Cloning, Expression, and Site-directed Mutagenesis of Chicken

THE JOURNAL 0 1988 by The American Society for Biochemistry

OF BIOLOGICAL CHEMISTRY and Molecular Biology , Inc

Val. 263, No. 5, Issue of Fehruary 15, pp. 2371-2376, 1988 Printed in U.S.A.

Cloning, Expression, and Site-directed Mutagenesis of Chicken Skeletal Muscle Troponin C*

(Received for publication, August 28, 1987)

Fernando C. ReinachS6 and Roger KarlssonllII From the $Departamento de Bioquimica, Instituto de Quimica, Universidade de Sao Paul0 C.P. 20.780 CEP 01498 Sao Paula SP, Brazil and the llLaboratory of Molecular Biology, Medical Research Council Centre, Hills Road, Cambridge CB2 2QH, United Kingdom

Skeletal muscle troponin C (TNC) is structured into two separate domains linked by a nine-turn a-helix (D/ E helix). It has been demonstrated that calcium binding to the regulatory sites within the N-terminal domain induces conformational changes in the C-terminal domain of isolated TNC. Since the only contact between the two domains is the long D/E helix, the transfer of information must involve conformational changes within this helix. The center of the helix is occupied by a glycine (Gly-92). A postulated mechanism for allowing interdomain interaction involves a conformational change of the D/E helix around Gly-92 (Herzberg, O., and James, M. N. G. (1985) Nature 312, 653-659). We tested this hypothesis using site-directed mutants of troponin C. Two separate mutants containing an alanine and a proline replacing Gly-92 were constructed and compared with wild type TNC. Calcium binding studies showed no significant differences among the TNC species. The different TNC were as- sembled into thin filaments and used to assay the calcium regulation of actin-activated ATPase of myosin. All TNC species were able to mediate the calcium regulation of ATPase. Under the conditions used for the assays, no differences were detected among the TNC species. These results show that Gly-92 is not essential for the proper interaction of the calcium regulatory sites with the other components of the thin filament, and therefore exclude a large rotation around Gly-92 as the mechanism of information transfer between the two domains of troponin C.

An increase in the cytosolic calcium concentration triggers skeletal muscle contraction within a few milliseconds. Con- traction is the result of actomyosin interaction leading to ATP hydrolysis and generation of tension. In vertebrate skeletal muscle this process is regulated by components associated to the thin filament. The regulatory system consists of a

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Supported by grants from Financiadora de Estudos e Projetos, FundasHo de Amparo a Pesquisa do Estado de SHo Paulo, Conselho Nacional de Desenvolvimento Cientifico e Tecnologico, and The Muscular Dystrophy Association. To whom correspondence should be addressed.

11 Supported by Post-doctoral Grant B-PD 1492-100 from the Swedish Natural Science Research Council. Present address: Dept. Zoological Cell Biology, Biology, House E5, University of Stockholm, S-10691, Stockholm, Sweden.

troponin molecule with three subunits (TNC,’ TNI, and TNT), a tropomyosin, and seven actin monomers. Calcium binding to TNC causes conformational changes of the regulatory system allowing the actomyosin interaction to occur. The molecular mechanism behind the regulatory process has been studied by a variety of methods including x-ray diffrac- tion (see Ref. 1 for review), fast kinetics, and spectroscopic methods (2). With the determination of the crystallographic structure of TNC (3, 4) and a low resolution x-ray structure of tropomyosin (28) and tropomyosin/troponin complexes (29), sufficient biochemical and structural data are available for the design and analysis of site-directed mutants of troponin and tropomyosin. We have cloned a full length cDNA for troponin C, expressed the protein in Escherichia coli, and studied the wild type protein and two mutants.

Troponin C has four metal-binding sites (5). There are two low affinity calcium-specific sites in the N-terminal half of the molecule. The other two sites, located toward the C- terminal half of the molecule, have a higher affinity but are thought to be occupied by magnesium in the relaxed muscle (5). The two sites at the N-terminal domain have “on rates” and “off rates” for calcium compatible with the kinetics of tension development and therefore are considered the regulatory sites (6-8). Proteolytic studies of TNC have shown that the C-terminal half of the molecule is necessary for proper regulation since the isolated N-terminal half cannot regulate the actomyosin ATPase (9). It was also shown that calcium binding to the N-terminal sites induce conformational changes in the C-terminal domain (10). This observation points toward a direct interaction between the two domains as a possible route of information transfer within the troponin complex (10).

The crystal structure of troponin C (3, 4) has shown that the two N-terminal metal-binding sites are separated from the C-terminal domain by a long nine-turn a-helix with the central 11-amino acid residues being exposed to the solvent. The presence of a glycine (Gly-92) in the center of this a- helix and the lack of any direct contact, other than the helix, between the C- and N-terminal domains led to the suggestion that a bending around Gly-92 would give rise to a direct interaction between the two domains (3).

To test the role of the putative flexibility provided by Gly- 92 to the long D/E helix and a possible function for this residue in the transfer of information between the N- and C- terminal domains, we mutated this residue into an alanine and a proline. Alanine was selected since it would decrease the flexibility of the helix, introducing a small side chain which should not interfere with the assembly of the troponin

TNT, troponin T; DTT, dithiothreitol; EGTA, [ethylenebis- The abbreviations used are: TNC, troponin C; TNI, troponin I;

(oxyethylenenitrilo)]tetraacetic acid.

2371

2372 Site-directed Mutants of Troponin C

complex. A proline residue would restrict to a larger extent the rotation around position 92. If Gly-92 was essential for TNC function, we would expect a large impairment in thin filament regulation in the presence of the mutant TNC mol- ecules. Although there is no "a priori" reason to expect changes in metal binding as a consequence of mutations in the long D/E helix, calcium binding in both classes of sites was assayed as a general measure of the folding properties of the mutants.

Our results show that the regulatory system containing either mutant could not be distinguished from the wild type structure, thus excluding a large rotation around Gly-92 as the route of information transfer between the two domains of troponin C.

MATERIALS AND METHODS

Isolation of Full Length cDNA-A total of 60,000 events from a gtl0 cDNA library constructed with poly(A)+ mRNA from 2-week post-hatch chicken pectoralis muscle (11) were screened with a 94- bp single strand probe encoding amino acids 72-102 of rabbit skeletal muscle TNC (12). The probe was prepared from M13 ssDNA using a standard DNA-sequencing primer and extension in the presence of dCTP, dGTP, dTTP, [3ZP]dATP, and the Klenow fragment of DNA polymerase. The reaction products were digested with EcoRI, and the labeled strand containing the insert was purified on a sequencing gel under denaturing conditions (13). Hybridization was performed in 5 X ssc (20 X ssc is 3 M NaC1, 0.3 M sodium citrate, pH 7.0), 0.1% sodium dodecyl sulfate, 0.1% Ficoll, 0.1% polyvinylpirrolidone, 1 mM EDTA at 67 "C. The filters were washed with a final stringency of 2 X SSC, 67 "C. A total of 150 positive signals were isolated. Ten clones giving strong signals had inserts longer than 500 bp; four clones with inserts larger than 850 bp were sequenced. Three of these contained the complete coding sequence of TNC and variable portions of the 5"untranslated region.

DNA and Protein Sequencing-All DNA sequences were determined using the dideoxy chain termination method (14). The cDNA was sequenced in both strands using a series of progressive deletions (15). After the initial sequence a series of oligonucleotides, evenly spaced along the sequence, were used to resequence the cDNA after each cloning or mutagenesis step.

The sequence of the N terminus of the E. coli produced TNC, after cleavage with factor Xa, was determined in a gas-phase sequencer using standard procedures.

Construction of the Plasmid Expressing Troponin C-The full length cDNA was cloned into the single EcoRI site of the M13 derivative M13K11RX (16), and a clone with the 5' end of the mRNA facing the universal primer was selected. Single strand DNA derived from this construct was obtained in an EcoK restriction minus, methylation minus strain (TG-2). An oligonucleotide coding for the factor Xa recognition sequence and the first three amino acids of TNC (ATCGAGGGTAGGATGGCGTCAATG) was used to delete the 5' end of the cDNA, and juxtapose the last codon of the factor Xa recognition sequence (Arg) to the first codon of TNC (Met). Hybridization and extension were performed as described (17, 18). The double strand product was transfected into an EcoK restriction plus strain (JM101) to select against the four EcoK sites and the 5'- untranslated region of TNC present in the single strand loop. A clone with the proper deletion was isolated and resequenced. The FXTNC portion of the insert was excised using BamHI and HindIII, cloned into pLCII (19), and transfected into QY13 (20). The final construct (pLCIIFXTNC) was tested for expression. The HindIII site is not present in the final construct due to the deletion of a single base pair.

Site-direct Mutagenesis-To avoid manipulations within the coding region of the fusion protein, the region encoding CIIFXTNC and the flanking sequences were excised with EcoRI and cloned into M13mp18. A clone containing the 5' end of the mRNA facing the universal primer (F884) was resequenced to completion before pro- ceeding with the mutagenesis. Two oligonucleotides (ACGCCA- AGCCCAAGTCTGA; ACGCCAAGGCCAAGTCTG) were used to mutate Gly-92 into Pro and Ala, respectively. Mutagenesis and screening were performed as described (17, 18). Discrimination between mutant and wild type TNC genes was possible with the labeled oligonucleotides after a 62 "C wash in 6 X SSC. The mutants were plaque purified, confirmed by sequencing, and the complete sequence for the coding region was determined. The Gly-92-Pro (F889) and the

Gly-92-Ala (F895) constructs were excised from M13mp18 with EcoRI and cloned into pLmplO (19). After selecting clones with the proper orientation, they were tested for expression.

Purification of Recombinant TNC and the Two Mutants Produced in E. coli-An overnight culture of QY13 containing the plasmid was grown at 30 "C in 2 X TY (16 g/liter tryptone, 10 g/liter yeast extract, 5 g/liter NaCl, pH 7.4) with 25 pg/ml ampicillin. Three liters of the same medium distributed in 6 X 2 liter flasks were each inoculated with 5 ml of the overnight culture and grown at 30 "C until OD, = 0.6. Induction was accomplished by immersing the flasks in a 85 "C bath while monitoring the temperature of the medium. As the temperature reached 42 "C (30-45 s) the flasks were transferred to a 42 "C bath for 15 min. The cultures were grown for another 4 h at 37 "C. The cells were collected at 6,000 X g for 10 min and kept frozen. Cells were resuspended in 15 ml of 50 mM Tris-C1, pH 8.0, 25% sucrose, 1 mM EDTA, 2 mg/ml lysozyme. After 15 min, on ice, MgC12, MnClZ, and DNAase I were added to a final concentration of

of 200 mM NaCl, 1% deoxycholic acid, 1.6% Nonidet P-40, 20 mM 10 mM, 1 mM, and 1 pg/ml. When the viscosity was reduced, 30 ml

Tris-C1, pH 8.0, 2 mM EDTA were added. The resulting extract was clarified for 15 min at 10,000 X g and trichloroacetic acid (100% w/v stock) was added under continuous stirring to a final concentration of 5%. The precipitate was collected (3000 X g; 10 min) and resuspended in 25 mM Tris-C1, pH 8.0, 1 mM MgClZ, 1 mM DTT, and dialyzed against the same buffer. After removing the material not

urea and loaded to a DEAE-cellulose (DE52 Whatman) 2 X 15-cm resuspended (10,000 X g; 20 min) the supernatant was made 6 M in

column equilibrated with 50 mM Tris-C1, pH 8.0, 6 M urea, 1 mM MgClz, 1 mM 2-8-mercaptoethanol at room temperature. The protein was eluted with a 100 X 100 ml gradient of 0-0.6 M NaCl in the same buffer. CIIFXTNC eluted at the end of the gradient. The protein was extensively dialyzed against 50 mM Tris-C1, pH 8.0, 1 mM MgC12 to remove traces of detergent and kept frozen. The fusion proteins were digested with factor Xa in 50 mM Tris-C1, pH 8.0, 100 mM NaCl, 1 mM MgClZ, 0.1 mM CaClZ with a substrate to enzyme ratio of 30:1, w/w. The presence of divalent cations protected an internal site which was cleaved by factor Xa at a lower rate. After digestion the protein was repurified on the same DE52 column to remove the N- terminal peptide.

Actin (21), myosin (22), troponin I and T subunits (23), and tropomyosin (24) were purified from rabbit back muscle using standard procedures. Troponin C was purified from chicken pectoralis muscle (23).

Calcium Binding-Calcium binding to isolated TNC was measured using a filtration assay (17). The measurements were made in free Ca2+ concentrations ranging from to lo-' M in 50 mM Imidazole- HCl buffer, pH 7.0, with or without 1 mM MgCL An apparent Ca/ EGTA binding constant at 25 "C, pH 7.0, of 3 X lo6 M" was used to calculate the free calcium concentrations. The final concentration of 45Ca/EGTA was in the 5-20 p M range (specific activity was 200-900 cpm/pmol).

Reconstitution of the Troponin Complex-The troponin complexes were reassembled using four different TNC species (wild type purified from chicken muscle; wild type produced in E. coli; Gly-92-Pro mutant and Gly-92-Ala mutant) with TNI and TNT purified from skeletal muscle. The reconstitution protocol is based on a standard procedure (23). The three subunits were mixed to a final concentration of 22 pmollpl of each component in 600 pl of 5 mM Imidazole- HCl, pH 7.0, 5 mM CaC12, 500 mM KCl, 2 mM sodium azide, 0.5 mM DTT. After 12 h of dialysis, the buffer had its CaCL concentration reduced to 2 mM and KC1 concentration to 250 mM. After another 12 h Imidazole was reduced to 2 mM, CaC12 to 20 p ~ , and KC1 to 100 mM. After another 12 h the samples were spun in an Eppendorf centrifuge (11,000 X g, 5 min) and a small precipitate (less than 1% of total protein) was observed in all four samples.

Calcium Regulation of the Actin-activated ATPase of Myosin-The troponin complex (80 pl) was premixed with tropomyosin (28 pl of 5.7 mg/ml solution in 20 mM Tris-C1, pH 7.5, 0.1 mM DTT). Actin (75 pl of a 5.1 mg/ml solution in 2 mM Tris-C1, pH 7.5, 0.1 mM CaATP, 1 mM MgCl,, 20 mM NaCl) was then mixed with the troponin-tropomyosin complex. After a 5 min incubation on ice,

buffer, pH 7.6, 0.6 M KC1, 4 mM azide, 0.2 mM DTT) was added. myosin (30 pl of a 10 mg/ml solution in 5 mM sodium phosphate

After mixing 1.73 ml of 40 mM KC1, 5 mM MgCl,, 0.1 mM EGTA, 0.25 mM DTT was added to the mixture. The 1.95 ml of the assay mixture was transferred to the pH-stat, equilibrated at 25 "C, and the end point titration, pH 7.6, was obtained with 10 mM NaOH. To measure the minus calcium rate 40 pl of 100 mM ATP was added, the

Site-directed Mutants of Troponin C 2373

pH drop was corrected, and a constant rate was measured. Calcium (10 pl of a 100 mM CaCI2) was added in the end of the experiment to ascertain the proper regulation of the filament. To measure the plus calcium rate, calcium was added before the ATP, and the same procedure was followed. A higher consistency from one experiment to the next could be obtained by measuring the plus and minus rate in separate experiments.

The final assay conditions were: 40 mM KC], 5 mM MgCI,, 0.1 mM EGTA, 0.25 mM DTT, 2 mM ATP, 4.5 p M actin, 1.28 pM troponin, 1.28 p M tropomyosin, 0.33 p M myosin. Calcium was 4 X M in the plus calcium situation and approximately lo-’ M in the minus calcium situation (4.6 p~ CaC12 contributed by the protein solutions and 100 p M EGTA from the salt solution). The above conditions were first established to give optimal regulation with both types of wild type TNC and were then used to assay all four troponin complexes in parallel.

RESULTS

Isolation and Sequence of the Full Length cDNA-We have isolated a complete cDNA encoding the chicken fast skeletal muscle troponin C. The sequence is shown in Fig. 1. All three cDNAs had identical sequences. The deduced protein sequence corresponds closely to the protein sequence (25). The numbering system used in this paper starts with the alanine as residue one, in order to be compatible with the system used for the crystallographic studies (3) and the original TNC protein sequence (25). The unknown order of the four N- terminal residues (25) was determined from the cDNA sequence (Fig. 1). Two differences were found when comparing the cDNA sequence with the protein sequence. In the cDNA Asp-100 was Asn and Thr-130 was Ile.

Expression, Mutagenesis, and Purification of TNC from E. coli-To express high levels of TNC in E. coli, we used a pL promotor under regulation of CI,,, to drive the synthesis of a fusion protein containing the first 31 amino acids of CII, separated from the TNC sequence by a factor Xa recognition sequence (Fig. 2). The fusion protein is cleaved with factor Xa to yield the final product. This system (19) was used to express myosin light chain (17). An EcoK selection vector and an oligonucleotide were used to make a precise deletion on

370 380 390 4 0 0 4 1 0 420 x I

GUC (196 PPC G C T GUT GGG T T C UTC GUC PTC GUG GPG C T G G G T GUG P T T C T C UGG G C C UCT U r p L y r Psn 911 P S P G l y P h e I L a R S P I l c G l u G l u L e u G l v GLu I l e L e u U r g P 1 a T h r

430 4 4 0 450 460 4 7 0 4 8 0

GGG GUG CUC GTC UTC GUG GUG GPC U T 0 GUU GUC C T C U T G PUG GUT TCU GPC UUG UPC OPT G l y G l u His V a l I l c G L u G l u P r o I l c G l u U S L LEU tic1 L Y S P s p 5 e r PIP L y r P r n P r n -

490 500 5 1 0 520 530 5 4 0

GDC GGC CGC UTT GUC T T C GPT GUG T T C C T G DUG PTG UTG GPG G G T G T G C O G TPD GGG UTC A S P G L Y DPg I l c U S P P h e U S P G L u P h e L e u L y s t i e l t i e l G l u G l y V a l G l n - - -

FIG. 1. Sequence of the cDNA encoding chicken skeletal muscle troponin C. The protein sequence encoded by the cDNA is also shown. The 2 residues where the sequence differs from the published protein sequence are underlined. Glycine 92 is boxed.

K u n Kt K t‘

@ TN C E H I

o l i g o n u c l e o t i d e

F X

Thr Ala Glu Gly Gly Ser U e Glu Gly Arg Met Ala Ser 2 9 30 31 1 2

FIG. 2. Diagram showing the two-step construction of the expression vector. First the factor Xa recognition sequence was linked in frame with the cDNA (A) , and second ( B ) , the FXTNC was cloned into the pLCII vector. A, the factor Xa recognition sequence is shown in the hatched box. The coding region for TNC is shown in the open box. An oligonucleotide was used to induce the formation of a loop which includes the four EcoK ( K ) sites and the 5’-untranslated sequences. B, the structure of the expression vector. The left promoter of lambda phage (pL) and the coding sequences for the first 31 amino acids of CII were separated from the coding region of troponin C (TNC) by the recognition sequence of factor Xa (FX). C, the DNA and protein sequence of the fusion protein flanking the factor Xa cleavage site. The uertical arrow indicates the cleavage site in the protein. Hind111 ( H ) , EcoRI ( R ) , and BamHI ( B ) sites used for the construction are also indicated.

the 5”untranslated region of the cDNA and juxtapose the factor Xa recognition sequence (FX) to the first methionine of TNC (Fig. 2). The number of recombinants containing the desired deletions was low due to the presence of an internal EcoK site (position 165) in the cDNA. Based on the protein sequence work (25) and the cDNA sequence, we presume that the N terminus structure of muscle TNC will contain a blocked alanine, in contrast with an unblocked methionine in our construct (Fig. 2). A high level of expression (20-40 mg of pure protein/liter of culture) was obtained with both the mutants and wild type TNC.

For mutagenesis, the whole coding region of the hybrid gene was recloned into M13 and then recloned back into the expression vector. We decided on this strategy to avoid un- necessary DNA manipulations within the coding region after resequencing the mutants.

In contrast to the myosin light chain (17) and globin (19), the fusion TNC was not sequestered into insoluble particles when expressed in the bacteria but was instead found in the soluble fraction. It was therefore purified from the lysate after the DNAase digestion by precipitation with trichloroacetic acid, a treatment from which the protein readily refolds. Additional contaminants were separated on a DE52 column. Factor Xa was found to produce a single cleavage on the fusion protein in the presence of divalent metals. The presence of a second site with a slower rate of cleavage was detected in the absence of divalent cations. Repurification on the same column removed the cleaved N-terminal peptide and traces of other contaminants (Fig. 3). The N-terminal sequence of the final product was determined. The first 49 residues with the expected sequence (starting with methionine) were obtained (data not shown).

Due to the difference in the N-terminal sequence and the possibility that other post-translational modifications had been introduced in the heterologous bacteria system, we have performed all assays using both the wild type TNC produced in E. coli and TNC purified from chicken muscle as controls.


FIG. 3. Polyacrylamide-SDS gel (15%) electrophoresis showing the total protein extract of the QY 13 cells containing the expression vector pLCIIFXTNC before ( A ) and after ( B ) the induction of the fusion protein. The arrow indicates the fusion protein. The final product after purification (C) is also shown. The band above the pure TNC (C) is the Mg2+ free form of the protein. It is not present if the gel is run in the presence of MP. Migration of molecular mass markers of 96,68,42,30, and 20 kDa are indicated (arrowheads).

Calcium Binding-Calcium binding in the presence and absence of magnesium (Fig. 4) failed to detect any differences among the four TNC species. The wild types presented similar binding curves which could not be distinguished from the results obtained with the two mutants. The calcium binding in the presence of magnesium, although more scattered a t high calcium concentrations, indicates that the lower affinity calcium-specific sites can be detected. The measurements using this assay are not made under equilibrium conditions and, thus, cannot be used to calculate binding constants or saturation values. The method is reproducible and a large number of assays can be done in parallel allowing a direct comparison of the different proteins. Large decreases in affinity, as produced by mutation on the metal coordination posi- tions of myosin light chain two (17), can be readily detected.

From these results we conclude that no gross changes in calcium binding or protein folding were introduced by chang- ing Gly-92 into either Ala or Pro. It still remains possible that small changes in binding were not detected.

Regulation of the Actin-activated Myosin ATPase-We have

0.:

0.:

0 E 0.

5 - 0

.‘ 0

0.: - 0

E,

0 .:

0.

B

FIG. 4. Calcium binding to the four TNC species used in this study. D, wild type TNC purified from muscle. 0, wild type TNC produced in bacteria. 0, Gly-92-Pro mutant of TNC. 0, Gly-92-Ala mutant of TNC. The data in the absence (A) and in the presence of 1 mM magnesium (€3) are averages of six experiments, except for the Ala-92 mutant where five measurements were used. Error bars and best-fit curves are omitted for clarity.

reconstituted the four different TNC with TNI, TNT, tropomyosin, and actin purified from skeletal muscle. We tested the regulatory function of these filaments measuring the steady-state actin-activated ATPase of myosin in the presence and absence of calcium. The results are shown in Table I. The plus and minus rates were measured in separate experiments since plus rates measured after the minus rates were not as reproducible probably due to the changes in ATP levels and enzyme denaturation. Pure myosin presented a low ATPase rate which increased 10-fold by the addition of actin and tropomyosin (Table I). The addition of TNT/TNI to this system inhibits the ATPase irrespective of the calcium concentration, showing the expected inhibitory effect of TNI (Table I). The use of whole troponin complex reconstituted with wild type TNC from either chicken muscle or E. coli restores regulation to the system as expected (Table I). The substitution of the wild type TNC for the two mutants of Gly- 92 did not alter the properties of the system. Both the values of the ATPase rates and the ratio of the plus to minus rates are indistinguishable from those of the wild type TNC (Table I).

These assays were done under conditions optimized for regulation. A 2-fold molar excess of troponin/tropomyosin over the normal Tl:l, actin/troponin/tropomyosin molar ratio was used to ensure proper saturation of the actin filaments with the regulatory proteins. A low myosin concentration was used to avoid the cooperative effect observed a t high myosin to actin ratios. Under these conditions we did not observe differences between the mutant and the wild type TNC. It still remains possible that small changes in calcium sensitivity remained undetected.

DISCUSSION

A cDNA-encoding chicken skeletal muscle troponin C has been cloned and expressed in E. coli. After purification, the

Site-directed Mutants of Troponin C 2375

TABLE I Regulation of actin-activated ATPase of myosin by different TNC species

Myosin ATPase rates measured over a linear, 3-min period in different preparations containing myosin (M), actin (A), tropomyosin (Tmy), troponin T (TNT), troponin I (TNI), and the four troponin C (TNC) species. Wild type troponin C from chicken muscle (c), wild type troponin C from bacteria (b), the Pro-92 mutant (Pro), and the Ala-92 mutant (Ala) were used to reconstitute the troponin complex. Minus (-lo-' M free calcium) and plus (-lo-* M free calcium) rates were measured in separate experiments, but calcium rates were also determined after the minus calcium rates (plus Ca2+ rates at end). A single set of data from one of two independent experiments is presented (see "Materials and Methods" for detailed assay conditions).

Steady state rate

-10" M CaZ+ -lo-' M Ca2+ -10" M CaZ+ at end

M M + A + T m y M + A + T m y + T N T + T N I M + A + T m y + T N T + T N I + T N C ( b ) M + A + T m y + T N T + T N I + T N C ( c ) M + A + Tmy + TNT + TNI + TNC (Pro) M + A + Tmy + TNT + TNI + TNC (Ala)

protein produced in bacteria was shown to bind calcium with the same properties as the muscle protein and to replace muscle TNC in regulated thin filaments. The production of functional TNC from a cloned cDNA allows a detailed site- directed mutagenesis study of the structure and function of this molecule.

We also produced and analyzed two mutants designed to test the importance of rotations around Gly-92 in the interaction between the N- and C-terminal domains of troponin C.

E. coli Produced TNC-The cDNA used for expression of TNC encodes a protein sequence corresponding to the fast skeletal muscle troponin C (25). The unknown sequence of the N-terminal 4 residues was determined from the cDNA sequence. Two other amino acid substitutions were found. The asparagine for aspartic acid change in position 100 is probably due to a deamidation of the protein during sequencing. The isoleucine for threonine change in position 130 is most likely an allelic variation. The major difference between the primary sequence of E. coli and chicken muscle TNC resides in the N-terminal structure. The chicken protein has a blocked N-terminal residue and lacks the initiation methionine (25). In our expression system, due to the cleavage of the fusion protein with factor Xa, the N-terminal residue remains unblocked and a methionine residue was left in the recombinant protein. In the crystal structure of TNC (3) the N-terminal residues cannot be resolved and appear to make no contacts with the remaining of the molecule. Although the wild type troponin C produced in bacteria could not be distinguished from the chicken muscle molecule, these differences should be considered in future studies.

The Mutants of Gly-92"The biochemical evidence for conformational changes in the C-terminal domain induced by calcium binding to the N-terminal domain are based on fluorescence intensity changes of a probe linked to Cys-98 (10). Since this residue is located toward the C-terminal end of the nine-turn D/E helix, it seemed unlikely that a conformational change restricted to the vicinity of the calcium- specific sites could alter the environment of the probe (10). The presence of a glycine residue in the center of the helix, suggested that a rotation around Gly-92 would bring the two domains into a closer configuration (3). We tested this hypothesis constructing two mutant TNC with alanine and proline in position 92.

If the rotation hypothesis is correct, an impairment in TNC function is expected characterizing the essential role of Gly-

56 570

729 729 729 755 -

nmol/min/mg

104 230 247 248 200

93 680 755 657 768 ~. ~~ ~ . "

92 in providing the flexibility necessary for information transfer between domains. If other amino-acid residues can replace Gly-92 without large changes in TNC function, an essential role for Gly-92 must be rejected. Small changes in function in the mutants cannot be used as an argument for the rotation hypothesis since they could be the result of differences in packing of the troponin complex or to a number of other possibilities. Small differences could only be interpreted if the crystal structure of the troponin complex was known.

Both mutants were designed with the aim of removing the putative flexibility of the D/E helix. Only a small increase in the stability of the helix due to restrictions on dihedral angles can be expected in the alanine-92 mutant. An effect of this mutant on regulation was only expected if a large change in the bond angles around the glycine were necessary for proper functioning of the complex.

Predicting the structure of the D/E helix with a proline in position 92 is difficult. Although prolines are usually thought as being helix breakers, prolines have been found in a-helices (e.g. 27). To analyze whether a proline could be accommodated in position 92 without a large disturbance in the helix, the coordinates of the proline present in the center of the 9- residue a-helix of T-4 lysozyme and the two adjacent residues (Lys-87, Pro-88, Val-89) were taken from the Brookhaven data bank (26, 27), and used to model the central part of the helix (Lys-91, Gly-92, Lys-93) in TNC. Using the TNC coordinates (3), the structure of the T-4 lysozyme tripeptide was superimposed in the corresponding TNC structure using a molecular graphics system. The N-terminal bonds of the two structures were first aligned. Small rotations around the first and second bonds were necessary to accommodate the proline in the structure without any disruption of the helix. In the final model, thf residue closest to Pro-92 was the side chain of Glu-88j3.02A). The C-terminal bond of the tripeptide was only 0.46A away from the position of the corresponding bond in the original structure. This result shows that Pro-92 can be accommodated within the D/E helix without any major disruption of the structure.

Both mutants were found to have normal calcium binding properties, to reassemble into thin filaments, and to regulate the myosin ATPase. This result demonstrates that glycine 92 is not essential for the proper functioning of the thin filament and that the conformational changes introduced by the substitutions can be accommodated within the complex. These results suggest that the Pro-92 mutant is incorporated in the helix as shown by the model presented above and hence does


not introduce any gross modification in the protein structure. Since both the Ala and Pro mutants probably increase the rigidity of the central helix it is unlikely that large rotations around this part of the molecule are involved in the function of TNC. An alternative explanation is that the stability of the a-helix provided by the mutations is small when compared with the energy involved in the conformational change. A detailed study of mutant proteins, including the determination of the “on” and “off” calcium rates and the influence of calcium binding to the N-terminal sites in the environment of Cys-98 (10) is necessary to establish whether the mutations increased the rigidity of the helix.

The expression of functional troponin C in E. coli and the analysis of site-directed mutants offers a new way of studying the molecular mechanism of muscle regulation and structure/ function relationships in the TNC molecule.

Acknowledgments-We would like to thank Scott Putney for the rabbit cDNA, Hans Thorgersen for factor Xa, and Heinz Nika for the protein sequencing. The discussions with Osnat Herzberg, John Moult, John Kendrick-Jones, Kiyoshi Nagai, and Larry Smillie were greatly appreciated.

REFERENCES

1. Squire, J. (1981) The Structural Basis of Muscle Contraction,

2. Leavis, P. C., and Gergely, J. (1984) CRC Crit. Reu. Biochem. 16 ,

3. Herzberg, O., and James, M. N. G. (1985) Nature 313,653-659 4. Sundaralingam, M., Bergstrom, R., Strasburg, G., Rao, S. T.,

Raychowdhory, P., Greaser, M., and Wang, B. C. (1985) Science

5. Potter, J. D., and Gergely, J. (1975) J. Biol. Chem. 2 5 0 , 4628-

6. Robertson, S. P., Johnson, J. D., and Potter, J. D. (1981) Biophys.

Plenum Press, New York

235-305

227,945-948

4633

J. 34,559-569

7. Johnson, J. D., Charlton, S. C., and Potter, J. D. (1979) J. Biol.

8. Drakenberg, T., ForsBn, S., Thulin, E., and Vogel. H. J. (1987) J. Chem. 254,3497-3502

9.

10.

11.

12.

13.

14.

15. 16.

17.

18.

19. 20.

21.

22.

23. 24. 25. 26.

27.

28.

29.

. . Biol. Chtm. 262,672-678

and Gergely, J. (1981) J. Biol. Chem. 2 5 6 , 13121-13127

- .

Grabarek, Z., Drabikowski, W., Leavis, P. C., Rosenfeld, S . S.,

Grabarek, Z., Leavis, P. C., and Gergely, J. (1986) J. Biol. Chem.

Reinach, F. C., and Fischman, D. A. (1985) J. Mol. Biol. 181,

Putney, S., Herlihy, W. C., and Schimmel, P. (1983) Nature 302 ,

Farrell, P. J., Deininger, P., Bankier, A., and Barrell, B. G . (1983)

Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl.

Henikoff, S. (1984) Gene (Amst.) 2 8 , 351-359 Waye, M. H. Y., Verhoyen, M. E., Jones, P. T., and Winter, G.

Reinach, F. C., Nagai, K., and Kendrick-Jones, J. (1986) Nature

Carter, P., Bedoulle, H., and Winter, G. (1985) Nucleic Acid Res.

Nagai, K., and Thorgersen, H. C. (1984) Nature 309,810-812 Naeai. K.. Perutz. M. F.. and Povart. C. (1985) Proc. Natl. Acad.

26 1,608-613

411-422

718-720

Proc. Natl. Acad. Sci. U. S. A. 77, 2979-2983

Acad. Sci. U. S. A. 74,5463-5467

(1985) Nucleic Acid Res. 13,8561-8571

322,80-83

13,4431-4443

- , . I

gci.’U. S . A. 82, 725217255 Pardee. J. D.. and SDudich. J. A. (1982) Methods Enzvmol. 85.

164-i81 ’

. .

Reinach, F. C., Masaki, T., Shafiq, S., Obinata, T., and Fischman,

Potter, J. D. (1982) Methods Enzymol. 8 5 , 241-263 Smillie, L. B. (1982) Methods Enzymol. 85, 234-241 Wilkinson, J. M. (1976) FEBS Lett. 7 0 , 254-256 Bernstein, F. C., Koetzle, T. F., Williams, G . J. B., Meyer, E. F.,

Jr., Brice, M. D., Rodgers, J. R., Kennard, O., Shimanouchi, T., and Tasumi, M. (1977) J. Mol. Biol. 112, 535-542

Remington, S. J., Anderson, W. F., Owen, J., Ten Eyck, L. F., Grainger, C. T., and Mattews, B. W. (1978) J. Mol. Biol. 118,

Phillips, G. N., Jr., Fillers, J. P., and Cohen, C. (1986) J. Mol.

White, S. P., Cohen, C., and Phillips. G. N.. Jr. (1987) Nature

D. A. (1982) J. Cell Biol. 9 5 , 78-84

81-98

Biol. 192,111-131

325,826-828

Documents

Cloning, Expression, and Site-directed Mutagenesis of Chicken