Distinct mechanisms regulating mechanical force-induced Ca²⁺ signals at the plasma membrane and the ER in human MSCs

1) The IP3 is elevated in responding to the mechanic force by optical laser tweezer-mediated bead traction and so it is not responsible for Ca²⁺ release from ER. The authors concluded that the force was deeply penetrated to the ER to trigger Ca²⁺ release. This is a very interesting discovery. However, the authors should provide some clues about what kind of signals in ER that may trigger the Ca²⁺ release in responding to mechanic force.

We thank the reviewers for this important comment. There are two possible signals in ER that may trigger the Ca²⁺ release in response to mechanical force. 1) IP₃R channels on the ER membrane are mechanosensitive and can be directly opened by transmitted mechanical force. Several lines of evidence supported this hypothesis. First, IP₃R is coupled to cytoskeleton and associated proteins allowing mechanical coupling. A direct binding between IP₃R and myosin II was discovered in C. elegans (Walker et al., 2002). In addition, IP₃R has linkage to actin mediated by an adaptor 4.1N protein (Fukatsu et al., 2004; Turvey et al., 2005). IP₃R also binds to ankyrins, which are adaptor proteins coupled to the spectrin-based cytoskeleton (Bourguignon et al., 1993; Joseph and Samanta, 1993). Second, IP₃R channel has an α-helix bundle at the pore forming region, similar to voltage-gated potassium and calcium channels (Schug et al., 2008), which are generally found to be mechanosensitive (Morris, 2011). The mechanism for their mechanosensitivity is possibly that the α-helix tilt angle tends to change when the membrane thins upon mechanical tension, in order to do proper hydrophobic matching with the interfacial region of the membrane, which leads to channel opening (Cheng et al., 2004; Kim and Im, 2010). Therefore it is likely that the IP₃R channel is also mechanosensitive. 2) Other mechanosensitive channels on ER, e.g. TRP family, may also contribute to this forced-induced Ca²⁺ release. A number of TRP channels have been found to express at ER membranes, such as TRPC1 (Berbey et al., 2009), TRPV1 (Gallego-Sandin et al., 2009), TRPM8 (Bidaux et al., 2007), TRPP2 (Polycystin-2) (Koulen et al., 2002). As some TRP channels have been shown to be mechanosensitive and have linkage to cytoskeleton (Barritt and Rychkov, 2005), it is likely that TRP channels located at ER may mediate, at least in part, the force-induced ER calcium release. Notably, these two possibilities are not mutually exclusive as more than one type of channels can be responsible for the force-induced ER calcium release. The discussion above has been incorporated in the manuscript (please see the third paragraph of the Results and Discussion section).

2) siRNA for TRPM7 almost completely abolished the Ca²⁺ release (Figure 2 E and F), indicating that it may play an essential role in mechanic force-induced Ca²⁺ release. The authors should explain how it is coupled to the Ca²⁺ in ER, because both pathways contribute to the Ca²⁺ flux.

Several possibilities may contribute to the observed results. 1) TRPM7 is functionally coupled to integrin, actomyocin contractility and cytoskeleton. As such, it may mediate and facilitate the force transmission to ER. Indeed, it has been shown that TRPM7 kinase can phosphorylate myosin II heavy chain (Clark et al., 2006) and regulate actomyocin contractility. 2) TRPM7 activity may have some downstream effect on IP₃R in ER. For example, TRPM7 can control the protease calpain (Su et al., 2006), which can regulate IP₃R degradation in ER (Diaz and Bourguignon, 2000). 3) TRPM7 may be directly coupled to IP₃R in the ER through adaptor proteins. Indeed, another TRP channel, TRPC1 has been shown to directly couple to IP₃R in the ER through an adaptor protein Homer (Yuan et al., 2003). The discussion above has been incorporated in the manuscript (please see the fourth paragraph of the Results and Discussion section).

3) If two pathways; ER and plasma membrane, contribute to the Ca²⁺ flux in responding to the mechanic force, please explain why the inhibition by Gd³⁺, La³⁺, or streptomycin, blocked all the Ca²⁺ signal (Figure 2–figure supplement 1)?

We performed Gd³⁺, La³⁺, and streptomycin inhibition experiments first. As they are board inhibitors for stretch-activated channels or store-operated calcium channels, we concluded that mechanosensitive channels are generally essential for the force-induced calcium release. We then conducted the TRPM7 siRNA experiment to reveal that TRPM7 as a potential mechanosensitive channel candidate is essential for the force-induced calcium release. Accordingly, this question is essentially similar to question #2: how are mechanosensitive channels at the plasma membrane coupled to ER calcium release, which has been addressed in responses to this question as shown above. In general, it is hypothesized that mechanosensitive channels at the plasma membrane can potentially mediate the force-induced calcium release at ER by modulating force transmission efficiency, mechanically altering IP₃R at ER through physical coupling, as well as affecting IP₃R functions at ER. It is of note that other mechanosensitive channels, including other TRP family channels besides TRPM7, may also be involved in this force transmission process.

4) The experiments demonstrate that the force transmission to the ER is an important prerequisite for ER calcium release. The finding that this effect can be blocked by inhibiting the plasma membrane channels is both intriguing and worrisome. It is hard to imagine how blocking the plasma membrane channels would affect force transmission to the ER, raising the prospect that there is some other form of signal transmission that causes plasma membrane channel activation to produce activation of the ER. If so, then this would counter all of the arguments supporting the role of direct force transmission to the ER as the primary mechanism for calcium release there. This is a point that deserves additional discussion.

We want to thank the reviewers for the question. This question is similar to questions #2 and #3, i.e. how can membrane channels affect force transmission to the ER. Accordingly, please see our responses to these questions above. Indeed, it becomes clear that membrane channels are not isolated entities floating in the plasma membrane. Instead, they are intimately coupled to integrins, cytoskeleton, actomyocin contractility and ER membrane channels (Cantiello et al., 2007; Deng et al., 2009; Matthews et al., 2007). Therefore, these structural and physical couplings enable membrane channels to participate in direct force transmission to ER. The discussion above has been incorporated in the manuscript (please see the fourth paragraph of the Results and Discussion section).

5) No data are presented regarding the delay time between force application and the first ensuing calcium transient. One might expect the channels to respond as soon as force is transmitted, which should be virtually instantaneous. Alternatively, a longer delay might be indicative of a biochemical signal mediated mechanism. This also deserves discussion.

We want to thank the reviewers for this inspiring comment. We have compiled the data of delay time in each of the three groups and added the graph as Figure 4–figure supplement 1C. (Figure caption has been added.) Our data indicated that in the absence of extracellular Ca²⁺, the average delay time of ER calcium release monitored by both cytosolic Ca²⁺ biosensor and ER Ca²⁺ biosensor are around 100 sec. Meanwhile, the average delay time of the calcium influx through plasma membrane in the presence of 2APB and extracellular calcium is much shorter, around 20 sec, indicating faster response to mechanical force stimulation.

Although the delay time of ER calcium release is longer than that of calcium influx across the plasma membrane, it does not contradict mechanical signal-mediated mechanism. Several reasons may contribute to longer time delay of ER calcium release. 1) The mechanical coupling machinery may need time for reinforcement to allow sufficient force transfer between the plasma membrane and the ER membrane. It is likely that mechanical force transmits through integrins and cytoskeleton (Perozo and Rees, 2003; Wang et al., 1993) as only Fn-coated, but not BSA-coated beads can induce ER calcium release (Figure 1C-F). However, the discrete network linkage of the existing cytoskeleton at the time of force application may not be sufficient for force focusing and transmission to ER, particularly because the apical integrins bound to the Fn-coated beads had not previously experienced force and therefore may have limited connection with cytoskeleton (Matthews et al., 2006). As force can unfold proteins, change their conformations to expose cryptic binding sites to allow the assembly of new/stronger network linkages (del Rio et al., 2009; Johnson et al., 2007), mechanical force can modulate the structural coupling of molecules and mechanical properties of the cells through cytoskeletal remodeling (Kaazempur Mofrad et al., 2005; Matthews et al., 2006; Rosenblatt et al., 2007). These modulation processes of mechanical coupling and reinforcement may need time to reach the threshold for sufficient mechanotransduction of ER calcium release; 2) the channels on ER membrane could have different kinetics and mechanosensitivity (Moe et al., 1998), therefore they may need larger focused and reinforced force to be developed at the site of ER to reach the threshold for physical opening. All these factors may contribute to the longer delay time for force-induced ER calcium release.

Again we want to emphasize that our model does not rule out biochemical signal-mediated mechanisms. As a matter of fact, biochemical signal-mediated mechanisms, such as protein-protein interaction and cytoskeletal remodeling under mechanical tension are important to mediate mechanical signals as discussed above. We have also incorporated the discussion of this time delay of ER calcium release in the revised manuscript (please see the eighth paragraph of the Results and Discussion section).