Cardiac troponin C-L29Q, related to hypertrophic cardiomyopathy, hinders the transduction of the protein kinase A dependent phosphorylation signal from cardiac troponin I to C Anja Schmidtmann1, Christopher Lindow1, Sylvie Villard2, Arnd Heuser3, Andreas Mu¨ gge1, Reinhard Geßner3, Claude Granier2 and Kornelia Jaquet1
1 Herz- und Kreislaufzentrum der Ruhr-Universita¨ t Bochum ⁄ Bergmannsheil, Forschungslabor Molekulare Kardiologie, Bochum, Germany 2 Centre de Pharmacologie et Biotechnologie pour la sante´ , Faculte´ de Pharmacie, Montpellier, France 3 Charite´ , Campus Virchow-Klinikum, Institut fu¨ r Laboratoriumsmedizin und Pathobiochemie, Berlin, Germany
Keywords cTnC-L29Q; hypertrophic cardiomyopathy; human cardiac troponin I phosphorylation; peptide arrays
Correspondence K. Jaquet, Herz- und Kreislaufzentrum der Ruhr-Universita¨ t Bochum ⁄ Bergmannsheil, Forschungslabor Molekulare Kardiologie, St Josef-Hospital, Gudrunstr. 56, 44791 Bochum, Germany Fax: +49 234 509 2363 Tel: +49 234 509 2722 E-mail: kornelia.jaquet@ruhr-uni-bochum.de
(Received 24 June 2005, revised 21 Sep- tember 2005, accepted 3 October 2005)
doi:10.1111/j.1742-4658.2005.05001.x
We investigated structural and functional aspects of the first mutation in TNNC1, coding for the calcium-binding subunit (cTnC) of cardiac tropo- nin, which was detected in a patient with hypertrophic cardiomyopathy [Hoffmann B, Schmidt-Traub H, Perrot A, Osterziel KJ & Gessner R (2001) Hum Mut 17, 524]. This mutation leads to a leucine–glutamine exchange at position 29 in the nonfunctional calcium-binding site of cTnC. Interestingly, the mutation is located in a putative interaction site for the nonphosphorylated N-terminal arm of cardiac troponin I (cTnI) [Finley NL, Abbott MB, Abusamhadneh E, Gaponenko V, Dong W, Seabrook G, Howarth JW, Rana M, Solaro RJ, Cheung HC et al. (1999) FEBS Lett 453, 107–112]. According to peptide array experiments, the nonphosphoryl- ated cTnI arm interacts with cTnC around L29. This interaction is almost abolished by L29Q, as observed upon protein kinase A-dependent phos- phorylation of cTnI at serine 22 and serine 23 in wild-type troponin. With CD spectroscopy, minor changes are observed in the backbone of Ca2+- free and Ca2+-saturated cTnC upon the L29Q replacement. A small, but significant, reduction in calcium sensitivity was detected upon measuring the Ca2+-dependent actomyosin subfragment 1 (actoS1)-ATPase activity and the sliding velocity of thin filaments. The maximum actoS1-ATPase activity, but not the maximum sliding velocity, was significantly enhanced. In addition, we performed our investigations at different levels of protein kinase A-dependent phosphorylation of cTnI. The in vitro assays mainly showed that the Ca2+ sensitivity of the actoS1-ATPase activity, and the mean sliding velocity of thin filaments, were no longer affected by protein kinase A-dependent phosphorylation of cTnI owing to the L29Q exchange in cTnC. The findings imply a hindered transduction of the phosphoryla- tion signal from cTnI to cTnC.
In the last decade, several mutations have been identi- fied that cause familial hypertrophic cardiomyopathy (FHC). Most of these mutations are located in genes
encoding sarcomeric proteins, such as myosin heavy and light chains or myosin-binding protein C [1]. To a lesser extent, mutations have been detected in genes
Abbreviations ActoS1, actomyosin subfragment 1; C- ⁄ I-membrane, cellulose membrane with covalently linked peptides of human cTnC ⁄ human cTnI; cTn, cardiac troponin; cTnT, cTnI, cTnC, cardiac troponin T, I, C, respectively; nH, Hill coefficient; pCa50, pCa required for half-maximum increase in actoS1-ATPase activity or thin filament sliding velocity; PKA, protein kinase A; p.a., pro analysis; p.p.m., parts per million; wt, wild type.
FEBS Journal 272 (2005) 6087–6097 ª The Authors Journal compilation ª 2005 FEBS
6087
A. Schmidtmann et al.
Cardiac troponin C L29Q
Results
Structural consequences of L29Q
exchange
encoding cardiac troponin T (cTnT), the tropomyosin- binding subunit, or cardiac troponin I (cTnI), the inhibitory subunit of the heterotrimeric cardiac tropo- nin complex (cTn) [1,2]. Recently, the first mutation in the TNNC1 gene encoding the human calcium-binding subunit of cTn (cTnC) was detected in a patient suffer- ing from hypertrophic cardiomyopathy. This mutation, a T–A exchange at position 86 of exon 3, leads to a the nonpolar leucine 29 in human replacement of cTnC by a polar glutamine residue [3] in the nonfunc- tional calcium-binding loop I. It is believed that the heart-specific N-terminal extension of cTnI binds in this region [4–6].
cardiac
The L29Q substitution in human cTnC was verified at the DNA level by sequence analysis and at the protein level by mass spectrometry. The determined molecular mass of cTnC-L29Q (18 401 Da) is identical to the calculated from its amino acid molecular mass sequence and differs, by 15 Da, from the molecular mass of cTnC-wt (18 416 Da) owing to the leucine– illustrated). Ca2+- glutamine (data not dependent motility shifts in SDS gels do not show any difference between cTnC-wt and cTnC-L29Q (Fig. 1A). Both proteins exhibit a longer migration dis- tance in the Ca2+-bound state than in the Ca2+-free state (Fig. 1A). The comparisons of the CD spectra of cTnC-L29Q with those of cTnC-wt showed that the minima at 208 nm and 222 nm, characteristic for an
A
[8]. Thus,
B
these serine residues alters
cTn is anchored to the thin filament via cTnT and contraction via reversible Ca2+ regulates binding to cTnC [7,8]. At nanomolar intracellular Ca2+ concentrations, cTnC is Ca2+ free and cTnI binds to actin, thus preventing myosin crossbridge cycling. However, when intracellular Ca2+ is raised to micromolar concentrations, Ca2+ binds to cTnC, whereby cTnI is released from actin to allow cross- bridge cycling and force development. This regulation of contraction by calcium is further fine-tuned by phosphorylation of several proteins, a major mechan- ism to adjust the pumping activitity of the heart to for momentary hemodynamic demands example, myosin-binding protein C, the l-type cal- cium channel, phospholamban, and cTnI are phos- phorylated by protein kinase A (PKA; EC 2.7.1.37) upon b-adrenergic stimulation. cTnI is phosphorylated at serine 22 and serine 23, which are located in the heart-specific N-terminal extension of cTnI [9]. Phos- phate binding at the calcium-binding properties of cTnC and affects the actin–myosin interaction [8,10,11]. At the molecular level a movement of the N-terminal cTnI arm is observed [10–12].
the
study
structural
Here we
Fig. 1. Effects of the L29Q exchange in cardiac troponin C (cTnC) on its mobility in SDS gels (A) and on its secondary structure (B). (A) Ca2+-dependent mobility in SDS gels of cTnC-wt (lanes 2 and 3) and cTnC-L29Q (lanes 4 and 5) in the Ca2+-bound (lane 2+, 4+) and the Ca2+-free (lane 3–, 5–) state. Bovine cTnC was used as a stand- ard (lane 1). Proteins were preincubated in a loading buffer for SDS gels, which contained either 200 lM CaCl2 or 4 mM EGTA. (B) CD spectra of cTnC-wt (black lines) and cTnC-L29Q (red lines) in either the Ca2+-bound (continuous lines) or Ca2+-free (dashed lines) state. The protein concentration was 0.25 mgÆmL)1. Theta designates the molar ellipticity relative to the mean molecular weight (i.e. the molecular mass divided by the number of amino acids).
and functional consequences of the L29Q substitution in cTnC. Only minor effects on its secondary structure are observed using CD spectroscopy. In accordance, few differences in the cTnC–cTnI interaction are detected with peptide arrays. We confirm that the N-terminal arm of cTnI interacts with the N-terminal domain of the wild-type human cTnC (cTnC-wt). This inter- action is almost abolished upon bisphosphorylation of cTnI and does not occur when leucine 29 is exchanged for glutamine in cTnC. The study of reg- ulatory properties using actomyosin fragment S1 (actoS1)-ATPase activity and in vitro motility assays that phosphorylation ⁄ dephosphorylation of reveals cTnI no longer affects the Ca2+ sensitivity of these processes.
FEBS Journal 272 (2005) 6087–6097 ª The Authors Journal compilation ª 2005 FEBS
6088
A. Schmidtmann et al.
Cardiac troponin C L29Q
observations were made for the Ca2+-bound and the Ca2+-free state of cTnC-L29Q and cTnC-wt, respect- ively (Fig. 2B). Titration of wild-type and mutant
a-helix, were both reduced by (cid:1) 14% as a result of the L29Q replacement, indicating a small decrease in the a-helical content of (cid:1) 2% (Fig. 2B). These
B
Fig. 2. Cardiac troponin I (cTnI)–cardiac troponin C (cTnC) interaction studies using peptide arrays. Membranes, after incubation with the biot- inylated ligand, streptavidine, coupled to alkaline phosphatase and substrate, are shown on the left panel. Spot intensities were determined by scanning densitometry and are given as integrated density value (IDV). The intensities of selected spots corrected for their background are shown in a bar diagram on the right panel of (A) and (B). Numbers indicated are spot numbers. SS designates the cTnI-peptide (amino acids 14–32) with nonphosphorylated serine 22 and 23. S* designates a phosphoserine residue in the same peptide. In (A) an I membrane is shown after incubation with cTnC-wt. Blue spots indicate bound cTnC-wt. Spot 1 contains the first 20 amino acids of human cTnI. Spots 8 and 98–101 contain PAPIRRRSSNYRAYATEPH in varying phosphorylation states, as indicated. The amino acids PIRRRSSN (bold letters, sequence list) are involved in the binding to cTnC. Spots 98¢)101¢ show the spots after incubation of the I-membrane with cTnC-L29Q. In (B), part of a C membrane incubated with nonphosphorylated cTnI-wt is shown. Spot 75 contains the cTnC-wt peptide with an amino acid sequence, as shown on the right panel. Spot 76 contains the same peptide but with Q in position 29, as indicated.
FEBS Journal 272 (2005) 6087–6097 ª The Authors Journal compilation ª 2005 FEBS
6089
A. Schmidtmann et al.
Cardiac troponin C L29Q
cTnC with urea revealed that the structural stability was not altered by the amino acid exchange because both proteins denatured at 6 m urea (not shown).
Fig. 3. 31P-NMR spectrum of human cardiac troponin (cTn) recon- stituted with recombinant cTnT-wt, cTnI-wt and cTnC-L29Q. The troponin complex was incubated with 70 mU of the catalytic sub- unit of protein kinase A (PKA) at room temperature for 2 h, result- ing in a mixture of mono- and bisphosphorylated cTnI-containing complexes. A total of 7 mg of cTn, with either cTnC-wt or cTnC- L29Q, in 2 mM NH4HCO3, was concentrated in vacuo and solved in 700 lL of D2O with 20 mM Mops, pH 7.4, 150 mM KCl, 2 mM dithiothreitol and 200 lL of CaCl2. The pH was controlled before and after each measurement. The measurements resulted in nearly identical spectra for both complexes. Thus, only one spectrum is given exemplarily. To obtain this spectrum, 30 000 scans were accumulated. The signals designated as -SS*- and -S*S- result from complexes containing monophosphorylated cTnI with phospho- serine at position 23 or 22, respectively. The signal designated as -S*S*- results from cardiac troponin containing bisphosphorylated cTnI with phosphate bound to serine 22 and serine 23.
[AAFDIFVLGAEDGCISTKEL(22–41)]
31P-NMR measurements cTn complexes used for (Fig. 3) contained a mixture of about 50% mono- and 50% bisphosphosphorylated cTnI, and nearly identical spectra were obtained. Thus, only one spectrum is shown in Fig. 3. Both 31P-NMR spectra showed a reso- nance at 4.99 and 4.87 p.p.m., corresponding to cTnI monophosphorylated at serine 22 and serine 23, respectively. The 31P-NMR signal at 4.65 p.p.m. results from bisphosphorylated cTnI and is probably caused by an interaction of the bisphosphorylated N-terminal arm with groups located in the holotroponin complex. An interaction probably occurs with cTnI itself, as des- cribed previously, and is not affected by the L29Q exchange in cTnC.
spot 76 was
As a result of the minor changes in the backbone structure of cTnC-L29Q, we also expected minor loc- ally restricted changes in the interaction with cTnC. To study the interaction of cTnI with cTnC, we used peptide arrays. First, a membrane, containing cova- lently linked cTnI peptides (I-membrane) was incuba- ted with biotinylated cTnC-wt (Fig. 2A). As expected, several spots (each spot contained a specific cTnI pep- tide) reacted with cTnC-wt. All known interaction sites between cTnC and cTnI were detected upon analysis of the sequence of the reacting spots. In addition, an interaction of the N-terminal arm of cTnI with cTnC- wt was observed (Fig. 2A, spots 3–9; list of the amino acid sequences). Within these peptides, amino acid resi- dues 17–24 (PIRRRSSN) were found to be involved in the interaction with cTnC-wt. This section, which con- tains the two PKA-dependent phosphorylation sites, serine 22 and serine 23, still interacted with cTnC when one phosphoserine residue, either in position 22 or 23, was present (Fig. 2A, spots 98–100). Even after subtracting background intensities, the intensities of spots 8 and 98–100 were very similar (Fig. 2A, right panel). However, when phosphate was bound to both serine residues, the interaction was nearly abolished (Fig. 2A, left and right panel: spot 101). After regener- ation, the I-membrane was incubated with cTnC- L29Q, which resulted in a very similar interaction pattern (data not shown). However, in contrast, the N-terminal peptide of cTnI did not show an inter- action with cTnC-L29Q, regardless of whether it was in the phosphorylated or nonphosphorylated state (Fig. 2A, spots 98¢)101¢). These data were verified by an additional set of experiments using the same mem- brane, a new I-membrane and also a membrane with covalently bound peptides of cTnC-wt and mutant cTnC (C-membrane). This latter membrane was incu- bated with nonphosphorylated cTnI (Fig. 2B). Spot 75 of the C-membrane, which contains the wild-type pep- tide reacted with nonphosphorylated cTnI. In contrast, the reaction of considerably weaker by (cid:1) 50% (Fig. 2B). Spot 76 contains the same peptide, but with a glutamine instead of a leucine in position 29.
Functional consequences
To investigate the consequences of
The reconstituted human holotroponin complexes containing cTnT, bisphosphorylated cTnI and either cTnC-wt or cTnC-L29Q were reconstituted with actin and tropomyosin to form thin filaments. For the gen- eration of dephosphorylated complexes, the thin fila- ments were incubated with protein phosphatase 2A
this altered cTnI–cTnC interaction on the bisphosphorylated N-terminus of cTnI and on function, human holotrop- onin complexes were reconstituted from recombinant cTnT, cTnI and either cTnC-wt or cTnC-L29Q, and phosphorylated by PKA. The phosphorylation degree was routinely checked by IEF (data not shown). Both
FEBS Journal 272 (2005) 6087–6097 ª The Authors Journal compilation ª 2005 FEBS
6090
A. Schmidtmann et al.
Cardiac troponin C L29Q
(cid:1) 50%, compared to that obtained with cTnC-wt (P ¼ 0.002; Table 1; Fig. 4A,C). Upon bisphosphory- lation of cTnI, the maximum actoS1-ATPase activity obtained with both cTnCs decreased to almost identi- cal levels. This decrease in maximum actoS1-ATPase activity was statistically significant for filaments with cTnC-L29Q (P < 0.0001, comparing phosphorylated and dephosphorylated filaments), but not for filaments with cTnC-wt (P ¼ 0.2).
(EC 3.1.3.16). cTnI was completely dephosphorylated, as demonstrated by ELISA (data not shown). The Ca2+-dependent actoS1-ATPase activation was used to investigate changes in the regulatory properties of these thin filaments that were dependent on the L29Q exchange in cTnC and on the phosphorylation state of cTnI. Sigmoidal curves were obtained for normalized (Fig. 4A,C) and relative (Fig. 4B,D) actoS1-ATPase thin filaments containing cTnC-wt or activities of cTnC-L29Q. At low [Ca2+], normalized actoS1-ATP- ase activities obtained with the dephosphorylated and phosphorylated forms of wild-type and mutant cTn did not differ significantly (Fig. 4A,C; Table 1). At high [Ca2+], however, the maximum actoS1-ATPase activity obtained with dephosphorylated thin filaments containing cTnC-L29Q was significantly enhanced, by
Figure 4B,D shows relative actoS1-ATPase activities and summarizes the Ca2+ sensitivity and steepness of the Ca2+ activation of the actoS1-ATPase activity. In the L29Q filaments with dephosphorylated cTnI, exchange in cTnC led to a small rightward shift of the activity–pCa relationship, resulting in a slightly, but significantly, reduced pCa required for a half-maximum
B
A
C
D
Fig. 4. Ca2+-dependent activation of the actomyosin subfragment 1 (actoS1)-ATPase in reconstituted thin filaments with human cardiac trop- onin C (cTnC)-wt or cTnC-L29Q and either phosphorylated or dephosphorylated cardiac troponin I (cTnI). The Ca2+-dependent actoS1-ATPase activity is shown of regulated thin filaments containing cTnC-wt (A, B) or cTnC-L29Q (C, D) in either the nonphosphorylated (grey circles) or the bisphosphorylated (black triangles) state. The actoS1-ATPase activities (mean ± SEM) were normalized to the actoS1-ATPase activity obtained with unregulated filamentous actin under corresponding conditions (A, C). The curves represent nonlinear least square fits to the data employing the Hill equation. In (B) and (D), relative actoS1-ATPase activities were obtained by normalizing the maximum normalized value to 1 and the minimum value to 0. Values of normalized minimum and maximum levels of actoS1-ATPase activities, of the pCa required for half-maximum increase in actoS1-ATPase activity or thin filament sliding velocity (pCa50), and of the Hill coefficient (nH), are summarized in Table 1.
FEBS Journal 272 (2005) 6087–6097 ª The Authors Journal compilation ª 2005 FEBS
6091
A. Schmidtmann et al.
Cardiac troponin C L29Q
Table 1. Summary of parameters from in vitro experiments. Data are expressed as the mean value ± SEM of n experiments using three dif- ferent preparations of cardiac troponin (cTn). Minimum and maximum levels (Y0, Ymax) of the actomyosin subfragment 1 (actoS1)-ATPase activity (ATPase) and the thin filament in vitro motility (IVMA) are expressed as normalized values. nH is the Hill coefficient and pCa50 the pCa required for a half-maximum increase in actoS1-ATPase activity or thin filament sliding velocity. de and PP refer to thin filament dephos- phorylated and phosphorylated at cTnI-serine 22 ⁄ 23, respectively, containing either wild-type cTnC (wt) or cTnC-L29Q (L29Q).
n
Y0
Ymax
nH
pCa50
ATPase wt-de wt-PP L29Q-de L29Q-PP
0.06 ± 0.06 0.01 ± 0.05 0.11 ± 0.06 0.02 ± 0.04
1.19 ± 0.1 0.85 ± 0.24 1.75 ± 0.09 0.94 ± 0.11
1.9 ± 0.2 1.0 ± 0.1 1.8 ± 0.1 1.6 ± 0.2
6.25 ± 0.02 5.73 ± 0.11 6.12 ± 0.02 6.08 ± 0.03
5 5 10 8
IVMA
wt-de wt-PP L29Q-de L29Q-PP
0.31 ± 0.09 0.34 ± 0.004 0.50 ± 0.01 0.41 ± 0.06
1.76 ± 0.12 1.63 ± 0.01 1.87 ± 0.02 1.89 ± 0.13
1.1 ± 0.1 1.4 ± 0.02 1.4 ± 0.1 1.0 ± 0.1
6.18 ± 0.04 5.76 ± 0.02 6.05 ± 0.01 6.08 ± 0.05
4 4 5 5
by (cid:1) 0.1 pCa units, from 6.18 ± 0.04 to 6.05 ± 0.01 (P ¼ 0.02), as a result of the L29Q exchange. The pCa50 value was not changed upon bisphosphorylation of cTnI using filaments containing mutant cTnC (Table 1), whereas the familiar rightward shift of the pCa curve occurred using filaments containing cTnC- wt. Here, pCa50 values were significantly reduced from 6.18 ± 0.04 to 5.76 ± 0.02 (Fig. 5B; Table 1, P ¼ 0.0002). These values were in good agreement with those obtained from Ca2+-dependent actoS1-ATPase activity measurements.
Discussion
increase in the actoS1-ATPase activity or thin filament sliding velocity (pCa50) value compared with the pCa50 (P ¼ 0.0013; Table 1; value obtained with cTnC-wt Fig. 4B,D). Upon bisphosphorylation of cTnI, the pCa curve of filaments containing cTnC-wt shifted towards lower pCa values, as expected (Fig. 4B). Thus, the pCa50 value was decreased significantly (P ¼ 0.0016), by (cid:1) 0.5 pCa, from pCa50 6.25 ± 0.02 (dephosphory- lated cTnI) to pCa50 5.73 ± 0.11 (bisphosphorylated cTnI). However, bisphosphorylation of cTnI in fila- ments with cTnC-L29Q did not induce a rightward shift of the pCa curve (Fig. 4D), as nearly identical pCa50 values of (cid:1) 6.1 were obtained with either bisphosphorylated or dephosphorylated cTnI (Table 1). All Hill coefficients (nH) ranged between 1 and 2, indi- cating co-operativity (Table 1).
amino
(Fig. 5A,C; Table 1). The
Leucine 29 of human cTnC is located at the transition of helix A to the nonfunctional Ca2+-binding loop I. According to Sia et al. [13] L29 is important for the integrity of helix A and thus of the pseudo Ca2+-bind- ing site I. As revealed by CD spectroscopy, only minor changes in the backbone structure are observed in cTnC-L29Q, leading to a loss in helicity of (cid:1) 2%. This indicates a locally restricted helix destabilization. It is tempting to assume that helix A, which directly pre- cedes leucine 29, is unwound by one to two turns owing to the leucine–glutamine substitution. This destabilization of helix A might affect the Ca2+-bind- ing properties of the adjacent regulatory Ca2+-binding site II and, more directly, might affect the interaction of the nonphosphorylated N-terminal arm of cTnI (amino acids 1–32) with cTnC [4,6,14,15]. In the wild- type cTnI–cTnC complex, this interaction occurs with residues adjacent to L29 [4,14] and is abolished upon phosphorylation of cTnI by PKA [4,6,14–17]. Accord- ing to Ward et al. [6], amino acid residues 16–29 of nonphosphorylated cTnI bind to the cTnC N-lobe. This finding is in good agreement with results obtained
Similar results were obtained from studies of the mechanical output of the interaction between thin filaments and myosin using in vitro motility assays (Fig. 5). The mean sliding velocities of the reconstitu- ted thin filaments with phosphorylated or dephosphor- ylated cTnI, and either cTnC-wt or cTnC-L29Q, were not significantly different at inhibitory Ca2+ concen- trations acid exchange in cTnC led to a slight enhancement of the maximum mean sliding velocity for dephosphorylated filaments, but this was not as pronounced as observed for maximum actoS1-ATPase activities. The maximum mean sliding velocity of filaments with cTnC-wt seemed to be lowered upon cTnI bisphosphorylation; however, the difference was not significant (Fig. 5A). No reduction in the maximum mean sliding velocity was observed with cTnC-L29Q upon bisphosphoryla- tion, both maximum mean sliding velocities being nearly identical (Table 1; Fig. 5C). As summarized in Fig. 5B,D, the pCa50 value was only slightly reduced,
FEBS Journal 272 (2005) 6087–6097 ª The Authors Journal compilation ª 2005 FEBS
6092
A. Schmidtmann et al.
Cardiac troponin C L29Q
A
B
C
D
Fig. 5. Ca2+-dependent in vitro motiliy of regulated thin filaments with human cardiac troponin C (cTnC)-wt or cTnC-L29Q and either phos- phorylated or dephosphorylated cTnI. Ca2+-dependent in vitro motility is shown of regulated thin filaments containing cTnC-wt (A, B) or cTnC-L29Q (C, D) in either the nonphosphorylated (grey circles) or the bisphosphorylated (black triangles) state. Mean sliding velocities of thin filaments (mean ± SEM) were normalized to that of phalloidin-tetramethylrhodamine-5-(6)isothiocyanate (TRITC)-labelled filamentous actin under corresponding conditions (A, C). The Ca2+ dependence of the relative mean sliding velocities is given in (B) and (D). Also see the legend of Fig. 4. Values of normalized minimum and maximum levels of mean sliding velocities, of the pCa required for half-maximum increase in actoS1-ATPase activity or thin filament sliding velocity (pCa50) and of the Hill coefficient (nH) are compiled in Table 1.
itself
using peptide array experiments presented in this report. They show that amino acids 17–24 (PIR- RRSSN) of cTnI interact with amino acids 24–41 (FDIFVLGAEDGCISTKEL) of cTnC. This interac- tion is nearly abolished upon bisphosphorylation of cTnI at serine 22 and serine 23. Furthermore, previous 31P-NMR measurements indicated that an interaction of the bisphosphorylated N terminus occurs within the cardiac troponin complex [12,17]. An interaction of the bisphosphorylated N-teminal arm with cTnC can be excluded, but may occur within cTnI [17]. This interaction of the bisphosphorylated N terminus of cTnI is not affected by the L29Q replacement in cTnC, as no changes in 31P-NMR resonances were observed.
The release of
the bisphosphorylated cTnI arm from cTnC upon bisphosphorylation of cTnI contri- butes to the reduced calcium affinity of cTnC and
thus to a reduced calcium sensitivity of the actin– myosin interaction [4,5,10,12,14,16,17]. Therefore, the hindered interaction of the N-terminal arm of non- phosphorylated cTnI with cTnC-L29Q should induce similar effects on the calcium sensitivity of the acto- myosin interaction, as does cTnI bisphosphorylation. Indeed, both in vitro experiments presented here revealed a small Ca2+ desensitization of (cid:1) 0.1 pCa units as a result of the L29Q replacement in cTnC, when using dephosphorylated thin filaments. How- ever, this shift is much smaller than that obtained when cTnI is bisphosphorylated in filaments contain- ing cTnC-wt. Hereby, a rightward shift of nearly 0.5 pCa units was observed. Thus, perhaps the inter- action of the N terminus of cTnI with cTnC is not completely abolished owing to the L29Q exchange in cTnC and ⁄ or the reversible binding of the N-terminal cTnI arm to cTnC is not solely responsible for the
FEBS Journal 272 (2005) 6087–6097 ª The Authors Journal compilation ª 2005 FEBS
6093
A. Schmidtmann et al.
Cardiac troponin C L29Q
cTnI-K206Q decreased the Bisphosphorylation of maximum activity, but to a level still higher than that obtained with bisphosphorylated cTnI-wt. Again, the Ca2+ sensitivity remained unchanged after bisphos- phorylation of cTnI-K206Q [20].
implies
In summary, the results indicate that besides alter- ing the dynamics of the actin–myosin interaction, the mutation mainly impairs the PKA-dependent signal- ling from TnI to TnC, leading to an elevated Ca2+ sensitivity under conditions when cTnI is phosphoryl- ated.
Experimental procedures
Site-directed mutagenesis
described, were
other
than
A T fi A exchange, leading to the L29Q mutation of human cardiac troponin C, was introduced by using the Quick-ChangeTM site-directed mutagenesis kit (Stratagene, Amsterdam, the Netherlands) and the primers 5¢-CATCTT CGTGCAGGGCGCTGAGGATGGCTG and 5¢)AGCGC CCTGCACGAAGATGTCGAAGGCTGC. The inserts of the wild type and resulting L29Q mutant vector (pET3c) were controlled by automatic DNA sequence analysis [19,20].
Expression of proteins
altered calcium response. The first explanation seems not to be likely because no difference in the inter- action of the N-terminal cTnI arm in either phos- phorylation state to cTnC-L29Q was detected in peptide array experiments. In addition, bisphosphory- lation of cTnI did not further reduce the calcium sen- the actin–myosin interaction when thin sitivity of filaments contained cTnC-L29Q. This that other parts of troponin might also contribute to the transduction of the phosphorylation signal and might be affected by the L29Q exchange in cTnC. However, the interaction site of the bisphosphorylated cTnI arm within cardiac troponin is not affected by the shown by 31P-NMR amino acid replacement, as measurements. According to peptide array experi- ments, reproducible changes in the cTnI–cTnC inter- not those action, identified. To detect such changes in the interaction between cTnI and cTnC, different methods, for exam- ple NMR spectroscopy, may have to be applied to study interactions within an intact troponin complex. Besides reducing the calcium affinity of cTnC, the L29Q replacement in cTnC induces an additional effect. The maximum actoS1-ATPase activity was enhanced by (cid:1) 30%, indicating an altered dynamics in the actin–myosin interaction. This effect was reversed upon bisphosphorylation of cTnI. Nearly identical maximum values were observed with mutant cTnC, as with TnC-wt.
To express the human subunits of cardiac troponin (T, I, C-wt and C-L29Q), Escherichia coli BL21 (DE3) (Strata- gene), transformed with the respective vectors, were cul- tured in 2 L of NZCYM-Medium (Difco Laboratories, Livonia, MI, USA) at 37 (cid:1)C containing 1 mgÆmL)1 of either kanamycin (cTnT) or ampicillin (cTnI, cTnC-wt and cTnC-L29Q). At an attenuance (D) of 600 nm in the range of 0.7–1.0, protein expression was induced by adding 0.4 mm isopropyl thio-b-d-galactoside (final concentration; AppliChem, Darmstadt, Germany). Cells were harvested after 2–3 h by centrifugation (8300 g, 4 (cid:1)C, 15 min), and the cell pellets were stored at )20 (cid:1)C.
Protein purification
As an increased actomyosin-ATPase rate can be explained by an increased crossbridge detachment rate [18], the maximum unloaded sliding velocity should also be increased [19,20]. Only a slight enhancement of the maximum sliding velocity for cTnC-L29Q was observed, however, which was not reduced upon PKA- dependent cTnI phosphorylation, in contrast to the maximum actoS1-ATPase activity. The reason why only maximum actoS1-ATPase activity is clearly affec- ted upon cTnI phosphorylation remains unclear, as in both assays the same batch of thin filaments was used, and in previous studies identical effects of cTnI phos- phorylation on the actoS1-ATPase activity and thin fil- ament sliding velocity were obtained [19,20]. No effect of cTnC-L29Q was observed at inhibitory [Ca2+] in any experiments.
The recombinant subunits of cTn-wt were isolated from E. coli according to standard protocols [19–22]. cTnC- L29Q was purified as cTnC-wt. Actin was isolated from rabbit skeletal muscle [23]. Polymerization of globular monomeric actin (60–90 lm) was induced by adding 5 mm triethanolamine, 100 mm KCl, 1 mm MgCl2, 0.2 mm CaCl2, 0.5 mm ATP, pH 7.5, at 4 (cid:1)C. Tropomyosin and myosin were isolated from rabbit skeletal muscle [24,25]. Myosin S1 fragments were obtained by limited proteolysis with papain [25] and stored at )80 (cid:1)C. Protein purity was con- trolled by SDS ⁄ PAGE.
In several, but not all, aspects, the L29Q exchange modifies the actoS1-ATPase activity and mean sliding velocity of thin filaments, as previously reported for other mutations in cardiac troponin subunits known to cause familial hypertrophic cardiomyopathy. For example, the well-characterized cTnI-K206Q exchange also shows an enhanced maximum actoS1-ATPase activity and nearly no effect on Ca2+ sensitivity [20].
FEBS Journal 272 (2005) 6087–6097 ª The Authors Journal compilation ª 2005 FEBS
6094
A. Schmidtmann et al.
Cardiac troponin C L29Q
Reconstitution of the troponin complex
Reconstitution of cardiac troponin was performed by mix- ing equimolar amounts of each subunit, as described previ- ously [19]. The composition of the complex was checked by SDS ⁄ PAGE. Native bovine cardiac troponin was used as a reference. The intensities of SDS bands were analysed using the gel documentation system, AlphaDigiDocTM (Biozym Diagnostik, Hess Oldendorf, Germany) and the alpha- easefc software, version 3.2.3 (AlphaInnotech, San Lean- dro, CA, USA).
Reconstitution of thin filaments
membrane, as described previously [31,32], with a frame- shift of two amino acids. Peptide synthesis was performed from the C terminus to the N terminus, whereby the carb- oxy group of the C-terminal amino acid of each peptide forms an ester bond with an OH-group of the cellulose. In addition, cTnI peptides (amino acids 15–32) containing nonphosphorylated serine 22 and serine 23, a phospho- serine in either position 22 or 23, or both serine residues phosphorylated, were synthesized onto the membrane con- taining cTnI peptides (I-membrane, spots 98–101). Addi- tional cTnC peptides (amino acids 22–41) with either the original sequence or the L29Q exchange, were synthesized onto the membrane with cTnC peptides (C-membrane, spots 75 and 76).
(TRITC) The membranes were treated mainly as described by Ferrieres et al. [31]. After incubation with blocking buffer SU-07–250 (Sigma Genosys, Sigma Aldrich Inc., Seelze, the I-mem- Germany) and several washing procedures, brane was incubated for 2 h at 37 (cid:1)C, with either biotinyl- ated cTnC-wt or cTnC-L29Q and the C-membrane with biotinylated nonphosphorylated cTnI [50 nm protein in NaCl ⁄ Tris (TBS)-Tween-saccharose]. Biotinylation was performed with a biotindisulfide-N-hydroxysuccinimide ester (Sigma-Aldrich, Seelze, Germany) according to the manufacturers’ manual.
To reconstitute thin filaments, troponin complexes contain- ing bisphosphorylated cTnI were used [26]. cTnI was bis- phosphorylated at S22 and S23 in the troponin complex with the catalytic subunit of PKA [27]. The degree of phos- phorylation was controlled by IEF [28]. Thin filaments were reconstituted from rabbit skeletal muscle actin, rabbit skel- etal muscle tropomyosin and cTn, as described previously [19]. For in vitro motility assays, regulated and unregulated filamentous actin was labelled with phalloidin-tetramethyl- rhodamine-5-(6)isothiocyanate (Fluka, Buchs, Germany) [19]. Half of the reconstituted thin filaments were dephosphorylated by incubation with protein phosphatase 2A, isolated from bovine heart [29]. The degree of phos- phorylation of cTnI present in thin filaments was controlled by ELISA, using mAbs specifically recognizing either non- phosphorylated or phosphorylated human cTnI [20]. (MTT;
Structure: spectroscopy
Italy;
31P-NMR spectra were recorded on a 400 MHz (1H res-
CD spectra [20] and 31P-NMR spectra [12] were recorded as described previously. CD spectra of human cTnC-wt and cTnC-L29Q were measured at 20 (cid:1)C on a Jasco 710 spec- tropolarimeter software (Jasco Europe, Gemella, spectra management V1.51.00) within a range of 190– 250 nm, with a bandwidth of 1 nm and 10 accumulations each, and a speed of 50 nmÆmin)1. The path length of the cuvette was 0.05 cm. Measurements were performed twice, using proteins from the same batch. The spectra were smoothed using the Savitsky–Golay function (Prism; GraphPad Software, Inc., San Diego, CA, USA). The con- tent of a-helix (in percentage) was calculated using the somcd software [30] (available online: http://geneura.ugr.es/ cgi-bin/somcd/index.cgi).
After washing the membranes with TBS (three washes, each of 3 min duration), peptide–protein complexes were detected using 0.2 lgÆmL)1 streptavidine coupled to alka- line phosphatase (EC 3.1.3.2) (Sigma Aldrich; 2 h at 37 (cid:1)C) and, as substrate solution, 50 lL of 45 mm methyl- Sigma- thiazolyl-diphenyl-tetrazolium bromide Aldrich) in 70% (v ⁄ v) dimethylformamide, 50 lL of 1 m MgCl2, 50 lL of 45 mm 5-bromo-4-chloro-3-inodyl phos- phate (BCIP) (Sigma-Aldrich) and 50 mL of TBS-Tween. At peptide spots with bound biotinylated protein, a blue precipitate was observed. The reaction was stopped by washing with TBS-Tween. For regeneration of the mem- branes, the blue precipitate was washed out by incubation with 100% dimethylformamide. Then, the membranes were incubated with 8 m urea, 1% (v ⁄ v) SDS, 2 mm b-merca- ptoethanol (3 · 10 min), subsequently with TBS (two incu- bations, each of 2 min duration), 50% (v ⁄ v) ethanol (P.A.H), 10% (v ⁄ v) acetic acid (three incubations, each of 10 min duration), then with TBS (two incubations, each of 2 min duration) and finally with 100% (v ⁄ v) methanol (three incubations, each of 10 min duration). This proce- dure was repeated until nearly no reaction occurred with streptavidine coupled to alkaline phosphatase. Regenerated membranes could be reused for further assays. onance) Bruker spectrometer at room temperature.
Peptide arrays
FEBS Journal 272 (2005) 6087–6097 ª The Authors Journal compilation ª 2005 FEBS
6095
To analyse the interaction, membranes were scanned and the spot intensities were determined using the alphaeasefc software (AlphaInnotech). Intensities were corrected for background intensities of spots before incubation with the ligands. 20-Mer peptides, comprising the amino acid sequence of either cTnI or cTnI, were synthesized on a cellulose
A. Schmidtmann et al.
Cardiac troponin C L29Q
Function: actoS1-ATPase activity
Herzstiftung, FoRUM DAAD and the Deutsche Fors- chungsgemeinschaft Mu662-2.
References
1 Richard P, Charron P, Carrier L, Ledeuil C, Cheav T,
Pichereau C, Benaiche A, Isnard R, Dubourg O, Burban M et al. (2003) Hypertrophic cardiomyopathy. Distribution of disease genes, spectrum of mutations, and implications for a molecular diagnosis strategy. Circulation 107, 2227–2232.
2 Gomes AV & Potter JD (2004) Molecular and cellular aspects of troponin cardiomyopathies. Ann N Y Acad Sci 1015, 214–224.
Measurements were performed according to Deng et al. [19,20] using the purine nucleoside phosphorylase (EC 2.4.2.1) ⁄ 2-amino-6-mercapto-7-methylpurine-riboside system, according to Webb [33]. The absorbance change at 360 nm was recorded for 3 min. The actoS1-ATPase activity was calculated as described previously [19,20]. As controls, the basal activity was recorded by using distilled water instead of thin filaments and the maximum activity (100%) by using unregulated actin filaments. Data analysis was per- formed as described by Deng et al. [19,20]. Briefly, the actoS1-ATPase activity was normalized to the reference (activity obtained with unregulated actin filaments). The mean values and standard errors were calculated (prism Software version 3.0; GraphPad Inc.) [19,20]. Data were fit equation (sigma plot 2001; SPSS Inc., to the Hill Mu¨ nchen, Germany): 3 Hoffmann B, Schmidt-Traub H, Perrot A, Osterziel KJ & Gessner R (2001) First mutation in cardiac troponin C, L29Q, in a patient with hypertrophic cardiomyopa- thy. Hum Mut 17, 524.
ðEqn 1Þ; y(pCa) ¼ a þ b 1 þ 10nH pCapCa50
4 Abbott MB, Dong WJ, Dvoretsky A, Dague B, Caprioli RM, Cheung HC & Rosevear PR (2001) Modulation of cardiac troponin C–cardiac troponin I regulatory inter- actions by the amino terminus of cardiac troponin I. Biochemistry 40, 5992–6001.
with pCa50, the pCa at half-maximum activation; nH, the level; and b the amplitude Hill-coefficient; a, the basal the Ca2+-dependent response. Relative values were of obtained by normalizing the maximum normalized value to 1 and the minimum to 0 and were fit to the equation:
ðEqn 2Þ; y(pCa) ¼ 1 1 þ 10nH pCapCa50 5 Heller WT, Finley NL, Dong WJ, Timmins P, Cheung HC, Rosevear PR & Trewhalla J (2003) Small-angle neutron scattering with contrast variation reveals spatial relationships between the three subunits in the ternary cardiac troponin complex and the effects of troponin I phosphorylation. Biochemistry 42, 7790–7800. 6 Ward DG, Brewer SM, Cornes MP & Trayer IP (2003)
To compare the mean ± SEM values, a two-way Student’s t-test was used. Values of P < 0.05 were considered as sig- nificant. Cross-linking study of the N-terminal extension of human cardiac troponin I. Biochemistry 42, 10324– 10332.
In vitro motility
7 Gomes AV, Potter JD & Szczesna-Cordary D (2002) The role of troponin in muscle contraction. IUBMB Life 54, 323–333. 8 Kobayashi T & Solaro RJ (2005) Calcium, thin fila-
ments, and the integrative biology of cardiac contracti- lity. Annu Rev Physiol 67, 39–67.
The assays were performed as described previously [19,20]. The movement of unregulated (control) and regulated fila- ments on myosin molecules fixed on a slide was recorded by a video camera at different [Ca2+]. Data were digitalized and analysed using the software retrac (Dr Nick Carter, MCRI, Oxted, UK) [19,20]. Data analysis was performed as described above. 9 Mittmann K, Jaquet K & Heilmeyer LMG (1990) A common motif of two adjacent phospho serines in bovine, rabbit and human cardiac troponin I. FEBS Lett 302, 133–137.
Acknowledgements
10 Layland J, Solaro RJ & Shah AM (2005) Regulation of cardiac contractile function by troponin I phosphoryla- tion. Cardiovasc Res 66, 12–21.
11 Metzger JM & Westfall MV (2004) Covalent and non- covalent modification of thin filament action. Circ Res 94, 146–158.
12 Jaquet K, Korte K, Schnackerz K & Heilmeyer LMG Jr (1993) Characterization of the cardiac troponin I phosphorylation domain by 31P nuclear magnetic reso- nance spectroscopy. Biochemistry 32, 13873–13878. 13 Sia SK, Li MX, Spyracopolous L, Gagne SM, Liu W, Putkey JA & Sykes BD (1997) Structure of cardiac
We thank Prof. Wegner, Dr Deng Yi and Prof. Her- berg for the gift of actin, myosin-S1 and the catalytic subunit of PKA, respectively. We further thank Dr Kuhlmann for providing the CD-spectrometer, Dr van den Boom and Sebastian Kruse for helpful discussions concerning in vitro motility assays and evaluation of CD-data, respectively. We are grateful to Dr A. Perrot and Prof. Osterziel for information on the cTnC-L29Q mutant. This work was supported by the Deutsche
FEBS Journal 272 (2005) 6087–6097 ª The Authors Journal compilation ª 2005 FEBS
6096
A. Schmidtmann et al.
Cardiac troponin C L29Q
of the barbed ends of actin filaments. Biochemistry 37, 7532–7538. muscle troponin C unexpectedly reveals a coiled regula- tory domain. J Biol Chem 272, 18216–18221. 24 Smillie LB (1992) Preparation and identification of 14 Finley NL, Abbott MB, Abusamhadneh E, Gaponenko
alpha- and beta-tropomyosins. Methods Enzymol 85, 234–241.
V, Dong W, Seabrook G, Howarth JW, Rana M, Solaro RJ, Cheung HC et al. (1999) NMR analysis of cardiac troponin C–troponin I complexes: effects of phosphorylation. FEBS Lett 453, 107–112. 25 Margossian SS & Lowey S (1982) Preparation of myo- sin and its subfragments from rabbit skeletal muscle. Methods Enzymol 85, 55–71.
15 Ward DG, Brewer SM, Gallon CE, Gao Y, Levine BA & Trayer IP (2004) NMR and mutagenesis studies on the phosphorylation region of human cardiac troponin I. Biochemistry 43, 5772–5781. 16 Sakthivel S, Finley NL, Rosevear PR, Lorenz JN, 26 Reiffert SU, Jaquet K, Heilmeyer LMG, Ritchie MD & Geeves MA (1996) Bisphosphorylation of cardiac tropo- nin I modulates the Ca2+-dependent binding of myosin subfragment S1 to reconstituted thin filaments. FEBS Lett 384, 43–47.
Gulick J, Kim S, Vanburen P, Martin LA & Robbins J (2005) In vivo and in vitro analysis of cardiac troponin I phosphorylation. J Biol Chem 280, 703–714. 17 Schmidtmann A, Lohmann K & Jaquet K (2002) The
interaction of the bisphosphorylated N-terminal arm of cardiac troponin I-A 31P-NMR study. FEBS Lett 513, 289–293.
27 Herberg FW, Bell SU & Taylor SS (1993) Expression of the catalytic subunit of the cAMP dependent protein kinase in Escherichia coli: multiple isozymes reflect dif- ferent phosphorylation states. Protein Eng 6, 771–777. 28 Ardelt P, Dorka P, Jaquet K, Heilmeyer LMG, Ko¨ rtke H, Ko¨ rfer R & Notohamiprodjo G (1998) Microanaly- sis and distribution of cardiac troponin I phospho spe- cies in heart areas. Biol Chem 379, 341–347. 29 Mumby SC, Russell KL, Garrad LJ & Green DD 18 Homsher E, Lee DM, Morris C, Pavlov D & Tobacman LS (2000) Regulation of force and unloaded sliding speed in single thin filaments: effects of regulatory pro- teins and calcium. J Physiol 524, 233–243. 19 Deng Y, Schmidtmann A, Redlich A, Westerdorf B,
(1987) Cardiac contractile protein phosphatases. Purifi- cation of two enzyme forms and their characterization with subunit specific antibodies. J Biol Chem 262, 6257– 6265. Jaquet K & Thieleczek R (2001) Effects of phosphoryla- tion and mutation R145G on human cardiac troponin I function. Biochemistry 41, 14593–14602.
30 Unneberg P, Merelo JJ, Chacon P & Moran F (2001) SOMCD: method for evaluating protein secondary structure from UV circular dichroism spectra. Proteins 42, 460–470.
20 Deng Y, Schmidtmann A, Kruse S, Filatov V, Heil- meyer LMG, Jaquet K & Thieleczek R (2003) Phos- phorylation of human cardiac troponin IG203S and K206Q linked to familial hypertrophic cardiomyopathy affects actomyosin interaction in different ways. J Mol Cell Cardiol 35, 1365–1374. 21 Babu A, Su H, Ryhu Y & Gulati J (1992) Determina-
31 Ferrieres G, Pugniere M, Mani JC, Villard S, Laprade M, Doutre P, Pau B & Granier C (2000) Systematic mapping of regions of human cardiac troponin I involved in binding to cardiac troponin C: N- and C-terminal low affinity contributing regions. FEBS Lett 479, 99–105. tion of the residue specifity in the EF-hand of troponin C of Ca2+ coordination by genetic engineering. J Biol Chem 267, 15489–15474.
FEBS Journal 272 (2005) 6087–6097 ª The Authors Journal compilation ª 2005 FEBS
6097
32 Frank R (1992) Spot-Synthesis: An easy technique for the positionally addressable, parallel chemical synthesis on a membrane support. Tetrahedron 48, 9217–9232. 33 Webb MR (1992) A continuous spectrophotometric 22 Reiffert SU, Jaquet K, Heilmeyer LMG & Herberg FW (1998) Stepwise subunit interaction changes by mono- and bisphosphorylation of cardiac troponin I. Biochem- istry 37, 13516–13525. 23 Teubner A & Wegner A (1998) Kinetic evidence for a readily exchangeable nucleotide at the terminal subunit assay for inorganic phosphate and for measuring phos- phate release kinetics in biological systems. Proc Natl Acad Sci USA 89, 4884–4887.
