| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Institute for Pharmacology and Toxicology, Philipps-University Marburg, Marburg, Germany
2Correspondence: Institut für Pharmakologie und Toxikologie, Philipps-Universität Marburg, Karl-von-Frisch-Str.1, 35043 Marburg, Germany. E-mail: chubanov{at}staff.uni-marburg.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-kinase domains, whereas TRPM2 is linked to an ADP-ribose phosphohydrolase (Nudix domain). At a molecular level, the evolutionary steps that gave rise to the structural and functional TRPM channel diversity remain elusive. Here, we provide phylogenetic evidence that Nudix-linked channels represent an ancestral type of TRPMs that is present in various phyla, ranging from protists to humans. Surprisingly, the pore-forming segments of invertebrate TRPM2-like proteins display high sequence similarity to those of Ca2+-selective TRPMs, while human TRPM2 is characterized by a loss of several conserved residues. Using the patch-clamp technique, Ca2+ imaging, and site-directed mutagenesis, we demonstrate that restoration of only two "ancient" pore residues in human TRPM2 (Q981E/P983Y) significantly increased (
4-fold) its permeability for Ca2+. Conversely, introduction of a "modern" sequence motif into mouse TRPM7 (E1047Q/Y1049P) resulted in the loss of Ca2+ permeation and a linear TRPM2-like current-voltage relationship. Overall, our findings provide an integrative view on the evolution of the domain architecture and the structural basis of the distinct ion permeation profiles of TRPM channels.—Mederos y Schnitzler, M., Wäring, J., Gudermann, T., Chubanov, V. Evolutionary determinants of divergent calcium selectivity of TRPM channels.
Key Words: transient receptor potential • TRPM2 • TRPM7 • TRPM8 • Nudix
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Several known mutations in TRP family members are the molecular cause of human diseases (2)
. However, in most instances, our knowledge about TRP channels remains insufficient to explain the relationship between key channel characteristics and distinct physiological roles of the gene family members.
Most TRP channels are poorly selective for permeating cations (PCa/PNa is in the range of 0.5–5) and can be activated by multiple extra- and intracellular signals (3)
. It is generally accepted that the architecture of the pore-forming segment in TRP channels has significant homology to tetrameric voltage-gated K+ channel complexes: a short stretch between transmembrane helices (TM) 5 and 6 contains a hydrophobic pore helix followed by a pore loop, and the loops from all four subunits of the tetramer are thought to contribute to a common selectivity filter (3)
. Studies on a subset of the vanilloid-receptor-related TRP channels (TRPV1/4/5/6) and the canonical Drosophila melanogaster TRP channel identified negatively charged residues located in the predicted pore loop as critical determinants of Ca2+ permeability (4
5
6
7
8)
. On the contrary, in two mammalian canonical TRP channels (TRPC1 and TRPC5), acidic residues located outside the putative pore loop were found to be important for their permeability to Ca2+ (9
, 10)
.
The melastatin-related TRP channels (TRPM channels) cover a broad spectrum of different cation permeation properties. TRPM4 and TRPM5 display unique properties among TRP channels in that they are selective for monovalent cations (11
12
13
14
15)
. In contrast, three other channels, TRPM3 (variant TRPM32), TRPM6, and TRPM7, have a noticeable preference for the permeation of divalent cations (16
17
18
19
20
21
22)
, whereas TRPM8 and TRPM2 discriminate poorly between monovalent and divalent cations (23
24
25
26
27
28)
. Recently, Nilius et al. (29)
demonstrated that conserved negatively charged residues in the predicted TRPM4 pore loop determine the permeation of monovalent cations, while Topala et al. (30)
reported that the corresponding acidic residues are also critically involved in the divalent cation permeability of TRPM6. Thus, the molecular mechanisms responsible for the divergent Ca2+ permeation through TRPM4 and TRPM6 remain elusive. Furthermore, Oberwinkler et al. (16)
identified an alternatively spliced variant of TRPM3 (TRPM31), which displays a diminished permeability for divalent cations due to an additional short stretch of amino acids located distal to the selectivity filter sequence proposed for TRPM4 and TRPM6.
TRPM channels also differ substantially in their domain architecture. Thus, TRPM6 and TRPM7 represent an unusual type of cation channel, since their C terminuses are fused to protein kinase domains (-type of Ser/Thr kinases; refs. 21
, 22
, 31
, 32
), while the C terminus of TRPM2 contains an ADP-ribose phosphohydrolase domain (nucleoside diphosphate linked moiety X-type motif 9 homology domain, NUDT9H or Nudix domain; refs. 24
, 33
34
35
). Until recently, phylogenetic analysis of the TRP gene family was restricted to a few vertebrate species and two ecdysozoan genetic model organisms, C. elegans and D. melanogaster (36
37
38)
. D. melanogaster has a single TRPM channel, while C. elegans contains three TRPM channels (Gon-2, GTL-1, and GTL-2; refs. 39
, 40
) and a large group of more distantly related proteins with a moderate homology to TRPM channels (41)
. Interestingly, the D. melanogaster and Caenorhabditis elegans NUDT9 homologs lack transmembrane segments, whereas TRPM proteins of these organisms are devoid of the Nudix- or -kinase domains. Therefore, it was generally assumed that a TRPM channel without enzymes represents the archetypal TRPM channel subunit and only a small subset of vertebrate TRPM channels has gained additional enzyme domains during evolution.
Here, we report that Nudix-linked TRPM proteins are present in major bilateral phyla and more "ancient" organisms such as sea anemones and unicellular protists. In contrast, TRPM channels lacking a Nudix or bearing an -kinase domain have emerged only in more modern bilateral species. Comparison of the pore-forming segments of ancient and modern TRPM channels revealed that a subset of mammalian channels with a lack of or low Ca2+ permeability is characterized by replacements of several conserved residues. We functionally tested the effect of these alterations in mammalian TRPM2, TRPM8, and TRPM7 and established that a short amino acid motif is the molecular determinant of the distinct cation permeation properties of these channels. These results highlight a direct link between evolutionary changes in the pore sequence and divergent functional characteristics of TRPM channels.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and inserted into in the pIRES2-EGFP vector (Clontech, Palo Alto, CA, USA). Human TRPM8 (NM_076985, in the pcDNA3.1-TOPO vector, Invitrogen, Carlsbad, CA, USA) was kindly provided by Dr. Carsten Strübing (Sanofi-Aventis GmbH, Frankfurt, Germany). Mutations in TRPM2, TRPM7, and TRPM8 were introduced by site-directed mutagenesis using the QuikChange system (Stratagene, La Jolla, CA, USA). All cDNA constructs used in the present work were confirmed by sequencing.
ClustalW (http://www.ebi.ac.uk/clustalw/) was used for multiple sequence alignment. A three-dimensional model of mbTRPM2-like 1 was generated using the annotated coordinates of the Aeropyrum pernix voltage-dependent K+ channel KvAP (42
, 43)
(PDB code 1orgC) and MODELLER (http://alto.compbio.ucsf.edu/modweb-cgi/main.cgi; ref. 42
). The calculated coordinates of mbTRPM2-like 1 were used to generate molecular graphic images of the TM5-TM6 segment with the UCSF Chimera package (44)
.
Human embryonic kidney (HEK) 293 cells were maintained at 37°C and 5% CO2 in Earle’s minimal essential medium supplemented with 10% fetal calf serum, 100 µg/ml streptomycin, and 100 U/ml penicillin (PAA Laboratories, Pasching, Austria). Cells were transiently transfected using the FuGENE6 (Roche Molecular Biochemicals, Pleasanton, CA, USA) or TransIT-LT1 (Mirus Bio Corporation, Göttingen, Germany) transfection reagent according to the manufacturer’s instructions.
Electrophysiological techniques
HEK293 cells grown in 35-mm dishes to 70% confluence were transfected with plasmid cDNAs encoding human TRPM2 (hTRPM2), human TRPM8 (hTRPM8), or mouse TRPM7 (mTRPM7) variants (1–2 µg/dish) 10–30 h before analysis. Whole-cell patch-clamp recordings were carried out at room temperature (22°C). Unless otherwise stated, cells were superfused with a standard physiological solution (Na+div): 140 mM NaCl, 5 mM CsCl, 2 mM CaCl2, 1 mM MgCl2, 10 mM glucose, and 10 mM HEPES, pH 7.4. In hTRPM8 measurements, the Na+div solution contained 10 µM icilin (Tocris Bioscience, Bristol, UK). The whole-cell recordings were performed with an EPC10 patch-clamp amplifier (HEKA, Lambrecht/Pfalz, Germany) using the Pulse software (HEKA). Patch pipettes were made of borosilicate glass (Science Products, Hofheim, Germany) and had resistances between 1.5 and 3.3 M
when filled with the following intracellular solution for hTRPM2 measurements: 130 mM CsCl, 4.046 mM CaCl2 (100 nM free Ca2+, calculated with CaBuf program; ftp://ftp.cc.kuleuven.ac.be/pub/droogmans/cabuf.zip), 10 mM BAPTA, 1 mM HEDTA, and 10 mM HEPES, pH 7.2). For stimulation of hTRPM2 currents, the intracellular solution contained 50–100 µM adenosine-5'-diphosphoribose (ADP-ribose). The liquid junction potential was +4.6 mV. Current-voltage (I-V) relations were recorded during voltage ramps from –100 to +100 mV with a slope of 0.5 V/s applied at a frequency of 2 Hz.
The intracellular solution for mTRPM7 and hTRPM8 measurements contained the following: 130 mM CsCl, 0.635 mM CaCl2 (5.5 nM calculated free [Ca2+]), 10 mM BAPTA, 1 mM HEDTA, and 10 mM HEPES, pH 7.2. The liquid junction potential was +5.1 mV and corrected for by the Pulse v8.7 software. In a subset of mTRPM7 measurements (
Fig. 5C
), the intracellular solutions contained between 100 µM and 3 mM MgCl2, while the free [Ca2+] was kept constant by adjusting the total CaCl2 concentration. To assess the dependence of mTRPM7 on extracellular Mg2+ (Fig. 6B
), concentrations of 10 µM to 20 mM MgCl2 were added to the nominally divalent cations free extracellular solution (Na+divf, described below) and mannitol concentrations reduced accordingly. To determine IC50 values, the data were fitted with the Hill equation.
|
|
|
|
|
|
For the determination of cation permeabilities of hTRPM2, hTRPM8, and mTRPM7, the following extracellular solutions were used: "Na+divf" (130 mM NaCl, 50 mM mannitol, and 10 mM HEPES, pH 7.4); "Ba2+" [10 mM BaCl2 and 150 mM N-methyl-D-glucamine (NMDG), pH 7.4]; "Ca2+" (10 mM CaCl2 and 150 mM NMDG, pH 7.4); and "Mg2+" (10 mM MgCl2 and 150 mM NMDG, pH 7.4). Permeability ratios (Pdiv/PNa) were determined using the following equation (45)
:
 |
With the use of the Pulse software, series resistance and capacitance were estimated and corrected automatically before each ramp. Series resistance compensation of 80% was used to reduce voltage errors in all experiments. Data were acquired at a frequency of 5 kHz after filtering at 1.67 kHz. The osmolality of all solutions was measured with the vapor osmometer Vapro 5520 (Wescor Inc., Logan, UT, USA) and was 300 mosmol/kg.
Aequorin-based intracellular [Ca2+] measurements
HEK293 cells cultured in 35-mm dishes were cotransfected with 2 µg/dish of plasmid DNAs encoding various mTRPM7 variants and 100 ng/dish of a pG5A cDNA construct encoding the enhanced green fluorescent protein fused in frame to the Aequorea victoria apo-aequorin (46)
. Twenty-four hours after transfection, cells were washed twice with PBS and incubated with 0.05% trypsin, and 1 mM EDTA in PBS for 2 min at room temperature. Cell suspensions were centrifuged twice at 600 g for 3 min and resuspended in Mg2+-free HEPES-buffered saline (Mg2+-free HBS: 140 mM NaCl, 6 mM KCl, 1 mM CaCl2, 10 mM HEPES, 5 mM glucose, and 0.1% BSA, pH 7.4). For reconstitution of aequorin, cell suspensions were incubated with 5 µM coelenterazin (Biaffin GmbH, Kassel, Germany) in Mg2+-free HBS for 30 min at room temperature. Cells were washed twice by centrifugation at 600 rpm for 3 min followed by resuspension of the pellet in Mg2+-free HBS and placed in a 96-well plate (105 cells per well). Luminescence was detected using a FLUOstar OPTIMA microplate reader at 37°C (BMG Labtech GmbH, Baden-Württemberg, Germany). To monitor the TRPM7-mediated Ca2+ influx, extracellular Ca2+ was increased by injection of Mg2+-free HBS containing 5 mM CaCl2 (final concentration). Experiments were terminated by lysing cells with 0.1% (v/v) Triton X-100 in HBS to record the total bioluminescence. Bioluminescence rates (counts/s) were analyzed at 1-s intervals and calibrated as intracellular [Ca2+] ([Ca2+]i) values using the following equation:
 |
Statistical analysis
Data are presented as means ± SE. Unless stated otherwise, data were compared by an unpaired Student’s t test if a Gaussian distribution was confirmed by applying a Shapiro-Wilk (normality) test. Significance was accepted at P < 0.05. For multiple comparisons, the one-way analysis of variance with Scheffé post hoc means comparison was applied.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
; see Supplemental Fig. 1). Surprisingly, we found that genes encoding Nudix-linked TRPM proteins, called "TRPM2-like" channels here due to their overall similarity in domain architecture and amino acid sequence to mammalian TRPM2 channels (Fig. 1
A), are invariably present in chordates, molluscs, echinodermates, and ancient organisms beyond the bilateral lineage-like sea anemones (cnidaria) and even in unicellular protists Monosiga brevicollis (choanoflagellata; Table 1
; Fig. 1C
). It is assumed that choanoflagellates and animals emerged from a common unicellular ancestor 1000 million years ago (47)
, indicating that Nudix-linked TRPM2-like proteins represent a very ancient type of TRP channels.
|
Remarkably, in M. brevicollis and Nematostella vectensis, TRPM2-like proteins are the only type of TRPM channels found (Fig. 1C
; Table 1
). We also found that Nudix-deficient TRPM proteins, called "TRPM3-like" proteins because of the highest sequence similarity within the pore-forming segment to human TRPM3, were present only in organisms within the bilateral lineage, suggesting that a loss of the Nudix domain was most likely a critical event during the evolution of TRPM3-like proteins. Interestingly, genes encoding -kinase-linked TRPMs emerged fairly recently, since they are present only in the genomes of vertebrate species but not in more ancient chordate relatives, including the lamprey Petromyzon marinus, the lancelet Branchiostoma floridae, and the sea squirt Ciona savignyi (Fig. 1C
; Table 1
). Thus, our phylogenetic analysis suggests that a Nudix-linked TRPM channel is most likely a common ancestor template that evolved into distinct types of TRPM channels in more recently emerged phyla.
To study the pore properties of TRPM channels, we generated a molecular model of the TM5-TM6 segment of mbTRPM2-like 1 based on a presumed overall structural and amino acid sequence similarity to the corresponding region of the A. pernix voltage-dependent K+ channel (KvAP) (43)
. In line with recent studies on mammalian TRPM4 (29)
and TRPM6 (48)
, we observed that the architecture of the TM5-TM6 segment in mbTRPM2-like 1 is similar to that of potassium channels, composed of a helical segment followed by a pore loop (Fig. 1B
). A characteristic feature of the TM5-TM6 of mbTRPM2-like 1 is a substantially longer loop segment comprising a highly conserved proximal region (which we call "proximal pore loop") and a distal segment (or "distal pore loop") devoid of significant amino acid similarity between TRPM channels (Fig. 1B
; see Supplemental Fig. 1A). Therefore, we reasoned that the proximal pore loop may represent a critical structure that determines the ion selectivity of the channel. Notably, the predicted proximal pore loop of invertebrate TRPM2-like proteins exhibits high sequence conservation across the phyla except for a short amino acid stretch altered specifically in the vertebrate TRPM2 orthologs (Fig. 1C
). Thus, a conserved E was replaced by a neutral Q at position 981 of human TRPM2, while an aromatic Y was changed to P at position 983. We reasoned that these evolutionary changes in hTRPM2 would not interfere with the general folding of the channel pore but would rather serve to fine-tune the ion permeability of distinct family members.
Restoration of ancient pore residues in hTRPM2 and hTRPM8 increases the Ca2+ permeability of the channels
To establish the functional consequence of the aforementioned evolutionary alterations in the pore-forming segment of hTRPM2, we exchanged Q981 and P983 for their ancient counterparts and assessed the biophysical properties of the mutants when transiently expressed in HEK293 cells using the patch-clamp technique (Fig. 2
A). In the whole-cell mode, intracellular application of ADP-ribose via the pipette elicited large inward and outward cation currents in cells expressing wild-type channels (TRPM2WT) as well as mutant variants. In line with previous reports (23
24
25)
, TRPM2WT-mediated currents displayed a characteristic linear I-V relationship and a Erev close to 0 mV in Na+div extracellular solution (Fig. 2A
). The TRPM2P983Y mutant was not significantly different from TRPM2WT. However, the TRPM2Q981E variant and, to a larger extent, the double-mutant TRPM2Q981E/P983Y resulted in I-V relationships that deviate from the characteristic linear relationship of TRPM2WT especially at negative potentials.
To compare the relative permeability of the hTRPM2 variants for different cations, we determined Erev values, when bath solutions contained 10 mM of different divalent cations (Ca2+, Mg2+, or Ba2+), while Na+ was substituted by NMDG (Fig. 2A
). Under these conditions, TRPM2Q981E and TRPM2Q981E/P983Y showed substantial shifts of Erev toward more positive values as compared with TRPM2WT. Accordingly, the calculated Pdiv/PNa revealed that TRPM2WT poorly discriminates between Na+ and divalent cations (PCa/PNa was 0.80±0.03; Fig. 2B
). However, TRPM2Q981E and, to a larger extent, TRPM2Q981E/P983Y were found to be significantly more permeable to divalent cations than TRPM2WT. Accordingly, the PCa/PNa ratio of TRPM2Q981E/P983Y was almost 4-fold higher than that of TRPM2WT (Fig. 2B
).
According to the primary amino acid similarity of mammalian TRPM channels, TRPM8 is the closest ortholog of TRPM2 (see Supplemental Fig. 1B). Human TRPM8 has been characterized as a nonselective cation channel (26
27
28)
. Furthermore, the putative pore loop of hTRPM8 contains the characteristic Q residue (see Supplemental Fig. 1A). We reasoned that in analogy to hTRPM2 introduction of an acidic residue in the pore of hTRPM8 would result in a channel variant with an increased permeability for Ca2+. To test this assumption, we generated the hTRPM8Q914E variant and evaluated the ion permeation properties of the channel in response to its specific synthetic agonist, icilin (Fig. 3
A). Similar to the situation with hTRPM2, Erev values determined in the presence of divalent cations were shifted in the mutant channel toward more positive values. The comparison of the Pdiv/PNa values of hTRPM8WT and hTRPM8Q914E revealed that the mutant channel was significantly more permeable for Ca2+ and Ba2+ (Fig. 3B
).
Together, we conclude that the ancient pore residues restored in either hTRPM2 or hTRPM8 facilitate the permeation of divalent cations through the channels.
Introduction of modern pore residues into mTRPM7 results in the loss of divalent cation permeability
TRPM7 has been characterized as a channel that is highly permeable to a broad range of divalent cations including Ca2+ and Mg2+ (20
, 21)
. We reasoned that if mutations in only two residues were sufficient to significantly increase the Ca2+ permeability of hTRPM2, reciprocal changes in the putative pore loop of mouse TRPM7 would cause an opposite effect, i.e., a decrease in divalent cation permeability. To test this assumption, we generated a set of mTRPM7 variants harboring the aforementioned hTRPM2-like signature residues and compared the biophysical properties of the mutated and wild-type channels. As illustrated in Fig. 4
A, in the presence of physiological concentrations of extracellular divalent cations (Na+div external bath solution) the wild-type channel (mTRPM7WT) yielded small inward currents and large outward currents with pronounced rectification at positive potentials. In contrast to mTRPM7WT, mTRPM7E1047Q and mTRPM7Y1049P mediated significantly larger inward currents (Fig. 4A
). As in hTRPM2, the double mutations in mTRPM7 had a stronger effect on channel properties than the single mutations (Fig. 4A
). The expression of mTRPM7E1047Q/Y1049P resulted in currents with a linear I-V relationship strikingly similar to hTRPM2.
We next asked whether the pore mutations affected the cation permeability of mTRPM7 and performed cation substitution experiments in analogy to the measurements with hTRPM2 and hTRPM8 variants. As shown on Fig. 4A, in
the presence of 10 mM Ca2+, Mg2+, or Ba2+ in the external bath solutions, the mutated mTRPM7 variants showed pronounced shifts of Erev toward negative values. We next compared the relative cation permeability of mTRPM7 variants determined from Erev values (Fig. 4B
). In contrast to the wild-type channel, all mTRPM7 mutants were significantly impaired in their ability to effectively discriminate between Na+ and Ca2+ or Mg2+. Remarkably, mTRPM7E1047Q was nearly 30-fold less permeable for Ca2+ or Mg2+ than mTRPM7WT.
To test whether the mutated mTRPM7 variants could still allow for Ca2+ influx when the intracellular ion composition was not manipulated by pipette solutions, we established an aequorin-based bioluminescence approach to monitor mTRPM7-mediated Ca2+ entry (Fig. 4C
). In the absence of Mg2+ in the bath solutions, an increase from 1 to 5 mM extracellular [Ca2+] resulted in a fast and pronounced rise of intracellular [Ca2+]i in mTRPM7WT expressing cells. In accord with the patch-clamp recordings, expression of mutant variants did not entail significant changes in [Ca2+]i (Fig. 4C
). We next examined whether mTRPM7E1047Q can still conduct Mg2+ when the latter cation is the only cation charge carrier available, i.e., in Mg2+-based extracellular solution (see Supplemental Fig. 2). In the whole-cell mode, we found that under these conditions mTRPM7E1047Q did not mediate significant inward Mg2+ currents. To test whether mutations of E1047 and Y1049 would interfere with protein expression, we performed Western blots using a mTRPM7-specific antibody (48)
and found that mTRPM7E1047Q, mTRPM7Y1049P, and mTRPM7E1047Q/Y1049P did not differ from the wild-type channel in their cellular protein levels (data not shown).
The effect of the negatively charged E1047 on the conduction properties of mTRPM7 is particularly noteworthy. A straightforward explanation for this phenomenon would be that the side chain carboxyl group of E1047 might constitute a high affinity-binding site for divalent cations and its elimination would result in a channel pore impermeable for divalent cations. To test this hypothesis, we substituted E1047 by an acidic residue with a shorter side chain length, i.e., D, and examined its effect on channel properties (Fig. 5
A, B). As expected, the E1047D mutation reduced but did not abolish the preference for divalent cations. However, introduction of N (mTRPM7E1047N) caused a dramatic loss of permeability to divalent cations as in the case of mTRPM7E1047Q (Fig. 5A, B
). In line with these findings, Ca2+ entry monitored by the aequorin-based approach was substantially lower in mTRPM7E1047D-expressing cells than in the presence of mTRPM7WT. The expression of mTRPM7E1047N did not elicit detectable changes in [Ca2+]i when compared with untransfected cells (Fig. 5C
).
In an additional set of experiments, we evaluated the role of another negatively charged mTRPM7 residue, E1052. Although E1052 is located in close vicinity to the decisive E1047, mTRPM7E1052Q displayed only modest changes in cation permeability ratios and Ca2+ entry when compared with other mutants examined (Fig. 5)
.
We next asked whether a hydroxyl group provided by the side chain of Y1049 is important for permeation of divalent cations through mTRPM7. To this end, we exchanged Y1049 to the nonpolar F (mTRPM7Y1049F) and investigated the channel characteristics as described above. The permeation properties of mTRPM7Y1047F were found to be indistinguishable from those of mTRPM7WT (data not shown), suggesting that the aromatic ring moiety of Y1049 rather then the hydroxyl group is critical determinant of mTRPM7 functional properties.
E1047 is essential for the sensitivity of mTRPM7 to extracellular Mg2+
TRPM7 is essential for cellular Mg2+ homeostasis (32
, 49)
, and intracellular and extracellular Mg2+ levels tightly regulate TRPM7 channel activity (20
, 21)
. External Mg2+ is a permeant blocker (20
, 21
, 50)
acting via a high affinity binding site in the TRPM7 pore region (51)
. Internal Mg2+ inhibits TRPM7 via a nucleotide-binding site in the kinase domain synergistically with another Mg2+-binding site extrinsic to the kinase domain (52)
. In light of these findings, we examined the possible effect of E1047 on the Mg2+ sensitivity of mTRPM7. To this end, we compared the effect of 10 and 20 mM extracellular Mg2+ concentration ([Mg2+]o), as well as of other divalent cations on the activity of mTRPM7WT and mTRPM7E1047Q (Fig. 6
A). In contrast to mTRPM7WT, the expression of mTRPM7E1047Q resulted in outward monovalent cation currents that were not suppressible by extracellular Mg2+, Ca2+, or Ba2+, indicating that mTRPM7E1047Q had completely lost the permeant block of monovalent currents by extracellular divalent cations.
We next examined the influence of [Mg2+]o on outward monovalent currents in a set of mTRPM7 variants carrying distinct mutations of the critical E1047 (Fig. 6B
). In accord with our previous experiments, neutralization of E1047 was sufficient to abrogate the sensitivity of mTRPM7 to [Mg2+]o (Fig. 6B
). In contrast, channel mutants containing either an acidic residue with shorter side chain at position 1047 or a charge-neutralized neighboring E1052 displayed only a reduced sensitivity to [Mg2+]o: the calculated IC50 values were 0.8 ± 0.1, 1.2 ± 0.1, and 2.8 ± 0.2 mM for mTRPM7WT, mTRPM7E1047D, and mTRPM7E1052Q, respectively.
Next, we compared the effect of intracellular Mg2+ levels on mTRPM7WT and mTRPM7E1047Q (Fig. 6C
) and found that the regulation of the channel by intracellular [Mg2+] ([Mg2+]i)was not affected by the E1047Q mutation. The determined IC50 values were 0.35 ± 0.03 and 0.33 ± 0.02 mM for mTRPM7WT and mTRPM7E1047Q, respectively. This observation is consistent with the idea (52)
that intracellular Mg2+ regulates mTRPM7 via other channel segments.
Collectively, we conclude that the highly conserved EVY sequence motif determines the divalent cation selectivity and key biophysical features of mTRPM7 such as the characteristic I-V relationship and the permeant block by external Mg2+.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Evolution of domain architectures of TRPM channels
The phylogenetic analysis presented here allocates the Nudix-linked TRPM channels to a central position within the TRPM evolutionary history. Thus, we noted that 1) TRPM2-like channels are present in various phyla, ranging from protists to humans; 2) a TRPM2-like channel is the only type of TRPM protein in two ancient phyla, choanoflagellata and cnidaria; and 3) only bilateral species contain additional TRPM genes lacking the Nudix domain. These findings strongly support the notion that an ancient TRPM2-like protein is the template that was modified during evolution to yield the diversity of TRPM channels. The phylogenetic data also suggest that a loss of Nudix domain-coding exons in one of the TRPM2-like genes was a critical evolutionary event giving rise to TRPM3-like channels in bilaterian organisms (Fig. 7
A).
|
According to amino acid similarity, domain architecture, and acquisition of the characteristic "QIP" pore signature, the mammalian TRPM gene family falls into two main branches (Fig. 7A
). The first group (TRPM1/3/6/7) most likely evolved from a common TRPM3-like ancestor. Within the latter subfamily, TRPM6/7 emerged in early vertebrates due to a fusion of genes encoding an -kinase and a TRPM3-like channel. The second family branch (TRPM4/5/8) most likely evolved after a secondary loss of the Nudix domain in a putative TRPM2-like ancestor already containing the characteristic QIP pore residues (Fig. 7A
).
To date, the biological importance of evolutionary modifications within the C-terminal segments of TRPM channels remains elusive. There is evidence to show that the Nudix domain serves as a ligand-binding site and is essential for ADP-ribose-mediated activation of hTRPM2 (24
, 34)
. Along these lines, we hypothesize that the loss of the Nudix domain and the acquisition of the new C-terminal domain in TRPM6/7 imply the development of novel regulatory mechanisms of TRPM channels.
Molecular determinants of distinct permeation properties of TRPM channels
The diversity of the functional properties of TRPM channels offers a unique possibility to use a comparative analysis of protein sequences to identify the molecular determinants of characteristic channel features. As a proof of principle, we searched for the molecular basis underlying the divergent ion permeabilities displayed by mammalian TRPM channels. To this end, our molecular modeling and comparative sequence analysis revealed that highly conserved acidic and aromatic residues located within a proximal segment of the putative pore loop are likely determinants of the cation selectivity of TRPM channels. This prediction was tested functionally in enzyme-linked channels representing distinct family branches, hTRPM2 and mTRPM7.
mTRPM7 is highly permeable to a broad range of divalent cations, including Ca2+ and Mg2+ (20
, 21)
. We observed that two residues, E1047 and Y1049, are critical for the divalent cation selectivity of the channel. Thus, the mTRPM7Y1049P variant displayed a very low preference for Ca2+ or Mg2+ as compared with Na+, whereas the mTRPM7E1047Q mutant was found to be essentially impermeable for divalent cations. Moreover, the double mutant E1047Q/Y1049P showed a linear I-V dependence resembling hTRPM2-mediated currents and a lack of permeant block of extracellular divalent cations. hTRPM2 and hTRPM8 are nonselective cation channels, i.e., they poorly discriminate between monovalent and divalent cations. The restoration of ancient pore residues in hTRPM2 and hTRPM8 resulted in a substantial increase in the permeability of the channels for Ca2+. Thus, several independent lines of evidence are consistent with the notion that the EVY motif in mTRPM7 is a molecular determinant of the high permeability of the channel for divalent cations. The EVY motif is highly conserved in TRPM channels, including ancient TRPM2-like proteins, suggesting that these amino acids play a similar role in related channels. Consequently, their loss in hTRPM2 and hTRPM8 resulting in the modern pore motif QIP may represent a critical evolutionary step in the development of the low cation selectivity of the latter channel.
Interestingly, as with mTRPM7, an essential role of acidic pore residues for the conductance of divalent cations has been shown for a subset of other TRP channels (4
5
6
7
8)
. Our findings, however, highlight two important aspects unrecognized in previous studies. First, we establish a critical role of the aromatic residue Y1049 for the divalent selectivity of mTRPM7. Second, we demonstrate that Y1049 and another crucial residue E1047 have a strong epistatic relationship, a readily observed phenomenon in the molecular evolution of new functional properties (53)
. We speculate that critical acidic pore residues in other TRP channels may also have a functional interaction with neighboring aromatic residues.
What could the function of E1047 and Y1049 in the predicted pore loop of TRPM7 be? We found that shortening the side chain of an acidic residue (E1047D mutation) reduced the permeability of mTRPM7 to divalent cations. In contrast, neutralization of E1047 (E1047Q or E1047N mutations) abolished the permeation of divalent cations through the channel. These findings suggest that the negatively charged oxygen atom of the carboxyl group of E1047 is involved in the formation of a high-affinity binding site for divalent cations, thus providing the molecular basis for the ion selectivity characteristics of mTRPM7. In search of a molecular explanation for the role of Y1049 in mTRPM7, we noted that all invertebrate TRPM channels harbor a nonpolar aromatic F at the corresponding position. The mTRPM7Y1049F variant was found to be functionally indistinguishable from the wild-type channel. Therefore, we assume that the side chain hydroxyl group of Y1049 has little importance, and most likely the presence of an aromatic ring is critical. It has been shown that the pore loop of potassium channels also contains an invariant Y that forms a specific packing interaction with an aromatic residue of the pore helix, while its carbonyl group derived from the peptide backbone interacts with permeating cations (43
, 54)
. We speculate that Y1049 in mTRPM7 (F in other TRPM channels) may play a similar role.
The 1047EVY1049 motif in mTRPM7 reported here differs substantially in length and location from a segment that has recently been suggested to be the possible selectivity filter in hTRPM6 (1028GEIDVC1033 in hTRPM6 corresponding to 1051YEIDVC1056 in mTRPM7; ref. 30
). While this article was in preparation, Li et al. (55)
reported that neutralization of E1052 or E1047 dramatically affects mTRPM7 selectivity for divalent cations. Although our studies are in overall agreement with Li et al. concerning the critical role of E1047, the mutation of E1052 in mTRPM7 did not affect the permeability to Ba2+ and resulted in only modest changes in the permeation of Ca2+ and Mg2+ in our experiments.
mTRPM7 is highly permeable to Ca2+ and Mg2+. However, the physiological relevance of Ca2+ vs. Mg2+ entry mediated by the channel is still poorly understood. The mTRPM7 pore mutants reported here could be instrumental to experimentally assess the role of Ca2+ vs. Mg2+ fluxes for cellular TRPM7 functions. A similar aspect is also applicable for hTRPM2, which is equally permeable for Na+ and Ca2+. Accordingly, hTRPM2Q981E/P983Y displaying a 4 fold higher permeability for Ca2+ can be used to evaluate the relevance of Na+ vs. Ca2+ fluxes mediated by hTRPM2.
Evolutionary switch of Ca2+ permeability in TRPM channels
Sequence comparison of the predicted pore loop segments shows that a subset of vertebrate-specific channels harbors the hTRPM2-like QIP motif. In contrast, protist TRPM2-like proteins and the majority of other TRPM channels contain the ancient or TRPM7-like motif. The members of the latter group characterized so far display a relatively high selectivity for Ca2+, while TRPM channels with modern residues are either completely Ca2+ impermeable (TRPM4/5) or poorly Ca2+ selective (TRPM2 and TRPM8). We demonstrate here that restoration of the two ancient residues in the predicted pore of hTRPM2 results in an 4-fold increase in the Ca2+ permeability of the channel. Collectively, these observations are very compatible with the idea that the putative TRPM2-like ancestor was a highly Ca2+-permeable channel. Consequently, a major evolutionary trend in vertebrate TRPM channels can be described as a reduction in Ca2+ selectivity due to the acquisition of the QIP motif in the pore loop (Fig. 7B
).
Two observations in the literature support this hypothesis. First, Ca2+-selective outwardly rectifying calcium channel currents (IORCa) (PCa/PNa is 60–70) were described in C. elegans epithelial cells (56)
. In terms of I-V relationship and Mg2+ dependence, the IORCa channel characteristics resemble those of TRPM3 and TRPM7 (56)
. Recent genetic data indicate that the C. elegans TRPM channel Gon-2 is the molecular substrate of IORCa (39)
. Second, an ADP-ribose-dependent "fertilization" cation channel has been described in oocytes of the sea squirt Ciona intestinalis (57)
. This channel displays a high permeability to Ca2+ and an I-V relationship strikingly similar to that of hTRPM2Q981E/P983Y. Therefore, we assume that a sea squirt TRPM2-like protein is the molecular candidate for the fertilization channel. Remarkably, both Gon-2 and the C. intestinalis TRPM2-like protein contain the ancient signature in their pore segments.
Finally, our studies with mTRPM7 show that the exchange of only a single amino acid residue in the putative pore segment is sufficient to yield a novel type of channel. For instance, the lack of Ca2+ permeation in mTRPM7E1047Q is strikingly reminiscent of the functional properties of TRPM4/5. It is imaginable that the monovalent selectivity of TRPM4/5 has evolved primarily due to the neutralization of the corresponding acidic residues in their pores. Indeed, Nilius et al. (29)
reported that the exchange of Q977 to E in hTRPM4 resulted in a channel that is moderately permeable to Ca2+. Thus, in terms of diversity in domain architectures and functional properties, the TRPM gene family is an outstanding example of how a particular ancestor protein can be evolutionary differentiated into multiple types of cation channels.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication September 13, 2007. Accepted for publication November 8, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. C. Hermosura, A. M. Cui, R. C. V. Go, B. Davenport, C. M. Shetler, J. W. Heizer, C. Schmitz, G. Mocz, R. M. Garruto, and A.-L. Perraud Altered functional properties of a TRPM2 variant in Guamanian ALS and PD PNAS, November 18, 2008; 105(46): 18029 - 18034. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Xia, Z.-Z. Mei, H.-J. Mao, W. Yang, L. Dong, H. Bradley, D. J. Beech, and L.-H. Jiang Identification of Pore Residues Engaged in Determining Divalent Cationic Permeation in Transient Receptor Potential Melastatin Subtype Channel 2 J. Biol. Chem., October 10, 2008; 283(41): 27426 - 27432. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
