HERG Is Protected from Pharmacological Block by α-1,2-Glucosyltransferase Function*

The HERG (human ether-à-go-go-related gene) protein, which underlies the cardiac repolarizing current IKr, is the unintended target for many pharmaceutical agents. Inadvertent block of IKr, known as the acquired long QT syndrome (aLQTS), is a leading cause for drug withdrawal by the United States Food and Drug Administration. Hence, an improved understanding of the regulatory factors that protect most individuals from aLQTS is essential for advancing clinical therapeutics in broad areas, from cancer chemotherapy to antipsychotics and antidepressants. Here, we show that the K+ channel regulatory protein KCR1, which markedly reduces IKr drug sensitivity, protects HERG through glucosyltransferase function. KCR1 and the yeast α-1,2-glucosyltransferase ALG10 exhibit sequence homology, and like KCR1, ALG10 diminished HERG block by dofetilide. Inhibition of cellular glycosylation pathways with tunicamycin abrogated the effects of KCR1, as did expression in Lec1 cells (deficient in glycosylation). Moreover, KCR1 complemented the growth defect of an alg10-deficient yeast strain and enhanced glycosylation of an Alg10 substrate in yeast. HERG itself is not the target for KCR1-mediated glycosylation because the dofetilide response of glycosylation-deficient HERG(N598Q) was still modulated by KCR1. Nonetheless, our data indicate that the α-1,2-glucosyltransferase function is a key component of the molecular pathway whereby KCR1 diminishes IKr drug response. Incorporation of in vitro data into a computational model indicated that KCR1 expression is protective against arrhythmias. These findings reveal a potential new avenue for targeted prevention of aLQTS.

I Kr , an important component of the cardiac delayed rectifier K ϩ current, is required for repolarization of the human cardiac action potential. Mutations in HERG (human ether-à-go-gorelated gene), which encodes the ␣-subunit of the channel complex, are responsible for the chromosome 7-linked form (LQT2) of the congenital long QT syndrome (LQTS), 2 an arrhythmia syndrome characterized by action potential prolongation and delayed cardiac repolarization. LQTS causes the atypical polymorphic ventricular tachycardia torsade de pointes, leading to sudden cardiac death (1)(2)(3). A far more commonly acquired form of LQTS (aLQTS) is precipitated by exposure to a wide range of therapeutic compounds that block HERG. aLQTS represents a significant problem for public health and the pharmaceutical industry 3 and is typically observed in predisposed patients receiving treatment for noncardiac illnesses. Indeed, about half of all drug withdrawals since 1998 were in response to potential pro-arrhythmic effects linked to aLQTS (5)(6)(7).
Numerous drug families bind the HERG channel and suppress I Kr in vitro, yet, surprisingly, few patients experience aLQTS. Clinically silent ion channel gene mutations have been characterized in families with cases of aLQTS, indicating that even the drug-induced form of the disorder may carry a genetic component (8 -15). However, mutations in the gene that encodes HERG are associated with only a small percentage of the aLQTS cases, suggesting that genetic variations in an array of other molecules modulate HERG pharmacology, either enhancing or reducing HERG block and aLQTS risk.
The K ϩ channel regulator KCR1 (38) protects heterologously expressed HERG current (I HERG ) from block by the antiarrhythmic drugs dofetilide, quinidine, and sotalol over clinically relevant concentration ranges (16). Moreover, a polymorphism associated with a reduced risk for aLQTS was shown to enhance KCR1-induced protection of I HERG from drug block in vitro (17). Here, we show that KCR1 protects I HERG through an unanticipated biochemical function, ␣-glucosyltransferase activity. The closest known homolog of KCR1, the endoplasmic reticulum (ER)-resident yeast ␣-glucosyltransferase Alg10 (Die2) (18), protects HERG from drug block in a manner identical to that of KCR1, and conversely, KCR1 functionally substitutes for Alg10 in yeast. Additionally, KCR1 modulation of HERG pharmacology is dependent upon a fully functioning cellular glycosylation machinery. These findings reveal an unprecedented mechanism for HERG pharmacological regulation and suggest an important new avenue for developing targeted interventions for preventing aLQTS.

EXPERIMENTAL PROCEDURES
Plasmid Constructs-The HERG cDNA was subcloned into the mammalian expression vector pSI (Promega Corp., Madison, WI) as described previously (19). The HERG(N598Q) mutation was introduced into the same backbone using the Stratagene QuikChange site-directed mutagenesis kit. The human KCR1 cDNA was subcloned into the pCGI vector for bicistronic expression with enhanced green fluorescent protein (GFP) (16). ALG10 (kindly provided by Dr. Markus Aebi, ETH Zurich, Zurich, Switzerland) was inserted into the same vector for mammalian cell expression. For expression in yeast, KCR1 and ALG10 were subcloned into the pGPD426 shuttle vector (a kind gift from Dr. Tony Weil, Vanderbilt University), which allows for selection on uracil-deficient medium (20). The 3Ј-end of the KCR1 cDNA was fused in-frame with cyan fluorescent protein by insertion into the AccI and SmaI sites of the pECFP-1 vector (Clontech). For in-frame fusion with the DsRed protein, KCR1 was subcloned into the XhoI/SacII sites of the pDsRed2-C1 vector (Clontech). pEYFP-ER was purchased from Clontech. The KCNE2 cDNA was amplified by PCR from human genomic DNA using the upstream primer 5Ј-TTAATTAAGCTTGGCCATGTCTACTTTATCCAATT-TCACACAG-3Ј and the downstream primer 5Ј-AACTGCA-GTCAGGGGACATTTTGAACCC-5Ј. For subcloning, Hin-dIII and PstI sites (boldface) were built into the primer sequence. The start and stop codons are shown in italics. The T8A mutation was included in the upstream primer by substituting the underlined A for G in a separate upstream primer. The cDNAs were subcloned into the pCGI vector for bicistronic expression with enhanced GFP.
A bath solution containing dofetilide was prepared from a 100 M dofetilide stock in water on the day of experimentation. Tunicamycin (TM) was purchased from Sigma. After HERG current was recorded under control conditions, dofetilide (at the concentrations listed in each figure legend) was applied to the bath solution, followed by repetitive depolarizing pulses to ϩ20 mV (2 s) from a holding potential of Ϫ80 mV (6 s) and by a hyperpolarizing pulse of Ϫ50 mV (2 s). The decline of relative tail currents during repetitive pulses in the presence of dofetilide was fitted to a single exponential function: y ϭ y 0 ϩ Ae Ϫt/ , where y 0 is the offset, A is the current amplitude under the control condition, t is the time, and is the time constant. All data are expressed as the means Ϯ S.E., and statistical comparisons were made using one-way analysis of variance (Origin, MicroCal, LLC, Northampton, MA), with p Ͻ 0.05 indicating significance.
Yeast Transformation-The culture and transformation of yeast were accomplished essentially as described previously (21). KCR1 or ALG10 cDNA contained in the pGPD426 vector or pGPD426 alone was transformed into the Alg10-deficient wbp1-2/alg10 yeast strain YG649. Transformed cells (2 ϫ 10 6 ) were counted in a hemocytometer and inoculated into uracildeficient synthetic dextrose medium plus 1 M sorbitol. In parallel, the same numbers of SS328 and untransformed YG649 control cells were grown in uracil-deficient synthetic dextrose medium plus uracil and 1 M sorbitol. 54.5 h post-transformation, yeast protein extracts were prepared as described below.
Western Blotting-Whole cell protein extracts were prepared from transfected CHO or Lec1 cells 2 days post-transfection as described previously (16). The membranes were probed with anti-HERG antibody (1:400 dilution; Alomone Labs) in phosphate-buffered saline (PBS) and 5% nonfat dry milk, and proteins were visualized with horseradish peroxidase-linked anti-rabbit F(abЈ) 2 antibody (1:10,000 dilution; Amersham Biosciences) and detected by ECL (Amersham Biosciences). Protein extracts from yeast cells were prepared by pelleting yeast cultures. The samples were heated to 70°C in lysis buffer (1% SDS, 50 mM Tris-HCl (pH 7.5), and 2 mM phenylmethylsulfonyl fluoride) for 10 min in the presence of glass beads, and vortexed. The pellet was extracted again by boiling, and supernatants were saved. 20 g of the sample was loaded onto 10% SDS-polyacrylamide gel; Western blot analysis was performed using anti-carboxypeptidase Y (CPY) monoclonal antibody (1:5000 dilution; Molecular Probes) in 1ϫ PBS/Tween plus 5% nonfat dry milk and 2% bovine serum albumin; and proteins were detected with horseradish peroxidase-linked anti-mouse secondary antibody (1:10,000 dilution; Sigma) using ECL.
Confocal Imaging-A triple-FLAG epitope was fused in-frame with amino acid 2 of the N terminus of rat KCR1 contained on the p3XFLAG-CMV vector (Sigma). Mouse atrial tumor myocyte HL-1 cells (22) were transfected with FLAG-KCR1 and split onto fibronectin-coated chamber slides 24 h post-transfection. 48 h post-transfection, cells were fixed in 4% paraformaldehyde in PBS and permeabilized for 20 min with 0.2% Triton X-100 in blocking solution (0.16 g of bovine serum albumin and 0.4 ml of goat serum in 10 ml of PBS). Cells were washed with PBS and incubated with fluorescein isothiocyanate-linked anti-FLAG monoclonal antibody M2 (Sigma) in PBS and 2% bovine serum albumin for 1 h. This was followed by three PBS washes and a 1-h incubation with rabbit anti-calnexin polyclonal antibody (1:200 dilution; Stressgen Bioreagents, Victoria, British Columbia, Canada) and by three PBS washes and a 45-min incubation with Cy5-conjugated Affi-niPure donkey anti-rabbit F(abЈ) 2 antibody (1:150 dilution; Jackson ImmunoResearch Laboratories, Inc. West Grove, PA) to reveal calnexin localization. The slides were washed with PBS and mounted in AquaMount (Lerner Laboratories). Images were acquired on a Zeiss Axiophot LSM 510 inverted confocal microscope with a Plan Apochromat ϫ63/1.4 oil differential interference contrast objective (ScanZOOM 1.0). The absorption/emission wavelengths used were 650/ 670 nm for Cy5 detection, 492/520 nm for fluorescein iso-thiocyanate detection, 495/505 nm for enhanced yellow fluorescent protein detection, and 427/468 nm for cyan fluorescent protein detection.

KCR1 Shares Sequence and Functional Characteristics with the Yeast Alg10
Protein-KCR1 (GenBank TM accession number AY845858) shows 28% amino acid identity (43% similarity) to ALG10 (DIE2; accession number D38049), which encodes the Saccharomyces cerevisiae ␣-1,2-glucosyltransferase ( Fig. 1). This enzyme is responsible for the addition of the terminal ␣-1,2-glucose residue to lipid-linked oligosaccharides in the yeast ER, a step that is thought to be necessary for the efficient transfer of oligosaccharides to nascent proteins (Fig. 2). Confocal imaging of FLAG-tagged KCR1 transfected into the mouse atrial tumor-derived cell line HL-1 (22) and the endogenous ER marker calnexin showed that the distribution pattern of KCR1 was consistent with ER localization (Fig. 3). Similar findings were obtained when DsRed-linked KCR1 and the ER marker enhanced yellow fluorescent protein were transfected into HL-1 cells (supplemental Fig. 1).
To test whether KCR1 and Alg10 also share functional characteristics in mammalian cells, we examined whether Alg10, like KCR1 (16), modulates drug block of I HERG . To this end, we transfected HERG alone or cotransfected HERG with KCR1 or ALG10 into CHO cells and assessed drug block of I HERG using the pulse protocol shown in Fig. 4A. As in our previous experiments, we added the antiarrhythmic dofetilide to the bath solution, applied repetitive pulses, and recorded I HERG tail currents (Fig. 4B). The pulse-dependent reduction of tail current amplitude was fitted to a single exponential function as described under "Experimental Procedures." Analogous to KCR1, coexpressing ALG10 and HERG significantly slowed drug block of I HERG (Fig. 4C), suggesting that KCR1 modulates drug block of I HERG via glycosylation.
We incorporated the data from Fig. 4C into the Luo-Rudy dynamic model of the mammalian action potential (23). The results show that HERG block sufficient to evoke action potential prolongation and development of early afterdepolarizations (the cellular precursor to torsade de pointes) was reversed by KCR1 effects on I HERG equivalent to those measured in vitro (Fig. 4D).
KCR1-dependent HERG "Protection" Requires Intact Glycosylation Pathways-To pursue this idea, we manipulated the cellular glycosylation pathway at multiple points (Fig. 2). TM is an antibiotic that blocks the first step of glycoprotein synthesis, thus inhibiting the synthesis of all N-linked glycoproteins. HERG-transfected or HERG/KCR1-cotransfected cells were  The green outline indicates the ER. TM blocks the initial step in glycoprotein synthesis by preventing the transfer of UDP-GlcNAc (black squares) to a dolichyl pyrophosphate lipid carrier (red ovals). In yeast, Alg10 acts by transferring the terminal glucose residue (triangles) to the pre-assembled oligosaccharide (4,18). The Alg10-catalyzed reaction is required for efficient transfer of the oligosaccharide to an asparagine residue on the nascent polypeptide. The magenta outline delimits Golgi-associated processes. Lec1 cells are deficient in GlcNAc glycosyltransferase, an enzyme important for Golgi-mediated glycoprotein processing. preincubated with or without TM (1 g/ml) for 24 h prior to current recording, and the drug response of I HERG to dofetilide was assessed using the pulse protocol shown in Fig. 4A. Preincubating HERG-transfected cells with TM did not influence the drug response of I HERG (Fig. 5A). The lack of effect of TM on HERG drug block allowed us to examine whether KCR1 is still able to influence I HERG drug block when N-linked cellular glycosylation is inhibited. Preincubating HERG/KCR1-cotrans-fected cells with TM prevented KCR1 from diminishing the drug sensitivity of I HERG , effectively abrogating the effects of KCR1 (Fig. 5B). Western blot analysis of total cell extracts derived from transfections done with or without 1 g/ml TM at various time points confirmed the effectiveness of the TM treatment in preventing the glycosylation of HERG, which was used as an indicator of glycosylation status (Fig. 5C).
Lec1 cells are derived from CHO cells with defective N-acetylglucosamine transferase activity in the Golgi apparatus, resulting in a defect in N-linked glycosylation and producing only truncated cell-surface glycoproteins (Fig. 2) (24 -26). If N-linked glycosylation is a requirement for the KCR1-mediated modulation of drug block of I HERG , the KCR1 effect should not be observed in Lec1 cells. HERG alone or with KCR1 was transfected into Lec1 cells, and drug block of I HERG was assessed using the pulse protocol shown in Fig. 4A. Coexpressing KCR1 and HERG in Lec1 cells did not influence drug block of I HERG (Fig. 6A). We used the glycosylation status of HERG to verify that Lec1 cells do not support glycosylation. As Western blot analysis showed (Fig. 6B), both the core and fully glycosylated forms of HERG are produced in CHO cells, whereas the fully glycosylated form is not produced in Lec1 cells. These findings further suggest that KCR1 modulates drug block of I HERG through glycosylation.
Human  (18). We subcloned the yeast ALG10 and human KCR1 cDNAs into a yeast expression vector with a URA selectable marker and transformed them into the YG649 yeast strain, selecting for growth on uracil-deficient medium. Fig. 7A shows that YG649 transformed with either KCR1 or ALG10 exhibited robust growth on uracil-deficient synthetic dextrose medium after 2 days at 30°C, whereas untransformed cells did not grow, and cells transformed with vector alone grew poorly. This suggests that, like Alg10, KCR1 can rescue the defect of the YG649 strain.
In yeast, Alg10 functions as an ␣-glucosyltransferase, the activity  (22) were transfected with FLAG-KCR1 and incubated with anti-FLAG and anti-calnexin antibodies to localize this endogenous ER marker. A, KCR1-associated staining shown in green pseudocolor; B, anti-calnexin staining shown in red pseudocolor; C, overlay of red and green pseudocolors shown in yellow. The cell marked by the white arrows expressed the FLAG-KCR1 protein, but was located in a different plane of focus than the surrounding cells. The calnexin staining thus appears brighter than that of the neighboring cells. Slides were imaged as described under "Experimental Procedures." FIGURE 4. Alg10, a yeast homolog of KCR1, protects I HERG from dofetilide block. A, shown is the pulse protocol used in this study. After I HERG was recorded under control conditions (pre-drug), dofetilide was applied to the bath solution, followed by repetitive depolarizing pulses to ϩ20 mV (2 s) from a holding potential of Ϫ80 mV (6 s) and by a hyperpolarizing pulse of Ϫ50 mV (2 s). B, HERG cDNA and the GFP vector (HERG alone), HERG and KCR1 cDNAs, or HERG and ALG10 cDNAs were transfected into CHO cells, and the response of I HERG to 40 nM dofetilide was assessed. Shown are representative tail currents for HERG alone (left panel) and HERG plus ALG10 (right panel) during repetitive pulses in the presence of dofetilide. C, shown are relative peak tail currents during repetitive pulses in the presence of dofetilide for HERG alone (f; n ϭ 9), HERG plus KCR1 (E; n ϭ 11), and HERG plus ALG10 (‚; n ϭ 11). Analogous to KCR1, coexpressing ALG10 and HERG also slowed I HERG drug block. The time constants () of the decline of peak tail currents were 556 Ϯ 65 s for HERG alone, 927 Ϯ 107 s (p Ͻ 0.05 versus HERG alone) for HERG plus KCR1, and 906 Ϯ 134 s (p Ͻ 0.05 versus HERG alone) for HERG plus ALG10. D, shown is the simulated action potential using the Luo-Rudy ventricular model (23) incorporating the HERG drug block data shown in C. HERG block sufficient to evoke action potential prolongation and development of early afterdepolarizations was reversed by KCR1 effects on I HERG . FEBRUARY 23, 2007 • VOLUME 282 • NUMBER 8

JOURNAL OF BIOLOGICAL CHEMISTRY 5509
of which can be assessed biochemically using the endogenous substrate CPY (18). CPY is a yeast vacuolar protease with four N-linked oligosaccharides (27) and therefore migrates at five different apparent molecular masses on a Western blot. We prepared yeast protein extracts from the wild-type (WT) strain SS328; YG649 mutants; and YG649 transformed with human KCR1, yeast ALG10, or the empty vector alone and performed Western blot analysis using anti-CPY antibody. Fig. 7B shows that, as expected, SS328 produced the mature form of CPY, whereas YG649 contained primarily the incompletely glyco-sylated Ϫ4, Ϫ3, and Ϫ2 forms of CPY. YG649 transformed with either KCR1 or ALG10 produced, in addition to the Ϫ4, Ϫ3, and Ϫ2 forms of CPY, the Ϫ1 and fully glycosylated forms, indicating that KCR1 rescues the alg10/wbp1-2 defect to the same extent as does ALG10 in yeast. This supports the conclusion that KCR1 supplies the same biochemical function as Alg10 and thus functions as an ␣-glucosyltransferase in yeast.

N-Linked Glycosylation-induced Protection of HERG Does Not Require Glycosylation of the I Kr Complex-
The HERG protein contains two consensus N-glycosylation sites, Asn 598 and Asn 629 , although a prior study indicates that only Asn 598 is glycosylated (28). Because dofetilide is known to access its binding site from the cytoplasmic aspect of the channel (29,30), whereas glycosylation occurs at the extracellular side, we did not expect HERG itself to be the target of the KCR1-mediated effects. Nevertheless, we assessed the effects of KCR1 on a glycosylation-defective HERG mutant and examined whether drug block of the N598Q mutant can be modulated by KCR1. HERG(N598Q) is assembled into functional channels and reaches the cell surface, although its membrane residency time is shortened compared with WT HERG (28).   (Lec1). A, HERG cDNA and the GFP vector (HERG alone) or HERG and KCR1 cDNAs were transfected into Lec1 cells, and drug block (60 nM dofetilide) of I HERG was assessed using the pulse protocol shown in Fig. 4A. Coexpressing KCR1 and HERG in Lec1 cells no longer modulated the drug response of I HERG . The time constants () of the decline of peak tail currents were 442 Ϯ 33 s (n ϭ 7) for HERG alone (f) and 540 Ϯ 66 s (n ϭ 8, p ϭ not significant) for HERG plus KCR1 (E). B, Western blot analysis of HERG showed both core (C) and fully (F) glycosylated forms in CHO cells. In contrast, a fully glycosylated form did not appear in Lec1 cells.

␣-Glucosyltransferase KCR1 Affects HERG Pharmacology
HERG alone, HERG(N598Q) alone, or HERG(N598Q) plus KCR1 were transfected into CHO cells, and drug block of I HERG was assessed using the protocol shown in Fig. 4A. As shown in Fig. 8A, drug block of HERG(N598Q) was still modulated by KCR1. To confirm the glycosylation status of HERG(N598Q), we performed Western blot analysis, which showed a lack of the fully glycosylated form (Fig. 8B). This indicates that, although N-linked glycosylation is essential for KCR1-mediated protection of I HERG block, glycosylation of HERG alone does not render this protection.
KCNE2 (MiRP1) co-assembles with HERG (31, 32), although the functional implications have been debated (33), and it has been difficult to demonstrate expression of KCNE2 in human heart (34). The KCNE2 polymorphism T8A, which abrogates complex glycosylation of KCNE2, mildly increases I Kr sensitivity to block by the antibiotic sulfamethoxazole (35). To address whether KCNE2 can modulate intracellular block of HERG by dofetilide in response to glycosylation, we first tested the HERG drug response in the presence of WT KCNE2 or KCNE2(T8A), but observed no significant differences (Fig. 9A). Moreover, KCR1 still attenuated drug block of I HERG regardless of whether WT KCNE2 or KCNE2(T8A) was coexpressed (Fig. 9B). Hence, N-linked glycosylation of neither HERG nor KCNE2 underlies the glycosylation-dependent protection of HERG from dofetilide block.

DISCUSSION
The most frequent cause of aLQTS is the unintended block of I Kr by commonly used pharmaceutical agents (7,36), leading to impaired repolarization of cardiac myocytes, action potential prolongation, and potentially fatal ventricular arrhythmias. It is now recognized that aLQTS has a genetic component, as patients with increased vulnerability for the syndrome and carrying mutations in LQTS-linked genes have been identified (11). However, not all family members who carry specific gene mutations develop aLQTS, suggesting "incomplete penetrance" and implying that hidden factors influence the pathol-ogy in individual patients (37). Moreover, the vast majority of patients do not experience aLQTS despite well documented in vitro block of HERG at clinically relevant concentrations of numerous therapeutic agents. Hence, it is likely that constitutive regulatory mechanisms usually protect patients from aLQTS, and disruption of such pathways could constitute risk.
We have shown that KCR1, a protein of previously unknown function, diminishes the effects of the HERG blockers dofetilide, sotalol, and quinidine on I Kr . In a follow-up study, the human "gain-of-function" KCR1 polymorphism I447V was associated with reduced risk of aLQTS (17). Here, we have identified a surprising mechanism whereby KCR1 influences the pharmacology of I HERG . The closest homolog of KCR1 in the public data bases is the yeast Alg10/Die2 protein (Fig. 1), the gene of which encodes an ER-resident ␣-1,2-glucosyltransferase (18,38). In fact, based on sequence homology to the yeast enzyme, KCR1 is identified as an ␣-1,2-glucosyltransferase (accession number CAC41349) in the NCBI Protein Database, although, to the best of our knowledge, biochemical evidence in support of this classification has not been reported. This enzyme is responsible for the addition of the terminal ␣-1,2-  ␣-Glucosyltransferase KCR1 Affects HERG Pharmacology FEBRUARY 23, 2007 • VOLUME 282 • NUMBER 8 glucose residue to the nascent lipid-linked oligosaccharide (Fig.  2). Lack of the terminal ␣-1,2-glucose residue impedes substrate recognition of the next enzyme in the pathway, oligosaccharyltransferase, thereby resulting in reduced transfer of the pre-assembled oligosaccharide trees to nascent polypeptides.
In this study, we approached the hypothesis that glycosylation impacts KCR1-mediated drug regulation of I HERG from multiple directions. (i) We showed that the yeast ␣-1,2-glucosyltransferase Alg10 modulated drug block of I HERG in the same manner as KCR1. (ii) Inhibition of the glycosylation pathway with TM abrogated the effects of KCR1 on drug block of I HERG , indicating that N-glycosylation is required for KCR1 to exert an effect on HERG. The drug sensitivity of HERG alone was not altered in the presence of TM, and KCR1 must be coexpressed with HERG for the diminished drug sensitivity to manifest. (iii) Similar to results in the presence of TM, KCR1 also did not influence the pharmacological properties of I HERG in Lec1 cells, a CHO cell line with a defect in the Golgi-resident N-acetylglucosamine transferase activity. This corroborates our conclusion that intact N-glycosylation is required for KCR1 to influence I HERG pharmacology.
Having demonstrated that the yeast ALG10 gene product is functionally homologous to KCR1 with respect to channeldrug interactions, we assessed the biochemical activity of KCR1 in alg10-deficient yeast. Our results showed that KCR1 was able to rescue the growth characteristics of an alg10-deficient yeast strain. Moreover, we demonstrated enzyme activity by providing evidence that KCR1 biochemically modified a test substrate of Alg10 in yeast (Fig. 7), leading us to conclude that KCR1 is a functional homolog of Alg10.
When we exogenously expressed epitope-tagged KCR1 in cardiac HL-1 cells (22), we noted expression in the ER (Fig. 3 and supplemental Fig. 1), consistent with its proposed function as an ␣-glucosyltransferase. However, KCR1 was not distributed evenly throughout the entire ER. This pattern may reflect compartmentalization within the ER, but is most likely a function of relatively weaker KCR1 expression levels compared with the ER marker.
The glycosylation status of HERG and its processing and maturation within the ER and Golgi are under intense scrutiny (39 -46) largely because of numerous congenital LQT2 mutations judged to be "trafficking-defective" (47,48). Such mutants never reach the plasma membrane because of abortive intracellular processing and cause a loss of I Kr function in afflicted LQT2 patients. Nevertheless, we have excluded the possibility that glycosylation of HERG alone or in complex with KCNE2 explains the ␣-1,2-glucosyltransferase effect of reducing I HERG drug block (Figs. 8 and 9).
In summary, our experiments strongly support the notion that KCR1 modulates drug block of I HERG through enhancing cellular glycosylation by acting as an ␣-1,2-glucosyltransferase. Incorporation of our data derived in vitro into a computational model of the ventricular action potential corroborated our conclusion that KCR1 can play a protective role in aLQTS (Fig. 4D).
We consider it unlikely that KCR1 simply causes a protective "umbrella" of increasingly glycosylated surface membrane proteins, which then keep drugs from accessing the interior of the cell, because a prior experiment indicated that KCR1 maintains its protective effect even if the drug (dofetilide) is applied through an intracellular pipette solution (16). Rather, our data suggest that the overall glycosylation efficiency of the cell imparts I HERG drug sensitivity. This implies that the status of the cellular glycosylation pathway may constitute a risk factor for aLQTS, motivating future studies of this biochemical pathway as a means of targeted prevention of drug-induced arrhythmias.