| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 278, Issue 41, 40373-40384, October 10, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
¶ 
From the
Graduate School of Bioagricultural Sciences and the ¶Bioscience and Biotechnology Center, Nagoya University, Nagoya 464-8601, the ||Faculty of Agriculture, Yamaguchi University, Yamaguchi 753-8515, and the **Department of Molecular Biology, Graduate School of Medical Science, Kyushu University, Fukuoka 812-8582, Japan
Received for publication, July 11, 2003 , and in revised form, July 24, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
The K+ channel ion conduction pore has two transmembrane segments (M1 and M2), a hydrophobic region responsible for K+ selectivity (P region), and N- and C-terminal cytoplasmic domains (Fig. 1). Recent high resolution structural data for the "membrane-pore-membrane" structure of the bacterial K+ channels, KcsA, MthK, KvAP, and KirBac 1.1, have revealed that the P region forms a loop structure, N-terminal to C-terminal, consisting of a helix, a turn, and a random coil (5-9). The overall structure of the pore-forming region is highly conserved not only in Na+ and Ca2+ channels but also in plant and bacterial Na+/K+ transporters (HKT, Ktr, and Trk) (10-14).
|
|
|
|
The arrangement of M1 and M2 is reported to differ between the bacterial K+ channel, KcsA, and the animal K+ channel, Kir 2.1. In KcsA, M1 and M2 interact intramolecularly, whereas in Kir 2.1, they interact both intra- and intermolecularly (17). Kir 2.1 forms a salt bridge between Glu and Arg in the P region, which is not present in KcsA or plant voltage-dependent K+ channels, and Kir 2.1 is therefore categorized as a structurally distinct class (17). It is not known whether these differences in the final membrane structure are due to differences in the topogenic functions of M1, the P region, and M2 in the two proteins. In ion transport proteins, the amino acid sequence forming the ion-conducting pathway is often surrounded by other segments of the polypeptide (18, 19). However, information on how the pore is formed and becomes embedded in the membrane is limited. In this study, we examined which topogenic features of M1, the P region, and M2 are shared by, or differ in, Kir 2.1 and KcsA and characterized the membrane biogenesis of K+ channels.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
|
|
Expression of Kir 2.1 in COS-7—cDNAs encoding full-length Kir 2.1 with or without the Q141N mutation were amplified by PCR with the HindIII site-containing primer, 5'-ATAAGCTTCCACCATGGGCAGTGTGAGAACCAAC-3', and the SpeI site-containing antisense primer, 5'-TTACTAGTTATCTCCGATTC-3'. The HindIII-SpeI-digested DNA was subcloned into the HindIII and XbaI sites of the pRcCMV-based plasmid constructed by Miyazaki et al. (25), and Kir 2.1 tagged C-terminally with influenza virus hemagglutinin (HA) was expressed from the resultant plasmid. Protein expression was performed essentially as described by Miyazaki et al. (25). COS-7 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum under a 10% CO2 atmosphere. Transfection was performed according to standard procedures using the FuGENE reagent (Roche Applied Science). After transfection, COS-7 cells in 3.5-cm culture dishes were cultured for 24 h and then lysed by sonication for 1 min in the denaturing buffer containing 0.5% SDS and 1% 
-mercaptoethanol. Where indicated, aliquots of the lysate were treated for 1 h at 37 °C with either endoglycosidase H (New England Biolabs) or peptide N-glycosidase F (New England Biolabs) using the conditions recommended by the supplier. The same amounts of protein were analyzed by SDS-PAGE and subsequent immunoblotting using a monoclonal anti-HA antibody (16B12, Covance, CA), bound antibody being detected using horseradish peroxidase-labeled anti-mouse IgG (BIOSOURCE) and ECL reagent (Amersham Biosciences).
Assay for N-Myristoylation and N-Acetylation of Kir 2.1—Plasmid pBDpro-TNF containing cDNA coding for the mature domain of TNF was constructed as described previously (26). Plasmids pBV2G,R3A-TNF and pBV2G,R3D-TNF (previously designated, respectively, as pBR3A-TNF and pBR3D-TNF) were constructed as described previously (27). Plasmid pBKir 2.1-TNF was constructed by PCR using pBDpro-TNF as template and the primers, 5'-ATATGGATCCATGGGCAGTGTGCGAACCAACCGCTACAGCGACAAGCCTGTAGCC-3' and 5'-GCCGGGATCCTAGGGCGAATTGGGTACC-3'. After digestion with BamHI and PstI, the amplified product was subcloned into pB at the BamHI and PstI sites. T3 polymerase was used to obtain transcripts of these cDNAs. The translation reaction was carried out using rabbit reticulocyte lysate (Promega) in the presence of [3H]leucine, [3H]myristic acid, or [3H]acetyl-CoA (Amersham Biosciences) using the conditions recommended by the manufacturer. The mixture (20 µl of rabbit reticulocyte lysate; 1 µl of 1 mM leucine-free amino acid mixture, or 1 mM complete amino acid mixture; 4 µl of [3H]leucine (5 µCi), [3H]myristic acid (25 µCi), or [3H]acetyl-CoA (2 µCi); and 4 µl of mRNA) was incubated for 90 min at 30 °C, and then the translation products were immunoprecipitated with polyclonal anti-hTNF antibody (R&D Systems) and protein G-Sepharose (Amersham Biosciences) and analyzed by SDS-PAGE and fluorography.
Yeast Complementation—To express the chimeric protein in yeast, cDNA encoding full-length Kir 2.1 was amplified by PCR using 5'-ATAAGCTTCCACCATGGGCAGTGTGAGAACCAAC-3'and 5'-TTGGATCCTCATATCTCCGATTCTCG-3', and the product was inserted into the pYES2-based vector used previously (22). The constructs were expressed under the control of the Gal1 promoter in the plasmid. Saccharomyces cerevisiae CY162, a gift from Dr. R. F. Gaber, was transformed as described previously (28). For complementation of S. cerevisiae CY162 with mutated Kir 2.1 (Q140N, or with M1 replaced with Met1-Leu28 of DPP IV), medium containing 6.7 g/liter of yeast nitrogen base (Difco), 2% galactose, 2% sucrose, 2% agar, and essential amino acids was used.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
The P region construct for either Kir 2.1 and KcsA did not give either glycosylated bands or proteinase K-resistant fragments (Fig. 2, B and C, lanes 4-6), showing that, although the P region is hydrophobic, it cannot integrate into the membrane on its own. This is consistent with the previously reported lack of topogenic function of the P region of the voltage-dependent K+ channel, KAT1 (15).
M2 in both Kir 2.1 and KcsA had SA-I function, as the constructs showed a glycosylated band (Fig. 2, B, lane 8, and C, lane 8). In addition, KcsA M2 showed a protease-protected band (Fig. 2C, lane 9), demonstrating that it possessed both SA-I and SA-II function.
Assessment of the Possibility of an Ncyt/Cexo Topology when the N-terminal of the Segment Is Located in the Cytoplasm—Although a transmembrane segment does not have SA-II function, it is possible that the segment initiates translocation when the N terminus of the segment faces the cytoplasm (21). We therefore evaluated the efficiency of the translocation initiation function of each transmembrane segment using the constructs shown in Fig. 3A; this function has been referred to as the internal SA-II function (21) or reinitiation function (15). For these experiments, the H1 segment of the E. coli leader peptidase (H1) was placed at the N terminus of M1; H1 acts as an SA-I transmembrane segment, and its function can be confirmed by the appearance of N-glycosylated bands (Fig. 3A) (21). If the inserted KcsA or Kir 2.1 transmembrane segment can integrate in the Ncyt/Cexo orientation, the C-terminal PL domain should become proteinase K-resistant (Fig. 3A, left). After proteinase K treatment, the constructs containing M1 of Kir 2.1 or M1 or M2 of KcsA gave substantial amounts of the truncated form (shown by an arrowhead; Fig. 3, B and C). These results show that, as long as the N-terminal end of M1 in Kir 2.1 or KcsA faced the cytoplasmic side, M1 could be correctly and effectively integrated into the membrane in the Ncyt/Cexo orientation. The translocation initiation function for each segment is shown in Fig. 3, B and C (right panels). M2 of Kir 2.1 showed low efficiency of membrane insertion, and the P region in both Kir 2.1 and KcsA did not show insertion.
Modification of the N-terminal Region of Kir 2.1—Post-translational modification of proteins assists in anchoring the protein in the membrane. The amino acid sequence of the N terminus of Kir 2.1 (MGSVRTNR) fits to the N-myristoylation consensus motif, MGXXX(S/T)XX (27, 30, 31). Protein N-myristoylation results from the cotranslational addition of myristic acid, a 14-carbon saturated fatty acid, to the Gly residue at the extreme N terminus after removal of the initiating Met by methionine aminopeptidase (32). The N-terminal sequence of Kir 2.1 also contains an N-acetylation consensus sequence, M(G/A/S/T) (32), which might be acetylated. Constructs coding for wild-type mature TNF (
pro-TNF), containing the sequence MVRSSSRTPS, mutants V2G, R3A-TNF, and V2G, R3D-TNF containing the indicated replacements site, and Kir2.1-TNF in which the first 10 amino acids were replaced by the N-terminal 10 amino acid residues of Kir 2.1 (Fig. 4A), were tested for their susceptibility to protein N-myristoylation and N-acetylation by metabolic labeling in an in vitro translation system (27). The mutant TNF proteins, V2G, R3A-TNF, and V2G, R3D-TNF, were used to confirm the occurrence of N-myristoylation and N-acetylation, respectively (27) (Fig. 4B). The Kir 2.1-TNF protein was acetylated but not myristoylated (Fig. 4B).
Assessment of the Stop-Transfer Function—The stop-transfer function interrupts ongoing translocation through the translocon. According to the "conventional insertion model" initially proposed by Blobel and co-workers (2-4), a stop-transfer function is essential for a transmembrane segment located after a Ncyt/Cexo-oriented transmembrane segment, such as M2. The stop-transfer function was examined as described previously (33) (Fig. 5A). The N-terminal signal peptide (S) on the fused protein initiates translocation, and then the following segment allows translocation into the membrane. If the inserted transmembrane segment shows stop-transfer function, the C-terminal half of the molecule (the second fused PL domain) is exposed on the cytoplasmic side of the membrane, and the domain is hydrolyzed by externally added proteinase K (Fig. 5A, left), and if the inserted transmembrane segment does not show stop-transfer function, the nascent polypeptide remains in the lumen space and is proteinase-resistant (Fig. 5A, right). Proteinase K treatment gave rise to truncated fragments of Kir 2.1 and KcsA (arrowheads; Fig. 5, B and C), indicating that M1 and M2 in both Kir 2.1 and KcsA showed a stop-transfer efficiency of almost 100% (Fig. 5, B and C). The P region did not have significant stop-transfer function and could not span the membrane.
The results for each segment of Kir 2.1 and KcsA in the present paper and for the voltage-dependent K+ channel, KAT1, (15) are summarized in Fig. 5D. KcsA M1 is probably able to integrate into the membrane without any support from the N-terminal cytoplasmic sequence. In contrast, the effective correct integration of the M1 in Kir 2.1, and the S5 in KAT1 is highly dependent on the localization of the preceding region. None of the three P regions have topogenic function. All M2/S6 have stop-transfer function to help stop their translocation and have SA-I and SA-II function with their different efficiency.
Exposure of the P Region on the Lumen Side of the ER during Membrane Integration—According to the crystal structure data for KcsA, MthK, KvAP, and KirBac 1.1 (5-7, 9), the P region exists as a loop in the membrane. During insertion into the membrane, the P region was thought to form the final structure inside the membrane space without ever being exposed in the lumen. To monitor the process of integration, we introduced potential N-glycosylation sites into the P region, because oligosaccharyltransferase can access polypeptide chains emerging from the translocon (34, 35). We generated an NXT sequence for N-glycosylation by a single amino acid substitution at the bottom of the loop in Kir 2.1 (substitution of Asn for Gln at position 140) and KcsA (substitution of Asn for Ala at position 73) (Fig. 6A). We also tested the P region of the plant Shaker-type K+ channel, KAT1, between S5 and S6 by mutating the corresponding position in the P region (Leu to Asn at position 258). All three constructs with glycosylation sites in the P region gave a glycosylated band in the presence of RM (Fig. 6B, lanes 5, 11, and 17), strongly suggesting that the P region was exposed on the lumen side of the ER during the membrane biogenesis process. We tested the effect of speed of peptide synthesis on integration of the P region into the membrane using the full-length Kir 2.1 (with and without Q140N mutation)-HA constructs shown in Fig. 6C and cultured COS-7 cells, in which synthesis and folding of membrane proteins proceed more rapidly than in the in vitro system (36). When expressed transiently in COS-7 cells, Kir 2.1-Q140N-HA, but not wild-type Kir 2.1-HA, showed a single glycosylated band, which was lost following treatment with endoglycosidase H or peptide N-glycosidase F (Fig. 6D). This shows that, as in the in vitro system, the P region was exposed on the lumen side of the ER before the protein arrived at the final membrane topology (Fig. 6E). Unexpectedly, the intensity of glycosylation was no less than in the in vitro system. It should be noted that, when the Q140N Kir 2.1 mutant was tested for functional complementation in yeast, it did not perform K+ uptake, presumably because the point mutation and/or the N-glycosylation disrupted K+ uptake activity (Fig. 6F).
By using the protease protection assay (Fig. 6B, lanes 3, 6, 15, and 18), M2 in Kir 2.1 and S6 in KAT1 were confirmed to take up the Nexo/Ccyt orientation in the membrane. With the KcsA constructs, although the isolated M2 showed strong stop-transfer function (Fig. 5), a protease-protected band was seen (Fig. 6B, lanes 9 and 12), suggesting that M2 in KcsA failed to completely stop its translocation. The contribution of the KcsA C-terminal region to membrane retention of M2 is examined below (Fig. 9).
The Influence of M1 on the Loop Structure of the P Region—We tested the role of M1 in the formation of the P region and K+ channel activity by replacement of M1 in Kir 2.1 with the unrelated Ncyt/Cexo-oriented hydrophobic transmembrane segment of DPP IV (Fig. 7A). Fig. 7B shows that the fusion bearing the Q140N replacement in the P region was glycosylated. The replacement of M1 did not affect the exposure of the P region in the ER lumen. The construct of wild-type Kir 2.1 in which M1 was replaced by the DPP IV peptide sequence did not confer K+ uptake activity in the yeast complementation assay (Fig. 6F).
Contribution of the Turn in the P Region to the Loop Structure—According to the x-ray crystal structure for KcsA, the hydrophobic P region is bent at the middle to form the loop (9). The amino acid sequence, TATT, constitutes the turn and is located at the C-terminal of the pore helix (8) (Fig. 1). The following random coil contains the K+-selective GYG signature sequence. We hypothesized that the turn helps to fold back the polypeptide in the P region. To examine this, we tried to extend the equivalent pore helix of Kir 2.1 (TQTT) by replacing all three Thr residues at positions 139, 141, and 142 with Leu, which has a high helical propensity (34) (Fig. 8A). Several secondary structure prediction methods (pbil.univ-lyon1.fr/) suggest that these replacements extend the 
-helix (Fig. 8A); in addition, these replacements make the region more hydrophobic, as shown by the Kyte and Doolittle method (37) (data not shown). However, the N terminus-P region construct (Kir 2.1-M1-P-PL) showed a protease-protected band, indicating translocation of the P region into the ER lumen (Fig. 8, B and C, lane 3). By using Kir 2.1-M1-P-M2-PL containing the 3 Thr replacements, this band was not seen because of correct membrane insertion (Fig. 8C, lane 6). The mutations in the TQTT sequence therefore did not affect the exposure of the P region in the ER lumen.
Effect of the Salt Bridge between Glu138 and Arg148 on the Formation of the P Region—In Kir 2.1, Glu138 and Arg148 in the P region make a salt bridge (17). We tested the effect of this bridge on the correct topogenesis of the Kir 2.1-M1-P-PL construct by replacing Glu138 in the Q140N mutant with Leu and found that the Q140N glycosylated band was seen, and translocating PL was detected (Fig. 8, D and E). In Kir 2.1-M1-P-M2-PL containing the same replacements, this band was not seen. The disruption of the salt bridge therefore did not induce a significant change in the membrane biogenesis of the P region.
Contribution of the C-terminal Region to M2 Membrane Integration in KcsA—In Fig. 6B, lanes 9 and 12, translocation of M2 in KcsA was detected. We assumed that the cytoplasmic C-terminal region of KcsA stabilized the retention of M2 in the ER membrane. To test this, we introduced an N-glycosylation site into the C-terminal PL of KcsA-M1-2-PL and KcsA-M1-2-A73N-PL used in Fig. 6B and into the full-length wild-type and A73N mutant constructs (Fig. 9A). As shown in Fig. 9B, M1-2-PLgly showed a glycosylated band (lane 5), and M1-2-A73N-PLgly gave both single- and double-glycosylated bands (lane 11). The glycosylated bands indicate translocation of M2. However, by using full-length wild-type KcsA-PLgly, the intensity of the glycosylated band was less (lane 17), and by using full-length-A73N-PLgly, the double-glycosylated band was not seen (lane 23). In addition, both constructs gave a weaker protease-protected band (Fig. 9B, lanes 18 and 24). These results show that the C-terminal region can help M2 to span the membrane.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
The M1 regions in Kir 2.1 and KcsA showed different topogenic function. M1 in KcsA possessed strong SA-II function, as would be expected for the first transmembrane region of a two-transmembrane region-type K+ channel. In contrast, M1 in Kir 2.1 had SA-II function but also had a stronger SA-I function, which led to the wrong orientation of insertion, and the cytoplasmic region preceding M1 helped in the correct Ncyt/Cexo orientation of M1 (Figs. 2C and 3). Interestingly, in the KAT1 channel, S5, which corresponds to M1, shows no SA-II function (15) (Fig. 5D). Because M1 is located further than M2 from the P region, forming the outer helix (5-7, 9), we tested whether M1 in Kir 2.1 could be replaced by other hydrophobic transmembrane domains by replacing it with another SA-II sequence from DDP IV unrelated to M1 (Fig. 7). The resultant chimeric protein showed the same membrane biogenesis of the P region as the wild-type protein, indicating that the topogenesis of the P region is not influenced by characteristics of amino acids in M1.
M2 is highly hydrophobic and forms the inner helix in the K+ channel (9). According to the previous study (15), S6 in KAT1 shows not only a stop-transfer function but also has a strong SA-I function (45%). Although we were unable to determine whether SA-I function is essential for S6, its SA-I function is likely to help the translocation of the P region into the lumen and the retention of S6 in the membrane. In contract, M2 in both Kir 2.1 and KcsA showed 
16% SA-I function, which is lower than that for S6 (Fig. 2, B and C). The weak SA-I function of M2 in KcsA may lead to the translocation of M2 seen in Fig. 6B, lanes 9 and 12, whereas the C-terminal cytoplasmic region helps to prevent M2 from translocating into the ER lumen (Fig. 9, A and B).
When we determined the topogenic function of the individual segments, the P region in Kir 2.1 and KcsA did not exhibit substantially any topogenic function (Fig. 5), a similar result to that previously observed for the KAT1 channel (15). The P region in K+ channels forms a loop consisting of a descending -helix and an ascending stretch containing the selective filter (5-9). The crystallography data show a loop structure for the P region, which is possibly formed within the membrane. To test this, we monitored the location of the P region by creating a novel N-glycosylation site in the P region of Kir 2.1 by a small change in a side chain (Gln140 replaced by Asn), resulting in a potential glycosylated site at the bottom of the loop far from the ER lumen. This residue was strongly glycosylated both in the in vitro system and in COS-7 cells (Fig. 6D). The latter result was unexpected, as the rate of protein synthesis and the speed of folding of membrane proteins is greater in living cells than in the cell-free translation/translocation system (36), and we therefore expected to see less glycosylation. Similar N-glycosylation at the corresponding position has been reported in another two-transmembrane segment-type K+ channel, ROMK (42). This attachment of carbohydrate to the P region shows that the P region is exposed to the lumen space before the channel arrives at the final topology (Fig. 6E).
We also examined the structural determinants that support the membrane biogenesis of the pore-forming structure in Kir 2.1 channel. The turn between the pore helix and the GYG sequence (TATT in KcsA and TQTT in Kir 2.1) is probably responsible for the bend in the middle of the P region. We predicted that, if the turn were removed, the P region might become transmembrane, and we therefore replaced TQTT in Kir 2.1 with LQLL, as Leu tends to form a helix. This replacement was predicted by various secondary structure models to give a longer helix than the original pore helix (Fig. 8A). However, the P region containing LQLL was not transmembrane (Fig. 8C). This failed attempt at converting the P region from a loop to a transmembrane segment suggests that, despite its hydrophobic properties, the P region shows little propensity to take up a transmembrane configuration.
We analyzed the effect of the salt bridge in Kir 2.1 between Glu138 and Arg148 in the P region which is involved in the stabilization of the pore structure, as shown by Yang et al. (43), who demonstrated that the reversal double mutant channel, E138R/R148E, exhibits a significant decrease in current amplitude and was converted to a Na+-permeable K+ channel. According to the KcsA structure, Glu138 is located at the bottom of the P region and at the end of the pore helix. A single mutation at the corresponding position in KAT1 also induces a change in ion permeability (22, 44). When we converted Glu138 to a hydrophobic residue, Leu, in order to break the salt bridge (Fig. 8, D and E), the glycosylated band within the P region was still seen, and M2 was correctly inserted into the membrane. The salt bridge therefore does not act as a determinant for the topogenesis of the P region.
In this study, we found that the P region showed no topogenic function, despite its high hydrophobicity and a high prediction score for its being transmembrane and even though it is embedded in the membrane in the final structure. We can infer that the non-topogenic function of the P region in Kir 2.1, KcsA, and KAT1 is one of the determinants contributing to the final membrane structure of the K channel. The topogenic functions of M1 and M2 differed in Kir 2.1 and KcsA, and the cytoplasmic region was found to help in achieving the final membrane topology.
|
FOOTNOTES |
|---|
Present address: Dept. of Molecular Biology, Graduate School of Medical Science, Kyushu University, Fukuoka 812-8582, Japan.
To whom correspondence should be addressed: Bioscience and Biotechnology Center, Nagoya University, Nagoya 464-8601, Japan. Tel.: 81-52-789-5202; Fax: 81-52-789-5206; E-mail: uozumi{at}agr.nagoya-u.ac.jp.
1 The abbreviations used are: ER, endoplasmic reticulum; DPP, dipeptidyl peptidase; H1, H1 segment of the E. coli leader peptidase; PL, prolactin; RM, rough microsomal membrane; SA-I, type I signal-anchor sequence; SA-II, type II signal-anchor sequence; HA, hemagglutinin; TNF, tumor necrosis factor.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  L. Zhang, Y. Sato, T. Hessa, G. von Heijne, J.-K. Lee, I. Kodama, M. Sakaguchi, and N. Uozumi Contribution of hydrophobic and electrostatic interactions to the membrane integration of the Shaker K+ channel voltage sensor domain PNAS, May 15, 2007; 104(20): 8263 - 8268. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C-W Lu, J-H Lin, Y S Rajawat, H Jerng, T G Rami, X Sanchez, G DeFreitas, B Carabello, F DeMayo, D L Kearney, et al. Functional and clinical characterization of a mutation in KCNJ2 associated with Andersen-Tawil syndrome J. Med. Genet., August 1, 2006; 43(8): 653 - 659. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Cuthbertson, D. A. Doyle, and M. S.P. Sansom Transmembrane helix prediction: a comparative evaluation and analysis Protein Eng. Des. Sel., June 1, 2005; 18(6): 295 - 308. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Bendahhou, M. R. Donaldson, N. M. Plaster, M. Tristani-Firouzi, Y.-H. Fu, and L. J. Ptacek Defective Potassium Channel Kir2.1 Trafficking Underlies Andersen-Tawil Syndrome J. Biol. Chem., December 19, 2003; 278(51): 51779 - 51785. [Abstract] [Full Text] [PDF]
|
|
| ||
