Journal of Muscle Research and Cell Motility 13, 308-314 (1992)

A monoclonal antibody that recognizes different conformational states of skeletal muscle troponin C and other calcium binding proteins PRISCILLA

F. S T R A N G *

and J A M E S

D. P O T T E R

Department of Molecular and Cellular Pharmacology, University of Miami School of Medicine (R-189), PO Box 016189, Miami, FL 33101, USA Received 25 July 1991; revised and accepted 20 October 1991

Summary Skeletal muscle troponin C contains four CaZ+-binding sites, two with a high affinity for Caz+ that also bind Mg 2+ competitively (Ca2+/Mg2+ sites) and two sites of lower affinity that are specific for Caz+. We have characterized a monoclonal antibody (BqD9) that was produced against rabbit skeletal troponin C. The binding of this antibody to rabbit skeletal troponin C is sensitive to the binding of Caz+. Increasing the Ca2+ concentration produces a decrease in the amount of antibody bound with a pK of approximately 6.9 which correlates with Ca2+ binding to the Ca2+/Mg2+ sites. Magnesium binding to rabbit skeletal troponin C had no effect on antibody binding. Thus the conformation of rabbit skeletal troponin C brought about by Ca2+ binding to these sites affects the antibody binding to its epitope. This epitope was unavailable for antibody binding in whole troponin. The antibody-binding site was localized in cyanogen bromide fragment CB9 of rabbit skeletal troponin C (residues 84-135). This antibody was also shown to cross-react with bovine cardiac troponin C, barnacle (Balanus nubilus) troponin C, bovine testis calmodulin and carp parvalbumin. In addition, the effect of Ca2+ on antibody binding seen with rabbit skeletal troponin C was also seen with bovine cardiac troponin C, and calmodulin. Thus these proteins appear to share a similar epitope and undergo similar structural changes. Wang and colleagues (1987) have presented evidence that rabbit skeletal troponin C at low pH has an elongated structure similar to that seen in the crystal structure and that at neutral pH its structure is more compact. We have found that in the absence of Caz+ this antibody binds best to rabbit skeletal troponin C at low pH and its binding is reduced with increasingly alkaline pH. It is possible that the structural alterations brought about by changes in pH may also be responsible for the reduction in antibody binding. Since pH and Ca2+ have the same effect on antibody binding, this may mean that Ca2+ binding to the Ca2+/Mgz+ sites may also make rabbit skeletal troponin C more compact.

Introduction A calcium dependence for actomyosin ATPase activity in the presence of tropomyosin and another protein called troponin was first described by Ebashi and Kodama (1965). Troponin was later shown by Greaser and Gergely (1971, 1973) to contain three subunits. TnT (tropomyosin binding), TnI (inhibitory) and TnC (the Caa+-binding subunit). STnC is the calcium-binding subunit of troponin from skeletal muscle of rabbit. It has four Ca2+-binding sites, two that have a high affinity for Ca 2+ that also bind Mg 2+ Competitively (Ca2+/Mgz+ sites) and two sites of lower affinity for Ca z+ (CaZ+-specific sites) (Potter & Gergely, 1975). *To whom correspondence should be addressed. 0142-4319 9

1992 Chapman & Hall

Studies of the crystal structure of STnC (grown at pH 5.0) suggest that the protein is elongated and consists of the low- and high-affinity metal-binding domains, separated by a nine turn alpha helix (Herzberg & James, 1985; Sundaralingam et al., 1985). Studies by Wang and colleagues (1987) employing energy transfer to measure the distance between Tb ~+ bound to the Ca2+-specific sites and 4-dimethylaminophenylazophenyl-4'-maleimide bound t o Cys 98, at pH 5.0, showed that the distances between these two Ca2+-binding domains were compatible with the elongated structure described for the crystal. However, as they raised the pH to 6.8, the measured distance decreased substantially (approximately 2.5 nm.), implying that at neutral pH the protein was more compact. It was suggested that this compactness could arise from a folding of the D-E helix. Solution studies

Monoclonal antibody recognizing conformational states

309

using low ionic strength conditions have shown that lowering the pH resulted in a tightening of the alpha helix in the protein as measured by increases in circular dichroism (Lehrer & Leavis, 1974). Tyrosol fluorescence (Leavis & Lehrer, 1978) has also been shown to change in parallel with these circular dichroism changes. Other studies using metal chelators to regulate the free metal concentration have demonstrated that metal binding causes similar tightening of the alpha helix. These conformational changes have been shown to occur mainly with Ca 2+ or Mg 2+ binding to the CaZ+/Mg 2+ sites (Leavis & Gergely, 1984) although some structural changes have been reported to occur with Ca 2+ binding to the Ca 2+specific sites (Johnson & Potter, 1978). In this paper we have used a monoclonal antibody produced against STnC to probe for structural changes in the protein. We present evidence that the binding of this antibody to STnC is affected by Ca 2+ and pH, and the binding region has been localized to the cyanogen bromide fragment, CBg, or STnC (Collins et al., I977). The antibody is cross-reactive with all of the Ca 2+binding proteins that we have tested, and is also able to detect similar CaZ+-dependent conformational changes in STnC, bovine cardiac troponin C (CTnC) and calmodulin suggesting a similar binding region in these calciumbinding proteins.

bromide fragments of STnC. The basic assay used is presented below. All the assays were carried out at room temperature. Purified proteins [10 gg per well] in buffer A (90 mM KC1, 120 mM MOPS (pH 7.0), 4 mM EGTA were adsorbed to microtitre wells for 2 h. Unbound protein was washed away by three washes (10 rain each) of buffer A. Primary antibody (purified IgM) in buffer A at the appropriate dilution was incubated with the antigen for 2 h. Unbound primary antibody was removed by three washes (10 min each) in 1% bovine serum albumin (BSA) in phosphate buffered saline (PBS, 150 mM NaC1, 10 mM NaH2PO 4, pH 7.2). Secondary antibody (goat anti-mouse IgM; Hyclone) conjugated to horseradish peroxidase (HRP) diluted 1:5000 in PBS was incubated in the wells for I h. Unbound secondary antibody was removed by three washes (10 min each) in PBS. The substrate 1.25 mM 4-amino-antipyrine (Sigma) in 85 mM phenol, 0.003% hydrogen peroxide in PBS was incubated with the antigen-antibody complexes for I h. The reaction was terminated by the addition of 0.01% sodium azide. The optical density was then read at 490 nm on a BioTec ELISA reader. Controls were routinely run to check for the non-specific effects of the secondary antibody and substrate. The specific variations to the basic ELISA are as follows. To study the effect of calcium on antibody binding to STnC, the amount of calcium added to the solutions to give the desired free Ca2+ concentration was determined using the programme of Robertson and Potter (1984). To study the effect of magnesium on antibody binding, 2 mM MgC12 was included in buffer A. To study the effect of pH, buffer A was changed from MOPS to either Pipes or K-PO4, and EDTA was used instead of EGTA. To study the amount of STnC adsorbed to the ELISA wells, I25I-STnC(50 000 c.p.m.) was added to each well at each of the pCa values used in the Ca2+-dependence assay. The STnC was labelled with ~zsI by New England Nuclear, Inc. Studies of antibody binding to the CB fragments of STnC used in inhibition ELISA. Fragments diluted in buffer A were added to the primary antibody which was then incubated with the adsorbed STnC.

Materials and methods Protein preparation Troponin subunits from rabbit skeletal muscle and bovine cardiac muscle were prepared according to the methods of Potter (1982). Cyanogen bromide (CB) fragments of STnC were prepared according to the methods of Collins and colleagues (1977), and were a gift from Dr J. Collins. Barnacle TnC (BTnC) was prepared as described elsewhere (Potter et al., 1987; Collins et al., 1991). Calmodulin was prepared from bovine testis according to the methods of Crouch and colleagues (1981). Parvalbumin was made from carp muscle according to the methods of Kretsinger and Nockolds (1973). Monoclonal antibody production The monoclonal antibody (BgD9) we prepared to STnC was made using the standard methods of Galfre' and Milstein (1981). Purified IgM was eluted from a Sephacryl S-500 (Pharmacia) column with 0.5 M NaC1, 20 mM Tris (pH 8.0), I mM EDTA, and 0.02% NaN 3. ELISA The antibody was tested for its binding to STnC as well as other troponin subunits and other selected calcium-binding proteins, using the ELISA (Voller et al., 1980). The ELISA was used in a variety of ways to: (1) study the effects of calcium and magnesium on antibody binding to STnC and other calcium-binding proteins, (2) study the cross-reactivity of the antibody with other troponin subunits and selected calciumbinding proteins, (3) study the effect of pH on antibody binding to STnC, and (4) study the binding of the antibody to cyanogen

Immunodot assay The methods of McDougal and colleagues (1983) were used for the immunodot assays. The immunological subtype of the antibody was determined using this assay. Also, the initial cross-reactivity studies with the calcium-binding proteins and other muscle proteins were done using this assay. A buffer containing 90 mM KC1, 120 mM Mops (pH 7.0), 4 mM EDTA was used for the dilution of the proteins. Results Characterization of the antibody The antibody (BgD9) was produced to purified STnC as described in Materials and methods and was found to be stable through the three limited dilution clonings. Using immunoglobulin isotype and light chain specific antibodies we were able to determine that the antibody was an IgM isotype containing kappa light chains. The results of the initial cross-reactivity experiments using the immunodot assay with other troponin subunits and calcium-binding proteins is shown in Table 1. The antibody bound to STnC, CTnC, BTnC~+2, calmodulin

310

STRANG and POTTER

Table 1. Cross-reactivity of antibody binding with calcium-binding proteins and troponin subunits Cross-reactivity of BgD9 with other proteins Calcium-binding proteins STnC CTnC BgD9 + + Troponin subunits w w BgD9 . . .

w

w

BTnC~+2 +

+

+

CTnT

CTnI

#BTnC1 +

#BTnC2 +

.

* The immunodot method of McDougal and colleagues (1983) was used for this experiment. Proteins (10 I~g per dot) in 90 mM KC1, I20 mM MOPS (pH 7.0), 4 mM EDTA were dotted and dried on nitrocellulose paper. The protein-containing nitrocellulose paper was then processed in a manner similar to the ELISA assay protocol outlined in Materials and methods with the following exceptions. The secondary antibody (goat anti-mouse IgM; Hyclone) was conjugated to alkaline phosphatase. The substrate for the secondary antibody visualization contained the substrate 5-bromo-4-chloroindoxyl phosphate and nitro blue tetrazolium (NBT) (Blake et al., 1984). #BTnC 1 and BTnC 2 are the two isoforms of TnC found in barnacle muscle (Collins et al., 1991) and BTnCI+z is a mixture of the two. w calmodulin; Parv, carp parvalbumin; STnT, rabbit skeletal troponin T; STnI, rabbit skeletal troponin I; CTnI, bovine cardiac troponin I.

and parvalbumin. There was no cross-reactivity with the other troponin subunits (TnI and TnT). Further studies of antibody binding to selected calcium-binding proteins were carried out in the presence and absence of calcium and/or magnesium and the results are summarized in Table 2. All of the values were normalized to STnC in the absence of Ca z+ (Buffer A). The results clearly show that there is a similar effect of Ca z+ on antibody binding to all of the calcium-binding proteins tested. The antibody bound best to each of the proteins and with essentially equivalent affinity in the absence of metal. The addition of magnesium had little effect on antibody binding to any of the proteins. Ca 2§ however, greatly reduced the amount of antibody binding and there was no additional effect of Mg 2+. The amount of binding in the presence of Ca2+ was reduced to 21-36% of that seen in the absence of metal. Also illustrated in Table 2, is the fact that antibody binding to whole troponin was essentially negligible under all of the conditions tested. Ca e+ dependence of antibody binding to STnC In order to determine the Ca 2+ dependence of antibody binding to STnC, the ELISA was carried out at different

Ca 2+ concentrations. The results of the binding of the antibody to STnC at the various pCa is shown in Fig. 1. The percentage total change in antibody binding was plotted as a function of free calcium concentration (note that in Table 2 in contrast to Figs 1-4, percentage bound rather than percentage total change is used). A single transition was observed and the data was fit to the Hill equation to estimate the pCas0. The value obtained was approximately 6.9. This value would correspond to Ca 2+ binding to the Ca2+/Mg 2+ sites (Potter & Gergely, 1975) although the transition is rather broad and it is not possible to rule out some contribution to this process from the CaZ+-specific sites. As similar results could have been obtained if STnC binding to the ELISA wells was Ca2+-dependent and antibody binding to STnC was not, we carried out additional experiments to make certain that this was not I00.0:

80.0

~'n

60.0

Table 2. Effect of Ca2+ and M g 2+ on the binding of B9D 9 to STnC and other CaZ+-binding proteins

o

40.0

Antibody binding to BgD9 Protein -Metal q-Ca 2+ + M g 2+ (%) (%) (%)

.c

20.0

+CaZ+/+Mg z+ (%)

STnC CTnC Calmodulin Troponin

21.5 36.4 27.3 5.5

100.0 99.5 99.1 2.i

28.5 32.9 31.2 3.3

I14.7 104.2 97.6 2.2

*To compare the antibody binding to selected calcium-binding proteins in the presence of different metals, the proteins (10 I.tg per well) were adsorbed to wells of ELISA plates in: (1) Buffer A (-metal), (2) Buffer A at pCa 4.0 (+Ca2+), (3) Buffer A plus 2 mM MgC12 (+MgZ+), or (4) Buffer A at pCa 4.0 plus 2mM MgC12 (+CaZ+/+Mg2+). All values are compared with that of STnC in Buffer A.

9

"

o ct~1 O

o~

0.0

10

I

t

i

i

9

8

7

6

--

5

4

pCa

Fig. 1. Ca2+ dependence of antibody binding to STnC. The general protocol for the ELISA assay outlined in Materials and methods was used here with the following exceptions. For each point, buffer A, containing the desired pCa, was used for the coating, washing and primary antibody-binding steps. After these additions, the protocol was identical to that described in Materials and methods. Each data point is an average of six determinations. The percentage total change in antibody binding is plotted on the y-axis as a function of pCa.

311

Monoclonal antibody recognizing conformational states the explanation for our results. To rule out this possibility the calcium titration experiment illustrated in Fig. 1 was repeated and 12SI-labelled STnC was used to coat the wells. The amounts of STnC bound to the wells was measured by cutting out the individual ELISA wells and counting them in a gamma counter. The results of this experiment clearly showed that there was no Ca2+ dependence of the binding of STnC to the ELISA wells and that the same amount of STnC was bound at each pCa (data not shown). To test this idea further and to also test whether the conformation of STnC, as sensed by antibody binding, could be changed once it was bound to the well, the following experiment was performed. In Fig. 1 the STnC coating, washing and primary antibody binding steps were all carried out at the pCa indicated. In Fig. 2, two variations of this procedure were employed. In the first case the STnC coating step was carried out at pCa 8.0 and the subsequent washing and primary antibody binding steps were carried out at the different pCa indicated on the abscissa of Fig. 2. This protocol gave essentially identical results to those illustrated in Fig. 1 and indicate that although the STnC was adsorbed at pCa 8.0 the amount of antibody bound was clearly related to the pCa at which the washing and primary antibody steps were carried out. Thus STnC could change its conformation once adsorbed to the well. In a second experiment STnC was adsorbed to the well at the pCa indicated on the abscissa of Fig. 2 and subsequent washes and primary antibody binding steps were carried out at pCa 8.0. Again, these results (Fig. 2) clearly show that the conformation of STnC could change in response to Ca 2+, once bound to the well. It also confirms the uSI-labelled STnC experiments, as the amount of antibody at pCa 8.0 was the same regardless of what pCa the STnC was adsorbed to the plate.

Effect of pH on antibody binding to STnC Since pH has been shown to affect the structure of STnC (Leavis & Lehrer 1978; Wang et al., 1987) it was of interest to see whether pH could affect antibody binding to STnC. The effects of pH on antibody binding to STnC are shown in Fig. 3. Antibody binding in either Pipes or K-PO4 buffer was maximal at pH 6.0. The percent maximal antibody binding (not percent binding) decreased as the pH became more alkaline and reached its lowest value at pH 8.0. The apparent pK D for these plots was 6.85 and 6.81. Control experiments similar to those carried out in Fig. 2 for Ca 2+ were also carried out for the pH experiments. To test whether the conformation change in STnC brought about by changes in pH was reversible once STnC was bound to the well and to test whether pH affected STnC binding to the plate, two variations of the experiment illustrated in Fig. 3 were carried out. In the first case the STnC coating step was carried out at pH 6.0 and for each point the subsequent washing and primary

100"0 L ~ " 0") E ._ "0 E ~ m

80.0

60.0

>.. "0 0 .(3 ~

40.0

E .< r"

t-

20.0 0.0

' 8

tO

' 7

' 6

' 5

pCa

Fig. Z. Ca2+ dependence of antibody binding to STnC. This experiment was carried out to test whether the conformation of STnC could change once it was bound to the well of the ELISAplate at different pCa and to test whether the pCa during the adsorption step affected the amount of STnC bound to the plate. The protocol is similar to that described in Fig. 1 with the following exceptions. In one case STnC was adsorbed to the ELISA plate at pCa 8.0 and the subsequent washings and primary antibody binding steps were carried out at the pCa indicated on the abscissa (O). In the other case, for each point the STnC was adsorbed to the wells of the plate at the pCa indicated and the subsequent washing and primary antibody binding steps were carried out at pCa 8.0 (I). The subsequent steps were carried out as described in the general protocol. Each data point is an average of six determinations. The percentage total change in antibody binding is plotted on the ordinate as a function of pCa.

100.0

S

0"~ C ._ 'XZ) (-

80.0

,_

cn "0 0

60.0

40.0

(-

A monoclonal antibody that recognizes different conformational states of skeletal muscle troponin C and other calcium binding proteins.

Skeletal muscle troponin C contains four Ca(2+)-binding sites, two with a high affinity for Ca2+ that also bind Mg2+ competitively (Ca2+/Mg2+ sites) a...
694KB Sizes 0 Downloads 0 Views