that PKC associates and, subsequently, penetrates PS:PC (1:1) lipid vesicles in the absence of divalent cations and D A G 9. In this report, the lipid specific effects on enzyme structure were examined. For analyses of lipid effects on spectral properties, the initial spectra was measured for PKC suspended in buffer containing 1 mM Mg 2 + as this is close to the cytosolic concentration of this divalent cation 17. Typical fluorescence changes in the tryptophan spectrum for native PKC in buffer (containing Mg 2 +) upon addition of lipid vesicles are shown in Figure 2a. When 6.25 #M SUV (PS/PC/DAG; 5/4/1, w/w) were added to the Mg 2 +" PKC complex at a protein:lipid ratio of 1:100 (mohmol), the intrinsic fluorescence was quenched (32 %) with no significant shift in the peak emission wavelength. Ca 2÷ (100#M free) quenched an additional amount (15%) in spite of the excess Mg 2 + in the buffer. This Ca 2 ÷ -induced quenching was Ca 2 +-specific and did not occur in the presence of the divalent cations, Mg 2 +, Zn 2 + or Mn 2 + (data not shown). Additional EGTA (1 mM) partially reversed the Ca 2 +induced quenching (approximately 50%). The peak emission wavelength (338+ 1 nm) did not significantly change for any of the additions. Thus, the Ca 2 ÷ specific change in structure occurred after lipid association. The quenching of intrinsic PKC tryptophan fluorescence was measured for a number of different phospholipid vesicle compositions (Figure 2b) at a fixed protein concentration (5 #g/mol, 0.0625 gM) and varying concentrations of lipid (0-6.25/tM) up to a maximum protein:lipid ratio of 1:100. The control or maximal fluorescence was obtained for PKC in the presence of Mg 2+ (1 mM). It is assumed that quenching of the intrinsic PKC fluorescence is an indication of ligand association. Phospholipases A 2, which has sequence homology with the prposed region of phospholipid
enzyme quenches the intrinsic tryptophan fluorescence up to 50% 9. The remaining tryptophans are in hydrophilic regions that are not affected by ligand binding. Initially, the effect of divalent cations on soluble enzyme was analysed. A typical fluorescence spectrum is presented in Figure la. Addition of either 100/UM Mg 2+ or 200#M Ca 2÷ resulted in maximal tryptophan quenching, 6 or 5 % respectively, at the peak emission wavelength of 338 rim. No shift in peak emission was observed. The decrease in fluorescence could not be reversed by the addition of divalent cation chelators, such as EGTA and EDTA, during the period of analysis (0.51 h). Monovalent cations (0.5 u NaCI) had no effect on the intrinsic spectra fluorescence of PKC in the presence or absence of divalent cations (data not shown). Analyses of the c.d. spectra of soluble PKC in the presence of these divalent cations demonstrated that either Ca 2 ÷ or Mg 2 4, at similar concentrations to those used for tryptophan fluorescence analyses, induced a significant rearrangement in the native conformation (Figure lb). Mg 2÷ reduced band intensity from - 17 000 to - 7000 deg" cm2/dM with a shift in peak wavelength from 218 nm to 222 nm. Ca 2÷ had a similar effect with a peak wavelength of 219 nm and a band intensity of - 8 5 0 0 I
I
I
t
a
I
I
b
_~ O
"<',x::~
/ :×
E u
D
n,
5 rY
-
f
I
I
300
340
380
j
-2~"
210
230
250
X(nm) Figure 1 Ca 2 + and Mg 2 ÷-induced changes in intrinsic spectral properties of protein kinase C. (a) Intrinsic tryptophan fluorescence quenching of PKC (10mg/ml) by Ca t+ (..... , 200ttM) or Mg 2+ ( - - - , 100/~M). (b) Changes in far u.v.-c.d. spectra of PKC (100mg/ml) by Ca 2+ ( , 2 0 0 / ~ M ) or Mg 2+ ( - - - , 100/tM). The Ca 2+ effect was gradual over a period of 20-30min while the Mg 2÷ effect was immediate. These analyses were repeated using three different preparations of enzyme AF
/~
, PKC÷ Mg
i
0.6
Sl
I
l
,
i
,
I
|
I
C
0.4b h
0.2 300
340 380 Emission (Anm)
0
~
20
¢
~
40
60
~
80
/Lipid/protein]
100
0
100
I 200
J 300
o
I 400
Free [Cc~l (JAM)
Figure 2 Ca 2 +-specific quenching of intrinsic fluorescence protein kinase C-lipid vesicle complexes. (A) Fluorescence spectra of soluble PKC and the active Ca 2+-protein-lipid complex. The initial spectra was protein (5/tg/ml) in the presence of I mM Mg2+ (PKC + Mg 2+). Lipid vesiclescomposed of PS/PC/DAG, 4:5:1 (DAG) were added and allowed to equilibrate. Ca 2 ÷ (100 #M) was added and the quenching recorded. This could be partially reversed by 0.5raM EGTA (broken line). (B) Quenching of intrinsic PKC fluorescence by various lipid mixtures. All measurements were at excitation and emission wavelengths of 280 nm and 340 nm, respectively. F is the fluorescence of PKC + Mg t+. AF is the change in fluorescence. The lipid mixtures are: O - - © , P C ; / k - - A , PS/PC (1:1); l-l--I--], PS; I I - - m , PS/DAG (9:1); A - - A , PS/PC/DAG (5:4:1). The protein concentration was 0.036/~M. Stock lipid solutions of 10/~M were prepared and diluted as desired. (C) Ca t +-specific quenching of PKC intrinsic fluorescence. The Mgt +-PKClipid complexes from B were used and increasing final concentrations of Ca t+ added. The standard error for fluorescence measurement was +2.0%. The change in pH upon Ca t+ addition was _0.1 pH units
Int. J. Biol. Macromol., 1990, Vol. 12, August
253
Structural changes in protein kinase C: D. S. Lester and V. Brumfeld Table 1 Comparative efficacy of lipid in supporting PKC activity, fluorescence quenching and changes in protein ellipticity Lipid
Activity a (counts/min)
PC PS PS/PC (1:1) PC/DAG (0.9:0.01) PC/TPA (0.99:0.01) PS/DAG (0.9:0.1) PsfrPA (0.99:0.01) PS/PC/DAG (0.5:0.4: 0.1) PS/PC/TPA (0.5:0.49:0.01) ° b c d
AF b
A022o c
(~)
(deg' cm2/dM)
-- C a 2 +
+ Ca 2 +
- Ca
+ Ca 2 +
- Ca
+ Ca z +
244___20 222___36 189 ___28 202 _ 24 238 + 17 6831 ___83 6473 + 53 1872 + 32 1243+118
334 + 43 8321 __+212 1042___87 387 _ 64 282 + 26 8625_ 127 8482__ 101 8983 ___63 9340+36
7 25 18 10 9 33 35 38 30
20 43 23 22 24 45 42 52 43
nd ~ 700 - 1600 nd nd 700 nd - 2700 -2750
nd 1350 - 250 nd nd 1350 nd - 850 -1860
P K C (40 ng) activity was m e a s u r e d as described in E x p e r i m e n t a l Intrinsic fluorescence of protein (5 # g / m l ) in buffer c o n t a i n i n g 2 mM MgCI2 was defined as 100% or A F = 0 . S t a n d a r d error was ___2.0% The 0220 value of P K C in 1 mM MgCI2 ( - 8600 deg. cm2/dM) was defined as 0 n d = not d e t e r m i n e d
binding xs, also undergoes intrinsic tryptophan quenching upon lipid association 19. Saturation or maximal quenching of PKC for each vesicle preparation was achieved at approximately 50-60 mol lipids/mol protein. Phosphatidylcholine (PC) was the least effective quencher of soluble PKC (10~o at 100 mol lipid/mol protein). Addition of the PKC activator, D A G (10 ~o w/w of total lipid), to the lipid compositions caused significant additional quenching of the lipid-protein complex for both PS (20-30~) and PS/PC (16-32~) vesicles indicating that D A G induced a different PKC conformation. Addition of Ca 2 + induced a further change in structure, irrespective of the composition of the lipid support, with a maximal effect (6--15 ~ quenching) at free Ca 2+ concentrations ranging between 100-200 #M. This Ca2+-specific effect resulted in quenching of the Mg 2+ ' P K C complex of 4 0 - 5 0 ~ upon addition of the lipid mixtures, PS, PS/DAG, and PS/PC/DAG, indicating a similar topography of the tryptophan residues for these PKC-lipid-Ca 2+ complexes. The relationship between the changes in intrinsic fluorescence and activity was analysed (Table 1). A constant protein :lipid ratio of 1:100 (mol:mol) was used for analyses of tryptophan fluorescence quenching and changes in optical elipticity. For activity measurements, a much greater lipid concentration was employed due to the substrate, histone III-S, requirement for lipid in order to support activity 2°. The enzyme is generally considered maximally activated at optimal concentrations of Ca 2 4, PS, and D A G or TPA. However, high concentrations of Ca 2÷, such as those used in this analysis (100#M), together with PS are capable of maximally activating PKC in the absence of D A G or TPA 2. Purified enzyme was partially activated (70~o) by incubation with PS vesicles containing the lipid activators, TPA or DAG, in the absence of Ca 2 ÷ (0.5 mM EGTA), but in the presence of Mg 2÷ (2.5mM). Ca2÷-stimulated activity was equivalent for PS vesicles in the presence or absence of D A G or TPA which was also reflected in the similar fluorescence quenching (42-45~o) of the soluble P K C ' M g 2÷ complex for these mixtures. In contrast, using PS/PC vesicles, Ca 2 ÷ -stimulated activity was only observed for PS/PC vesicles containing either D A G or TPA. Maximal quenching of the control, P K C ' M g 2÷ (43 and 52~), was similar only for the active enzyme (PS/PC/DAG and PS/PC/TPA). This suggests a similar
254
Int. J. Biol. Macromol., 1990, Vol. 12, August
configuration of tryptophan residues in the active enzyme state. To further enforce this notion, the non-active enzyme-lipid complexes (PC, plus or minus activator, or PS/PC) had similar maximal tryptophan quenching of soluble PKC in the presence of lipid and Ca 2 + (20-30~o). It should be noted that addition of Ca 2 ÷ to PC vesicles in the presence or absence of PKC activators quenched intrinsic PKC fluorescence to a similar value as observed for the inactive lipid mixture of PS/PC, but had no effect on activity (Table 1). Interestingly, significant quenching of intrinsic PKC fluorescence in the absence of Ca 2 + was observed for PS/PC vesicles containing the PKC activators, D A G and TPA, of 38 and 30 ~ , respectively. However, as the fluorescence quenching was less than 42 ~o, the complexes were not active, further supporting the notion of a critical topography of PKC residues. Far ultraviolet (200-250nm) c.d. spectra of polypeptides and proteins change when the molecule inserts into a lipid matrix 21. However, the data for protein secondary structure evaluation have been derived from c.d. spectra of globular proteins 22'23. Thus, the modifications in the PKC c.d. spectra upon addition of lipid vesicles are due to structural changes in the protein, but quantitative assumptions of these modifications cannot be made. The vesicle concentration (50/~M) was the same for all lipid mixtures used. In relation to the soluble PKC ( - 8 6 0 0 deg" cm2/dM) the lipid-associated enzyme significantly increased its molar ellipticity at 208 and 220 nm indicating that PKC had more at-helix in the presence of PS/PC vesicles (compare Figure 3A solid line to the 3B broken line). The secondary structure as indicated by the band intensity and shape was similar for the PKC associated with PS/PC vesicles, plus or minus lipid activators (Figures 3B, C and D). In contrast, PS (Figure 3A) and PS/DAG (Table 1) rendered the enzyme to a similar extent slightly less helical or rigid as indicated by the decrease in negative ellipticity at 0220 (Table 1). Irrespective of the lipid vesicle composition, addition of Ca 2÷ induced the PKC-lipid complexes to undergo a further structural change (Figure 3A, B, C and D, Table 1). For all vesicle mixtures, plus or minus lipid activators, Ca 2 + resulted in less ~-helix in the enzyme (Figure 3C, Table 1). This Ca 2 ÷ -induced change in band intensity was significantly greater for interactions with P S / P C / D A G than PS/PC/TPA ( - 1850 compared to 950 deg- cm2/dM, respectively) (Table 1). Ca 2 ÷ had the smallest effect on the
Structural changes in protein kinase C: D. S. Lester and V. Brumfeld I
I
I
I
I
I
I
I
I
I
A
I
B
_
~..--
+Ca2+.,.. ~:.....=,,~"'~"~"
~'XN.
P S/PC
oJ
E u
i
I
I
I
I
I
I
D
C
v
'©
I
t
(D
0
Q~ -1
+ c0 2+ .\'"'~..'.--'__,,.-7.; ..... " \ -.,..~ + C 0 2 + + ATP ~ o PS/PC/DAG
-
i
210
I
i
1
~x+Ca
"<~.....~
__
.i~'~"
PS/PC/TPA 1
220 230 240 250
I
210
I
I
I
I
220 230 240 250
X (nrn) Figure 3 Changes in the far u.v.-c.d, spectra of protein kinase C upon association with ligand activators. PKC (60/xg/ml) suspended in buffer containing 1 mM Mg2÷ was considered the initial or control spectra (see A, PKC). The lipid addition is indicated by the broken line. Ca 2÷ [100/xM] (dot-dashed line) was added to the Mg 2+-PKC-lipid complex. ATP (dotted line) was added to only the one lipid mixture (see C). Four lipid vesicle mixtures: phosphatidylserine, PS (A); phosphatidylserine/phosphatidylcholine, PS/PC, 1:1 ( B ) ; phosphatidylserine/ phosphatidylcholine/diacylglycerol, PS/PC/DAG, 1:0; 0.8:0.2 (C); and phosphatidylserine/ phosphatidylcholine/phorbol myristate acetate, PS/PC/TPA, 1.0:0.98:0.02 (D)
secondary structure of the PKC-PS complexes plus or minus DAG (Figure 3A, Table 1), however, this relative change ( - 6 5 0 deg'cm2/dM) was observed in three different protein preparations. Upon addition of ATP (50 #~l) to the activated PKC complex (PS/PC/DAG and Ca 2 ÷) there was a small decrease in eliptical intensity of - 6 5 0 deg.cm2/dM suggesting a further interaction (Figure 3C). From Table 1 and Figure 3 it can be seen that it is difficult to correlate the activity state of the enzyme to the secondary structure as indicated by the c.d. spectra and the 0220. This suggests that the enzyme may be active in a number of different conformations. Discussion
Extrinsic fluorescence probes have been used to determine interactions of PKC with lipid vesicles 24-26. These analyses depend on an energy transfer reaction from the protein such as tryptophan to some fluorescent residue. Thus, in order to register an interaction, the position of the label (lipid headgroup 2.26 or hydrocarbon chain 9) must be in the vicinity of the energy emitter, which experiences some change in its environment (hydrophobic or hydrophilic27). In this report, we monitored changes in the intrinsic spectral properties of PKC in order to determine how its activators affect its conformation. PKC has eight tryptophan residues of which at least four are in the proposed regulatory domain 7. We have previously demonstrated that half of the tryptophans are in hydrophobic regions and are quenched upon interaction with lipid and divalent cation activators 9. Association of divalent cations with soluble PKC induced a small
decrease in the intrinsic tryptophan fluorescence. In contrast, there was a strong divalent cation-induced change in c.d. spectra represented by a loss of ~t-helicity. Even though Ca 2÷ or Mg 2÷ induced this structural change, the concentration of divalent cation required was within the physiological range of cytosolic unbound Mg 2÷ (Ref. !7). We assume that Mg 2÷ binds to the cytosolic enzyme reorganizing its structure as a 'priming step' for further interactions with functional ligands. Considering that the cytosolic Mg 2÷ concentration is above that required for the observed in vitro conformational change, we consequently assume that enzyme will be in this 'primed' state in the cell. A similar process has been proposed with the multidomain Ca 2 +/ calmodulin-dependent protein kinase 2s. Thus, this divalent cation effect was not Ca 2 ÷-specific which argues against the proposal that Ca 2 ÷ binds to cytosolic PKC inducing a conformational change resulting in translocation from the cytosol to the membrane 6. Consequently, we analysed lipid vesicle interactions with the enzyme in this Mg 2÷-induced conformational state. PKC tryptophan residues were quenched upon interaction with liposomes. Maximal quenching was at a lipid:protein ratio of 50450:1 (mol:mol). Considering the lipid surface is a bilayer, this would suggest that PKC associates with or covers an area of 30 lipid molecules on the membrane surface. The magnitude of quenching was dependent on the lipid composition, i.e. whether the lipid was charged or uncharged. Consequently, the acidic phospholipid, PS, which is the most potent lipid support for phosphorylating activity 1,2, was the strongest quencher. Only when the enzyme was in contact with lipid was it capable of undergoing a Ca2÷-specific
Int. J. Biol. Macromol., 1990, Vol. 12, August
255
Structural chan#es in protein kinase C: D. S. L e s t e r and V. Brumfeld
conformational change. As the Ca2+-specific effect on protein conformation occurred with all liposome systems analysed, this suggests that the determining factor for activity is the lipid composition. The role of activators in PS mixtures is generally considered to lower the affinity for Ca 2+ (Refs 1, 2). As can be seen in Table 1, the tryptophan quenching in the absence of Ca 2 ÷ is greater for PS vesicles containing activators than pure PS. However, as a relatively high Ca 2÷ concentration (100#M) was used for activity analyses in this study, the stimulation by lipid activators in pure PS vesicles was not significant. Partial irreversibility of Ca 2÷ binding to PKC-lipid complexes has been reported using the extrinsic fluorescent probe, dansylphosphatidylethanolamine 24'26. We also observed that the C a 2 +-specific effect on intrinsic P K C fluorescence was partially reversible. The c.d. measurements provided a different overview as they measured gross changes in P K C secondary structure. However, certain features were similar for changes in both c.d. and intrinsic tryptophan fluorescence spectra. Changes in secondary structure were also dependent on the specific lipid compositions. For example, PS/PC vesicles containing D A G or T P A had a stronger c.d. band intensity than PS/PC vesicles alone. Also, the Ca 2÷ effect on PKC-lipid activator complexes was more significant for PS/PC mixed liposomes. These Ca 2 ÷-indiced changes in the protein may either be a direct rearrangement of the membranepenetrating P K C secondary structure or may be due to some influence on the protein-lipid interaction resulting in changes in secondary structure. This Ca 2 ÷ effect was not significantly reversible in c.d. measurements (data not shown), which may have been due to insufficient time for the reversal to be observed or the irreversibility of subsequent changes upon Ca 2 + binding. The activity measurements were made using the exogenous substrate, histone Ill-S, which in order to undergo phosphorylation requires all of the P K C activating ligands. In contrast, the substrate, protamine, is phosphorylated by P K C in the absence of these ligands 2° and their presence has no significant effect on protamine phosphorylation 9. The fluorescence changes as observed by tryptophan quenching correlated to histone III-S phosphorylating activity. The intrinsic fluorescence of active PKC-lipid complex was 50-60 ~o of the native soluble enzyme. This indicated that the topography of the tryptophan residues was important in the active conformation. The tryptophan residues affected by the ligand association are in the regulatory domain of P K C 9. Thus, it may be concluded that portions of the regulatory domain must be in a specific conformation to obtain an active enzyme state. In contrast, the changes in c.d. spectra were difficult to relate to the activity as each active protein-lipid complex had a different secondary structure. The c.d. spectra differed according to the phospholipid and lipid activator composition. Thus, these differences in structure may have some significance in the in vivo cellular system where alternative PKC-substrate reactions may occur. The differences in c.d.-determined conformation of the various active forms (Figure 3 A , C and D) cannot be attributed to specific regions in the protein. Therefore, it appears from the c.d. analyses that there are regions of the protein secondary structure undergoing change which remain to be clarified and identified. For example, from
256
Int. J. Biol. Macromol., 1990, Vol. 12, August
our studies, it is proposed that the enzyme rearranges upon lipid association to form a Ca 2 ÷-specific site. Such a suggestion requires a different approach than searching for consensus regions. However, this provides a possible explanation for the lack of any identifiable Ca 2 +-binding site consensus region (e.g. 'E-F hand') in the c D N A deduced sequence of the P K C isozymes 7's. The direct spectral measurements of changes in protein structure of P K C would suggest that the proposed pathway of P K C activation based on biochemical analyses requires re-examining. F r o m our biophysical analyses, we suggest that the enzyme associates with lipid before it is capable of specifically binding Ca 2 + to obtain its final (active) configuration. These structure-function analyses provide an important alternative outlook on the changes in secondary structure that P K C undergoes On the pathway to activation.
Acknowledgements This work was supported in part by a grant from the Mazer Center for Structural Biology (D.S.L.)and a US Naval Research Aid (V.B.). We would like to express our gratitude to Martin Welch for his critical comments.
References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
Bell, R. M. Cell 1986, 45, 631 Nishizuka,Y. Science 1986, 233, 305 Ashendel,C. L., Staller, J. M. and Boutwell, R. K. Cancer Res. 1973, 43, 4327 Sando,J. J. and Young, M. C. Proc. Natl. Acad. Sci. USA 1983, 80, 2642 Kondo, H., Kinoshita, J., Matsuba, T. and Sunamoto, J. FEBS Lett. 1986, 201, 94 Ashendel,C. L. Biochim. Biophys. Acta 1985, 822, 219 Parker, P. J., Coussens, L., Totty, N., Rhee, L., Young, S., Chen, E., Stabel, S., Waterfield, M. D. and Ullrich, A. Science 1986, 233, 853 Ono, Y., Fujii, T., Igarashi, K., Kuno, T., Tanaka, C., Kikkawa, U. and Nishizuka, Y. Proc. Natl. Acad. Sci. USA 1989, 86, 4868 Brumfeld,V. and Lester,D. S. Arch. Biochim. Biophys. 1990,277, 318 Nishizuka,Y. Nature 1989, 334, 661 Huang,K.-P., Chan, K.-F. J., Singh, T. J., Nakabayashi, H. and Huang, F. L. J. Biol. Chem. 1986, 261, 12134 Bach,D. Personal communication Lester,D. S. J. Neurochem. 1989, 52, 1950 Pershadsingh, H. A. and McDonald, J. M. J. Biol. Chem. 1980, 255, 4087 Freifelder, D. 'Physical Biochemistry. Applications to Biochemistryand Molecular Biology'.W. H. Freeman and Co., San Francisco, 1976, pp. 415-421 Bradford,M. M. Anal. Biochem. 1976, 72, 248 Arslan,P., DiVirgilio,F., Beltrane, M., Tsien, R. Y. and Pozzan, T. J. Biol. Chem. 1985, 260, 2719 Maraganore,J. M. TIBS 1987, 12, 176 Yang,C. C. and Chang, L. S. Int. J. Biol. Macromol. 1989, 11, 13 Bazzi,M. D. and Nelsestuen, G. L. Biochemistry 1988, 26, 1974 Wallace,B. A. and Teeters, C. L. Biochemistry 1987, 26, 65 Chang,C. T., Wu, C.-S. C. and Young, J. T. Anal. Biochem. 1978, 91, 13 Hennessey,J. P. and Johnson, W. C. Jr Biochemistry 1981, 20, 1085 Bazzi,M. D. and Nelsestuen, G. L. Biochemistry 1987, 26, 115 Rodriguez-Paris,J. M., Shoji, M., Yeola, S., Liotta, D., Volger, W. R. and Kuo, J. F. Biochem. Biophys. Res. Comm. 1989, 159, 495 Epand,R. M., Stafford,A. R., Bottega, R. and Ball, E. H. Biosci. Reports 1989, 9, 315 Calhoun,D. B., Vanderkooi, J. M., G. R. and Englander, S. W. Proteins 1986, 1, 109 King,M. M. J. Biol. Chem. 1988, 263, 4754
(Received 5 September 1989; revised 21 January 1990) The changes in intrinsic spectral properties of protein kinase C were monitored upon association with its divalent cation and lipid activators in a model membrane system. The enzyme demonstrated changes in both its intrinsic fluorescence and far ultraviolet circular dichroism spectra upon association with lipid vesicles in the absence of calcium. The acidic phospholipid, phasphatidylserine, sionificantly quenched the intrinsic tryptophan fluorescence and was also the most potent lipid support for the phosphorylating activity of the enzyme. The enzyme was fully activated by a number of Ca2+-lipid combinations which correlated with maximal fluorescence quenching (40-50 %) of available tryptophan residues in hydrophobic domains. The circular dichroism structure of the associated active-protein Ca 2 +-lipid complexes suooested different active enzyme secondary structures. However, the Ca2+-dependent changes in fluorescence and circular dichroism spectra were observed only after the enzyme associated with the lipid vesicles. These data suggest that protein kinase C has the properties of a complex multidomain protein and provides an additional perspective into the mechanism of protein kinase C activation. Keywords: Protein kinaseC; lipids; diacylglycerol;phorbol ester; conformation;tryptophan;circular dichroism
Introduction The Ca 2 +/phospholipid-dependent protein kinase, protein kinase C (PKC), is considered to play a significant role in numerous cellular processes including transmembrane signal transduction and tumour promotion 1'2. PKC is activated in vitro by Ca 2÷, phosphatidylserine (PS), and lipid activators such as diacylglycerol (DAG) or the exogenous class of tumour promoting agents (TPA) known as the phorbol esters t. In addition, Mg 2+ is required for optimal tritiated phorbol ester binding a'4, while Mg2+'ATP is the preferred substrate for the phosphotransferase activity2,s. Ca z+ ions are generally considered to induce the translocation of the enzyme from the cytosol to the plasma membrane ~,2,6, possibly by binding to the soluble enzyme and eliciting a conformational change promoting translocation6. However, examination of the primary structure deduced from cDNA sequence does not reveal any of the motifs typifying a Ca 2+ binding site 7'8. Additionally, PKC has been shown to associate and penetrate lipid bilayers in the absence of divalent cations 9. The enzyme is a single polypeptide, however, it has been subdivided into two functional domain structures. The N-terminal region is considered to be the regulatory domain as it has specific binding regions for Ca 2 +, PS and the lipid activator, while the C-terminal region is the catalytic domain containing the Mg 2+ "ATP and the substrate binding site 7'a'1 o. Different regions of each domain have been proposed to be involved in binding of the various ligands to PKC ~°. We have recently shown that half of the PKC tryptophan residues * To whomcorrespondenceshouldbe addressedat: NIH, Sectionon Neural Systems,Laboratoryof Molecularand CellularNeurobiology, Park 5 Bldg,Room 435, Bethesda,MD 20892, USA. 0141-8130/90/040251-06 © 1990 Butterworth-HeinemannLimited
are in regions of the protein that undergo conformational change(s) upon addition of the activating lipids 9. In this report, we have examined the contribution of the various activating cofactors by monitoring changes in the intrinsic spectroscopic properties of the protein. The data obtained suggest that PKC has a number of conformational states and the process of PKC activation may occur in a different sequence than that which has been proposed according to the models based on biochemical analyses.
Experimental Cytosolic PKC was purified from male Wistar rat brains (6 weeks) according to the procedure of Huang et a l ) I. Generally, 50 such brains yielded an average of 1.0 mg of pure enzyme (single band on a Coomassie and silver stained an SDS-polyacrylamide gel) containing all three isozymes (Types I, II and III). The specific activity of the enzyme preparation was approximately 1000 units/mg proteins. Lipids were purchased from Lipid Products Inc. or Serdary Research Laboratories. Phorbol myristate acetate (TPA), histone III-S and all other chemicals were purchased from Sigma Chemical Co. [7-32p]ATP (3500 Ci/mmol) was obtained from New England Nuclear. All buffers were prepared from double distilled, deionized water that was further washed against Chelex 100 (Bio-Rad) and Amberlite MB-3 mixed bed monovalent resin. All buffers were corrected to pH 7.5 after divalent cation addition. Lipid vesicle preparations
The grade 1 phospholipid sources were from natural sources; phosphatidylserine (PS) from bovine brain and
Int. J. Biol. Macromol., 1990, Vol. 12, August
251
Structural changes in protein kinase C: D. S. Lester and V. Brumfeld phosphatidylcholine (PC) from egg (lecithin). 1,2Dioleoyl-glycerol (18:1, DAG) was of synthetic origin. The phospholipids were used above their phase transition temperatures as measured by differential scanning calorimetry, thus, they were in the liquid-gel state/2. Lipid mixtures were all prepared according to weight ratios. PS and PC were assumed to have an Mr of 800, while synthetic DAG had an Mr of 620.9 (Serdary Research Laboratories). Lipid mixtures were dried under a stream of nitrogen and further under a high vacuum for at least 4 h to reduce residual organic solvents. Small unilamellar vesicles (SUV) were prepared by suspending the dried mixtures in 20 mM Tris buffer, pH 7.5, 0.5 mM EGTA and exposed to ultrasonic irradiation in a bath sonicator for 30 min. SUV were used for all spectral measurements to minimize problems of light scattering by the relatively high lipid vesicle concentration.
Activity measurements Phosphorylating activity of the enzyme was measured using the exogenous substrate, histone I (Sigma histone III-S) as previously described la. The reaction mixture routinely contained 20mM Tris/HCl, pH 7.5, 0.5mM EGTA, 0.5 mM D T r , histone I (0.05 mg/ml), 20 #M [Va2p]ATP (1000 counts/min/pmol ATP), leupeptin (1 #M), 2.5 mu MgC12, plus or minus 100 #M 'free' Ca 2 +. Free Ca 2+ was defined according to calculations for Ca 2 +-EGTA buffers 14. The various lipid vesicle mixtures (50#M) as indicated in the figures and tables were prepared as previously described. Enzyme samples (40 ng) were added to the reaction mixture and incubated at 30°C for 4 min. The reaction was stopped by the addition of 1 ml 25 % trichloroacetic acid. The mixtures were kept at 4°C for 30 min and then filtered through MFS cellulose nitrate filters (0.45 #m) and washed with 4 x 2 ml of 5 % trichloroacetic acid. Scintillant (10 ml) was added to the dry filters and the radioactivity estimated on a Kontron fl-scintillation counter. Intrinsic fluorescence Changes in intrinsic tryptophan fluorescence were monitored by recording the fluorescence emission spectra between 300 and 400nm on a Perkin Elmer L5 Luminescence Spectrophotometer (spectral pathwidth of 5 nm for excitation and emission) at an excitation wavelength of 280 nm. Changes in the environment of tryptophans were monitored as changes in the peak emission wavelength and the relative fluorescence, i.e. quenching of or increased emission fluorescence1s. These changes were considered to be a result of the binding of the activators 9. A protein concentration of 5 #g/ml was used. All measurements were done at 25 + I°C. The effect of divalent cations upon intrinsic fluorescence were measured by addition of Ca 2 + (200 #M free) or Mg 2+ (100/~M free) to the protein in 20 mM Tris buffer (pH 7.5) containing divalent cation chelators (0.5 mM EGTA and 0.5 mM EDTA). These samples were briefly stirred and then left for 30 min before recording of the fluorescence spectra. This equilibration period had no significant effect on the structure of the native enzyme (data not shown). For lipid effects, the initial protein spectrum was measured in the presence of 1 mM free Mg 2+ and 100 mM KCI in the same buffer used for the divalent cation effects. Lipid vesicle mixtures (compositions and concentrations as indicated in the figure and table legends) were added
252
Int. J. Biol. Macromol., 1990, Vol. 12, August
and the spectra recorded. Free C a 2 +, at concentrations indicated in Figure 2c, was mixed with the PKC-lipid vesicle mixture and changes in intrinsic fluorescence monitored. The lipid vesicles and Ca 2 + were allowed to interact with the protein before recording the fluorescence spectrum. The peak fluorescence was between 338340 nm for all measurements. Subsequently, for Figure 2b and 2c the emission was monitored at 338nm. For comparative analyses of lipid-dependent fluorescence changes and activity measurements a fixed lipid vesicle concentration of 6.25#M was used, i.e. a protein:lipid ratio of 1:100. The pH of the samples did not significantly change (___0.1 pH unit) upon addition of the divalent cations.
Circular dichroism Circular dichroism (c.d.) spectra were measured on a Jasco 500A. Spectropolarimeter. Protein concentrations were 50-100 #g/ml (,-, 0.6-1.2 #M). The samples showed approximate u.v. absorption values of 0.3-0.6 at 200 nm measured on a Hewlett Packard Diode Array Spectrophotometer at room temperature (25_ 1°C) in a 1 - m m pathlength cuvette. This absorption value enabled an optimal c.d. signal to be measured between 200250 nm. All the c.d. measurements were done in a 1-mm pathlength cylindrical cuvette. C.d. spectra presented are an average of 8 scans (20nm/min) between 200 and 250 nm. Digitally recorded curves were fed through the Jasco J-DPY data processor for signal averaging and base-line subtraction. The estimated experimental error was less than _+600 deg.cm2/dM between different protein preparations, however, the relative response to the addition of the ligand was equivalent. No significant protein concentration changes due to ion or lipid addition were recorded, thus, correction for dilution was not necessary. The addition of divalent cations and lipids was as described in the previous section on intrinsic fluorescence. Only one lipid vesicle concentration (50 #M) was used so as to limit problems of light scattering Titrations were not performed due to the extensive time required for recording each spectrum. One concentration of Ca 2+ (100#M free) was used for addition for the Mg 2 +-protein-lipid complex to demonstrate the Ca 2 + specific effect. Molar elipticity [0], was calculated from the observed elipticity, 0, according to the following formula: OM E°] = lOlc where M is the molecular weight (80kD), 1 is the pathlength (1 mm), and c is the concentration (mg/ml). Estimations of secondary structure were not made due to lipid effects (flattening) on the spectra. Protein concentrations were measured using the Coomassie dye protein binding procedure of Bradford 16. Results
The potential role of established activators in changes in PKC conformation was monitored by measuring changes in spectral properties of the native enzyme using c.d. and intrinsic tryptophan fluorescence. We have previously shown that half of the PKC tryptophan residues (4) are in hydrophobic regions of the protein that upon association with ligands which optimally activate the
Structural changes in protein kinase C: D. S. Lester and V. Brumfeld deg. cm2/dM. There was considerable reduction of ~helicity as seen by the decrease at wavelengths of 208 and 220 nm. These changes were irreversible for both divalent cations. Subsequent addition of Ca 2÷ upon Mg 2÷ induced changes had no further effects on either the c.d. or fluorescence spectra (data not shown). Thus, K P C undergoes strong conformational changes upon association of Ca 2 ÷ and Mg 2 +. No Ca 2 +-specific change was detectable using these techniques. In a previous report from our laboratory, it was shown that PKC associates and, subsequently, penetrates PS:PC (1:1) lipid vesicles in the absence of divalent cations and D A G 9. In this report, the lipid specific effects on enzyme structure were examined. For analyses of lipid effects on spectral properties, the initial spectra was measured for PKC suspended in buffer containing 1 mM Mg 2 + as this is close to the cytosolic concentration of this divalent cation 17. Typical fluorescence changes in the tryptophan spectrum for native PKC in buffer (containing Mg 2 +) upon addition of lipid vesicles are shown in Figure 2a. When 6.25 #M SUV (PS/PC/DAG; 5/4/1, w/w) were added to the Mg 2 +" PKC complex at a protein:lipid ratio of 1:100 (mohmol), the intrinsic fluorescence was quenched (32 %) with no significant shift in the peak emission wavelength. Ca 2÷ (100#M free) quenched an additional amount (15%) in spite of the excess Mg 2 + in the buffer. This Ca 2 ÷ -induced quenching was Ca 2 +-specific and did not occur in the presence of the divalent cations, Mg 2 +, Zn 2 + or Mn 2 + (data not shown). Additional EGTA (1 mM) partially reversed the Ca 2 +induced quenching (approximately 50%). The peak emission wavelength (338+ 1 nm) did not significantly change for any of the additions. Thus, the Ca 2 ÷ specific change in structure occurred after lipid association. The quenching of intrinsic PKC tryptophan fluorescence was measured for a number of different phospholipid vesicle compositions (Figure 2b) at a fixed protein concentration (5 #g/mol, 0.0625 gM) and varying concentrations of lipid (0-6.25/tM) up to a maximum protein:lipid ratio of 1:100. The control or maximal fluorescence was obtained for PKC in the presence of Mg 2+ (1 mM). It is assumed that quenching of the intrinsic PKC fluorescence is an indication of ligand association. Phospholipases A 2, which has sequence homology with the prposed region of phospholipid
enzyme quenches the intrinsic tryptophan fluorescence up to 50% 9. The remaining tryptophans are in hydrophilic regions that are not affected by ligand binding. Initially, the effect of divalent cations on soluble enzyme was analysed. A typical fluorescence spectrum is presented in Figure la. Addition of either 100/UM Mg 2+ or 200#M Ca 2÷ resulted in maximal tryptophan quenching, 6 or 5 % respectively, at the peak emission wavelength of 338 rim. No shift in peak emission was observed. The decrease in fluorescence could not be reversed by the addition of divalent cation chelators, such as EGTA and EDTA, during the period of analysis (0.51 h). Monovalent cations (0.5 u NaCI) had no effect on the intrinsic spectra fluorescence of PKC in the presence or absence of divalent cations (data not shown). Analyses of the c.d. spectra of soluble PKC in the presence of these divalent cations demonstrated that either Ca 2 ÷ or Mg 2 4, at similar concentrations to those used for tryptophan fluorescence analyses, induced a significant rearrangement in the native conformation (Figure lb). Mg 2÷ reduced band intensity from - 17 000 to - 7000 deg" cm2/dM with a shift in peak wavelength from 218 nm to 222 nm. Ca 2÷ had a similar effect with a peak wavelength of 219 nm and a band intensity of - 8 5 0 0 I
I
I
t
a
I
I
b
_~ O
"<',x::~
/ :×
E u
D
n,
5 rY
-
f
I
I
300
340
380
j
-2~"
210
230
250
X(nm) Figure 1 Ca 2 + and Mg 2 ÷-induced changes in intrinsic spectral properties of protein kinase C. (a) Intrinsic tryptophan fluorescence quenching of PKC (10mg/ml) by Ca t+ (..... , 200ttM) or Mg 2+ ( - - - , 100/~M). (b) Changes in far u.v.-c.d. spectra of PKC (100mg/ml) by Ca 2+ ( , 2 0 0 / ~ M ) or Mg 2+ ( - - - , 100/tM). The Ca 2+ effect was gradual over a period of 20-30min while the Mg 2÷ effect was immediate. These analyses were repeated using three different preparations of enzyme AF
/~
, PKC÷ Mg
i
0.6
Sl
I
l
,
i
,
I
|
I
C
0.4b h
0.2 300
340 380 Emission (Anm)
0
~
20
¢
~
40
60
~
80
/Lipid/protein]
100
0
100
I 200
J 300
o
I 400
Free [Cc~l (JAM)
Figure 2 Ca 2 +-specific quenching of intrinsic fluorescence protein kinase C-lipid vesicle complexes. (A) Fluorescence spectra of soluble PKC and the active Ca 2+-protein-lipid complex. The initial spectra was protein (5/tg/ml) in the presence of I mM Mg2+ (PKC + Mg 2+). Lipid vesiclescomposed of PS/PC/DAG, 4:5:1 (DAG) were added and allowed to equilibrate. Ca 2 ÷ (100 #M) was added and the quenching recorded. This could be partially reversed by 0.5raM EGTA (broken line). (B) Quenching of intrinsic PKC fluorescence by various lipid mixtures. All measurements were at excitation and emission wavelengths of 280 nm and 340 nm, respectively. F is the fluorescence of PKC + Mg t+. AF is the change in fluorescence. The lipid mixtures are: O - - © , P C ; / k - - A , PS/PC (1:1); l-l--I--], PS; I I - - m , PS/DAG (9:1); A - - A , PS/PC/DAG (5:4:1). The protein concentration was 0.036/~M. Stock lipid solutions of 10/~M were prepared and diluted as desired. (C) Ca t +-specific quenching of PKC intrinsic fluorescence. The Mgt +-PKClipid complexes from B were used and increasing final concentrations of Ca t+ added. The standard error for fluorescence measurement was +2.0%. The change in pH upon Ca t+ addition was _0.1 pH units
Int. J. Biol. Macromol., 1990, Vol. 12, August
253
Structural changes in protein kinase C: D. S. Lester and V. Brumfeld Table 1 Comparative efficacy of lipid in supporting PKC activity, fluorescence quenching and changes in protein ellipticity Lipid
Activity a (counts/min)
PC PS PS/PC (1:1) PC/DAG (0.9:0.01) PC/TPA (0.99:0.01) PS/DAG (0.9:0.1) PsfrPA (0.99:0.01) PS/PC/DAG (0.5:0.4: 0.1) PS/PC/TPA (0.5:0.49:0.01) ° b c d
AF b
A022o c
(~)
(deg' cm2/dM)
-- C a 2 +
+ Ca 2 +
- Ca
+ Ca 2 +
- Ca
+ Ca z +
244___20 222___36 189 ___28 202 _ 24 238 + 17 6831 ___83 6473 + 53 1872 + 32 1243+118
334 + 43 8321 __+212 1042___87 387 _ 64 282 + 26 8625_ 127 8482__ 101 8983 ___63 9340+36
7 25 18 10 9 33 35 38 30
20 43 23 22 24 45 42 52 43
nd ~ 700 - 1600 nd nd 700 nd - 2700 -2750
nd 1350 - 250 nd nd 1350 nd - 850 -1860
P K C (40 ng) activity was m e a s u r e d as described in E x p e r i m e n t a l Intrinsic fluorescence of protein (5 # g / m l ) in buffer c o n t a i n i n g 2 mM MgCI2 was defined as 100% or A F = 0 . S t a n d a r d error was ___2.0% The 0220 value of P K C in 1 mM MgCI2 ( - 8600 deg. cm2/dM) was defined as 0 n d = not d e t e r m i n e d
binding xs, also undergoes intrinsic tryptophan quenching upon lipid association 19. Saturation or maximal quenching of PKC for each vesicle preparation was achieved at approximately 50-60 mol lipids/mol protein. Phosphatidylcholine (PC) was the least effective quencher of soluble PKC (10~o at 100 mol lipid/mol protein). Addition of the PKC activator, D A G (10 ~o w/w of total lipid), to the lipid compositions caused significant additional quenching of the lipid-protein complex for both PS (20-30~) and PS/PC (16-32~) vesicles indicating that D A G induced a different PKC conformation. Addition of Ca 2 + induced a further change in structure, irrespective of the composition of the lipid support, with a maximal effect (6--15 ~ quenching) at free Ca 2+ concentrations ranging between 100-200 #M. This Ca2+-specific effect resulted in quenching of the Mg 2+ ' P K C complex of 4 0 - 5 0 ~ upon addition of the lipid mixtures, PS, PS/DAG, and PS/PC/DAG, indicating a similar topography of the tryptophan residues for these PKC-lipid-Ca 2+ complexes. The relationship between the changes in intrinsic fluorescence and activity was analysed (Table 1). A constant protein :lipid ratio of 1:100 (mol:mol) was used for analyses of tryptophan fluorescence quenching and changes in optical elipticity. For activity measurements, a much greater lipid concentration was employed due to the substrate, histone III-S, requirement for lipid in order to support activity 2°. The enzyme is generally considered maximally activated at optimal concentrations of Ca 2 4, PS, and D A G or TPA. However, high concentrations of Ca 2÷, such as those used in this analysis (100#M), together with PS are capable of maximally activating PKC in the absence of D A G or TPA 2. Purified enzyme was partially activated (70~o) by incubation with PS vesicles containing the lipid activators, TPA or DAG, in the absence of Ca 2 ÷ (0.5 mM EGTA), but in the presence of Mg 2÷ (2.5mM). Ca2÷-stimulated activity was equivalent for PS vesicles in the presence or absence of D A G or TPA which was also reflected in the similar fluorescence quenching (42-45~o) of the soluble P K C ' M g 2÷ complex for these mixtures. In contrast, using PS/PC vesicles, Ca 2 ÷ -stimulated activity was only observed for PS/PC vesicles containing either D A G or TPA. Maximal quenching of the control, P K C ' M g 2÷ (43 and 52~), was similar only for the active enzyme (PS/PC/DAG and PS/PC/TPA). This suggests a similar
254
Int. J. Biol. Macromol., 1990, Vol. 12, August
configuration of tryptophan residues in the active enzyme state. To further enforce this notion, the non-active enzyme-lipid complexes (PC, plus or minus activator, or PS/PC) had similar maximal tryptophan quenching of soluble PKC in the presence of lipid and Ca 2 + (20-30~o). It should be noted that addition of Ca 2 ÷ to PC vesicles in the presence or absence of PKC activators quenched intrinsic PKC fluorescence to a similar value as observed for the inactive lipid mixture of PS/PC, but had no effect on activity (Table 1). Interestingly, significant quenching of intrinsic PKC fluorescence in the absence of Ca 2 + was observed for PS/PC vesicles containing the PKC activators, D A G and TPA, of 38 and 30 ~ , respectively. However, as the fluorescence quenching was less than 42 ~o, the complexes were not active, further supporting the notion of a critical topography of PKC residues. Far ultraviolet (200-250nm) c.d. spectra of polypeptides and proteins change when the molecule inserts into a lipid matrix 21. However, the data for protein secondary structure evaluation have been derived from c.d. spectra of globular proteins 22'23. Thus, the modifications in the PKC c.d. spectra upon addition of lipid vesicles are due to structural changes in the protein, but quantitative assumptions of these modifications cannot be made. The vesicle concentration (50/~M) was the same for all lipid mixtures used. In relation to the soluble PKC ( - 8 6 0 0 deg" cm2/dM) the lipid-associated enzyme significantly increased its molar ellipticity at 208 and 220 nm indicating that PKC had more at-helix in the presence of PS/PC vesicles (compare Figure 3A solid line to the 3B broken line). The secondary structure as indicated by the band intensity and shape was similar for the PKC associated with PS/PC vesicles, plus or minus lipid activators (Figures 3B, C and D). In contrast, PS (Figure 3A) and PS/DAG (Table 1) rendered the enzyme to a similar extent slightly less helical or rigid as indicated by the decrease in negative ellipticity at 0220 (Table 1). Irrespective of the lipid vesicle composition, addition of Ca 2÷ induced the PKC-lipid complexes to undergo a further structural change (Figure 3A, B, C and D, Table 1). For all vesicle mixtures, plus or minus lipid activators, Ca 2 + resulted in less ~-helix in the enzyme (Figure 3C, Table 1). This Ca 2 ÷ -induced change in band intensity was significantly greater for interactions with P S / P C / D A G than PS/PC/TPA ( - 1850 compared to 950 deg- cm2/dM, respectively) (Table 1). Ca 2 ÷ had the smallest effect on the
Structural changes in protein kinase C: D. S. Lester and V. Brumfeld I
I
I
I
I
I
I
I
I
I
A
I
B
_
~..--
+Ca2+.,.. ~:.....=,,~"'~"~"
~'XN.
P S/PC
oJ
E u
i
I
I
I
I
I
I
D
C
v
'©
I
t
(D
0
Q~ -1
+ c0 2+ .\'"'~..'.--'__,,.-7.; ..... " \ -.,..~ + C 0 2 + + ATP ~ o PS/PC/DAG
-
i
210
I
i
1
~x+Ca
"<~.....~
__
.i~'~"
PS/PC/TPA 1
220 230 240 250
I
210
I
I
I
I
220 230 240 250
X (nrn) Figure 3 Changes in the far u.v.-c.d, spectra of protein kinase C upon association with ligand activators. PKC (60/xg/ml) suspended in buffer containing 1 mM Mg2÷ was considered the initial or control spectra (see A, PKC). The lipid addition is indicated by the broken line. Ca 2÷ [100/xM] (dot-dashed line) was added to the Mg 2+-PKC-lipid complex. ATP (dotted line) was added to only the one lipid mixture (see C). Four lipid vesicle mixtures: phosphatidylserine, PS (A); phosphatidylserine/phosphatidylcholine, PS/PC, 1:1 ( B ) ; phosphatidylserine/ phosphatidylcholine/diacylglycerol, PS/PC/DAG, 1:0; 0.8:0.2 (C); and phosphatidylserine/ phosphatidylcholine/phorbol myristate acetate, PS/PC/TPA, 1.0:0.98:0.02 (D)
secondary structure of the PKC-PS complexes plus or minus DAG (Figure 3A, Table 1), however, this relative change ( - 6 5 0 deg'cm2/dM) was observed in three different protein preparations. Upon addition of ATP (50 #~l) to the activated PKC complex (PS/PC/DAG and Ca 2 ÷) there was a small decrease in eliptical intensity of - 6 5 0 deg.cm2/dM suggesting a further interaction (Figure 3C). From Table 1 and Figure 3 it can be seen that it is difficult to correlate the activity state of the enzyme to the secondary structure as indicated by the c.d. spectra and the 0220. This suggests that the enzyme may be active in a number of different conformations. Discussion
Extrinsic fluorescence probes have been used to determine interactions of PKC with lipid vesicles 24-26. These analyses depend on an energy transfer reaction from the protein such as tryptophan to some fluorescent residue. Thus, in order to register an interaction, the position of the label (lipid headgroup 2.26 or hydrocarbon chain 9) must be in the vicinity of the energy emitter, which experiences some change in its environment (hydrophobic or hydrophilic27). In this report, we monitored changes in the intrinsic spectral properties of PKC in order to determine how its activators affect its conformation. PKC has eight tryptophan residues of which at least four are in the proposed regulatory domain 7. We have previously demonstrated that half of the tryptophans are in hydrophobic regions and are quenched upon interaction with lipid and divalent cation activators 9. Association of divalent cations with soluble PKC induced a small
decrease in the intrinsic tryptophan fluorescence. In contrast, there was a strong divalent cation-induced change in c.d. spectra represented by a loss of ~t-helicity. Even though Ca 2÷ or Mg 2÷ induced this structural change, the concentration of divalent cation required was within the physiological range of cytosolic unbound Mg 2÷ (Ref. !7). We assume that Mg 2÷ binds to the cytosolic enzyme reorganizing its structure as a 'priming step' for further interactions with functional ligands. Considering that the cytosolic Mg 2÷ concentration is above that required for the observed in vitro conformational change, we consequently assume that enzyme will be in this 'primed' state in the cell. A similar process has been proposed with the multidomain Ca 2 +/ calmodulin-dependent protein kinase 2s. Thus, this divalent cation effect was not Ca 2 ÷-specific which argues against the proposal that Ca 2 ÷ binds to cytosolic PKC inducing a conformational change resulting in translocation from the cytosol to the membrane 6. Consequently, we analysed lipid vesicle interactions with the enzyme in this Mg 2÷-induced conformational state. PKC tryptophan residues were quenched upon interaction with liposomes. Maximal quenching was at a lipid:protein ratio of 50450:1 (mol:mol). Considering the lipid surface is a bilayer, this would suggest that PKC associates with or covers an area of 30 lipid molecules on the membrane surface. The magnitude of quenching was dependent on the lipid composition, i.e. whether the lipid was charged or uncharged. Consequently, the acidic phospholipid, PS, which is the most potent lipid support for phosphorylating activity 1,2, was the strongest quencher. Only when the enzyme was in contact with lipid was it capable of undergoing a Ca2÷-specific
Int. J. Biol. Macromol., 1990, Vol. 12, August
255
Structural chan#es in protein kinase C: D. S. L e s t e r and V. Brumfeld
conformational change. As the Ca2+-specific effect on protein conformation occurred with all liposome systems analysed, this suggests that the determining factor for activity is the lipid composition. The role of activators in PS mixtures is generally considered to lower the affinity for Ca 2+ (Refs 1, 2). As can be seen in Table 1, the tryptophan quenching in the absence of Ca 2 ÷ is greater for PS vesicles containing activators than pure PS. However, as a relatively high Ca 2÷ concentration (100#M) was used for activity analyses in this study, the stimulation by lipid activators in pure PS vesicles was not significant. Partial irreversibility of Ca 2÷ binding to PKC-lipid complexes has been reported using the extrinsic fluorescent probe, dansylphosphatidylethanolamine 24'26. We also observed that the C a 2 +-specific effect on intrinsic P K C fluorescence was partially reversible. The c.d. measurements provided a different overview as they measured gross changes in P K C secondary structure. However, certain features were similar for changes in both c.d. and intrinsic tryptophan fluorescence spectra. Changes in secondary structure were also dependent on the specific lipid compositions. For example, PS/PC vesicles containing D A G or T P A had a stronger c.d. band intensity than PS/PC vesicles alone. Also, the Ca 2÷ effect on PKC-lipid activator complexes was more significant for PS/PC mixed liposomes. These Ca 2 ÷-indiced changes in the protein may either be a direct rearrangement of the membranepenetrating P K C secondary structure or may be due to some influence on the protein-lipid interaction resulting in changes in secondary structure. This Ca 2 ÷ effect was not significantly reversible in c.d. measurements (data not shown), which may have been due to insufficient time for the reversal to be observed or the irreversibility of subsequent changes upon Ca 2 + binding. The activity measurements were made using the exogenous substrate, histone Ill-S, which in order to undergo phosphorylation requires all of the P K C activating ligands. In contrast, the substrate, protamine, is phosphorylated by P K C in the absence of these ligands 2° and their presence has no significant effect on protamine phosphorylation 9. The fluorescence changes as observed by tryptophan quenching correlated to histone III-S phosphorylating activity. The intrinsic fluorescence of active PKC-lipid complex was 50-60 ~o of the native soluble enzyme. This indicated that the topography of the tryptophan residues was important in the active conformation. The tryptophan residues affected by the ligand association are in the regulatory domain of P K C 9. Thus, it may be concluded that portions of the regulatory domain must be in a specific conformation to obtain an active enzyme state. In contrast, the changes in c.d. spectra were difficult to relate to the activity as each active protein-lipid complex had a different secondary structure. The c.d. spectra differed according to the phospholipid and lipid activator composition. Thus, these differences in structure may have some significance in the in vivo cellular system where alternative PKC-substrate reactions may occur. The differences in c.d.-determined conformation of the various active forms (Figure 3 A , C and D) cannot be attributed to specific regions in the protein. Therefore, it appears from the c.d. analyses that there are regions of the protein secondary structure undergoing change which remain to be clarified and identified. For example, from
256
Int. J. Biol. Macromol., 1990, Vol. 12, August
our studies, it is proposed that the enzyme rearranges upon lipid association to form a Ca 2 ÷-specific site. Such a suggestion requires a different approach than searching for consensus regions. However, this provides a possible explanation for the lack of any identifiable Ca 2 +-binding site consensus region (e.g. 'E-F hand') in the c D N A deduced sequence of the P K C isozymes 7's. The direct spectral measurements of changes in protein structure of P K C would suggest that the proposed pathway of P K C activation based on biochemical analyses requires re-examining. F r o m our biophysical analyses, we suggest that the enzyme associates with lipid before it is capable of specifically binding Ca 2 + to obtain its final (active) configuration. These structure-function analyses provide an important alternative outlook on the changes in secondary structure that P K C undergoes On the pathway to activation.
Acknowledgements This work was supported in part by a grant from the Mazer Center for Structural Biology (D.S.L.)and a US Naval Research Aid (V.B.). We would like to express our gratitude to Martin Welch for his critical comments.
References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
Bell, R. M. Cell 1986, 45, 631 Nishizuka,Y. Science 1986, 233, 305 Ashendel,C. L., Staller, J. M. and Boutwell, R. K. Cancer Res. 1973, 43, 4327 Sando,J. J. and Young, M. C. Proc. Natl. Acad. Sci. USA 1983, 80, 2642 Kondo, H., Kinoshita, J., Matsuba, T. and Sunamoto, J. FEBS Lett. 1986, 201, 94 Ashendel,C. L. Biochim. Biophys. Acta 1985, 822, 219 Parker, P. J., Coussens, L., Totty, N., Rhee, L., Young, S., Chen, E., Stabel, S., Waterfield, M. D. and Ullrich, A. Science 1986, 233, 853 Ono, Y., Fujii, T., Igarashi, K., Kuno, T., Tanaka, C., Kikkawa, U. and Nishizuka, Y. Proc. Natl. Acad. Sci. USA 1989, 86, 4868 Brumfeld,V. and Lester,D. S. Arch. Biochim. Biophys. 1990,277, 318 Nishizuka,Y. Nature 1989, 334, 661 Huang,K.-P., Chan, K.-F. J., Singh, T. J., Nakabayashi, H. and Huang, F. L. J. Biol. Chem. 1986, 261, 12134 Bach,D. Personal communication Lester,D. S. J. Neurochem. 1989, 52, 1950 Pershadsingh, H. A. and McDonald, J. M. J. Biol. Chem. 1980, 255, 4087 Freifelder, D. 'Physical Biochemistry. Applications to Biochemistryand Molecular Biology'.W. H. Freeman and Co., San Francisco, 1976, pp. 415-421 Bradford,M. M. Anal. Biochem. 1976, 72, 248 Arslan,P., DiVirgilio,F., Beltrane, M., Tsien, R. Y. and Pozzan, T. J. Biol. Chem. 1985, 260, 2719 Maraganore,J. M. TIBS 1987, 12, 176 Yang,C. C. and Chang, L. S. Int. J. Biol. Macromol. 1989, 11, 13 Bazzi,M. D. and Nelsestuen, G. L. Biochemistry 1988, 26, 1974 Wallace,B. A. and Teeters, C. L. Biochemistry 1987, 26, 65 Chang,C. T., Wu, C.-S. C. and Young, J. T. Anal. Biochem. 1978, 91, 13 Hennessey,J. P. and Johnson, W. C. Jr Biochemistry 1981, 20, 1085 Bazzi,M. D. and Nelsestuen, G. L. Biochemistry 1987, 26, 115 Rodriguez-Paris,J. M., Shoji, M., Yeola, S., Liotta, D., Volger, W. R. and Kuo, J. F. Biochem. Biophys. Res. Comm. 1989, 159, 495 Epand,R. M., Stafford,A. R., Bottega, R. and Ball, E. H. Biosci. Reports 1989, 9, 315 Calhoun,D. B., Vanderkooi, J. M., G. R. and Englander, S. W. Proteins 1986, 1, 109 King,M. M. J. Biol. Chem. 1988, 263, 4754