Determining the Crystal Structure of TRPV6

Calcium ions play important roles in many physiological processes, including neurotransmitter release, excitation-contraction coupling, cell motility, and gene expression [1]. Cellular calcium levels are precisely tuned by various channels and transporters. Transient receptor potential (TRP) channels, which are generally nonselective cation channels, conduct Ca in response to disparate activators, including sensory stimuli such as temperature, touch, and pungent chemicals [2]. Members 5 and 6 of vanilloid subfamily (TRPV5 and TRPV6, previously named ECaC1 and ECaC2/CaT1, respectively) are uniquely Ca-selective (P/P > 100) [3,4] TRP channels, both of which were identified in 1999 by expression cloning strategies utilizing cDNA libraries from rabbit kidney [5] and rat duodenum [6], respectively. While TRPV5 expression is mainly restricted to the kidney, TRPV6 is expressed in various tissues including the stomach, small intestine, prostate, esophagus, colon, and placenta. Genetic knockout of TRPV5 or TRPV6 in mice suggests the importance of these channels for Ca homeostasis. TRPV5 knockout mice showed defects in renal Ca reabsorption and reduced bone thickness [7], and the knockout of TRPV6 resulted in defective intestinal calcium absorption, decreased bone mineral density, reduced fertility, and hypocalcemia when challenged with a low Ca diet [8]. Further support for the role of these channels in Ca absorption and homeostasis stems from the robust regulation of their expression by the calciotropic hormone vitamin D [9–12] (see Chapter 13). TRPV6 has been shown to be aberrantly expressed in numerous cancer types, including carcinomas of the colon, prostate, breast, and thyroid [13–18]. The correlation between TRPV6 expression and tumor malignancy and its potential contribution to cancer cell survival has highlighted TRPV6 as a target for cancer diagnosis and treatment [15,17,19,20]. Indeed, a selective inhibitor of TRPV6 activity derived from northern short-tailed shrew venom [21] has entered phase I clinical trials in patients with advanced solid tumors of tissues known to express TRPV6, including the pancreas and ovary [22]. Structurally, TRPV5 and TRPV6 share ∼75% sequence identity with each other and are ∼25% identical to the founding member of the TRPV subfamily TRPV1. The transmembrane (TM) domain has the same topology as tetrameric K channels [23], with six TM helices (S1–S6) and a pore-forming re-entrant loop between the S5 and S6. Importantly, this loop contains a conserved aspartate residue that is critical for the calcium permeability of TRPV5 and TRPV6 [24,25], suggesting that this residue at least in part comprises the selectivity filter. Flanking the TM domain are relatively large intracellular N- and C-termini. The N-terminus, which includes six ankyrin repeats [26], is critical for proper channel assembly and function [27,28], while the C-terminus contains domains involved in Ca/calmodulin-dependent inactivation [29–31] (see Chapter 13). To help understand the functional mechanisms of TRPV5/6 and potentially inform rational drug design, we sought to obtain a high-resolution structure of an intact channel. Until several years ago, the only viable method of obtaining such a structure was x-ray crystallography. However, producing well-diffracting crystals can be a notoriously difficult and resource-consuming process because membrane proteins typically have low expression and purification yields, poor stability in detergent, and inherent flexibility [32]. However, structural biologists are now able to circumvent this major bottleneck, owing to recent advances in single-particle cryo-electron microscopy (cryo-EM), which have facilitated the determination of membrane protein structures at near-atomic resolutions without prior crystallization [33]. These advances have had a particularly profound effect on the TRP channel field, as atomic-level cryo-EM structures have been determined for TRPV1 [34–36]; TRPV2 [37,38]; and ankyrin subfamily member [39]. The ability to computationally select specific conformational states from a heterogeneous cryo-EM sample can be especially powerful when studying mechanisms of gating, as exemplified by studies of TRPV1 in various ligand-induced conformations [34–36]. Cryo-EM will surely continue to be exploited with great effect to elucidate structures of TRP channels and other membrane proteins. As yet, there are several benefits that may make obtaining an x-ray structure desirable over cryo-EM. First, crystallographers can use true statistical approaches such as the Free R value [40] to evaluate the accuracy of atomic models against experimental data, while analogous methods in cryo-EM [41–43] are still relatively nascent. Second, the resolution of a cryo-EM map usually varies widely across a single reconstruction, with more flexible regions, typically in the periphery, being less resolved or completely absent. For example, in TRPV1, while the TM domain is well resolved, the first two ankyrin repeats at the distal N-terminus are missing from the electron density maps [34–36], presumably due to their flexibility (Figure 14.1a). In x-ray structures, the resolution obtained from the diffraction data is more representative of the structure as a whole, and peripheral or flexible regions may be stabilized by crystal contacts and thus adequately resolved (Figure 14.1b). Third and perhaps most importantly, the position of anomalous scatterers, such as selenium atoms in selenomethionine-labeled protein, sulfur atoms in native cysteine or methionine side chains, or heavy atoms bound to the protein, can be accurately identified with little ambiguity. Anomalous scattering can therefore be utilized to robustly aid or validate sequence registry (Figure 14.1c), which is especially important for low-resolution structures and/or regions with poor electron density. Using anomalous scattering to identify bound ions is particularly useful for studying ion channel structures, as ion binding at specific locations is vital for understanding permeation and ion channel block. For TRPV6, we used these techniques to identify binding sites for the permeant cations Ca and Ba, as well as the channel blocker Gd (Figure 14.1d through g). Methods to unambiguously identify specific atoms or small labels in cryo-EM electron density maps have yet to be developed. We were motivated by each of these factors as we attempted to crystallize TRPV5/6 in the midst of the cryo-EM “resolution revolution.” In 2016, we reported the crystal structure of intact rat TRPV6 at 3.25 Å resolution [44]. To our knowledge, this represented the first crystal structure of a TRP channel and the second crystal structure of a naturally occurring Ca-selective channel, after the structure of the calcium release-activated calcium (CRAC) channel Orai reported in 2012 [45]. A detailed description of the TRPV6 structure can be found elsewhere [44]. In this chapter we will focus on the multiyear journey taken to determine the structure, in which >150 constructs were purified and subjected to crystallization screening, and thousands of crystals were tested for diffraction. We will summarize the methods used to screen constructs and precrystallization conditions, express and purify protein, grow and optimize crystals, and collect and analyze diffraction data. Finally, we will briefly compare the structural bases of Ca-selective permeation in TRPV6 and Orai.