A Rare Mutation in ABCC8/SUR1 Leading to Altered ATP-Sensitive K⁺ Channel Activity and β-Cell Glucose Sensing Is Associated With Type 2 Diabetes in Adults

A total of 187 adult subjects diagnosed with type 2 diabetes or hyperglycemia before the age of 40 years (all of French Caucasian origin, except one subject with Antillean black ancestry) entered the study for gene screening. Thirty-nine exons of the ABCC8 gene were sequenced from genomic DNA in patients, as previously described (4).

cDNA encoding mouse Kir6.2 (CoreNucleotide NM_010602) or hamster SUR1 (CoreNucleotide L40623) were subcloned into plasmids pcDNA3 and pIRES2, respectively. Nucleotide substitutions were introduced into SUR1 cDNA using a Quick-Change site-directed mutagenesis kit (Stratagene). The primers used for the mutagenesis are given in online appendix Table 1 (available at http://dx.doi.org/10.2337/db07-1547). We used pIRES2–enhanced green fluorescent protein (EGFP) and/or pIRES2-dsRed2 (Clontech) vectors to allow channel-independent expression of reporter proteins, EGFP (mutant SUR1), and dsRed2 (wild-type SUR1).

HEK293 or INS1(832/13) (16) cells were plated (1 × 10⁵ cells/35-mm dish), cultured overnight, and cotransfected with pcDNA3-Kir6.2 and pIRES2-SUR1 cDNA in 7:3 ratio (HEK293 cells) or pIRES2-SUR1 on its own [INS1(832/13) cells], using Lipofectamine2000 (Invitrogen). Cells were studied 2 days later.

Currents were recorded using an EPC9 patch-clamp amplifier controlled by Pulse acquisition software (HEKA Elektronik, Lambrecht/Pfalz, Germany). Inside-out patches excised from the membrane of HEK293 cells were recorded in response to 3-s voltage ramps from −110 to +100 mV (holding potential, 0 mV) (Fig. 2A, inset), filtered at 0.15 kHz and digitised at 0.5 kHz. If the level of expression was low, K_ATP channel currents were recorded as single-channel events at a constant holding potential of −60 mV, filtered at 1 kHz and digitized at 2 kHz. To control for possible rundown, the control conductance (G_c) was taken as the mean of that in nucleotide-free solution before and after the application of ATP. For each recording, ATP concentration-inhibition curves were fitted to the Hill equation: G/G_c = 1/(1 + [{ATP}/IC₅₀]^h), where IC₅₀ is the concentration at which inhibition is half maximal and h is the Hill coefficient. Given ATP inhibition values are the means of the fitted parameters for individual patches.

The plasma membrane potential of INS1(832/13) β-cells was recorded in perforated-patch whole-cell configuration. The pipette tip was dipped into pipette solution and then back-filled with the same solution containing 0.24 mg/ml amphotericin B. Recordings were initiated after a 30-min exposure to substrate-free solutions at 37°C, and the duration of exposure to each concentration of effector(s) was ≥2 min. The frequency of action potentials was measured from recording intervals ≥1 min, corresponding to different extracellular solutions. For firing cells, the membrane potential was determined as the baseline potential between the spikes. Cells that were not responsive to tolbutamide were excluded from analysis. Series resistance and cell capacitance were compensated automatically by the acquisition software. Experiments were carried out by periodically switching from current-clamp to voltage-clamp mode, thus obtaining pseudosimultaneous recordings of cell membrane potential (V_m) and K_ATP channel conductance (G_KATP) (9). This controlled for the leaks of the patch and verified that the depolarization (hyperpolarization) of the membrane was linked to K_ATP channel closure (opening). The current clamp protocol involved continuous recording, without electrical stimulation. In the voltage clamp, the membrane potential was held at −70 mV and whole-cell currents were evoked by ±10-mV 0.5-Hz pulses. Data were filtered at 0.2 kHz and digitized at 0.5 kHz.

For inside-out patch recordings, the pipette solution contained (in mmol/l) 140 KCl, 10 HEPES (pH 7.2 with KOH), 1.1 MgCl₂, and 2.6 CaCl₂. The intracellular (bath) solution contained (in mmol/l) 107 KCl, 1 CaCl₂, 2 MgCl₂, 11 EGTA, 10 HEPES (pH 7.2 with KOH), plus MgATP, as indicated. Mg²⁺-free intracellular solutions contained (in mmol/l) 107 KCl, 1 CaCl₂, 0.5 EDTA, 11 EGTA, 10 HEPES (pH 7.2 with KOH), and ATP, as indicated. pH was measured after ATP addition and readjusted if required. For perforated-patch experiments, the pipette solution contained (in mmol/l) 76 K₂SO₄, 10 NaCl, 10 KCl, 1 MgCl₂, and 5 HEPES (pH7.35 with KOH). No ATP was added. The bath solution contained (in mmol/l) 137 NaCl, 5.6 KCl, 10 HEPES (pH 7.4 with NaOH), 2.6 CaCl₂, and 1.1 MgCl₂. All experiments were conducted at 21–23°C, and the bath solution was perfused continuously.

Cells were preloaded by a 40-min incubation with 2 μmol/l Fura-Red acetoxy-methyl ester (Invitrogen) (17,18) dissolved in modified Krebs’ Ringer bicarbonate solution comprising (in mmol/l) 130 NaCl, 3.6 KCl, 0.5 NaH₂PO₄, 0.5 MgSO₄, 2.0 NaHCO₃, 3 glucose, 10 HEPES (pH7.4 with NaOH), and 1.5 CaCl₂ equilibrated with O₂/CO₂ (95:5, vol/vol) at 37°C. Changes in cytoplasmic free Ca²⁺ concentration ([Ca²⁺]_cyt) were monitored at 0.2 Hz, using a Cell∧R (Olympus) epifluorescence imaging system, based around an Olympus IX-81 inverted optics microscope fitted with a ×40 oil immersion objective. Cells were continuously perifused in modified Krebs’ Ringer bicarbonate solution at the glucose concentrations indicated. The “KCl” solution comprised modified Krebs’ Ringer bicarbonate solution in which 50 mmol/l NaCl was substituted by KCl. [Ca²⁺]_cyt was expressed as the ratio of fluorescence intensity (λ_em = 597 nm, λ_ex = 440 nm) to that at λ_ex = 490 nm, after subtraction of background fluorescence. All experiments were performed at 37°C. To account for the differences in [Ca²⁺]_cyt between individual cells, the fluorescence ratios at 20 mmol/l glucose, in each trace, were normalized to the ratios observed at 3 mmol/l glucose. Cells were selected by expression of the reporter proteins.

Cells were transfected with pIRES2-SUR1–c-myc or pIRES2-SUR1–c-myc plus pcDNA3-Kir6.2. Forty-eight hours posttransfection, cells were stained with mouse anti–c-myc antibody clone (9E10; Roche). Cells transfected with SUR1 were fixed with 4% paraformaldehyde and permeabilized [methanol/acetone for HEK293 cells, Triton X-100 for INS1(832/13) cells], whereas cells transfected with SUR1 and Kir6.2 were fixed and directly stained for surface expression. After a 2-h incubation with primary antibodies, cells were washed and labeled with goat anti-mouse Alexa 586 for 60 min, mounted using Prolong Gold antifade mounting media (Invitrogen), and observed with a Zeiss LSM510 confocal microscope.

Data were analyzed using Clampfit (Axon Instruments), Cell∧R (Olympus), and Excel (Microsoft) software. Unless specified otherwise (Fig. 4D), statistical significance was estimated using Mann-Whitney U test or Student's t test with Bonferroni correction for multiple sampling. Differences with P < 0.05 were considered statistically significant.

We screened for mutations in the ABCC8 gene in 204 diabetic subjects with disease onset before age 40 years. One of the patients with normal BMI, diagnosed with hyperglycemia at age 39 years, having developed overt diabetes at age 45 years, presented an ABCC8 missense mutation causing a substitution of tyrosine 356 with cysteine (Y356C) in the SUR1 subunit of the K_ATP channel (Fig. 1B and online appendix Fig. 1). The mutation was also identified in two offspring of the patient, aged 33 and 35 years, who showed normal fasting blood glucose levels but displayed a mild decrease of insulin secretion during an oral glucose tolerance test (online appendix Table 1). The disposition index, as a measure of β-cell function related to insulin sensitivity status, was low in the two offspring, although they were not diagnosed with diabetes. The Y356C mutation was not found in 170 unrelated normoglycaemic individuals of European Caucasian origin.

The two other ABCC8 mutations that we found to be associated with adult-onset diabetes were R248Q (type 2 diabetic patient diagnosed at age 39 years without familial cosegregation) (online appendix Fig. 1) and K1521N (two type 2 diabetic patients diagnosed at age 37 and 42 years). The amino acids affected by the three mutations are highly conserved among species (online appendix Fig. 2).

To test whether the mutations associated with type 2 diabetes might affect stimulus-secretion coupling in β-cells, we next measured the sensitivity to ATP of recombinant K_ATP channels carrying SUR-Y356C, -R248Q, and -K1521N and compared these to the ATP sensitivity of TND-associated mutants (4), L582V, H1023Y, and R1379C. cDNAs encoding the above SUR1 mutants were generated by site-directed mutagenesis and coexpressed with wild-type Kir6.2 in HEK293 cells (see research design and methods). Analysis of the ATP sensitivity of the resulting K_ATP channel complexes in inside-out excised patches revealed a clear correlation with the two different forms of diabetes (Fig. 1C). Thus, all three TND-associated mutations tested caused a substantial (>40-fold) decrease of ATP sensitivity. By contrast, the type 2 diabetes–associated mutations had no effect or a much smaller effect on ATP sensitivity. The concentration-inhibition curves for K_ATP channels carrying SUR1-R248Q and SUR1-K1521N were practically identical to the wild type, suggesting either that these mutations affected other properties of the channel or were not responsible for diabetes (Fig. 1C). K_ATP channel conductance of the inside-out patches expressing SUR1-K1521N was not different from wild type (11.3 ± 5.6 ns and 12.5 ± 5.9 ns, respectively), as measured in nucleotide-free solution. Patches with SUR1-R248Q channels exhibited much smaller conductances of 1.2 ± 0.8 ns. Neither of these two mutations can therefore be directly linked to the downregulation of insulin secretion.

By contrast, K_ATP channels carrying SUR1-Y356C showed an approximately fourfold decrease in ATP sensitivity (Fig. 1C). This prompted us to investigate in detail how the Y356C mutation affected the ATP sensitivity and/or surface expression of K_ATP channels. We also assessed the impact of this and other mutants on stimulus-secretion coupling. Given the limited magnitude of the type 2 diabetes–associated mutant's effects, we used a TND-associated SUR1 mutation, L582V (4), as a positive control.

Since all of the patients carried only a single copy of the mutated allele, we deemed it important to mimic expression of the channel in this heterozygotic form in single cells. To achieve this, we coexpressed the recombinant wild-type and mutant channel subunits (19,20). cDNAs encoding wild-type or mutant SUR1 were cloned into plasmids (pIRES) from which reporter proteins that have nonoverlapping fluorescent emission, namely EGFP (λ_em = 507 nm) or dsRed2 (λ_em = 582 nm), were expressed from an internal ribosome entry site on the same message as SUR1. Although EGFP-to-dsRed2 emission intensity ratio may have been used, in principle, to quantify the relative expression of mutant and wild-type K_ATP channels in any given cell, we assumed every EGFP⁺dsRed2⁺ cell to be heterozygous (online appendix).

The ATP sensitivity of heterozygous SUR1-Y356C (hetY356C) (Fig. 2A and B) (Table 1) was higher than that of homozygous SUR1-Y356C (homY356C) channels. Thus, hetY356C and homY356C channels were each inhibited by ATP with IC₅₀ = 61 μmol/l and IC₅₀ = 95 μmol/l, respectively, compared with IC₅₀ = 24 μmol/l for the wild-type channel. Heterozygous channels expressing SUR1-L582V (hetL582V) were also more ATP sensitive than homozygous SUR1-L582V channels (homL582V): IC₅₀ = 869 μmol/l and IC₅₀ = 1,140 μmol/l for het582V and hom582V, respectively (Fig. 2A and D) (Table 1).

ATP has a dual effect on the activity of K_ATP channels: it inhibits when binding to Kir6.2 but activates, in a Mg²⁺-dependent manner, when binding to nucleotide-binding domains (NBDs) of SUR1 (21–25). In the wild-type channels, the former effect dominates over the physiological range of free ATP concentrations (26). Gain-of-function mutations in either subunit frequently act by enhancing the Mg²⁺-dependent activation (4,27), so we tested if this was the case for SUR1-Y356C. When Mg²⁺ was removed from the intracellular (bath) solution, thereby abolishing Mg²⁺-dependent activation, ATP blocked the wild-type channels with IC₅₀ = 8 μmol/l. The sensitivity of hetY356C and homY356C channels increased to IC₅₀ = 25 μmol/l and IC₅₀ = 38 μmol/l, respectively (Fig. 2A and C) (Table 2). HetL582V and homL582V were both blocked with IC₅₀ = 17 μmol/l, which represents almost a 100-fold shift compared with the Mg²⁺-containing solution (Fig. 2A and E) (Table 2). Thus, the gain-of-function effect of L582V mutation was mediated via Mg²⁺-dependent activation, while the effect of Y356C apparently occurred through a different mechanism.

Transfection of HEK293 cells with cDNA encoding wild-type Kir6.2 and SUR1 subunits resulted in significant accumulation of SUR1 in the cytoplasm or cytoplasmic structures, as detected using anti–c-myc antibodies in permeabilized cells (Fig. 3A). Examined in intact cells, SUR1 could also be detected on the plasma membrane (Fig. 3B), consistent with previous findings in β-cells (28,29).

Introduction of the Y356C mutation into SUR1 did not affect cytoplasmic (Fig. 3C) or membrane (Fig. 3D) localization. Similarly, the cytoplasmic disposition of SUR1-L582V was not different from that of the wild-type SUR1 (Fig. 3E). Interestingly, we did note a tendency toward lower cell surface expression of SUR1-L582V (Fig. 3F) versus wild type.

In the absence of a metabolic stimulus, the membrane potential of pancreatic β-cells is largely set by the K⁺ conductance of K_ATP channels (G_KATP) (30). Metabolic inhibition of G_KATP depolarizes the membrane logarithmically, in agreement with Goldman-Hodgkin-Katz formalism (9). Alterations in the metabolic inhibition of K_ATP channels may therefore affect glucose-induced electrical activity of β-cells as well as Ca²⁺ influx via l-type Ca²⁺ channels. We therefore studied the effect on these two phenomena of the Y356C and L582V mutations in SUR1.

To this end, we overexpressed wild-type or mutant SUR1 subunits in INS1(832/13) β-cells. Overexpression of SUR1 had no effect on the level of cell surface expression of K_ATP channels (data not shown). We incubated cells in glucose-free extracellular solution for 30 min and measured changes in membrane potential (V_m) and K_ATP channel conductance (G_KATP), in response to increasing levels of glucose. The resting V_m of cells expressing the mutant channels (het- or homY356C and het- or homL582V cells, respectively) was not different from that of cells expressing only wild-type SUR1: V_m = −67.4 mV (wild type), V_m = −67.1 mV (homY356C), V_m = −67.1 mV (hetY356C), V_m = −70.5 mV (homL582V), and V_m = −68.4 mV (hetL582V). Although addition of 1 mmol/l glucose had no effect on V_m in any of the three groups, further addition of 5 and then 10 mmol/l glucose depolarized the membrane of wild-type cells and hetY356C cells (Fig. 4A–C). By contrast, homY356C and both hom- and hetL582V showed a markedly inhibited response to increasing glucose concentrations. On the other hand, 0.2 mmol/l tolbutamide depolarized all the cells (see research design and methods). Subsequent perifusion with the mitochondrial uncoupler carbonyl cyanide 4-(trifluoromethoxy)-phenylhydrazone (2 μmol/l) repolarized the plasma membrane (Fig. 4B and C).

Despite the presence of a clear tendency (Fig. 4B), we did not observe statistically significant differences in glucose-induced changes of the baseline membrane potential between wild-type and hetY356C cells. However, overall, glucose affected the electrical activity differently in the two groups. In wild-type cells, addition of 5, 10, and 20 mmol/l glucose depolarized the β-cell membrane and increased the frequency of action potentials (0.62 ± 0.12 Hz in 10 mmol/l glucose). Addition of 0.2 mmol/l tolbutamide produced further depolarization, which lead to a reduction of the firing frequency (0.37 ± 0.07 Hz) and, in some cases, degeneration of the oscillations (Fig. 4A and D). This effect is not typical for native β-cells, where the firing frequency increases monotonously, with the K_ATP channel conductance decreasing (31). The nonmonotonous dependence has been described experimentally in α-cells (32) (and can be explained theoretically in terms of the Morris-Lecar model of interacting ion conductances [33]). In hetY356C cells, the glucose-induced depolarization evoked action potentials with lower frequency (0.35 ± 0.07 Hz), which was further increased by application of tolbutamide (0.41 ± 0.10 Hz). It should be stressed that 10 mmol/l glucose did not produce a larger depolarization in hetY356C cells. Thus, despite a relatively small shift in the ATP sensitivity of K_ATP channels, β-cell lines expressing SUR1-Y356C demonstrated impaired coupling between nutrient stimulation and electrical activity.

To further explore the effects of the identified mutants on metabolic signal transduction, we deployed single-cell Ca²⁺ imaging of INS1(832/13) cells (Fig. 5A–C) (see research design and methods). Examined at 3 mmol/l glucose (basal) [Ca²⁺]_cyt was stable in both wild-type and homY356C cells (Fig. 5A). The addition of 20 mmol/l glucose had different effects on the two groups: in the majority of wild-type cells we observed oscillations of [Ca²⁺]_cyt, which were not detected in Y356C cells. INS1(832/13) cells overexpressing SUR1-K1521N channels showed an unchanged [Ca²⁺]_cyt response to glucose compared with wild-type cells (Fig. 5B and C). By contrast, glucose-stimulated [Ca²⁺]_cyt increases were essentially abolished in cells overexpressing any of the three TND mutants (Fig. 5B and C).

An important goal of the present study was to determine whether mutations in ABCC8, identified in patients with type 2 diabetes or mild disturbances in insulin secretion in adulthood, lead to changes in the molecular properties of K_ATP channels and impaired stimulus-secretion coupling. In the case of SUR1-Y356C this was indeed observed. Thus, we show that a relatively small change in channel ATP sensitivity 1) affects glucose signaling in vitro and 2) is associated with adult-onset type 2 diabetes. Importantly, we observed a clear correlation between the magnitude of the shift in ATP sensitivity for channels incorporating different SUR1 mutants, and the severity of impaired glucose metabolism in carriers, as previously demonstrated for KCNJ11/Kir6.2 mutations (6). While the present study was under review, Patch et al. (34) reported several mutations in ABCC8 that appear to be associated with diabetes onset in adults of median age 27.5 years, consistent with the present findings. However, the Y356C mutation that we describe appears to cause diabetes outside the range described (34) (maximum age recorded 35 years) and may thus represent a milder effect on channel opening.

The relatively small shift in ATP sensitivity (from 24 to 95 μmol/l as measured in inside-out patches) caused by the Y356C mutation in ABCC8/SUR1, clearly affected glucose-induced changes in β-cell electrical activity. This result strongly suggests that the Y356C mutation may lead to a diabetic phenotype. Indeed, the oral glucose tolerance test and euglycemic-hyperinsulinic clamp performed in the two nondiabetic carriers of the Y356C mutation showed a mild decrease of insulinogenic and disposition indexes (Table 1). This mild alteration of insulin secretion could thus lead to overt diabetes, as diagnosed in their father. Similar results were reported previously for mutations in KCNJ11/Kir6.2, which affected glucose homeostasis (metabolic sensitivity) (9,10,13,14).

We observed that two mutations (Y356C and L582V) that are associated with phenotypes of different severity in heterozygous patients cause different shifts in the ATP sensitivity of the K_ATP channel (Fig. 2B and D). While clear differences were observed between the glucose-induced changes in V_m when either channel was expressed in INS1(832/12) heterozygously, the homozygous expression of either mutant led to a near-complete suppression of depolarization in response to glucose (Fig. 4A–C). By contrast, glucose-induced [Ca²⁺]_cyt changes (Fig. 5) were still clearly different for the two types of channel, even after homozygous expression. Thus, the Y356C mutant lead to a substantially less marked inhibition of glucose-induced [Ca²⁺]_cyt increase than L582V (Fig. 5B). This suggests that subtle differences in V_m may be translated into more pronounced differences in [Ca²⁺]_cyt and, possibly, exocytosis. Alternatively, this apparent discrepancy may reflect the fact that there may be differences in the generation of K_ATP channel regulators including ATP but also the substrates/products of adenylate (23,35,36) and creatine (26,37,38) kinases, fatty acids (39), and inositol-phosphates (40,41) during electrophysiological recordings in whole-cell perforated patch (at 23°C) and in Ca²⁺ imaging experiments (35–37°C).

Although the function of K_ATP channels is well described, there is little direct experimental information on their three-dimensional structures. Predictions based on a previously described homology model of the Kir6.2 subunit are consistent with experimental data on its interaction with ligands (42,43), and this model has proved to be useful for interpreting the effects of mutations in KCNJ11 associated with neonatal diabetes (44). However, the structure of SUR1 is less well understood, because its overall level of sequence identity with the most closely related prokaryotic homologue of known high-resolution structure, the Staphylococcus aureus transporter Sav1866 (45), is only ∼21%. In addition, the resolution of a structural model of the channel complex, obtained by cryoelectron microscopy of a Kir6.2-SUR1 fusion protein, is only 18Å (46).

The interaction between the Kir6.2 and SUR1 subunits in this complex is predicted to be mediated by the C-terminus of Kir6.2 and transmembrane domain (TMD)0 of SUR1 (47). Mutations in TMD0 (48) or the TMD0-TMD1 linker (49) can affect the K_ATP channel gating (50) or amplify the stimulatory effect of Mg-nucleotides on SUR1 (49), thereby causing a severe diabetic phenotype in neonatal patients. However, the low-resolution structure of the K_ATP channel complex suggests that TMD1 and/or TMD2 are also likely to interact with the Kir6.2 subunit (46).

Given that tyrosine 356 is known, from the experimentally determined topology of SUR1 (51), to be located in TMD1, it was of interest to predict its potential structural and functional roles within this domain. To this end, a homology model of a portion of SUR1 (lacking TMD0) was created, using the bacterial multidrug transporter Sav1866 (45) as a template (online appendix). In the model, Y356 is located at the extracellular end of the second transmembrane helix in TMD1, with the side-chain oriented toward the outside of the helix bundle (online appendix Fig. 3). It would thus, at least in principle, be in a position to interact with TMDs of Kir6.2 (or TMD0 of the same SUR subunit). Such an interaction, leading to an effect of the mutation on ATP sensitivity via an allosteric effect on Kir6.2 rather than involving the SUR1 NBDs, would be consistent with the observation that the removal of Mg²⁺ (which abolishes the activatory effect of adenine-nucleotides on NBDs) (24) did not abolish the activatory effect of Y356C (Fig. 3C). The functional importance of Y356 is also suggested by the fact that an aromatic residue is conserved at the corresponding location not only in human SUR1 but also in other ABC transporters, e.g., multidrug resistance proteins 1, 2, 3, 4, and 6. Moreover, the TMD1 helix shows greater evolutionary conservation than the corresponding helix in TMD2, consistent with a role in protein-protein interaction rather than interaction with the lipid bilayer (data not shown).

In contrast to Y356C, the activatory effect of mutations L582V (Fig. 3D) and H1023Y (4) was not observed under Mg²⁺-free conditions, suggesting that these mutations exert their effects via the SUR1 NBDs. In the model, these residues, in the sixth transmembrane helix of TMD1 and the first transmembrane helix of TMD2, respectively, are more deeply buried within the protein structure and are predicted to interact with residues in TMD1 helix 3 and TMD2 helix 6, respectively (Fig. 1 and online appendix Fig. 3). Disruption of such interactions by mutation might exert an allosteric effect on the nucleotide binding activity of the NBDs: the TMDs are known to influence the ATPase activity of these domains in SUR1 (52).

In conclusion, we demonstrate that a weakly activating mutation in ABCC8 is the likely underlying cause of a heritable form of type 2 diabetes and insulin secretory insufficiency. This observation provides further evidence that quantitative shifts in the ATP sensitivity of single K_ATP channels, in this case mediated by SUR1, can lead to broadly proportional changes in whole body glucose homeostasis (44).

Supported by Wellcome Trust Programme Grants 067081/Z/02/Z and 081958/2/07/Z and a Divisional PhD scholarship from Imperial College to T.N.

We thank Dr. Martin Spitaler (the Facility for Imaging by Light Microscopy) for live-cell imaging and microscopy, Gao Sun for invaluable technical assistance, and Aurélie Dechaume for help in gene sequencing.