HIV-Tat Exacerbates the Actions of Atazanavir, Efavirenz, and Ritonavir on Cardiac Ryanodine Receptor (RyR2)

The incidence of sudden cardiac death (SCD) in people living with HIV infection (PLWH), especially those with inadequate viral suppression, is high and the reasons for this remain incompletely characterized. The timely opening and closing of type 2 ryanodine receptor (RyR2) is critical for ensuring rhythmic cardiac contraction–relaxation cycles, and the disruption of these processes can elicit Ca2+ waves, ventricular arrhythmias, and SCD. Herein, we show that the HIV protein Tat (HIV-Tat: 0–52 ng/mL) and therapeutic levels of the antiretroviral drugs atazanavir (ATV: 0–25,344 ng/mL), efavirenz (EFV: 0–11,376 ng/mL), and ritonavir (RTV: 0–25,956 ng/mL) bind to and modulate the opening and closing of RyR2. Abacavir (0–14,315 ng/mL), bictegravir (0–22,469 ng/mL), Rilpivirine (0–14,360 ng/mL), and tenofovir disoproxil fumarate (0–18,321 ng/mL) did not alter [3H]ryanodine binding to RyR2. Pretreating RyR2 with low HIV-Tat (14 ng/mL) potentiated the abilities of ATV and RTV to bind to open RyR2 and enhanced their ability to bind to EFV to close RyR2. In silico molecular docking using a Schrodinger Prime protein–protein docking algorithm identified three thermodynamically favored interacting sites for HIV-Tat on RyR2. The most favored site resides between amino acids (AA) 1702–1963; the second favored site resides between AA 467–1465, and the third site resides between AA 201–1816. Collectively, these new data show that HIV-Tat, ATV, EFV, and RTV can bind to and modulate the activity of RyR2 and that HIV-Tat can exacerbate the actions of ATV, EFV, and RTV on RyR2. Whether the modulation of RyR2 by these agents increases the risk of arrhythmias and SCD remains to be explored.


Effects of Pretreating RyR2 with HIV-Tat on Equilibrium [ 3 H]ryanodine Binding to ATV, RTV, and EFV
Having found that ATV, RTV, EFV, and HIV-Tat alter equilibrium [ 3 H]ryanodine binding to RyR2, we then investigated if pretreating RyR2 with a low HIV-Tat to mimic inadequate HIV-suppression would alter the ability of ATV, EFV, and RTV to modulate [ 3 H]ryanodine binding to RyR2.
Preincubating junctional SR vesicles with 14 ng/mL HIV-Tat for 10 min at 37 • C prior to the addition of ATV, EFV, or RTV resulted in leftward shifts in their equilibrium [ 3 H]ryanodine binding curves. The EC 50 for ATV in the presence of HIV-Tat was 5.3-fold lower than that in the absence of HIV-Tat (70.8 ± 17.6 ng/mL; 0.1 ± 0.0 µM; Figure 1B (∆)), with an increase in the Hill slope (1.7 ± 0.1). The EC 50 for RTV in the presence of HIV-Tat was 3.1-fold lower than that in the absence of HIV-Tat (167.3 ± 29.6 ng/mL; 0.08 ± 0.01 µM; Figure 1B ( )), with no change in the Hill slope (0.8 ± 0.1). The IC 50 for EFV in the presence of HIV-Tat was also 3.1-fold lower than that in the absence of HIV-Tat (214.5 ± 37.5 ng/mL; 0.7 ± 0.1 µM; Figure 1B ( )), with a decrease in the Hill slope (0.9 ± 0.1). These data indicate that when HIV-Tat is present in the binding medium, the affinities of ATV, EFV, and RTV for RyR2 are enhanced.

Effects of ATV, EFV, RTV, and HIV-Tat on Ca 2+ Homeostasis in Primary Rat Cardiac Myocytes
Since ATV, EFV, RTV, and HIV-Tat modulated the equilibrium binding of [ 3 H]ryanodine to RyR2, we then investigated if ATV, EFV, RTV, and HIV-Tat could also elicit Ca 2+ transients/waves in rat primary cardiac myocytes. For this, live-cell confocal imaging was conducted with freshly isolated rats that were loaded with the Ca 2+ -sensitive dye Fluo-3.
When antiretroviral-naïve primary rat ventricular myocytes were exposed to 5.0 µM ATV, Ca 2+ transients/waves were seen within one minute after the addition of the drug. The Ca 2+ waves elicited by ATV persisted for the duration of the recording (60 s). The myocyte contractions and relaxations repeated during the recording. In the four chambers of cells (>10 myocytes per 10× frame per chamber) investigated, the Ca 2+ waves and contraction/relaxation stopped 240 s after ATV exposure. Figure 2A(i) shows the ATVinduced-Ca 2+ waves in a representative myocyte (also see Supplemental Video S1 (60 s of recording)). Figure 2A(ii) shows a summary of the cells in the four separate chambers before and after ATV treatment. Pretreating the antiretroviral-naïve myocytes with 50 µM for 10 min to block RyR2 significantly attenuated the Ca 2+ waves induced by ATV (see Supplemental Video S2 (60 s of recording)), suggesting that ATV elicits Ca 2+ release from the SR.
Exposing antiretroviral-naïve primary rat ventricular myocytes to 5.0 µM EFV also leads to Ca 2+ transients/waves within one minute of being administered. In the four chambers of the cells (>10 myocytes per 10× frame per chamber) investigated, the Ca 2+ waves and contraction/relaxation stopped 140-150 s after the addition of EFV. The myocytes exposed to EFV also showed elevated Ca 2+ originating from regions close to the plasmalemma prior to "permanent cell shortening" within 60 s after EFV addition to the culture medium. Figure 2B(i) shows the EFV-induced Ca 2+ waves in a representative myocyte (also see Supplemental Video S3 (60 s of recording)). It is of interest to note that while the number of Ca 2+ waves elicited by EFV (per unit of time) reduced when compared to ATV, their amplitudes were 25 ± 5% more than that of ATV (comparing Figure 2A(ii) and B(ii)). Figure 2B(ii) shows a summary of the data of the cells in the four separate chambers before and after EFV treatment that exhibited the Ca 2+ waves. Pretreating antiretroviral-naïve myocytes with 50 µM for 10 min to block RyR2 significantly attenuated the Ca 2+ waves induced by EFV (see Supplemental Video S4 (60 s of recording)). However, we did notice an elevation in Ca 2+ originating from those regions close to the plasmalemma of the myocytes prior to "permanent cell shortening", suggesting that EFV elevates cytoplasmic Ca 2+ by nonryanodine receptor mechanisms as well. Time-dependent changes in intracellular Ca 2+ in primary rat ventricular myocytes following the addition of a bolus dose of HIV-Tat, ATV, EFV, and RTV. For this, primary ventricular myocytes isolated from rat hearts were incubated with the Ca 2+ -sensitive Fluo-3; these were placed on the head stage of a laser confocal microscope equipped with 10X and 40 X objectives (Zeiss Confocal LSM 710, equipped with an Argon-Krypton Laser, 25 mW argon laser, 2% intensity, Thornwood, NJ, excitation wavelength 488 nm, and emission wavelengths of 515 nm). Panel A(i) shows the timedependent changes in intracellular Ca 2+ from a representative rat ventricular myocyte when exposed to ATV (5.0 μM). Panel A(ii) shows the mean ± SD for >40 cells in the four separate chambers before and after ATV treatment. Panel B (i) shows the time-dependent changes in intracellular Ca 2+ from a representative rat ventricular myocyte when exposed to EFV (5.0 μM). Panel B(ii) shows the mean ± SD for >40 cells in the four separate chambers before and after EFV treatment. Panel C(i) shows the time-dependent changes in intracellular Ca 2+ from a representative rat ventricular myocyte when exposed to RTV (5.0 μM). Panel C(ii) shows the mean ± SD for >40 cells in the four separate chambers before and after RTV treatment. Panel D(i) shows the time-dependent changes in intracellular Ca 2+ from a representative rat ventricular myocyte when exposed to HIV-Tat (25 ng/mL).

Figure 2.
Time-dependent changes in intracellular Ca 2+ in primary rat ventricular myocytes following the addition of a bolus dose of HIV-Tat, ATV, EFV, and RTV. For this, primary ventricular myocytes isolated from rat hearts were incubated with the Ca 2+ -sensitive Fluo-3; these were placed on the head stage of a laser confocal microscope equipped with 10× and 40× objectives (Zeiss Confocal LSM 710, equipped with an Argon-Krypton Laser, 25 mW argon laser, 2% intensity, Thornwood, NJ, excitation wavelength 488 nm, and emission wavelengths of 515 nm). Panel A(i) shows the timedependent changes in intracellular Ca 2+ from a representative rat ventricular myocyte when exposed to ATV (5.0 µM). Panel A(ii) shows the mean ± SD for >40 cells in the four separate chambers before and after ATV treatment. Panel B(i) shows the time-dependent changes in intracellular Ca 2+ from a representative rat ventricular myocyte when exposed to EFV (5.0 µM). Panel B(ii) shows the mean ± SD for >40 cells in the four separate chambers before and after EFV treatment. Panel C(i) shows the time-dependent changes in intracellular Ca 2+ from a representative rat ventricular myocyte when exposed to RTV (5.0 µM). Panel C(ii) shows the mean ± SD for >40 cells in the four separate chambers before and after RTV treatment. Panel D(i) shows the time-dependent changes in intracellular Ca 2+ from a representative rat ventricular myocyte when exposed to HIV-Tat (25 ng/mL). Panel D(ii) shows the mean ± SD for >40 cells in four separate chambers before and after HIV-Tat treatment. * denotes significantly different (p < 0.05) from the control (before the addition of the drug).
Exposing antiretroviral-naïve primary rat ventricular myocytes to 5.0 µM RTV also resulted in Ca 2+ transients. However, the pattern of Ca 2+ elevation was distinct from that of ATV and EFV. After the addition of RTV, the cytosolic Ca 2+ in most of the myocytes in the four chambers investigated (>10 myocytes per 10× frame per chamber) increased. But the increase was in multiple distinct regions throughout the myocyte without significant Ca 2+ wave generation, which is reminiscent of increases in spontaneous Ca 2+ sparks. This was followed by global Ca 2+ transients with and without wave formation. Figure 2C (i) shows the RTV-induced Ca 2+ waves in a representative myocyte (also see Supplemental Video S5, (240 s of recording)). Figure 2C(ii) shows the summary data of the cells in the four separate chambers before and after RTV treatment that exhibited Ca 2+ waves. Pretreating antiretroviral-naïve myocytes with 50 µM for 10 min to block RyR2 significantly attenuated the Ca 2+ waves induced by RTV (see Supplemental Video S6 (120 s of recording)).
Exposing antiretroviral-naïve primary rat ventricular myocytes to 25 ng/mL HIV-Tat resulted in Ca 2+ transients/waves within one minute after administration to the culture medium in the four chambers of the cells (>10 myocytes per 10× frame per chamber) investigated. In many of the cells, HIV-Tat also elicited a global elevation in myocyte Ca 2+ , which was followed by irreversible cell shortening. Figure 2D(i) shows the HIVinduced Ca 2+ waves in a representative myocyte (also see Supplemental Video S7, (240 s of recording)). Figure 2D(ii) shows the summary data of the cells in the four separate chambers of the myocytes before and after HIV-Tat treatment. Pretreating antiretroviralnaïve myocytes with 50 µM for 10 min to block RyR2 significantly attenuated the Ca 2+ waves induced by HIV-Tat (see Supplemental Video S8). Heat-inactivated HIV-Tat did not alter Ca 2+ homeostasis in rat ventricular myocytes.

Effects of ATV. EFV, RTV, and HIV-Tat on the Gating and Conductance of RyR2
Next, single-channel studies were conducted to gain insights into the effects of ATV, EFV, RTV, and HIV-Tat on the gating and conductance of RyR2. In this study, all agents were added cumulatively to the cis chamber of the bilayer, equivalent to the cytosolic side of RyR2.

Effects of ATV on RyR2
The mean open probability (P o ) and current amplitude of the RyR2 channels used prior to the addition of ATV was 0.08 ± 0.01 and 21.4 ± 1.3 pA (a representative channel is shown in Figure 3; conductance = 712 ± 36 pS; 1.0 µM cis Ca 2+ at +35 mV holding potential (HP); n = 13 channels, respectively). These channels were kept at a low P o to assess the ability of ATV to both open and close RyR2. The cumulative addition of ATV to the cis chamber (704-8448 ng/mL, 1.0-12.0 µM) dose-dependently increased the P o of RyR2 from 0.08 ± 0.03 to 0.65 ± 0.11 within 15 s of addition ( Figure 3A

Effects of EFV on RyR2
The mean P o of RyR2 with 1.0 µM cis Ca 2+ prior to the addition of cis EFV was 0.08 ± 0.02 at +35 mV HP (a representative channel is shown in Figure 3A; top). The mean current amplitude was 21.4 ± 1.6 pA ( Figure

Effects of RTV on RyR2
The mean P o of the 11 RyR2 channels before the cis addition of RTV was 0.06 ± 0.01 at +35 mV HP, and their mean current amplitude was 21.5 ± 1.4 pA, which is equivalent to a conductance = 715.5 ± 45 pS (a representative channel is shown in Figure 4A). The addition of 721 ng/mL (1.0 µM) RTV to the 1.0 µM cis Ca 2+ chamber increased the P o of RyR2 3.8-fold to 0.20. The increase in P o arose from an increase in the dwell time in the opened state ( Figure 4D). Lower concentrations of RTV and the solvent (1% ethanol) did not have any quantitative effect on the P o of RyR2. Increasing the cis concentration of RTV to 2163 and 4326 ng/mL increased the P o of RyR2 further to 0.46 ± 0.05 and 0.82 ± 0.06, respectively, by increasing the dwell time in the opened state ( Figure 5A-D). Higher concentrations of RTV in the cis chamber (8652-25,956 ng/mL) dose-dependently decreased the P o of RyR2 to closure (0.006 ± 0.001; Figure 5A-D).

Effects of HIV-Tat on RyR2
The mean P o of the RyR2 channels with 1.0 µM cis Ca 2+ at + 35 mV (HP) was 0.03 ± 0.00 (a representative channel is shown in Figure 6A; top). The mean current amplitude was 21.8 ± 1.8 pA, which is equivalent to a conductance of 726 ± 60 pS. The addition of HIV-Tat (7 ng/mL) to the cis chamber increased the P o of RyR2 to 0.14 ± 0.01 at + 35 mV HP in 10 out of the 11 channels, respectively ( Figure Figure 6D). At 28 and 42 ng/mL HIV-Tat, a reversible state was induced from the reduced conductance (R1) (50% of maximum; 346 ± 15 pS) that lasted for ≥ 200 ms ( Figure 6A). At 56 ng/mL cis, HIV-Tat increased the P o of RyR2 up to 0.85 ± 0.05 at + 35mV HP. Heat-inactivated HIV-Tat (boiled for 10 m) at 7-42 ng/mL had no quantitative effects on the P o of RyR2 (Supplemental Figure S1). However, heat-inactivated HIV-Tat at 56 ng/mL decreased the conductance of RyR2 (Supplemental Figure S1). We suspect that this may be due, in part, to heat-inactivated HIV-Tat nonspecifically interacting with the pores of RyR2.

Effects of Luminal Ca 2+ on HIV-Tat Actions on RyR2
We also investigated whether the ability of HIV-Tat to activate RyR2 was dependent on trans Ca 2+ content (i.e., the SR Ca 2+ content). In this experiment, when cis Ca 2+ was kept at 100 nM, and when and trans Ca 2+ was progressively increased from 100 nM to 3000 µM, the P o of RyR2 increased from 0.001 to 0.051 (a representative channel is shown in Figure 7). The increase in P o was significant (p < 0.05) and arose from increases in the gating frequency and not dwell times in the opened state (0.22 ± 0.01 ms). The addition of 14 ng/mL HIV-Tat to the cis chamber increased the P o to 0.45, similar to that with a low trans Ca 2+ ( Figure 7A: third panel from the top; see also Figure 5A: third panel from top). Increasing trans Ca 2+ further to 5000 µM in the presence of 14 ng/mL cis HIV-Tat did not increase the P o of RyR2 further. These data suggest that the ability of HIV-Tat to activate RyR2 was independent of the SR Ca 2+ content.
RyR2 as a function of the cis concentrations of the EFV channels for n = 12 channels. Recordings shown are at +35 mV (upward deflections) in a symmetric KCl buffer solution (0.25 mM KCl; 20 mM K/HEPES; pH 7.4). O, open; C, closed. * denotes significantly different (p < 0.05) from the control (before the addition of the drug).

Effects of HIV-Tat Pretreatment on the Actions of ATV, EFV, and RTV on RyR2
Experiments were conducted to investigate if pretreating RyR2 with a low HIV-Tat (7 ng/mL), which exhibited a small effect on the P o of RyR2, would synergize and/or antagonize the actions of ATV, EFV, and RTV on RyR2. We found that when a RyR2 with 1.0 µM cis Ca 2+ was pretreated with 7 ng/mL HIV-Tat (cis), the ability of ATV and RTV to activate RyR2 was enhanced ( Figure 8A,E; see also raw data in Supplemental Figures S2  and S4). The exaggerated activation of RyR2 by ATV and RTV arose from an increase in the dwell time in the opened state ( Figure 8B,F). When RyR2 in 1.0 µM cis Ca 2+ was pretreated with 7 ng/mL HIV-Tat, the ability of EFV to activate RyR2 was blunted and the closure of RyR2 by EFV was enhanced, primarily by decreasing the dwell time in the open state and the gating frequency ( Figure 8C,D; see also raw data in Supplemental Figure S3).

In Silico Molecular Docking of HIV-Tat on RyR2
Having established that HIV-Tat binds to and activates RyR2, and that pretreating RyR2 with a low HIV-Tat (7 ng/mL) exacerbated the actions of ATV, EFV, and RTV, in silico molecular docking studies were then conducted using Schrodinger's Prime proteinprotein docking software to identify the thermodynamically favored interacting sites for HIV-Tat on a reconstructed human RyR2. Figure Figure 9A and are color-coded to easily view these sites on the 3D structure of RyR2 [52][53][54]. After mixing HIV-Tat with RyR2 in silico and performing 70,000 rotations and energy evaluations, three thermodynamically favored interaction sites for HIV-Tat on RyR2 were identified ( Figure 8E-G, showing views from the membrane, cytoplasm, and lumen of the SR, respectively). Site 1 (yellow) was the most thermodynamically favored, with a binding energy of −315.16 AU; site 2 (brown) had a binding energy of −251.50 AU, and site 3 (magenta) had a binding energy of −102.67 AU. When using Schrodinger's Prime protein-protein docking software, the more negative the binding energy, the better the fit. The most thermodynamically favored site of interaction (site 1: brown) for HIV-Tat resides in the region within the junctional solenoid region (JSOL: amino acids (AA) 1702-1963) [56,57], with 14 hydrogen bonding, 3 salt bridges, and 1 pi stack (See Table 1). Note, the numbering of the AA for HIV-Tat that is listed is derived from the crystal structure provided by Protein Data Bank (PDB id; 1TIV). The second thermodynamically favored site of interaction (site 2: yellow) resides within the SPRY1 (AA 647-659), with 2 hydrogen bonding and 2 salt bridges, the SPRY2 domain (AA 1151-1249), with 5 hydrogen bonding, and the SPRY 3 domain (AA 1335-1465), with 5 hydrogen bonding and 1 salt bridge ( Table 2). The least thermodynamically favored of the three sites (site 3: magenta) interacts with the N-terminal domain of RyR2 between AA 201 and 554, with 13 hydrogen bonding and 1 pi stack, and the JSOL domain between AA 1739-1816, with 6 hydrogen bonding and 2 salt bridges (Table 3).

Discussion
PLWH, especially those with inadequate viral suppression and a low CD4 + T-cell count (<200 cells/mm 3 ), are 2.5-4 times more likely to succumb to SCD compared to uninfected individuals [2][3][4][5]10]. The available data suggest nonarrhythmic and arrhythmic causes for SCD, with the former attributed to occult drug overdose [10]. To date, the arrhythmic causes of SCD are thought to arise principally from the prolongation of the QT interval by HIV proteins, select antiretroviral and prescription medications, recreational and illicit drugs, and from myocardial fibrosis.
Here, we identify another potential mechanism that could be increasing the risk of arrhythmias and SCD, i.e., the disruption of the opening and closing of RyR2 by HIV-Tat and select antiretroviral drugs. This conclusion is based on several novel findings. First, we show that HIV-Tat, ATV, and RTV enhanced equilibrium [ 3 H]ryanodine binding to RyR2 and that EFV attenuated equilibrium [ 3 H]ryanodine binding to RyR2. The other antiretroviral drugs investigated, ABC, BIC, TDF, and RPV, had no measurable effects on equilibrium [ 3 H]ryanodine binding to RyR2. In this study, we also show, for the first time, that when RyR2 was preincubated with HIV-Tat (reminiscent of inadequate viral suppression), the abilities of ATV, EFV, and RTV to modulate equilibrium [ 3 H]ryanodine binding to RyR2 were enhanced as indicated by leftward shifts in their binding curves. However, since SR membrane vesicles were used, it was not clear whether the abilities of HIV-Tat, ATV, EFV, and RTV to alter equilibrium [ 3 H]ryanodine binding arose from direct interactions with RyR2 or from their interactions with endogenous proteins that are known for their RyR2 activity.
Compounds that alter equilibrium [ 3 H]ryanodine binding to RyR2 in SR vesicles should also modulate the activity of RyR2 in vivo. As such, we investigated whether HIV-Tat, ATV, EFV, and RTV can perturb intracellular Ca 2+ homeostasis in cardiac myocytes. For this, freshly isolated primary rat ventricular myocytes were loaded with the Ca 2+ -sensitive dye Fluo-3, and live-cell imaging was conducted before and after exposure to these agents. In these studies, we found that HIV-Tat, ATV, EFV, and RTV elicited Ca 2+ transients and/or waves in myocytes within one minute (after exposure). However, the Ca 2+ transient/wave patterns were distinct for each compound, with ATV eliciting repeated Ca 2+ waves and EFV and HIV-Tat eliciting less frequent but larger global Ca 2+ transients. RTV also triggers Ca 2+ release, firstly, at multiple distinct regions within the myocyte, which then transitions to short-lasting global Ca 2+ elevation. Persistent Ca 2+ release from the SR can activate Ca 2+ -sensitive signaling pathways, which may disrupt cellular functions. The increases in intracellular Ca 2+ elicited by HIV-Tat, ATV, EFV, and RTV were also blunted when the myocytes were pretreated with ryanodine to close RyR2. Although these data are consistent with the notion that HIV-Tat, ATV, EFV, and RTV modulate RyR2, it was still not clear whether their effects on myocyte Ca 2+ homeostasis arose from their direct interactions with RyR2 or from their interactions with the endogenous regulatory modulators of RyR2.
In order to address this question, lipid bilayer single-channel studies were conducted. For this, the junctional vesicles were solubilized with the mild detergent CHAPS to remove the attached RyR2-associated/modulatory protein. Purified RyR2 that was reconstituted into proteoliposomes was then fused to artificial lipid membranes, and single-channel assays were conducted. Using this approach, we show for the first time that HIV-Tat, ATV, EFV, and RTV dose-dependently increase the P o of RyR2, primarily by increasing its dwell time in the open state. The effects of HIV-Tat were not dependent on trans Ca 2+ levels, i.e., the Ca 2+ content inside the SR. These data provide the first direct evidence that HIV-Tat, ATV, EFV, and RTV interact directly with the RyR2 isolated from bovine hearts. Because of the high degree of sequence homology among species, we expect HIV-Tat, ATV, EFV, and RTV will also bind directly to and modulate RyR2 from other species.
In this study, we also found that HIV-Tat at concentrations 26 and 42 ng/mL also induced a reversible state of reduced conductance that was 50% of the maximum opening. This state of reduced conductance (R1) was not seen with lower or higher concentrations of HIV-Tat. At present, the specific molecular reasons for the induction of a reversible state of reduced conductance by 28 and 42 ng/mL HIV-Tat are not known. In an attempt to answer this question, a in silico molecular docking study was conducted to identify the thermodynamically favored regions on RyR2 where HIV-Tat could be binding. Using Schrodinger's Prime protein-protein docking software, three thermodynamically favored sites of interaction for HIV-Tat on RyR2 were identified. The highest interacting site for HIV-Tat on RyR2 was mapped to the junctional solenoid region (JSOL: AA 1702-1963). Other thermodynamically favored sites of interaction reside in the SPRY domains, the N-terminal domain, and the JSOL domain [56,57]. All these sites are critical for stabilizing the RyR2 monomers and for regulating the gating/conductance of RyR2. The JSOL region, equivalent to the "handle domain" [56], is involved in intra-and intersubunit interactions [57], and the destabilization of these interactions could alter the activity and conductance of RyR2. The SPRY domains are involved in fine-tuning the activity of RyR2 [57,58]. The NTD interacts with itself to promote channel closure, as well as to support the formation of functional tetrameric RyR2 channels. Immunophilin FKBP12.6, for which the binding site has been mapped to AA 305-1937, helps prevent Ca 2+ leaking from RyR2 [59][60][61], and its dissociation by HIV-Tat could have deleterious clinical consequences, including SCD. In an earlier report, Kaftan et al. [62] showed that when rapamycin was used to deplete FKBP12.6 from RyR2, the conductance of RyR2 was reduced. Jayaraman et al. [63] also showed that solubilization with detergent was not sufficient to remove immunophilins from ryanodine receptors. Since our "purified" RyR2 contained FKBP12.6 (see Supplemental Figure S5) and a thermodynamically favored docking site for HIV-Tat that resides within the binding site for FKBP12.6 (AA 305-1937 on RyR2), [60,61], it stands to reason that HIV-Tat could be reversibly binding to FKBP12.6. However, this hypothesis must be experimentally tested.
Here, we also show for the first time that pretreating RyR2 with a dose of HIV-Tat that does not significantly alter the P o of RyR2 exacerbated the ability of ATV and RTV to activate RyR2 and EFV to close RyR2. These findings are of significant importance and need to be expanded upon for multiple reasons. First, activating RyR2 could deplete the Ca 2+ content inside the SR, which may reduce the force of myocardial contractions. "Leaky" RyR2 can also trigger myocyte contractions independent of membrane depolarization. When the latter occurs, the Na + -Ca 2+ exchanger (NCX) on the plasma membrane will extrude the intracellular Ca 2+ in exchange for an influx of Na + , resulting in delayed after depolarizations (DADs), aberrant myocyte contractions, and arrhythmias, some of which may be fatal [37,64]. A sudden tachycardia by common stressors that activate the sympathetic nervous system can also trigger arrhythmias. Common stressors include physical exercise, emotional stress, and recreational and illicit drugs [65]. Second, when RyR2 is unable to be sufficiently activated by the Ca 2+ that enters the cell following depolarization, the amplitude of Ca 2+ released from the SR will be reduced, and the force of the myocardial contractions will also be reduced. Since a smaller amount of Ca 2+ is released from the SR following depolarization, the repeated refilling of the SR via SERCA could increase the Ca 2+ content inside the SR. At some maximal SR Ca 2+ load, depolarization could result in a decreased but sustained release of Ca 2+ from the SR. A prolonged rise in intracellular Ca 2+ will supercharge the electrogenic NCX, generating a depolarizing inward Na + current as Ca 2+ is extruded. This aberrant depolarization, known as early after depolarization (EAD), can serve as a trigger an arrhythmia, a precursor for SCD [45,66,67]. Third, animal models to assess the arrhythmia risks posed by ATV, EFV, RTV, and other antiretroviral agents should include, at a minimum, HIV-Tat. Whether other HIV-1 auxiliary proteins (gp120 Nef) and inflammatory cytokines are also needed remains to be evaluated. Without these components in animal models, it would be challenging to decipher why PLWH, especially those with inadequate viral suppression, are at significant risk of SCD [3].
This study is not without limitations. First, only a limited number of available antiretroviral agents and HIV-Tats were tested. Of the seven antiretroviral drugs tested, only ATV, EFV, and RTV bind to and modulate RyR2; the others did not. Although these drugs are not generally used as first-line drugs in developed economies, they continue to be used in developing countries where the incidence of HIV-1 infection is high [48]. EFV at a dose of 400 mg in combination with an NRTI is recommended as the alternative first-line regimen for antiretroviral naïve adults and adolescents living with HIV who are initiating ART treatment. ATV is an alternate second-line protease inhibitor used in combination with ritonavir to increase efficacy [48]. Second, although we show that HIV-Tat, ATV, EFV, and RTV can elicit Ca 2+ waves and global Ca 2+ transients in primary rat cardiac myocytes, we have not specifically demonstrated the development of arrhythmia and SCD using animal models. Third, the in silico mutation studies that identify the potential docking sites of HIV-Tat to RyR2 is the first step. These interacting sites must be validated using mutation studies in silico and in cell culture models.
In summary, the present study shows for the first time that low HIV-Tat and therapeutic levels of ATV, EFV, and RTV can bind to and modulate the function of RyR2. ABC, BIC, RPV, and TDF had no significant effect on the equilibrium binding of [ 3 H]ryanodine to RyR2. In the presence of low HIV-Tat (analogous to inadequate HIV-1 viremia control), the abilities of ATV and RTV to activate RyR2 were potentiated, while the ability of EFV to close RyR2 was enhanced. We also show that HIV-Tat, ATV, EFV, and RTV can perturb intracellular Ca 2+ homeostasis in primary rat ventricular myocytes via a RyR2-sensitive mechanism and that there are potential interacting sites for HIV-Tat on RyR2. Although these data indicate that HIV-Tat, ATV, EFV, and RTV can modulate RyR2 in cardiac myocytes, additional studies in clinically relevant models must be conducted to determine if these drugs are capable of triggering arrhythmias and even SCD.

Preparation of Stock Solutions of HIV-Tat, EFV, ATV and RTV
HIV-Tat was prepared by dissolving 100 µg in 100 µL of apirogenic sterile water, and stock solution was aliquoted and stored at −80 • C in silanized vials until use. ATV (17 mg), EFV (8 mg), and RTV (14 mg) were dissolved in 2.5 mL of ethanol and the suspensions were vortexed for 5 min at room temperature (22 • C) and centrifuged at 20,000× g for 30 min. The supernatants were removed and diluted 10× with methanol, and high-performance liquid chromatography (using a YMC Octyl C8 column, Waters Inc., Milford, MA) with a mobile phase consisting of 52% 25 mM KH 2 PO 4 , pH 4.15/48% acetonitrile at a flow rate of 0.4 mL/min and a UV/V is detector at 212 nm) was used to determine the concentrations of the drugs in solution [50]. ABC (14.31 mg), BCT (22.46 mg), RPV (14.36 mg), and TDF (18.32 mg) we dissolved in dimethyl sulfoxide (DMSO) to generate 50 mM stock solutions. These drugs were further diluted 1:10 in DMSO for working solutions and 1:100 for binding assays.

Preparation SR Vesicles and Enrichment of RyR2 from Bovine Hearts
Bovine hearts were generously donated by the Greater Omaha Packaging Company (Omaha, NE, USA).
(a) Sarcoplasmic reticulum vesicles: Sarcoplasmic reticulum (SR) membrane vesicles were prepared using the procedure described earlier [53,68]. Briefly, after removal of outer fat, ventricular tissues were cut into small pieces, placed in ice-cold isolation buffer (10 mM NaHCO 3 , 230 µM phenylmethylsulfonyl fluoride and 1.1 µM leupeptin, pH 7.4) and homogenized for 3 × 30 s using a Kinematica PT-600 Polytron at setting 4.5. Homogenates were then placed in Beckman type JA-10 centrifuge tubes (250 mL) and centrifuged for 20 min at 7500× g av . The pellets were collected and rehomogenized (3 × 30 s interval) in isolation buffer at a setting of 6.0. Homogenates were centrifuged at 11,000× g av for 20 min, and supernatants were filtered through four layers of cheese cloth, and microsomal vesicles were obtained by sedimentation at 85,000× g av for 30 min at 4 • C in a Beckman type 45Ti rotor. The pellets containing the microsomal vesicles were resuspended in buffer (0.3 M sucrose, 10 mM histidine, 230 µM PMSF, and 1.1 µM leupeptin, pH 7.4), flash frozen, and stored at −80 • C.
(b) Junctional SR vesicles: discontinuous sucrose gradient centrifugation was used to separate junctional SR vesicles from microsomal vesicles, as described in [53,69]. For this, microsomal membranes (5 mL) were layered onto centrifuge tubes containing discontinuous sucrose gradients (bottom to top: 5 mL of 1.5 M, 7 mL of 1.2 M, 7 mL of 1.0 M, and 7 mL of 0.8 M sucrose in 30 mL tubes). The tubes were then centrifuged in a Beckman SW-28 swinging bucket-type rotor at 110,000× g av for 2 h at 4 • C. The membrane fractions that sedimented at the 1.2 M/1.5 M sucrose interface were collected and resuspended in 10 mM histidine, 230 µM PMSF, and 1.1 µM leupeptin, and the junctional SR vesicles were recovered by sedimentation at 100,000× g av for 30 min at 4 • C. The pellet was resuspended in 0.3 M sucrose, 10 mM histidine, 230 µM PMSF, and 1.1 µM leupeptin, flash frozen, and stored at −80 • C.

Statistical Analysis
Paired student t-tests were used to compare data before and after drug treatment using Microsoft Excel (Microsoft Corporation, Seattle, WA, USA). One-way analysis of variance (ANOVA), followed by the Bonferroni's post hoc test, were also used for some analyses using GraphPad Prism 7.0 (La Jolla, CA, USA). Data are presented in text, and graphs as the mean standard error of mean (S.E.M). Significance was determined at the 95% confidence interval.