Basic residues within the cardiac troponin T C terminus are required for full inhibition of muscle contraction and limit activation by calcium

Striated muscle is activated by myosin- and actin-linked processes, with the latter being regulated through changes in the position of tropomyosin relative to the actin surface. The C-terminal region of cardiac troponin T (TnT), a tropomyosin-associated protein, is required for full TnT inactivation at low Ca2+ and for limiting its activation at saturating Ca2+. Here, we investigated whether basic residues in this TnT region are involved in these activities, whether the TnT C terminus undergoes Ca2+-dependent conformational changes, and whether these residues affect cardiac muscle contraction. We generated a human cardiac TnT variant in which we replaced seven C-terminal Lys and Arg residues with Ala and added a Cys residue at either position 289 or 275 to affix a fluorescent probe. At pCa 3.7, actin filaments containing high-alanine TnT had an elevated ATPase rate like that obtained when the last TnT 14 residues were deleted. Acrylodan-tropomyosin fluorescence changes and S1-actin binding kinetics revealed that at pCa 8, the high-alanine TnT-containing filaments did not enter the first inactive state. FRET analyses indicated that the C-terminal TnT region approached Cys-190 of tropomyosin as actin filaments transitioned to the inactive B state; that transition was abolished with high-alanine TnT. High-alanine TnT-containing cardiac muscle preparations had increased Ca2+ sensitivity of both steady-state isometric force and sinusoidal stiffness as well as increased maximum steady-state isometric force and sinusoidal stiffness. We conclude that C-terminal basic residues in cardiac TnT are critical for the regulation of cardiac muscle contraction.

Activation of striated muscle occurs both through myosinlinked (1, 2) and actin-linked processes (3)(4)(5). The actin-linked portion of regulation works primarily through changes in the position of tropomyosin relative to the surface of actin (6 -8).
Tropomyosin is in an inactivating state when Ca 2ϩ levels are low (9,10). Saturation of the regulatory binding site(s) of troponin C (TnC) 3 with Ca 2ϩ opens a hydrophobic cleft to which the switch region of TnI (troponin I) can bind (11), permitting detachment of the inhibitory region of TnI from actin. Tighter binding of the switch region to TnC has been associated with Ca 2ϩ sensitization (12) and with movement of tropomyosin into a mixture of the C or Ca 2ϩ state and M or myosin-stabilized states. Fig. 1 defines the various states. The C state, like the B state, appears to be inactive (13). Full activation occurs when in addition to Ca 2ϩ a fraction of actin protomers has tightly bound myosin that is either nucleotide-free or contains only bound ADP (14,15). Stimulation of S1 ATPase activity by Ca 2ϩ and "activating" forms of myosin S1 (rigor and S1-ADP) are adequately described by changes in the population of the states of regulated actin (15).
The degree to which Ca 2ϩ stimulates the ATPase rate depends on the equilibrium between the C and M states, and that depends on the composition of the actin filament. For example, substitution of acanthamoeba actin into skeletal thin filaments enhances the degree of activation by Ca 2ϩ (16,17). An enhanced activation by Ca 2ϩ occurs in some disease-causing mutations of TnT, most notably with deletion of the last 14 C-terminal residues of TnT (⌬14 TnT) (18,19). When both ⌬14 TnT and A8V TnC were present in the same thin filaments, there was full activation with Ca 2ϩ alone (no myosin-ADP binding required) (19). Other changes in the thin filament, such as protein kinase C phosphomimetic mutants of TnI (20), decrease the response to Ca 2ϩ . Stabilization of the intermediate C state of actin filaments can also occur with another mutation that causes familial hypertrophic cardiomyopathy (13). The C terminus of TnT is particularly interesting because deletion of the last 14 residues has dual effects of eliminating the inactive B state at very low Ca 2ϩ and enhancing the active M state at saturating Ca 2ϩ (19,21).
The preponderance of positively charged residues in the C terminus of TnT raises the possibility that electrostatic interactions are responsible for the functions of this region. We pro-duced a construct of human cardiac TnT that maintained the length of WT TnT but had the terminal 7 basic amino acids replaced with Ala. We call this construct HAHA TnT (high Ala, high activation). Both the HAHA TnT and WT contained either an added Cys at position 289 as a terminal residue or a Ser to Cys mutation at position 275 for placement of a fluorescent probe.
By measuring rates of ATP hydrolysis, changes in fluorescence of a probe on tropomyosin, kinetics of S1 binding to actin filaments, and FRET studies, we show that the C-terminal basic residues of TnT are critical for proper regulation. Actin filaments containing HAHA TnT were unable to enter the inactive B state at low-Ca 2ϩ conditions. At saturating Ca 2ϩ levels, those same filaments moved more fully to the M state and had enhanced activity. The effects of HAHA TnT on muscle fiber mechanics were consistent with the solution results. Furthermore, the C-terminal region of TnT underwent a Ca 2ϩ -dependent conformational change that was eliminated in the case of HAHA TnT. These results suggest that the C-terminal region of TnT, or its binding partners, are potential targets for manipulating cardiac contractility and improving the function of diseased hearts.

Results
Saturating Ca 2ϩ normally gives about 30% of the maximum possible ATPase activity (19), but deletion of the last 14 C-terminal residues of human cardiac TnT gives a further doubling of activity (i.e. doubling of the population of actin in the active M state). Because the C-terminal region of TnT is rich in basic amino acids, replacing the basic residues with uncharged Ala residues was expected to increase the fraction of actin filaments in the active state at saturating Ca 2ϩ . Fig. 2 shows the effect of different types of TnT on the actinactivated ATPase rate of S1 at saturating Ca 2ϩ and subsaturating actin concentrations. In the absence of tropomyosin or troponin, the rate was 1.2/s; this is typical for these conditions (19).
Human cardiac troponin normally increases the activity of actin-tropomyosin, and it increased the rate here to 2.2/s. When WT troponin was replaced with TnT having either S275C or S289C mutations, the rates remained similar to WT (2.0 and 2.1/s for S275C and S289C, respectively). In contrast, 289C-HAHA TnT increased the rate to 3.8/s. That is similar to the increase from 1.9/s to 4/s noted earlier when TnT missing the last 14 residues was analyzed (19). The fraction of actin in the active state increased from 34% in WT to 69% in 289C-HAHA TnT based on the results of Fig. 2.
We showed earlier that stepwise truncation of the 14 C-terminal residues of TnT progressively decreases the population of the inactive B state of regulated actin (21). We have now measured changes in the distribution of the inactive B that occur with the 289C-HAHA mutant of TnT. The first method used is based on the change in the rate of binding of ATP-free S1 to actin-tropomyosin-troponin at low free Ca 2ϩ relative to the rate at high Ca 2ϩ (10,22). In studies reported here, low Ca 2ϩ is 10 Ϫ8 M and high Ca 2ϩ is generally 0.2 mM free. The rate of binding at high Ca 2ϩ , where the C and M states are populated, is normally about 3 times that rate in the virtual absence of Ca 2ϩ , where the B state predominates. Fig. 3 shows binding isotherms for rigor S1 binding to actintropomyosin-troponin at a very low free Ca 2ϩ concentration. Curve 1 is for actin filaments containing WT human cardiac TnT. S1 binding to pyrene-labeled actin occurred with a monoexponential decrease in fluorescence with an apparent rate constant of 1.0/s. The apparent rate constant for filaments containing C289-TnT was similar at 1.2/s (curve 2). Curve 3 was obtained using actin filaments containing 289C-HAHA TnT. The apparent rate constant of binding was 4.1/s. That value is similar to the WT rate at saturating Ca 2ϩ , where there is no B state.
Similar measurements were made at a saturating Ca 2ϩ concentration and are shown in Fig. 4. Mono-exponential traces were obtained for WT (curve 1), 289C-TnT (curve 2), and 289C-HAHA TnT. The apparent rate constants for WT and 289C-TnT were similar (3.7 and 3.9/s, respectively). These rate constants were faster than those measured in the virtual absence of Ca 2ϩ and are generally regarded to represent the maximum possible rate (10). However, the apparent rate con-  (9), has no bound Ca 2ϩ and is inactive. State C, equivalent to state 1 Ca in the Hill model, has bound Ca 2ϩ but is inactive. State M (equivalent to state 2) is active and is stabilized by binding of both Ca 2ϩ and rigor myosin to regulated actin filaments.  Basic residues within the cardiac troponin T C terminus stant of binding of S1 to actin filaments containing 289C-HAHA TnT was approximately double that rate (6.8/s). A similar rate of binding was observed with ⌬14 TnT (21). This suggests that the rate of rigor S1 binding to the M state is faster than to the C state.
We also measured changes in the occupancy of the inactive B state by monitoring changes in the fluorescence of acrylodanlabeled tropomyosin (23). The active M and inactive B states have high fluorescence, whereas the inactive C state has low fluorescence. Actin-acrylodan tropomyosin-troponin is maintained in the M state by rigor S1 binding even at very low free Ca 2ϩ , giving rise to a high fluorescence state. Upon mixing rapidly with ATP, the S1 dissociates, and the regulated actin transitions rapidly to the C state and then more slowly to the B state. Because the transition to the C state is very rapid and virtually complete, the magnitude of the final fluorescence rise is proportional to the occupancy of the B state. TnT had an amplitude of 0.28 relative fluorescence units and an apparent rate constant of 29/s. However, eliminating the basic residues in the C-terminal region of TnT produced a large change. No increase in fluorescence was observed in the case of 289C-HAHA TnT (curve 3). That curve is similar to that observed previously for ⌬14 TnT (23,24), indicating absence of the B state.
Because the C-terminal region of TnT appears to be involved in regulation, we expected to see changes in the location or orientation of that region with changes in the state of activation. A donor molecule, IAEDANS, was placed on a thiol group engineered into the C-terminal region of TnT. FRET was measured between an IAEDANS donor and a DABMI probe on Cys-190 of tropomyosin during the transition from the active M state to the inactive C and B states using the ATP chase protocol used in Curve 1 in Fig. 6A shows 289C-TnT IAEDANS (WT with an added Cys for labeling) fluorescence in the absence of an acceptor as the filaments transitioned from the M state to the C state and finally to B state at 10°C. Curve 2 shows the same transition in the presence of the acceptor DABMI on tropomyosin. Here, a single exponential decay was observed with an apparent rate constant of 6.2 Ϯ 0.5/s. This rate is similar to that measured in Fig. 5 for the transition from the C to the B state. The same measurement was made at 18°C (B) and 25°C (C). D-F show . Rate of binding of S1 to an excess of pyrene-labeled actin filaments containing tropomyosin and troponin in the absence of ATP at high Ca 2؉ . Plots are averages of five traces of binding to actin-tropomyosin containing WT (curve 1), 289C (curve 2), and 289C-HAHA TnT (curve 3). Dashed lines, single-exponential fits. The apparent rate constants were 3.7 Ϯ 0.6, 3.9 Ϯ 0.6, and 6.8 Ϯ 0.8 per second for WT, 289C, and 289C-HAHA TnT, respectively. Conditions were the same as in Fig. 3, except 0.2 mM CaCl 2 was substituted for EGTA.

Basic residues within the cardiac troponin T C terminus
the calculated changes in FRET efficiency with time for the three temperatures. The increase in FRET efficiency indicates that in going from the C to the B state, the distance from the C terminus of TnT to Cys-190 of tropomyosin decreased. Table  1 shows the calculated distance changes in going from the C state to the B state at a very low concentration of Ca 2ϩ . Table 2 gives values of the anisotropy of several of the probes at different free Ca 2ϩ concentrations. The anisotropies of the donor molecules changed only slightly with changes in free Ca 2ϩ . The values were near 0.1, which is between the limits for free rotation (Ϫ0.2) and for immobile probes (0.4) (25). Because the acceptor probe, DABMI, is nonfluorescent, we measured the anisotropy of a fluorescein probe on Cys-190 of tropomyosin, which was also near 0.1.
We know from earlier work that the transition between the inactive C and B states is temperature-dependent (23). The apparent rate constants for the distance change attributed to the transition from the C to the B state are shown in the form of an Arrhenius plot in Fig. 7 along with those earlier acrylodantropomyosin data. The similarity of the temperature dependences is evidence that the energy transfer observed corresponds to the C to B transition. Fig. 8 shows that no energy transfer occurred in the presence of IAEDANS-labeled 289C-HAHA TnT where the C-terminal basic residues have been substituted with Ala. The transfer efficiency, and thus the distance between the probes, remained constant at a value similar to that of the WT in the C state. In the case of 289C-HAHA TnT, there is no transition to the B state. Furthermore, the movement shown in Fig. 6 appears to require the presence of the basic residues in the C-terminal region of TnT.
To further define the movement of the C-terminal region of TnT, we measured the change in FRET efficiency with IAE-DANS placed on an engineered Cys at position 275 of human cardiac TnT. The IAEDANS probe on Cys-275 of TnT produced an increase in FRET efficiency in going to the B state ( Fig.  9, A and B). This change differed in two respects from that observed with Cys-289 of TnT. First, the difference in effi-ciency between the C and B states was smaller. Second, the initial and final values of the efficiency were larger than with 289C-TnT, indicating that residue 275 is closer to Cys-190 of tropomyosin than is residue 289 of TnT. The transition was slower (3.5 Ϯ 0.2/s), possibility due to the probe at position 275.
Panels C and D of Fig. 9 show energy transfer between IAE-DANS at position 143 of TnI and Cys-190 of tropomyosin. This was measured because it is well-known that TnI moves toward the actin-tropomyosin filament in the relaxed B state.     Our earlier work showed that deletion of the last 14 residues of human cardiac TnT increased the Ca 2ϩ sensitivity of muscle fibers (18). We exchanged the native troponin in permeabilized cardiac muscle preparations (CMPs) with recombinant human cardiac WT TnT or with human cardiac 289C-HAHA TnT. To determine the efficiency of the exogenous TnT to displace the native troponin complex, we measured the unregulated tension at low Ca 2ϩ concentration (pCa 8) after TnT incubation, as described under "Experimental procedures" (26,27). We did not observe a significant change in the ability of the 289C-HAHA TnT to displace the native troponin complex compared with its respective control (86.96 Ϯ 1.13% versus 81.80 Ϯ 3.79%; Table 3). This finding indicates that the 289C-HAHA TnT does not display an altered affinity for the thin filament because its incorporation into the CMPs was comparable with the WT TnT. Fig. 10A is a plot of relative force against pCa for CMPs containing WT or 289C-HAHA TnT. Replacement of basic residues with neutral Ala in the C-terminal region of TnT produced myofilament Ca 2ϩ sensitization measured by steadystate isometric force with a shift of 0.24 pCa units in the midpoint of activation from FpCa 50 5.38 Ϯ 0.03 to 5.62 Ϯ 0.02 (Table 3). No changes in cooperativity of thin filament activation (n Hill ) were observed in CMPs containing 289C-HAHA versus WT TnT ( Fig. 10 and Table 3). The other notable feature of these curves is that the maximum force recovery was increased 24.6% in CMPs containing 289C-HAHA TnT ( Fig.  10B and Table 3). Table 3 reports the absolute steady-state isometric force values before (P) and after (P 0 ) recombinant troponin incorporation.
Measurements of sinusoidal stiffness in CMPs were made to estimate changes in the number of force producing crossbridges. Fig. 11A shows sinusoidal stiffness as a function of pCa. 289C-HAHA TnT increased sinusoidal stiffness at all levels of Ca 2ϩ activation (including maximum sinusoidal stiffness, SS max ), indicating an increase in the number of force-producing cross-bridges (Table 3). When viewed as a function of the force normalized to WT (Fig. 11B), there was an increase in sinusoi-

Basic residues within the cardiac troponin T C terminus
dal stiffness for the 289C-HAHA over the WT at the same level of force. 289C-HAHA TnT apparently recruited more forceproducing cross-bridges relative to WT at the same level of force. Furthermore, CMPs containing 289C-HAHA displayed an increased Ca 2ϩ sensitivity of pCa-sinusoidal stiffness relation compared with WT ( Fig. 12 and Table 3). The leftward shift of 0.19 log units observed in the sinusoidal stiffness (SSpCa 50 ) was correlated to those obtained from the corresponding pCa-force relations (Table 3).

Discussion
We describe here the production of a mutant of human cardiac TnT that eliminated the B state, the major state that actintropomyosin-troponin occupies in the virtual absence of Ca 2ϩ . That mutant also caused an increase in activation by Ca 2ϩ in both solution and permeabilized cardiac muscle preparations and an enhanced response to Ca 2ϩ . Fluorescent probes on Cys residues placed within the C-terminal region of that TnT reported conformational changes occurring in response to changes in the state of activation. The HAHA mutation of TnT mimicked all of the effects of deleting the C-terminal region of TnT (19,21,24). This study demonstrates the importance of positive charges in the C-terminal region of TnT for normal regulation of contraction and for a conformational change in the C-terminal region of TnT that occurs in moving between the two inactive states (state C to B). Cys-labeled HAHA TnT is a useful tool for studying contractile regulation.
It is well-known that rigor myosin binding to actin can stabilize the active M state and increase ATPase activity at both low and high Ca 2ϩ . It now appears that troponin can give the same kind of activation of ATPase activity as occurs with rigor myosin binding. Understanding how the C-terminal region of TnT modulates the degree of activation by Ca 2ϩ could have important health implications as this is a previously unrecognized target.
HAHA TnT has Ala in place of the Lys and Arg residues within the last 16 residues of TnT. We used a 16-residue rather than a 14-residue stretch in an attempt to increase the Ca 2ϩ activation beyond that seen with ⌬14 TnT. We were limited to the last 16 residues by the presence of the I-T helix. Extending the modification beyond 14 residues did not result in complete activation.
Elimination of the inactive B state by the HAHA mutation was demonstrated here by the disappearance of the acrylodan tropomyosin signal. We showed earlier that an increase in acrylodan fluorescence occurs as the B state becomes populated (19,23,24). The absence of a signal indicates that the B state was virtually eliminated with 289C-HAHA TnT.
Additional evidence for loss of the B state comes from an increase in the rate of binding of S1 to pyrene-labeled actin in the virtual absence of Ca 2ϩ , from 1/s to 4.1/s (Fig. 3), a rate equal to that observed at saturating Ca 2ϩ (4.1/s in Fig. 4).
Strong support for the idea that the C-terminal region of TnT is critical for forming the B state came from the observation that the conformational change associated with forming the B state was eliminated when the basic residues in the C-terminal region of TnT were eliminated (see below).
The primary evidence for stabilization of the M state by the basic residues within the C-terminal region of TnT comes from ATPase measurements. Replacement of WT TnT with HAHA TnT doubled the actin-activated ATPase activity at saturating Ca 2ϩ . Although this is a large effect, it does not represent complete stabilization of the active state. At saturating Ca 2ϩ , activation by N-ethylmaleimide-labeled S1 or by a combination of two troponin mutants (A8V TnC and ⌬14 troponin) tripled the activation of ATPase activity by Ca 2ϩ (19). At saturating Ca 2ϩ , HAHA TnT gave ϳ70% of that maximum ATPase rate. Put another way, deleting the basic residues in the C-terminal

Basic residues within the cardiac troponin T C terminus
region of TnT increased the population of actin in the active M state, at saturating Ca 2ϩ , from Ͻ35% to about 70% (see also Baxley et al. (19)).
HAHA TnT has advantages over the use of rigor S1 to stabilize the active state. HAHA TnT allows measurements to be made at saturating ATP. Also, corrections for changes in available myosin-binding sites (19) are not necessary. Furthermore, when one of the Cys-containing varieties of HAHA TnT is used, both the probe and the function can be monitored. As we show here, HAHA TnT can be used in fiber studies as well as in solution.
Rigor S1 stabilizes the active state by displacing tropomyosin on actin (8). An interesting question is whether troponin accomplishes this activation also by repositioning tropomyosin in the same manner as occurs with rigor S1 binding to actin. It is interesting that the ⌬28 deletion of TnT does not appear to cause changes in the position of tropomyosin (28). The ⌬28 deletion extends into the I-T helix and weakens binding of troponin to actin-tropomyosin.
Another indication that basic residues within the C-terminal region of TnT impact the function of regulated actin at saturating Ca 2ϩ comes from the rate of rigor S1 binding to actin. The rate of binding to WT actin filaments has been assumed to be at its maximum at saturating Ca 2ϩ (10). In other words, the rates of rigor S1 binding to the active M and inactive C states are assumed to be equal. At saturating Ca 2ϩ , HAHA TnT shifts the equilibrium between the C and M states toward the active M state. Thus, replacing HAHA TnT for WT should not affect the kinetics of S1 binding at high Ca 2ϩ . However, we observed an increase from the WT rate of 3.7/s to 6.8/s with HAHA TnT (Fig. 4). We observed a similar behavior with deletion mutants of TnT (21). The simplest explanation is that the rate of rigor S1 binding is Ͼ2-fold faster to the M state than to the C state with cardiac regulatory proteins. Alternatively, these TnT mutants could produce a state that is normally unpopulated. If the latter is true, then that state must be identical to the M state in stimulating ATP hydrolysis.
In situations where the rate of rigor S1 binding to the M state is greater than to the C state, one cannot calculate the fraction of actin in the B state by comparing the rates of binding in the presence and absence of Ca 2ϩ as described by others (10).
Despite this limitation, measurements of S1 binding kinetics do give an accurate measure of the change in occupancy of the B state; the magnitude of acrylodan-tropomyosin fluorescence gives a similar measure of the change (21). Quantitation by those methods requires a standard that is 100% in the B state. Regulated actin filaments containing S45E TnI had an actinactivated ATPase rate, at low Ca 2ϩ , equal to 0.61 times the WT rate (20). The acrylodan tropomyosin fluorescence amplitude for S45E TnI containing filaments was 1.3 times that of WT (24). This is a useful benchmark, although it is unclear that it represents 100% B state.
To begin to understand how the C-terminal basic residues of TnT function, we measured changes in position of that region relative to Cys-190 of tropomyosin in the transition from the inactive C state to the inactive B state. The location of Cys-190 relative to the core region of troponin is shown in Fig. 13, which is based on Takeda et al. (29). Earlier evidence suggested that Cys-190 of tropomyosin was near TnT residues 197-239 (30) or residues 272-288 (31,32). That location is not certain, however, as a fragment of TnT (residues 262-288) was shown to interact with whole TnI (33).
In moving from the C state to the B state, probes on positions 275 and 289 of TnT moved toward Cys-190 of tropomyosin. Fig. 13 illustrates the location of the donor probe on position 289 of TnT (d1) and the acceptor on tropomyosin in that relaxed or B state. Positions 289 and 275 moved similarly relative to Cys-190 of tropomyosin.
As a control, we also measured changes in a probe on position 143 of TnI, shown as d2 in Fig. 13. That probe is within the inhibitory region of TnI that binds to actin in the B state. That inhibitory region was seen to move closer to actin in going from the C state to the B state as expected.
The results seen here point to interesting structural differences between the two inactive states, B and C. Although neither state supports productive actomyosin interactions, they differ in their orientations of both TnI and TnT relative to Cys-190 of tropomyosin. We are currently exploring these changes and those that occur between the active M state and the inactive C state.
No change in distance between the C-terminal region of TnT and Cys-190 of tropomyosin was observed in the case of 289C-HAHA TnT. This is strong evidence that the B state does not form in the absence of a basic patch at the C-terminal end of TnT. The distance between residue 289 of HAHA TnT and 190 of tropomyosin remained at ϳ45 Å, which seems to be the distance characteristic of the C state. The basic residues within the C-terminal region of TnT appear to be required for tropomyosin movement into the B state.
The changes in FRET reported here have been interpreted in terms of changes in distance between the donor and acceptor probes. Changes in energy transfer efficiency can also occur as a result of changes in the quantum yield of the donor or probe orientation ( 2 term of Equation 4). In the present case, changes in the quantum yields between high and low Ca 2ϩ were insig- Figure 13. Schematic of the key interaction sites between troponin and actin (yellow) and tropomyosin (black) in the relaxed state based on Takeda et al. (29). TnC (blue) is positioned between two helices, one from TnI (red) and the IT helix involving both TnI and TnT (green). In the relaxed state, TnI interacts with actin-tropomyosin through the inhibitory region (IR) and the C-terminal region (C-TnI). Between these regions is the switch region (SW) that binds to TnC at saturating Ca 2ϩ . TnT is shown in green. Adjacent to the IT helix is the C-terminal region of TnT, shown as a thin green line. That C-terminal region has a fluorescent donor (d1). An acceptor of fluorescence (a) is present on Cys-190 of tropomyosin. The distance between d1 and a is at its minimum in the relaxed B state. The black arrow shows a type of change in the C-terminal region of TnT when Ca 2ϩ binds to TnC. Ca 2ϩ binding also increases the distance between the donor in the IR region of TnI (d2) as shown by the red arrow. The interactions among troponin and actin-tropomyosin, shown by the columns of black dashes, are also disrupted, and the switch region ultimately binds to TnC. For clarity, the donor at position 275 is not shown.

Basic residues within the cardiac troponin T C terminus
nificant. Anisotropy measurements showed that there were no large changes in probe mobility, so the value of 2 was apparently unchanged. It seems likely that the changes in transfer efficiency reported here are due to distance changes.
Actin filaments containing human cardiac 289C-TnT and 289C-HAHA TnT behaved, in solution, like WT and ⌬14 TnTcontaining filaments, respectively. The 289C-HAHA troponin construct also behaved like ⌬14 TnT when introduced into CMPs. In solution, the ATPase activity was increased, and the B state was eliminated when the basic residues were removed from the C-terminal region of TnT. Similarly, in CMPs, the sinusoidal stiffness was increased at low Ca 2ϩ concentrations. Eliminating those basic residues also caused an increase in ATPase activity and a faster rate of rigor S1 binding at high Ca 2ϩ . Similarly, cardiac muscle preparations containing 289C-HAHA TnT produced a greater level of steady-state isometric force and had an increased number of force-producing crossbridges at all Ca 2ϩ levels.
289C-HAHA TnT produced a shift of 0.24 pCa units in the myofilament Ca 2ϩ sensitivity measured by steady-state isometric force in skinned CMPs. This compares well with the 0.2 pCa leftward shift observed with ⌬14 TnT in skinned trabeculae strips (18). The increase in force was associated with an increase in the number of force-producing cross-bridges (Fig. 10).
One may argue that the state of actin obtained by using HAHA TnT, or other such mutants (21,24), could be different from the M state. The M state is defined as the state obtained when actin-tropomyosin-troponin has sufficient rigor-type myosin bound to achieve full activity at saturating Ca 2ϩ . We described earlier the ATPase activity of regulated actin in the presence of N-ethylmaleimide-labeled S1 with various types of regulatory proteins bound to actin (19). The use of various mutants of troponin achieved the same level of activity as obtained with full activation with modified S1 (19,21). This is evidence that the same state is stabilized in both cases. The change in pCa-force relationship observed here is also expected if the M state is stabilized. We analyzed other mutants of troponin and found them to affect the equilibria among the B, C, and M states (34). So the idea that mutations alter the distribution of actin states seems reasonable. The most straightforward explanation is that HAHA TnT as well as other mutants of TnT stabilize the M state.
All of the results shown here, both in solution and in organized muscle preparations, support the idea that the C-terminal region of TnT serves a critical role in contraction. Elimination of the basic residues in the C-terminal region of TnT eliminates one of the inactive states (the B state). The other function of the basic residues is to limit the extent of activation by Ca 2ϩ . This leads to the possibility that the extent of activation may be regulated by an unknown mechanism in addition to Ca 2ϩ binding to TnC. Fig. 10 shows that altering the C-terminal region of TnT can produce increases in cardiac muscle force of contraction at physiological Ca 2ϩ concentrations (e.g. pCa 5. 8 -5.4). If these same changes can be created in a WT fiber using a drug or other intervention, it could be possible to improve the performance of a diseased heart.

Proteins
Actin was prepared from a bulk dissection of the back muscles of a rabbit (35). New Zealand White rabbits were sacrificed in accordance with National Institutes of Health guidelines and the animal care protocol approved by the Animal Care and Use Committee of East Carolina University.
F-actin was labeled with N-(1-pyrene)iodoacetamide (36,37). The modification buffer was 1 mM Tris, pH 8, 0.1 mM CaCl 2 , and 0.5 mM ATP. The reaction was stopped with excess DTT, and the actin was centrifuged in a Ti 50 rotor for 20 min at 30,000 rpm to remove precipitated protein and excess probe. Pyrene-labeled actin was dialyzed against a minimum of three changes of 4 mM MOPS, 2 mM MgCl 2 , 1 mM DTT buffer at 4°C. The extent of pyrene labeling was determined using an extinction coefficient of 22,000 M Ϫ1 cm Ϫ1 at 344 nm. The extent of labeling with pyrene was generally 70%.
Rabbit skeletal myosin was prepared from back muscle (38) and was digested with chymotrypsin to prepare the soluble catalytic fragment, S1 (39). Tropomyosin was prepared from bovine cardiac left ventricles (40) and was labeled at Cys 190 with acrylodan using a 10:1 ratio of acrylodan to tropomyosin (24). The extent of labeling was ϳ70% using an extinction coefficient of 14,400 M Ϫ1 cm Ϫ1 at 372 nm for acrylodan (41). Tropomyosin was labeled at Cys 190 with DABMI using a similar procedure. The molar ratio of DABMI to tropomyosin was 5:1, and the extent of labeling was 68% using an extinction coefficient of 24,800 M Ϫ1 (42). Human cardiac troponin components were expressed in Escherichia coli and purified as described earlier (19). TnT (isoform 2) was expressed in pSBETa, TnI, and TnC in pET3d. The purified troponin components were reconstituted and purified by ion-exchange chromatography (19).
The mutants of human cardiac TnT used in this study are shown in Table 4. They were synthesized and cloned into a pMK vector by the Invitrogen GeneArt Gene Synthesis Service. Two restriction enzyme sites, NdeI and BamHI, were added to the 5Ј and 3Ј termini, respectively. Each cloned DNA sequence was verified and excised from the recombined plasmid by digestion with appropriate restriction enzymes and subcloned into the NdeI-BamHI site of the vector pSBETa (a kind gift from Dr. H.-H. Steinbiss) to yield an expression plasmid designated pSBET-289C-HAHA TnT. Clones containing the correct recombinant plasmid were confirmed by NdeI and BamHI double digestion and DNA sequencing. The following primers were used for TnT sequencing: TNNT2FW, ATCCGGAATGAGC-GGGAGAA; TNNT2RV, ATTCAGGTCCTTCTCCATGCG. BL21(DE3)pLysS strain was used as the expression host for 289C-HAHA TnT. The masses of tryptic fragments of the 289C-HAHA TnT were analyzed by LC-MS. The masses obtained were consistent with the sequences.

Basic residues within the cardiac troponin T C terminus
The concentrations of actin and S1 were determined by absorbance at 280 nm after correction for light scattering at 340 nm using the following extinction coefficients (⑀ 0.1% ): 1.15 for actin and 0.75 for S1. Tropomyosin and troponin subunits were quantified by the Lowry protein assay using a BSA standard. Molecular weights were assumed to be 120,000 for myosin S1, 68,000 for tropomyosin, 42,000 for actin, 35,923 for TnT, 24,000 for TnI, and 18,400 for TnC.

ATPase assays
The rate of liberation of 32 P i from [␥-32 P]ATP was measured as described earlier (43). The phosphate concentration was determined from a 0.05-ml aliquot taken at 3, 6, and 9 min to ensure linearity of the reaction. Initial rates were determined from a linear least-squares analysis. All reactions were run at 25°C at the conditions indicated in the figure legends. The fraction of actin in the active M state was calculated from the ATPase rate (19).

Rapid kinetic measurements
Pre-steady-state rates were measured using a SF20 sequential mixing stopped-flow spectrometer equipped with an LED light source (Applied Photophysics, Leatherhead, UK). Excitation wavelengths were determined by the wavelength of the LED light source. The emission wavelength was set with a highpass filter. Reactions were generally run at high Ca 2ϩ (ϳ0.2 mM free) and low Ca 2ϩ (ϳ10 Ϫ8 M free).
The formation of the B state was determined by monitoring acrylodan-tropomyosin fluorescence (21,23). Acrylodan was excited with a 390-nm LED, and emission was measured with a 451-nm high-pass filter. The B state was also determined by comparing the rate of binding of rigor S1 to actin filaments in the virtual absence of Ca 2ϩ with that at saturating Ca 2ϩ (10). Pyrene probes on actin were excited using a 360-nm LED, and fluorescence was monitored through a 400-nm midpoint highpass filter.
For FRET studies, the single Cys residue of donor proteins 275C of TnT or 289C of both TnT and HAHA TnT or 143C of TnI were labeled with IAEDANS. The acceptor molecule DABMI was placed on Cys-190 of tropomyosin. A 340-nm LED was used for IAEDANS excitation, and emission was monitored through a high-pass filter with a cut-on midpoint of 451 nm.
IAEDANS emission was measured in both the absence and presence of the DABMI acceptor probe. The absorbances of all solutions at 336 nm were Ͻ0.01 to avoid inner filter effects. The traces were corrected using a baseline generated from the same experiment in the absence of donor and acceptor probes. The initial and final fluorescence were recorded for the donor with (I DϩA ) and without acceptor (I D ). The efficiency of transfer, E, was calculated from Equation 1, where I DϩA (t) is the time course of the fluorescence intensity of the donor in the presence of the acceptor, I D is the time course of the donor alone, and F A is the fraction of the acceptor protein, tropomyosin, labeled with DABMI.
The distance (Å) between the donor and acceptor probes, r, where R 0 is the Förster distance for the donor-acceptor pair on their respective proteins, and R 0 6 , in units of Å 6 , was calculated from Equation 3. 2 , the orientation factor, was assumed to be 2 ⁄ 3 for random orientations of the probes. An alternate assumption of 0.476 for rigid probes and a random ensemble (45) resulted in distances ϳ10% smaller. The refractive index of the medium, n, was assumed to be 1.33. The quantum yield of the donor probe on the reconstituted troponin complex, Q D was measured using a standard of 1 M quinine sulfate in 0.1 N H 2 SO 4 (46). Q S , in Equation 3, is the quantum yield of the standard and equals 0.55. A S /A D is the ratio of absorbance values of the standard and the donor, and F D /F S is the corresponding ratio of the integrated fluorescence intensities of the donor and the quinine sulfate standard, respectively. The overlap integral, J(), was calculated from the emission spectrum of the donor (IAEDANS) normalized to a maximum value of 1 and the absorption spectrum of the acceptor (DABMI) normalized to a maximum value equal to the extinction coefficient of DABMI (24,800 M Ϫ1 ) (42) using the program a͉e (FluorTools, www.fluortools.com). 4 Fluorescence emission spectra were measured with a Fluoromax-4 spectrofluorometer (Horiba Scientific, Edison, NJ) and corrected for lamp output and radiometric corrections. The baseline was corrected, prior to integration, using the Horiba software or by fitting two or more Gaussian curves to the data using Mathematica (Wolfram Research). The absorbances of all solutions at 336 nm were Ͻ0.01 to avoid inner filter effects.
Anisotropy measurements were made on the same instrument at the same conditions used for distance measurements.
ATPase measurements at high Ca 2ϩ showed that the effects of the probes were modest. The actin-activated ATPase rate in the presence of tropomyosin was 0.46 Ϯ 0.06/s. That rate reduced to 0.37 Ϯ 0.02/s when tropomyosin was labeled. The rate measured with actin-tropomyosin-troponin was 1.9 Ϯ 0.08/s. That compares with 1.4 Ϯ 0.02/s when TnI was labeled, 1.8 Ϯ 0.07 when TnT was labeled at Cys-275, and 1.6 Ϯ 0.04/s when TnT was labeled at Cys-289.
Mathematical analysis of the temperature dependence of apparent rate constants, k(T), was done with MLAB (Civilized Software, Bethesda, MD). The Arrhenius equation is shown in Equation 5, where A is the pre-exponential factor, E A is the energy of activation with units of R⅐T, R is the gas constant 8.314 J/(mol⅐K), and T is in degrees K.