Micromechanical Thermal Assays of Ca2+-Regulated Thin-Filament Function and Modulation by Hypertrophic Cardiomyopathy Mutants of Human Cardiac Troponin

Microfabricated thermoelectric controllers can be employed to investigate mechanisms underlying myosin-driven sliding of Ca2+-regulated actin and disease-associated mutations in myofilament proteins. Specifically, we examined actin filament sliding—with or without human cardiac troponin (Tn) and α-tropomyosin (Tm)—propelled by rabbit skeletal heavy meromyosin, when temperature was varied continuously over a wide range (∼20–63°C). At the upper end of this temperature range, reversible dysregulation of thin filaments occurred at pCa 9 and 5; actomyosin function was unaffected. Tn-Tm enhanced sliding speed at pCa 5 and increased a transition temperature (Tt) between a high activation energy (Ea) but low temperature regime and a low Ea but high temperature regime. This was modulated by factors that alter cross-bridge number and kinetics. Three familial hypertrophic cardiomyopathy (FHC) mutations, cTnI R145G, cTnI K206Q, and cTnT R278C, cause dysregulation at temperatures ∼5–8°C lower; the latter two increased speed at pCa 5 at all temperatures.


Introduction
Biomolecular motors have clear promise for transport applications in micro-and nanofabricated devices although control of motion remains a major challenge [1][2][3][4]. Toward that end, we constructed a microfabricated thermoelectric controller for rapid and reversible actuation using myosin, the biomolecular motor of muscle, and its partner filament, actin [5,6]. Thermal control of kinesin-microtubule motility using thermally responsive polymers has also been demonstrated as a possible temperature-dependent control mechanism for biomolecular motors [7]. We present here novel applications of our thermoelectric control system to investigate changes in actin-myosin motility when Ca 2+regulated thin (actin) filaments are used, and when thinfilament Ca 2+ -regulatory protein mutations are introduced that underlie an inherited cardiovascular disease, familial hypertrophic cardiomyopathy (FHC) [8][9][10]. Results provide insights into mechanisms of altered function when thin filaments are reconstituted with human cardiac Ca 2+ -regulatory proteins troponin (Tn) and tropomyosin (Tm), and further modulated by FHC-related mutations in cardiac Tn.
In the heart-and in vertebrate striated muscles in general-contraction is physiologically controlled by binding of cytoplasmic Ca 2+ to thin filaments. Thin filaments are comprised of three major proteins: actin, Tm, and Tn that Journal of Biomedicine and Biotechnology is itself a complex of three polypeptides [11]. Ca 2+ activates thin filaments by binding to Tn, leading to structural changes that ultimately expose myosin-binding sites on actin subunits [11]. Our previous work on the thermoelectric controller utilized only myosin and unregulated F-actin, that is, Ca 2+ -independent motility achieved using actin filaments without Tn or Tm [5,6]. This study was initiated because Ca 2+ -regulated thin filaments could have several advantages over unregulated F-actin in a thermoelectric nanoactuator. First, Ca 2+ would provide a secondary control mechanism that is separate from temperature. An additional benefit is that the Ca 2+ -regulatory proteins Tn and Tm at saturating Ca 2+ can enhance function of the actomyosin system [12][13][14][15][16][17][18][19][20]. These effects of Tn and Tm on actomyosin function, however, appear to vary markedly over the limited temperature ranges studied previously [16,18,21] and thus could introduce nonlinearities to a thermoelectric nano-actuator.
Temperature has obvious physiological relevance as a determinant of mammalian striated muscle contraction. It affects several fundamental aspects of muscle structure and function [22][23][24] although temperature effects on muscle function appear to be greater at low temperatures used in many experiments than at physiological temperature; examples are isometric tetanic tension [25][26][27] and maximum mechanical power output [28]. Temperature also modulates the control of striated muscle contraction by Ca 2+ regulatory proteins Tn and Tm, although there are both qualitative and quantitative discrepancies on this point in the striated muscle literature [22,26,29,30]. Interestingly, mild hypothermia has been reported to be a positive inotropic effector in living cardiac muscle due to interplay between temperature and cardiac thin filament Ca 2+ regulation [31]. Our thermoelectric controller allows us to address these issues and more in a single experiment through rapid exploration of broad temperature regimes surrounding physiological temperature.
Some mutations in Tn and Tm alter thin-filament function in ways that might make the mutant proteins useful in biomolecular motor-based transport applications. In particular, mutations associated with familial hypertrophic cardiomyopathy (FHC)-one of a growing number of diseases caused by mutations in the genes for sarcomere proteins including thin-filament Ca 2+ -regulatory proteins [8][9][10]may be especially useful in bionanomechanical systems even though those mutations are devastating for patients [32][33][34][35][36][37][38][39][40][41]. The molecular basis by which these mutations alter thin-filament function and bring about cardiovascular disease is unknown [8,10,42]. Our thermoelectric controller could provide important insight into mechanisms by which these mutations in myofilament proteins alter contractile function, and directly or indirectly signal changes in gene expression that underlie the progression to pathological cardiac hypertrophy in FHC. Use of Laboratory Animals. Briefly, rabbits were injected IM with 3 mg Acepromazine + 10 mg/kg Xylazine + 50 mg/kg Ketamine. Following verification of appropriate surgical depth of anesthesia, the rabbits were exsanguinated via laceration of the carotid artery. The euthanized animals were skinned, eviscerated, and chilled on ice. Back and leg muscles were removed for acetone powder, or back muscles only for myosin preparation [12,13,43]. Myosin and the soluble chymotryptic digest fragment heavy meromyosin (HMM) were prepared from fast skeletal muscle as described [12,13,43,44]. HMM was made from myosin stored <5 weeks in glycerol (1 : 1) at −20 • C; HMM was used within 3 days after preparation. Rabbit muscle actin was purified from muscle acetone powder as described [12,13,43,45]. F-actin was labeled with rhodamine phalloidin (RhPh; Molecular Probes) for fluorescence microscopy (Figure 1(a)). F-actin (33 μg/mL) and RhPh (1 μM) were incubated in actin buffer (AB) with 1 mM DTT at 4 • C overnight; fluorescently labeled F-actin was stable at 4 • C for at least 2 weeks.

Materials and Methods
Recombinant human α-Tm was expressed in E. coli as a homodimeric fusion protein with maltose binding protein (MBP); α-Tm was purified following removal of the MBP affinity tag via thrombin cleavage [13,33]. After removal of the MBP tag, recombinant α-Tm has two extra amino acids (GS-) at the N-terminus; GS-is a conservative alternative to the AS-dipeptide in bacterially expressed Tm that substitutes functionally for acetylation of Tm's N-terminus [46,47]. Purified Tn from human cardiac muscle (cTn) was obtained from Research Diagnostics (Flanders, NJ), or coexpressed recombinantly (rHcTn) in E. coli as a fusion protein with glutathione S-transferase (GST); the ternary rHcTn complex was purified following removal of the GST affinity tag via cleavage with TEV protease [13]. FHC mutations of rHcTn were introduced via site-directed mutagenesis to the bacterial coexpression plasmid; changes were verified by DNA sequencing. Proteins were assessed by coomassiestained Tricine-SDS PAGE [48] and quantitative analysis with a Kodak EDAS 290 digital imaging system.

In Vitro Motility
Assays. The speed of RhPh F-actin over HMM-coated surfaces was measured to assess the kinetics of actomyosin interactions and, for reconstituted thin filaments, their regulation by Ca 2+ . All aspects of the motility experiments with unregulated F-actin, such as flow cell assembly, solution preparation, assay procedures, and data collection were conducted as described [5,13,43,49].
[HMM] applied to flow cells was 250 μg mL −1 in the majority of assays to achieve high density of HMM (ρ) on the nitrocellulose-coated motility surface; to reduce ρ in a limited set of experiments, 75 μg mL −1 HMM was applied to the flow cell. Temperature ( • C) ε 1 ε 2 Figure 1: Temperature dependence of HMM-driven sliding of RhPh-labeled actin filaments determined using a microfabricated thermoelectric heater. (a) Modified flow cells were constructed as described [5]. Flow cells consisted of a microscope slide (light blue) and a modified cover slip (gray and gold), separated by two, thin spacers and held together by nonfluorescent vacuum grease. Solutions (Section 2.2) are pipetted through the flow cell as indicated by the blue, dashed arrows. To warm the flow cell, current is injected via circuit ε 1 into a microfabricated thermoelectric heater (gold) on the coverslip's outer surface; flow cells are cooled by circulating chilled water through a copper coil wrapped around the microscope objective. Temperature is monitored via a linear, resistive thermometer (dark gold structure) that was microfabricated on the inner face of the coverslip (i.e., on the same side where actomyosin motility occurred, and opposite to the side with the thermoelectric heater); temperature is obtained from a calibrated voltage readout (V) with bias current set by circuit ε 2 .
Motion of fluorescently labeled filaments on the inner side of the coverslip (top micrograph) is observed and recorded through a window (black, dashed arrow) located immediately adjacent to the thermometer. (b) Time course of temperature (gray line; right ordinate) and sliding speed (solid symbols; left ordinate) for regulated thin filaments (F-actin plus native cTn and α-Tm) at pCa 5 during the heating phase of three cycles of heating and cooling (Section 2). Points represent mean ± S.D. of the sliding speed for 3-8 filaments during 1 s. Note that the upper limit of the temperature-dependent increase in sliding speed occurs at ∼54 • C (t > 4.5 min). (c) The data from (b) were replotted to eliminate time: cycle 1 ( ), cycle 2 ( ), and cycle 3 ( ). An additional dataset for unregulated F-actin sliding is shown: cycle 1 ( ) and cycle 2 ( ). Note that the sliding speed of regulated thin filaments at the highest temperature is almost the same as for unregulated F-actin.  [12,14]. Methylcellulose (final concentration 0.3%), dialyzed against sodium azide (0.02%), was added to MB to prevent actin filament diffusion away from the assay surface [12,49,51]. Solutions for assays with reconstituted thin filaments were the same as for unregulated F-actin, except for the last two solutions: the buffer applied to the flow cell before MB contained 75 nM each of α-Tm and either native cTn or recombinant hcTn and was incubated for three minutes before infusing MB containing the same concentrations of Tn and Tm [12,13,32,50]. The minimum concentrations of native or WT Tn and Tm added to MB to obtain wellregulated filaments at 30 • C were determined by titrations and applying the following criteria: filament sliding was [Pi] in columns 3 and 5 were estimated using K eq = 0.06 for the reaction sucrose + Pi ↔ α-D-glucose-1-phosphate + D-fructose [52,54].
[ADP] estimates are based on the equilibrium constant (K eq ) for the reaction ATP + Cr ↔ ADP + CP, and are given in columns 3-4 with lowest and highest value corresponding to the lowest (27 • C) and highest (42 • C) temperatures employed, respectively. The estimate of [ADP] in column 2 was based on Chase and Kushmerick [53]. 1 Motility buffer (MB) employed in the majority of experiments (data in Figures 1-4) did not include sucrose phosphorylase and creatine kinase. 2 Modified motility buffers were used for experiments shown in Figure 5.
In a limited set of motility experiments, standard MB was modified to vary metabolite concentrations (Table 1) as follows: either 5.3 μg mL −1 sucrose phosphorylase and 10 mM sucrose were added to reduce inorganic phosphate (Pi) [52] or acetate was partially replaced by Pi to achieve the high Pi condition (sucrose phosphorylase was not added); contaminating ADP was reduced to <1 μM by addition of creatine phosphate (CP) and creatine phosphokinase (CK) [53]. Motility data were collected while varying temperature by employing modified flow cells containing microfabricated Au heater and thermometer elements (Figure 1(a)), where the thermometer is located immediately adjacent to the viewing region [5]. In experiments using the thermoelectric heater, cooling was achieved by circulating chilled water through a copper coil wrapped around the microscope objective; in a limited set of experiments, assays at constant temperature were achieved by circulating temperaturecontrolled water through the coil surrounding the objective [13,43]. Motility speed was analyzed using MetaMorph software (Universal Imaging) as described [5]. Stacks of frames (one stack for each second of temperature transient data, or 10-12 stacks for each constant temperature experiment from one flow cell) were created from digitized movies.

Statistical Analyses.
Averages and error estimates were calculated with Microsoft Excel 2000. Averages are given as unweighted mean ± SD. Linear regression analyses were performed using Microsoft Excel 2000 or SigmaPlot (Windows versions 8.0 and 11.2; SPSS Inc., Chicago, IL); nonlinear regression analyses were performed using SigmaPlot.
A transition temperature (T t ) was estimated from Arrhenius transformed speed versus temperature data that exhibited a change in slope with temperature. To obtain T t , the data were divided into two parts (high-and lowtemperature regimes) that were separated by two adjacent points on the Arrhenius plot's temperature axis; several such divisions of each dataset were processed using the following algorithm. Linear regressions were obtained for each regime, and the intersection point determined for those regressions. In general, the intersection point does not fall between the two data points that separate the high and low temperature regimes. The small number of intersection points that did fall between the high-and low-temperature regimes were considered to be candidates for T t ; in the few instances when there was more than one candidate point, the median value was chosen for T t .

Temperature Dependence of Unregulated F-Actin and Regulated Thin-Filament Sliding.
To examine sliding of Ca 2+regulated thin filaments over a broad range of temperatures, we employed a thermoelectric controller (Figure 1(a)) that enables both rapid and continuous variation of a flow cell's temperature [5]. Figure 1(b) shows sliding speed (solid symbols) as a function of time for thin filaments reconstituted with native cTn·α-Tm at pCa 5 during three heating phases of cyclic heating and cooling (∼20-60 • C). Note that the peak temperature (gray line) increased with each cycle. Speed increased with temperature during the heating phase of each cycle over most of the range examined. As temperature increased during the third cycle, however, sliding speed of regulated actin reached a maximum at ∼54 • C, and then speed declined above ∼58 • C (Figure 1(b)). This anomalous decline in speed for regulated thin filaments is particularly evident in Figure 1(c), solid symbols where data from all three heating cycles of Figure 1(b) were combined to show the temperature dependence of sliding speed. Superposition of the data from three cycles suggests that brief exposure to elevated temperature over at least the first two cycles was fully reversible.
Comparison of regulated versus unregulated F-actin sliding speed at high temperature suggests that the anomalous decline for regulated thin-filament speed is more likely due to an effect of temperature on Tn·Tm rather than on actin or HMM. Regulated filament speed at pCa 5 (solid symbols) was at least as fast as that for unregulated Factin (open symbols) at each temperature over the range examined (∼20-60 • C) (Figure 1(c)), consistent with previous reports over more limited temperature ranges [12,13,16,20]. The speed of unregulated F-actin, however, increased continuously over the entire temperature range-even at the highest temperatures examined, where regulated thinfilament speed declined. To test the hypothesis that Tn·Tm is responsible for the anomalous decline in speed (Figure 1), we examined the effect of temperature on regulated thinfilament motility at pCa 9, that is, in the absence of activating Ca 2+ . Figure 2(a) shows sliding speed (red symbols) as a function of time for thin filaments reconstituted with rHcTn·α-Tm at pCa 9 during five heating phases of cyclic heating and cooling. Note that the peak temperature (gray line) increased with each cycle, as in Figure 1(b). Data from all five heating cycles of Figure 2(a) were combined to show the temperature dependence of sliding speed in Figure 2(b). As expected in the absence of Ca 2+ , regulated thin filaments exhibited little or no motion (note red symbols at speed ∼0) over a broad range of temperatures (∼20-43 • C). Above ∼43 • C, filament sliding was observed indicating loss of Ca 2+ regulation (red symbols; Figure 2) although the average speed was generally less than or similar to that of unregulated F-actin (open symbols; Figure 2(b)). As for native cTn (Figure 1(c), solid symbols), rHcTn WT regulated thin filaments at pCa 5 ( Figure 2(b), solid symbols) were faster than unregulated F-actin ( Figure 2(b), open symbols); an anomalous decline in speed with WT rHcTn (Figure 2(b), solid black circles) occurred at slightly a higher temperature (by ∼4-5 • C) than for native cTn (Figure 1(c), solid symbols), suggesting increased stability of the recombinantly expressed protein or its interactions with F-actin-α-Tm. The apparent loss of Ca 2+ -regulation, indicated in Figure 2 by thin filament sliding (nonzero speeds) at pCa 9 for temperatures above ∼43 • C, is completely reversible because thin-filaments stopped moving after temperature was lowered during the cooling phase of each cycle at pCa 9 ( Figure 2(a)). Arrhenius analysis of the data in Figure 1(c) reveals that the activation energy (E a ) for regulated thin-filament speed at pCa 5 exhibited two distinct values, with a change in the slope at T t ∼ 38 • C (Figure 3, solid symbols and black solid lines). For regulated thin filaments (Figure 3, solid symbols and black solid lines), E a at temperatures below physiological (100.4 kJ/mol) was >2-fold greater than E a for the highertemperature regime (41.8 kJ/mol), while the latter value was closer to that of unregulated F-actin (61.9 kJ/mol) (Figure 3, open symbols and black dashed line). Temperature has a dramatic, nonlinear effect on the viscosity of water [55], and the speed of filament sliding varies inversely with solvent viscosity [43,56]; we therefore asked whether temperaturedependent changes in solvent viscosity could explain the nonlinear Arrhenius relation (Figure 3). The data in Figure 3 were processed to remove the temperature dependence of solvent viscosity by first normalizing viscosity with respect to that of water at 37 • C (i.e., body temperature), and second assuming that speed varies inversely with solvent viscosity [43,56]. After removing the effects of viscosity, slopes of the Arrhenius plots were reduced (Figure 3, blue lines).  Figure 1), with data omitted for regulated thin filaments above ∼54 • C and filament sliding speeds expressed in μm s −1 . Dashed lines (black) are linear least squares regressions on the data before correction for temperature-dependent changes in solvent viscosity. Note the change in slope at ∼38 • C (T t ) for regulated thin filaments. Effects of temperature-dependent changes in solvent viscosity (normalized to that at body temperature, 37 • C) were removed by assuming that speed varies inversely with viscosity, as demonstrated experimentally [43]; the net result of this correction in all instances is a decrease in apparent E a .
speed and viscosity [43,56]. Despite the reduction of slopes for regulated thin-filaments, the ratio of E a 's between the two temperature regimes increased from 2.4 to 3.1.
Below 38 • C, the ratio of E a 's for regulated thin-filament sliding speed with respect to that of unregulated F-actin increased from 1.6 to 1.8. It is clear from this analysis that it is important to consider temperature-dependent changes in solvent viscosity when evaluating actin filament sliding speed. Nonlinearities in the Arrhenius plot (Figure 3), however, cannot be explained by temperature-dependent changes in solvent viscosity.

Influence of FHC Mutations in Troponin on the Temperature Dependence of Thin-Filament Sliding.
Because the temperature dependence of maximum filament sliding speed is influenced by the presence and function of Tn·Tm ( Figures  1-3), and FHC-related mutations in troponin profoundly influence the Ca 2+ dependence of myofilament function [8][9][10]57], we utilized the thermoelectric controller to investigate temperature effects on thin filaments reconstituted with rHcTn·α-Tm when one of three FHC-related mutations-TnI R145G, TnI K206Q, or TnT R278C-was incorporated into the troponin complex (Section 2). At pCa 5, filament sliding speed for two mutants (TnI K206Q and TnT R278C) was faster than WT over the range of temperatures examined, consistent with previously reported data for TnI K206Q using recombinant rat cardiac Tn [32]. In contrast, speed with TnI R145G was similar to WT ( Figure 4). The micromechanical thermal assays with thin filaments containing FHC mutations in rHcTn suggest that either the mutant troponins are intrinsically less thermally stable, or the regulatory interaction of mutant troponin with F-actin-α-Tm is less stable, or both (Figure 4(a)). All three of the FHC mutants exhibited anomalous declines in speed at lower temperatures than WT rHcTn, with the cTnT R278C and cTnI R145G mutations resulting in ∼4-5 • C decreases, and the cTnI K206Q mutation leading to a larger, ∼8-9 • C decrease in the temperature at which the anomalous decline in speed occurred as temperature continued to rise (Figure 4(a)).
Arhennius analysis suggests that there was little effect of the FHC mutations on T t for thin filaments reconstituted with WT or mutant rHcTn, with values of 38-41 • C where they could be unambiguously determined (Figure 4(b)). This was similar to the value of T t = ∼38 • C obtained with native cTn (Figure 3). The slopes of the high-and lowtemperature regimes were more similar for the two fastest mutants (cTnT R278C and cTnI K206Q) compared with WT or the cTnI R145G mutant (Figure 4(b)); in fact, it was not possible to unambiguously identify a value of T t from data with TnI K206Q by the algorithm used (Section 2). When the data for unregulated F-actin from Figure 2(b) (open symbols) were plotted in Arrhenius form (Figure 4(b), gray dashed lines), a value of T t ∼ 33 • C was obtained in contrast to the dataset in Figures 1 and 3, where T t for unregulated F-actin could not be determined because there were not sufficient points below T t to define a lowtemperature regime.

Influence of Factors that Affect Cross-Bridge Number and Kinetics on Temperature Dependence of Unregulated F-Actin and Regulated Thin-Filament Sliding Speed.
To address whether changes in the number of cross-bridges and distribution of cross-bridge states could be responsible for differences in sliding speed and E a between unregulated Factin and regulated thin filaments at saturating Ca 2+ , we studied the effects of altered HMM density, and metabolite concentrations which are known to affect the fraction of   Figure 4: Influence of FHC-linked mutations in human cardiac troponin on temperature dependence of thin-filament sliding at pCa 5. Thin filaments were reconstituted with α-Tm and either rhcTn WT (circles, black), rhcTn TnI,R145G (triangles, green), rhcTn TnI,K206Q (squares, yellow), or rhcTn TnT, R278C (diamonds, blue). Data were obtained during the heating phases of multiple heating/cooling cycles, as in Figure 1; note that WT data were replotted from Figure 2 weakly versus strongly bound cross-bridges. Solutions were modified in the following ways, and compared with a control condition that was itself modified in comparison with assays reported in Figures 1-4 Figure 5). In these experiments only, motility assays were performed at steadystate temperatures of 27, 32, 37, or 42 • C so that we could calculate metabolite concentrations in the presence of CP/CK to buffer ATP/ADP and reduce ADP (Section 2), conditions that did not apply to the nonsteady state temperature assays reported in Figures 1-4. Sliding speed increased with temperature for each condition tested for both regulated thin filaments at pCa 5 and unregulated F-actin ( Figure 5) as expected for the temperature range examined. Sliding speed with respect to the Control condition was reduced significantly for the low ATP condition at the two highest temperatures for unregulated F-actin ( Figure 5(a)) and at the three highest temperature points for regulated thin filaments ( Figure 5(b)). Sliding speed at high Pi ( Figure 5) was higher than in Control conditions at the two lower temperatures for both regulated thin filaments and unregulated F-actin. Decreasing ρ had little effect on unregulated F-actin sliding speed ( Figure 5(a)) but decreased regulated thin-filament speed significantly at the three lowest temperatures (27,32, and 37 • C, Figure 5(b)). Addition of Tn·Tm was typically associated with increased speed, but not at low ATP, which is in good agreement with the results of Homsher et al. [16]. At reduced ρ, the enhancing effect of Tn·Tm was greater at higher temperatures. The fraction of motile filaments at reduced ρ increased over the three lowest temperatures (27,32, and 37 • C) for unregulated F-actin (inset, Figure 5(a)), but over the entire temperature range for regulated thin filaments (inset Figure 5(b)).

Discussion
In this study, we utilized microfabricated thermoelectric devices to rapidly and reversibly investigate the influence of temperature on myosin-driven sliding of Ca 2+regulated, muscle thin filaments over a wide temperature range (∼20-60 • C). To investigate molecular mechanisms of normal cardiac function and pathophysiology, thin filaments were reconstituted with human cardiac Tn and Tm, or with FHC-related mutants of human cardiac Tn. The major results of this study are as follows.
(a) At pCa 5, sliding speed of regulated thin filaments increased reversibly with increasing temperature (∼20-∼54 • C) as expected and was faster than unregulated F-actin over that range.
(b) Reversible dysregulation-loss of Ca 2+ -regulationoccurred at both pCa 5 and pCa 9 at the upper end of the temperature range examined; this suggests thermal unfolding of the cardiac regulatory proteins and/or their dissociation from actin. (c) Arrhenius analysis indicates two temperature regimes with different E a ; the transition temperature between those regimes was higher for regulated thin filaments than unregulated F-actin, and this is explained by a simple model.

(d) Factors that alter cross-bridge number and kinetics-[ATP], [Pi]
, and ρ HMM -alter the temperature dependence of sliding speed of both regulated thin filaments and unregulated F-actin in qualitatively similar ways that provide insight into rate limiting factors.
(e) Three FHC mutants in cTn (cTnI R145G, cTnI K206Q, and cTnT R278C) destabilize regulatory protein structure and/or interactions with F-actin-Tm and may also alter the kinetics of thin filament sliding at pCa 5.

Functional Assays for Thermal Stability of Regulated
Thin Filaments and Ca 2+ -Regulatory Proteins. A custom thermoelectric controller [5] that was used in this study (Figure 1(a)), and miniaturized versions thereof [6], provide a novel approach to evaluate contractile protein function over a wide range of temperatures and multiple temperature cycles. The multiple datasets collected this way with native and WT regulatory proteins (Figures 1-3) complement and expand the range of information that can be obtained using conventional motility assays [16,58,59] and traditional approaches to studying protein structure in solution [60][61][62][63]. Our device provides a simpler readout of temperature than relying on intensity of the fluorescent label [64]. Speed of unregulated F-actin increased monotonically over the temperature range used (Figures 1-2, open symbols), indicating that HMM and actin are not affected detrimentally at temperatures below 60 • C; this is consistent with previous functional measurements [5,64] and structural assays showing that the structure of Ph-stabilized F-actin in solution is thermally stable up to 78 • C [62], while unfolding of subfragment 1 (S1) was reported for temperatures greater than 58 • C in the presence of analogs of ATP or ADP·Pi [63]. Dysregulation of thin filaments at elevated temperatures was observed at both pCa 5 (Figures 1-2) and pCa 9 (Figure 2), suggesting that altered function was likely due to dissociation of Tn·Tm from actin or denaturation of Tn and/or Tm. At pCa 5, the anomalous decline in sliding speed for regulated thin filaments above ∼54 • C suggests a loss of regulation (dysregulation) because as temperature increased, speed decreased toward that of unregulated Factin (Figure 1(c)). At pCa 9, filament sliding was observed at temperatures above ∼43 • C; speed increased with temperature and was generally lower than, but approached that of unregulated F-actin (Figure 2), indicating that cTn stabilizes actin·Tm at pCa 5. Taken together, these results from microthermal heater assays provide functional correlates that are consistent with previous solution studies showing half-maximal dissociation of Tm from Ph-stabilized F-actin·Tm complex (no Tn) at 46-48 • C, and unfolding of Tm alone in solution that was resolved into two separate calorimetric domains with maxima of 42 and 51 • C [61,62]. More generally, our results demonstrate that microthermal Journal of Biomedicine and Biotechnology Temperature ( • C) Figure 6: Temperature dependence of two apparent rate constants that may limit actin filament sliding speed at pCa 5. Data in Figure 3 (T < ∼54 • C) for sliding speed of regulated thin filaments were fitted to (2). The inverse of each of the exponential rate terms in the denominator of (2) is plotted: for WT regulated thin filaments, one rate exhibits only slight temperature sensitivity (solid black line) while the other is markedly temperature sensitive (red line). The rate-limiting step for filament sliding is the smaller of the two rates, and T t is the intersection of the two lines. The less temperature sensitive rate is also shown for unregulated F-actin (dashed black line; data from Figure 3). Note that by assuming the temperature-sensitive rate (red line) is similar for both instances, decreasing the less temperature sensitive rate (black arrow, as observed experimentally with unregulated F-actin) decreases T t .
heater assays can provide unique, molecular level insights into muscle thin-filament function, particularly when the maximum temperature is limited such that the effects of temperature on Ca 2+ regulation are fully reversible (Figure 2(a)).

Temperature Dependence of Ca 2+ -Regulated Thin-Filament Sliding at pCa 5.
Comparison of temperaturedependence of regulated thin-filament sliding at pCa 5 with that of unregulated F-actin provides insights into the roles of cardiac troponin and tropomyosin. This is important in light of the observation that not only do Tn-Tm modulate the thin filament side of actomyosin interactions but cTn can also directly accelerate myosin ATPase activity even in the absence of actin and Tm [20]. This direct effect of cTn on myosin kinetics is presumed to at least partially underlie Tn·Tm's enhancement of maximum unloaded sliding speed under a wide variety of conditions (e.g., Figures 1(c), 2(b), 3, 4(b), and 5), although it is also possible that there are differences in the myosin binding interface due to Tm·S1 interactions [65] or regulatory protein-induced structural changes in actin itself [15]. While it is clear that Tn·Tm modulate filament sliding speed primarily by affecting the number of cross-bridges at submaximal Ca 2+ activation [12,14,50], the most straightforward mechanism by which Tn·Tm could directly enhance actomyosin kinetics at saturating Ca 2+ is through a reduction in the time that cross-bridges spend in the strongly bound state (t s ) under the unloaded conditions of a motility assay, as observed experimentally [33]. The maximum sliding speed (s m ) for regulated thin filaments at saturating Ca 2+ or for unregulated F-actin is simply related to myosin step size (d x ) and t s as proposed by Uyeda et al. [51]: To accommodate the observed, biphasic Arrhenius plots (Figures 3 and 4(b)), we assume that d x = 0.0055 μm [66] for all conditions examined and rewrite (1) to express t s as the sum of two exponential terms: which, when fit to the data for regulated thin filaments in Figure 3, yields s m = 0.0055/(0.301e −0.1855T + 0.00145e −0.0277T ) with R 2 = 0.97. Separating the two terms and plotting them as rates, we can see that one varies little with temperature (black solid line; Figure 6) and while the other is highly temperature sensitive (red solid line; Figure 6). Assuming that the more highly temperaturedependent term is similar for both regulated thin filaments and unregulated F-actin, regression estimates were obtained for unregulated F-actin sliding in Figure 3, c = 0.0225 and d = 0.0676 (R 2 = 0.97), and the corresponding rate prediction was plotted in Figure 6 (black dashed line).
If the intersections of the red line with the two black lines in Figure 6 are related to discontinuities in the Arrhenius plots (Figures 3 and 4(b)) and thus values of T t , then the simple analysis of (1), (2) is in accord with observations that the transition from high to low E a will occur at higher temperature for regulated thin filaments (red and solid black lines) than for unregulated F-actin (red and dashed black lines). In this model, the rate limitation for filament sliding at a given temperature would be the lower of the two rates from the red line and one black line: at low temperatures, the steeper, red value is limiting, while the shallower, black value is limited at elevated temperatures ( Figure 6). The temperature dependence of unloaded filament sliding speed has been reported to vary as a function of temperature in some but not all conditions [16,58,59]. For unregulated F-actin, Arrhenius plots have been previously reported to be linear within 20-60 • C [5] and 7-25 • C [16], but nonlinear within 3-42 • C with a transition temperature (T t ) of 15.4 • C [59], and 10-35 • C with a T t ∼ 25 • C [58,67]. T t depends not only on the experimental conditions but also on whether Ca 2+ -regulatory proteins are present. The Arrhenius plot of fish myofibrillar ATPase at saturating Ca 2+ exhibited a discontinuity [68], while this was not the case for in vitro motility sliding speed and actin-activated MgATPase activity with fish myosin and unregulated F-actin [69]. For regulated thin filaments, Homsher et al. [16] reported a break in the Arrhenius plot around 12.5 • C that was not observed for unregulated F-actin within the 7-25 • C temperature range. In this study using micromechanical thermal assays, discontinuities in Arrhenius plots for regulated thin filaments (Figures 3 and 4(b)) occurred at a higher temperature than when T t could be identified for unregulated F-actin (Figure 4(b)). While there is no universal agreement in the limited data available on absolute values for T t for filament sliding, there does appear to be general agreement that Tn·Tm modulates the temperature dependence of unloaded filament sliding. Our data suggest that this modulation of T t by human cardiac Tn·Tm occurs around physiological temperature.
The existence of discontinuities in Arrhenius plots suggests that different factors are rate-limiting to filament sliding above and below T t . What factors could contribute to this difference? Prior experiments with regulated thin filaments at subsaturating levels of Ca 2+ suggested that the number of cross-bridges is limiting to sliding speed at 30 • C [12,50]. The data in Figure 5 indicate that sliding speed is markedly sensitive to a reduction in the density (ρ) of HMM on the flow cell surface at 27 and 32 • C for both unregulated F-actin and especially regulated thin filaments at maximal Ca 2+ activation. At higher temperatures (37 and 42 • C), however, there was little or no effect of the same reduction in ρ on sliding speed for either unregulated Factin or regulated thin filaments ( Figure 5). In addition, the fraction of moving filaments at low ρ increased with temperature for regulated thin filaments at pCa 5 over the entire temperature range ( Figure 5(b), inset), while for unregulated F-actin, the fraction started at a lower value and plateaued ( Figure 5(a), inset). Taken together, these observations suggest an increase in strong cross-bridges with temperature for both unregulated F-actin and regulated thin filaments, consistent with previous reports using in vitro assays and permeabilized muscle fibers [70][71][72]. Thus, it appears that Tn·Tm can modulate at least one rate-limiting factor, and also T t , around physiological temperature: once a temperature (T t ) is attained at which the availability of crossbridges is great enough to guarantee continuous filament sliding, further increase of speed with temperature will only depend on less temperature-sensitive kinetic rates that are associated with transitions between strongly bound crossbridge states, and will therefore be characterized by low E a (Figure 6, cross-over from red to black lines).
In addition to the number of cross-bridges, filament sliding speed can be altered by variations in the substrate and products of the ATPase reaction through modulation of actomyosin kinetics. According to the model implicit in (1), changes in metabolite concentration would primarily influence t s , and therefore s m , and also the number of attached cross-bridges. For example, Tn·Tm could affect the kinetics of ATP-induced cross-bridge dissociation that has been suggested to limit filament sliding speed, at least in fast skeletal muscle at low temperature [16,73]. When [ATP] was reduced to 200 μM-a substrate concentration that is similar to the apparent K d for sarcomere shortening velocity and filament sliding speed [16,53,74,75]inhibition of speed was greater for regulated thin filaments at pCa 5 than unregulated F-actin ( Figure 5), as reported by Homsher et al. [16]. Inhibition of s m at low ATP levels is explained by slower ATP-induced detachment of crossbridges at the end of attached phase the cycle, which increases t s (1); the fraction of strongly attached cross-bridges would also increase, although the dominant factor in determining s m in this instance is slower cross-bridge cycling. The relatively greater inhibition of regulated thin-filament sliding was accentuated at higher, more physiological temperatures ( Figure 5). The net result was a generally shallow slope of speed versus temperature for both regulated thin filaments and unregulated F-actin ( Figure 5) that corresponds to a regime of relatively low E a . The expectation would then be that reduced [ATP] should shift T t to a lower temperature ( Figure 6) although it was not resolved in this experiment, likely because of the lower limit of temperature examined.
There are some interesting parallels with reduced [ATP] effects, along with significant differences, when the concentration of product Pi was increased from micromolar levels to 4 mM (Table 1), that is, the highest concentration that we could achieve without necessitating an increase in ionic strength. Temperature has previously been shown to modulate the effect of Pi on isometric force in muscle fibers, although there is disagreement about whether Pi is either more [29,76] or less [77] inhibitory of isometric force as temperature increases. While Pi inhibits isometric force, elevated Pi increases sliding speed for both regulated thin filaments at pCa 5 and unregulated F-actin, although the effect predominates at lower temperatures and is more pronounced for regulated thin-filaments ( Figure 5). Maximum velocity of sarcomere shortening [52,70,78,79] and unloaded sliding of unregulated F-actin [80] have previously been reported to increase with elevated Pi. This is generally understood to be the result of reversal of crossbridge attachment, which decreases t s as demonstrated at the single molecule level by Baker et al. [81], thus reducing the number of strongly attached cross-bridges and increasing s m (1). The relatively greater enhancement of sliding speed at lower temperatures results in a generally shallow slope of speed versus temperature for both regulated thin filaments and unregulated F-actin ( Figure 5) that corresponds to a regime of relatively low E a ; in contrast to the result with low ATP, the expectation is that elevated Pi should shift T t to a higher temperature ( Figure 6). Taken together, the data in Figure 5 illustrate that the relative importance of factors which influence cross-bridge cycling kinetics and/or filament sliding speed varies with temperature around the physiological range, and which are rate-limiting also varies as implicated in Figure 6. This suggests that energetic state may impact cardiac function when core body temperature varies, such as in mild hypothermia or with exercise, in ways that may seem counterintuitive [31]. Figure 4 illustrates that the microthermal heater can be used to study effects of disease-related mutations in human cardiac Ca 2+ regulatory proteins on function of individual thin filaments. This clearly goes beyond, and will ultimately help explain, prior work that examined physical properties of the mutant proteins in bulk solution, and the Ca 2+dependence of function. cTn mutations may influence not just s m (TnI K206Q and TnT R278C), but also the temperature dependence of E a (TnI K206Q and TnT R278C) and possibly T t (TnI K206Q). The mutations may also influence the temperature at which thin-filament structure is affected to the extent that dysregulation occurs and Ca 2+ dependent function is altered, whether the disease-related change in function is most relevant in the myofilaments, or perhaps in the nucleus [82]. For all three Tn mutants examined (TnI R145G, TnI K206Q, and TnT R278C) mutations destabilized thin-filament structure and function at lower temperatures than WT; this effect was particularly noticeable for an FHC mutant of human cardiac Tm (E180G) that causes extreme shifts in Ca 2+ sensitivity [33,40]. Thus, we anticipate that the thermoelectric heater will be generally useful for characterization of clinically relevant mutants of thin filament protein structural stability through assessment of function at the level of individual thin filaments.

Conclusion.
Results demonstrate the utility of our thermoelectric controller for investigating molecular mechanisms underlying biomolecular motor function, and cardiovascular diseases related to altered biomechanics of cardiac myofilament proteins. This assay provides a novel view of structure-function relationships for cardiac thin filaments.