Evidence for Presence and Functional Effects of Kv1.1 Channels in β-Cells: General Survey and Results from mceph/mceph Mice

Background Voltage-dependent K+ channels (Kv) mediate repolarisation of β-cell action potentials, and thereby abrogate insulin secretion. The role of the Kv1.1 K+ channel in this process is however unclear. We tested for presence of Kv1.1 in different species and tested for a functional role of Kv1.1 by assessing pancreatic islet function in BALB/cByJ (wild-type) and megencephaly (mceph/mceph) mice, the latter having a deletion in the Kv1.1 gene. Methodology/Principal Findings Kv1.1 expression was detected in islets from wild-type mice, SD rats and humans, and expression of truncated Kv1.1 was detected in mceph/mceph islets. Full-length Kv1.1 protein was present in islets from wild-type mice, but, as expected, not in those from mceph/mceph mice. Kv1.1 expression was localized to the β-cell population and also to α- and δ-cells, with evidence of over-expression of truncated Kv1.1 in mceph/mceph islets. Blood glucose, insulin content, and islet morphology were normal in mceph/mceph mice, but glucose-induced insulin release from batch-incubated islets was (moderately) higher than that from wild-type islets. Reciprocal blocking of Kv1.1 by dendrotoxin-K increased insulin secretion from wild-type but not mceph/mceph islets. Glucose-induced action potential duration, as well as firing frequency, was increased in mceph/mceph mouse β-cells. This duration effect on action potential in β-cells from mceph/mceph mice was mimicked by dendrotoxin-K in β-cells from wild-type mice. Observations concerning the effects of both the mceph mutation, and of dendrotoxin-K, on glucose-induced insulin release were confirmed in pancreatic islets from Kv1.1 null mice. Conclusion/Significance Kv1.1 channels are expressed in the β-cells of several species, and these channels can influence glucose-stimulated insulin release.


Introduction
It is established that voltage-dependent K + channels (Kv) mediate repolarisation of bcell action potential and thereby abrogate insulin secretion.Elucidation of the molecular mechanisms and the regulation of Kv channels are however complicated by their multitude of forms, each class being recognized to have several subcategories.Alpha subunits form the actual conductance pore and have been grouped into 12 classes named Kv1-12.Among these channels, Kv2.1 has been shown to be robustly functional in b-cells, to the extent that in a number of species, it is regarded as the most influential Kv channel on insulin secretion [1].A considerable impact of this channel at the level of electrophysiology has also been documented [2].
It is also recognized that Kv2.1 is not the only Kv channel of importance [2].Kv1.1 has in animal models been found to be functional in brain, gastrointestinal muscle cells, renal tissue and heart [3,4,5].A functional role for the Kv1.1 channel in b-cells has been questioned and its expression in this tissue has not been consistently detected [6].However, a systematic search for the Kv.1.1 channels has not been performed, and a role for Kv1.1 in b-cell function and/or development has not, to our knowledge, been tested rigorously.
In this study, we first obtained preliminary results showing that isolated islets of Langerhans from BALB/cByJ-Kv1.1 mceph/mceph mice (called mceph/mceph), for which the Kv1.1 protein is truncated and non-functional secreted more insulin than islets from wild type mice to a glucose challenge.We then searched for and found evidence for Kv1.1 expression in islets from several strains of mice as well as in islets from rats and humans.We continued to test for functional and developmental effects of Kv1.1 on b-cells.We found, as expected that a full-length Kv1.1 protein was missing in islets from mceph/mceph mice.Upon finding evidence for mutation-linked effects on insulin secretion we performed electrophysiological measurements to test for traits associated with Kv channel deficiency.Finally, analyses of mice in which the Kv1.1 gene was knocked out (BALB/cByJ.C3 HeB/FeJ-Kv1.1 2/2 ) confirmed some of the key results in this study.

Ethics statement
All animal studies were approved by the Northern Stockholm Ethical Committee on Experimental Animal Care (376/03; 348/ 06) and performed in accordance with guidelines from the Swedish National Board for Laboratory Animals.Human islets were obtained from a Nordic network under conditions specified elsewhere [7].

Reverse-transcription polymerase chain reaction (PCR)
Total RNA was isolated using RNeasy kits, reverse transcribed and analyzed by conventional and real-time PCR as described in Supporting Information.

Microarray
Human islets were obtained from three normal individuals and cultured for 48 h in 5.5 mmol/l glucose before RNA was isolated.Labelled cRNAs were synthesized and hybridised to the Human Gene 1.0 ST Array (Affymetrix, Santa Clara, CA) at the Novum Affymetrix core facility, Karolinska Institutet.

Histopathological and immunohistochemical techniques
Pancreatic glands were dissected away from the surrounding tissues in their entirety.Specimens from the corpus/cauda regions were used for histopathological staining procedures and immunohistochemical (IHC) analyses.For further details see Supporting Information.

Western blotting
Protein extracts were separated on SDS-PAGE gels, transferred to nitrocellulose filters and probed with monoclonal antibodies as described in Supporting Information.

Blood collection and isolation and culture of pancreatic islets
Blood collection and isolation of pancreatic islets as well as tissue culture of MIN6 and INS-1 cells, were carried out as described in Supporting Information.

Insulin release measurements
Batch-type incubations.Following culture and preincubation as described in Supporting Information equal sized islets were incubated in groups of three for 60 min at 37uC in 300 ml KRB containing 3.3 or 16.7 mmol/l glucose, each with or without 10 nM dendrotoxin-K.Each experimental condition consisted of three or four individual groups of three islets.The insulin accumulated in the incubation medium was measured as previously described [15].Islet insulin content was measured after acid-ethanol extraction [16] of islets retrieved from the batch incubations.
Perifusion of islets and dispersed islet cells.After culture and preincubation, 60-80 islets or dispersed islet cells from the same amount of islets (see Supporting information) were added to each of two or three perifusion chambers and perifused as previously described [17].Briefly, islets or cell suspension were layered between polystyrene beads (Bio-Rad) and perifused by use of a peristaltic pump (Ismatec SA, Zu ¨rich, Switzerland).Samples of the perifusate were collected every minute, frozen and stored at 220uC.

Electrophysiology
Whole-cell currents and membrane potentials were recorded with patch-clamp technique [18] by using an EPC-10 patch-clamp amplifier (HEKA Electronics, Lambercht, Germany).All experiments were performed at room temperature, approximately 22uC.For details see Supporting Information.

Statistical analysis
Results are expressed as mean 6 SEM.Significant differences in the means of sample groups from mceph/mceph and wild-type mice were tested using Student's paired t-test (two-sided).A P value ,0.05 was considered significant.
We detected Kv1.1 expression both in islets from ob/ob and lean mice, as well as in islets from rat (Fig. 1C).Kv1.1 expression was also detected in the INS-1 cell line, but not in MIN6 cells (Fig. 1C).Importantly we in addition detected Kv1.1 expression in human islets isolated from both non-diabetic donors and donors with type 2 diabetes (Fig. 1D).Using real-time PCR no differences in Kv1.1 expression levels (relative to 18S) were observed between islets from three non-diabetic and three diabetic individuals.However, the results demonstrated a greater variability in Kv1.1 expression in islets from donors with type 2 diabetes (1.060.05 vs. 1.560.8,P = 0.7, n = 3, Fig. 1E).
Kv1.1 expression was also verified in human islets from three non-diabetic individuals by microarray (Affymetrix, results not shown).
Hence, Kv1.1 expression was detectable in normal islet tissue from rats, humans and mice.

Detection of Kv1.1 protein
By Western blotting using a monoclonal antibody against the Cterminal of Kv1.1 protein, the full-size Kv1.1 protein was demonstrated to be present in islets from wild-type mice and rats (Fig. 2A), but, as expected, could not be detected in islets from mceph/mceph mice which express only the N-terminal of Kv1.1 protein (Fig. 2A).
To investigate in which islet cell types the Kv1.1 protein was present we performed double immunostaining for Kv1.1 and insulin, as well as for glucagon and somatostatin.The Kv1.1 antibody used was against the N-terminal region present in both the normal and the truncated proteins.A clear-cut Kv1.1-like immunoreactivity was found to be co-localized with insulin in the bcells both in wild-type and in mceph/mceph mice.The immunoreactivity was moderate in islets of wild-type mice (Fig. 2B, upper panel, green), but strong in islets of mutated mceph/mceph mice (Fig. 2B, lower panel).Hence, the IHC data suggest over-expression of the truncated protein in mceph/mceph islets.
We also found co-localization of Kv1.1 with glucagon in the acells as well as with somatostatin in d-cells (data not shown).
The peptide, used to generate the N-terminal Kv1.1 antibody, shares an identity with Kv1.2 consisting of a stretch of 6 out of 24 amino acids.No signal was elicited when islets from mceph/mceph mice were incubated with an antibody against Kv1.2.This observation indicates that the signal elicited by the Kv1.1 antibody was due to the truncated (over-expressed) Kv1.1 protein and not due to cross-reactivity with Kv1.2 (Fig. 2C).

Islet structure in mceph/mceph mice
The basic structure of the pancreatic islets of Langerhans was similar for mutant, and wild-type mice, as detected using histopath- ological staining and IHC analyses.Thus, the typical ''mantle'' structure of rodent islets was seen, with insulin-immunoreactive cells predominating in the central areas of islets, and glucagon-and somatostatin-imunoreactive cells at the periphery.The proportions of these cell types were 85% insulin-imunoreactive cells, 10% glucagon-imunoreactive cells, and 5% somatostatin-imunoreactive cells (Fig. 3A).There was no evidence of necrosis, inflammation, hypertrophy or atrophy of the three islet parenchymal cell types.

Body weights and blood glucose
Body weights were lower in mceph/mceph vs. wild-type mice at corresponding ages (Table 1).Levels of blood glucose (non-fasting) did not differ between mceph/mceph and wild-type mice (Table 1).

Islet insulin contents and secretion
Insulin contents of isolated islets did not differ between mceph/ mceph and wild-type islets (16566194 vs.15096111 mU/islet, n = 11).showing-all at the same magnification x400-serial sections of an islet of Langerhans in the pancreatic parenchyma from mice belonging to the three strains (A).The fact that the islet chosen for this particular set of photomicrographs happens to be larger from the Kv1.1 null mouse strain than those from the other two strains, is by pure incidence only, and does not, by no means, imply that the Kv1.1 null islets in general would be larger than those in a normal murine pancreas.Over-expression of truncated Kv1.Islets from mceph/mceph and wild-type mice secreted similar amounts of insulin during culture conditions (596658 vs. 588663 mU/islet/24 h).In batch incubations however, glucose (16.7 mmol/l) -induced insulin secretion was moderately although significantly enhanced in islets from mceph/mceph mice (7169.5 vs. 5267.5mU/islet/h, respectively, P,0.03, n = 11, Fig. 4A).We also performed experiments in the presence of the Kv1.1-specific blocker dendrotoxin-K [19].Addition of dendrotoxin-K to 16.7 mmol/l glucose-containing media tended to increase insulin secretion from wild type islets (from 4969.6 to 68614 mU/islet/ h), whereas there was a decrease in release from mceph/mceph islets (from 60613 to 50610 mU/islet/h, n = 7, Fig. 4B).The difference in response to dendrotoxin-K in wild-type and mceph/mceph mice was significant, P = 0.02 (Fig 4C ).

Perifusion experiments
Islets from wild type and mceph/mceph mice were perifused with 3.3 mmol/l glucose followed by 16.7 mmol/l glucose and finally 10 nmol/l dendrotoxin-K was added to the 16.7 mmol/l glucose in one perifusion chamber whereas the other chamber continued to be perifused with 16.7 mmol/l glucose only.The response to dendrotoxin-K was calculated as follows: The average insulin response to 16.7 mmol/l glucose was calculated for min 29-49 and subtracted from the average insulin response at min 69-89 making it possible to calculate the pure effect of dendrotoxin-K in each channel.In wild type islets dendrotoxin-K elicited a significant insulin response of 0.2960.1 mU/islet/min, P = 0.05, Fig. 4D and G) whereas no such response was seen in islets from mceph/mcpeh mice (20.160.1 mU/islet/min, P = 0.45, Fig. 4E and  G).The difference in response between the strains was significant (P = 0.035).Also the insulin response to elevating glucose from 3.3 to 16.7 mmol/l during perifusion was significantly increased in islets from mceph/mceph vs. wild-type mice (0.3460.04 vs. 0.5160.04mU/islet/min, P = 0.01, Fig. 4D, E and G).
We also performed experiments in dispersed islet cells from wild type mice (Fig. 4H).In these experiments basal insulin secretion was similar before stimulation with dendrotoxin-K and TEA (Tetraethylammonium, a non-selective K + channel inhibitor).In the presence of 16.7 mmol/l glucose TEA increased insulin secretion by 148664%, n = 3, Fig. 4H.A much smaller but still significant effect was obtained by the addition of dendrotoxin-K which increased insulin secretion by 3569%, P, 0.04.

Electrophysiology
Previous studies in brain tissue found that electrophysiological parameters were affected by the mceph/mceph mutation [10].Furthermore, other Kv deficiencies give rise to excitatory abnormalities in b-cells [2].Our aim was to see if abnormalities indicative of Kv channel deficiencies and specifically those of Kv1.1, were present in mceph/mceph islets.
Whole-cell currents.Whole-cell currents were recorded from pancreatic b-cells isolated from wild-type and mceph/mceph mice, using the whole-cell configuration of the patch-clamp technique (Fig. 5A-E).The cells were voltage-clamped at 280 mV and subsequently depolarized in steps of 20 mV every s (Fig. 5A and B).After establishment of the whole-cell configuration, current recordings showed a marked 'run-down'.To compensate for this 'run-down', whole-cell current was measured before and after addition of the Kv1.1 blocker dendrotoxin-K and the effect of the toxin was compared to the estimated whole-cell current.In wild-type b-cells, the K + current was decreased by 20 nmol/l dendrotoxin-K with 2065% (n = 6; P,0.01) and in mceph/mceph b-cells with 12611% (n = 6, n.s.).No significant difference could be monitored between the wild-type and mceph/mceph b-cell whole cell currents.Compiled data of the effect of dendrotoxin-K on normalized whole-cell K + currents in wild-type and mecph/mceph beta cells are shown in Fig. 5E.Membrane potentials.Membrane potentials were monitored using the perforated-patch technique.The membrane potential at 3 mmol/l glucose was 25864 mV (n = 4) in wild-type b-cells, and increased to 24162 mV (n = 5) in the presence of 15 mmol/l glucose (measured as the lowest point between two action potentials).Addition of 20 nmol/l dendrotoxin-K did not affect membrane potential (23967 mV, n = 3, n.s.).In mceph/ mceph b-cells, membrane potential was 25965 mV (n = 6) and 23264 mV (n = 3) in the presence of 3 mmol/l and 15 mmol/l glucose, respectively.In the presence of 15 mmol/l glucose and 20 nmol/l dendrotoxin-K, the membrane potential was 23666 mV (n = 3, n.s).The mceph/mceph mice islets were thus more depolarised in 15 vs. 3.3 mmol/l glucose compared to wild type islets (23264 vs. 24162 mV), P = 0.018.
Action potentials.The traces of membrane potential recordings from wild-type (Fig. 6A) and mceph/mceph b-cells are shown in Fig. 6B.On an expanded time scale (Fig. 6C-D) details of action potential frequency can be monitored.The frequency of action potentials in wild-type b-cells was significantly lower, compared to that observed in mceph/mceph b-cells, whereas the frequency of action potentials decreased in the presence of dendrotoxin-K in wild-type b-cells but remained unaltered in mceph/mceph b-cells (Fig. 6G).
Detailed analysis of action potential characterizes.Fig. 6E-F show the summary of 20 action potentials obtained in the presence of 15 mmol/l glucose 6 dendrotoxin-K.In wild-type b-cells (Fig. 6E), addition of 20 nmol/l dendrotoxin-K prolonged the repolarisation phase (ii and iv).This effect was reversible upon wash-out of the toxin (iii).Augmentation of the repolarization phase was not seen in mceph/mceph b-cells (Fig. 6F), in which action potentials remained unaltered in the presence of dendrotoxin-K (Fig. 6F, i-iii).Fig. 6H shows compiled data of the repolarisation phase in wild-type and mceph/mceph b-cells, measured as a time lag from the action potential peak and 2/3 of the action potential amplitude (Fig. 6I).The action potential duration was significantly prolonged in b-cells from mceph/mceph mice, compared to those from wild-type mice.Action potential duration was also significantly increased in wild-type b-cells treated with dendrotoxin-K, while dendrotoxin-K did not affect the action potential duration in mceph/mceph b-cells.Collectively, the electrophysiological recordings show that dendrotoxin-K is able to block a fraction of the total K + current in wild-type b-cells, whereas this effect is not present in b-cells from the mceph/mceph mouse.Detailed analyses of these action potentials showed that mceph/mceph b-cells display a higher action potential frequency and a prolongation of repolarization, compared to wild-type b-cells.This effect on action potential duration could be mimicked by dendrotoxin-K in wild-type b-cells.

Results in Kv1.1 null mice
With limited access to Kv1.1 null mice we tested for confirmation of some of the findings in mceph/mceph.There were no differences in islet structure in Kv1.1 null mice, compared to wild-type and mceph/mceph (Fig. 3A).In perifusion experiments with islets from wild-type and Kv1.1 null mice, in which the glucose concentration was increased from 3.3 to 7 mmol/l, insulin secretion was significantly increased in Kv1.1 null islets vs. wildtype islets (0.4760.06 vs. 0.1960.05mU/islets/min, n = 3 and 4 experiments, P = 0.037).The average insulin response from islets when glucose levels rose from 3.3 to 16.7 mmol/l in perifusion experiments was significantly elevated in both mceph/mceph and in Kv1.1 null islets vs. wild type islets (Fig. 4E, F and G).

Discussion
We demonstrate the presence of Kv1.1-specific transcripts in mouse, rat and human islets.Previous reports are discrepant as to the expression of Kv1.1 in primary pancreatic b cells [1,20,21].Discrepancies may be due to the sensitivity of the techniques used.Here we examined Kv1.1 expression by an optimized PCR methodology with appropriate negative controls in all cases.Consistent Kv1.1 amplification from cDNA prepared from rodent islets required 40 cycles of amplification from template levels corresponding to 50 ng total RNA.This indicates that this gene is expressed, albeit at low levels, in mouse and rat islets.Notably, in human islets prepared from six individuals (three controls, three diabetics), Kv1.1 expression was detectable in each sample at template levels corresponding to 10 ng total RNA, and these observations were confirmed by TaqMan real-time PCR and cDNA microarray (data not shown).A previous study did not detect Kv1.1 expression in human islets, however, in that study islets from only a single individual were analyzed [22].
We found that pancreatic islets from mceph/mceph mice secrete more insulin in response to glucose compared to islets from wild type mice.The confirmation of these findings in Kv1.1 null mice indicates that the increase in insulin secretion is indeed due to the lack of functional Kv1.1 channels.The effects of the Kv1.1 blocker dendrotoxin-K on insulin release also strongly indicate that an enhanced insulin response to glucose in the mceph/mceph mice is due to a lack of functional Kv1.1 activity.Thus, both in static incubations (islets) and in perifusion experiments (islets and dispersed islet cells), dendrotoxin-K augmented insulin secretion from wild-type mice but not from mceph/mceph mice.Perifusion experiments with islets from Kv1.1 null animals further confirmed these findings.
Our experiments give further indications of the specific impact of the Kv1.1 channel mutation on membrane electrophysiology of b-cells from mceph/mceph mice.Importantly, glucose-induced action potentials were increased in spiking frequency and duration in mceph/mceph mice when measured in the absence of dendrotoxin-K.The fact that mceph/mceph mice b-cells showed an increased action potential frequency at the same time as the action potential duration was increased may seem hard to understand since one would expect a slower frequency (2).At 15 mmol/l glucose there was an increase in membrane potential in mceph/mceph vs. wild type b-cells (23264 vs. 24162 mV, P = 0.018) which could at least in part explain this finding.Increased action potential duration has also been observed in ventricular myocytes from the hearts of Kv1DN mice with overexpression of a truncated Kv1. 1 [23].Our results with dendrotoxin-K were in further agreement with the Kv1.1 channel being functional at the level of electrophysiology.Thus, in wild type b-cells dendrotoxin-K obliterated the differences between wild-type and the mceph/mceph mice regarding b-cell action potential duration.
Our results concerning expression of the Kv1.1 mRNA and protein support a key role for the Kv1.1 mutation in the abnormal function of b-cells from mceph/mceph mice.Kv1.1 expression was thus readily detected in wild-type islets, whereas a truncated form was found in mceph/mceph islets.Interestingly, the Kv1.1 protein seemed, in its truncated form in mceph/mceph islets, more abundant than the normal protein in islets of wild-type mice.These observations are similar to those reported for some regions in the brain of mceph/mceph mice.Accumulation of truncated protein could theoretically have untoward and toxic effects.However, we note that there were, by histological analysis, no signs of b-cell toxicity in mceph/mceph mice.Furthermore, there was no sign of change in the expression of Kv2.1 protein.Ablation of Kv2.1 also affected action potentials in a way that was not observed here [2].Dendrotoxin-K increases insulin secretion in wild type but not in mceph/mceph and Kv1.1 null mouse islets.Incremental insulin release from batch-type incubations of islets from wild-type (wt) and mceph/mceph (m/m) mice in response to 10 nmol/l dendrotoxin-K (dtx), n = 7 (B).The difference in response to dendrotoxin-K for wild-type and mceph/mceph mice was significant, *P = 0.02 (C).Each batch-type experiment was performed in triplicate or quadruplicate with three equal-sized islets per tube.Equal amounts of islets from wild-type (D, n = 4), mutant mceph/mceph (E, n = 4) or Kv1.1 null (F, n = 3) mice were loaded into separate perifusion chambers.For typographical reasons only every fifth error bar is shown in (E and F).The response to 10 nmol/l dendrotoxin-K in wild-type islets was significant (*P = 0.05).For details of how the dendrotoxin-K-effect was calculated see results.(G) summarises average insulin response to glucose level elevation from 3.3 to 16.7 mmol/l as well as addition of dendrotoxin-K in the islet perifusion experiments.(The insulin response to an increase in glucose concentration from 3.3 to 16.7 mmol/l was significantly increased in islets from mceph/mceph and Kv1.1 null vs. wild-type mice (*P = 0.01).The response to dendrotoxin-K in wild-type islets was significant ({P = 0.05), while that was not the case in mceph/mceph and Kv1.1 null islets.Increased insulin response to dendrotoxin-K and TEA in dispersed islet cells from wild type mice.Perifusion with 200 nmol/l dendrotoxin-K or 15 mmol/l TEA.0.5 mmol/l IBMX was present from min 5 to 51.One typical experiment out of three is shown (H).doi:10.1371/journal.pone.0018213.g004 Immunostaining showed that Kv1.1 was present also in aand d-cells.The functional role, if any, for Kv1.1 in these cells remains to be investigated.
The significantly lower levels of blood glucose and the tendency for higher serum insulin levels in Kv1.1 null mice vs. wild type seem logical.In contrast mceph/mceph mice exhibit blood glucose and insulin levels that are similar to those seen in wild-type mice.mceph/mceph and Kv1.1 null mice have a similar behavioural phenotype with body tremor, jittering and occasional forelimb paddling, and cramps.However, the phenotype of mceph/mceph mice appears more severe than that of the Kv1.1 null mice (Lavebratt, unpublished observations).More marked severity of Kv1.1 deficiency phenotype could lead to a higher stress level which in turn could result in higher blood glucose levels and increased insulin resistance.In vivo stress could also activate alpha-adrenergic receptors on the b-cell and thereby inhibit insulin secretion thus counteracting the increased insulin secretion from islets when tested in vitro.
Another possibility is that lack of Kv1.1 in the CNS induces problems of glucose sensing in the hypothalamus that could secondarily affect the endocrine pancreas in different ways in the mceph/mceph and Kv1.1 null mice.The truncated Kv1.1 in mceph/ mceph mice can assemble with other Kv1 subunits and potentially trap them in endoplasmatic reticulum with subsequent degradation [11].
Which importance should then be assigned to the Kv1.1 channel for b-cell function vis-a `-vis other Kv channels?A large body of evidence indicates that the Kv2.1 channel is the major K +rectifying channel in b-cells.Nevertheless, as highlighted by a recent study [2] Kv2.1 is not the only Kv channel of importance.Hence a supplementary role for Kv1.1 channel (as well as perhaps other Kv channels) can be envisaged.Notably, our results indicate that this notion can be extended also to human beta cells.
In summary Kv1.1 channels display an inter-species expression pattern.Furthermore, our results in mouse b-cells indicate that these channels are of functional importance.

Figure 1 .
Figure 1.Kv1.1 is expressed in mceph/mceph mice.Kv1.1 mRNA expression in islets from wild-type and mceph/mceph mice by RT-PCR with the S2645-AS3455 (A) and S2688-AS2912 (B) primer pairs, which amplify sequences not including, and including, the mceph mutation respectively.Marker = pUC18 Msp I digest.Kv1.1 is expressed in mice, rat and human islets as well as INS-1 cells but not MIN6 cells.Kv1.1 mRNA expression by RT-PCR with the S2645-AS3455 primer pair (C).Marker = pUC18 Msp I digest.(D) Kv1.1 mRNA expression in islets isolated from three human control and three diabetic individuals by RT-PCR with the S1926-AS2092 primer pair.Marker = pUC18 MspI digest.(E) Kv1.1 mRNA expression levels relative to 18S RNA by real-time PCR in islets from human non-diabetic (n = 3) and diabetic individuals (n = 3).Samples were run in duplicates.All analyses included a corresponding negative control (RT-ve).doi:10.1371/journal.pone.0018213.g001 Over-expression of truncated Kv.1.1 protein in the mceph/mceph could affect other Kv channels.To investigate that we performed Western blot for the most important Kv channel involved in repolarisation of action potentials in the beta cell, namely Kv 2.1.No sign of change in the expression of Kv 2.1 channel protein was seen in mceph/mceph vs. wild type islets (108610 vs. 100%, n = 3, Fig.3B+C).

Figure 2 .
Figure 2. Kv1.1 protein is present in islets from wild type mice and rats.Western blotting using a monoclonal antibody against the Cterminal of Kv1.1 protein (A).Brain from wild type and mceph/mceph mice was used as controls.mceph/mceph served as specificity control as these mice express only the N-terminal of Kv1.1.b-actin was used as loading control.Kv1.1 protein is present in b-cells.Islets from wild type mice (wt) (B upper panel) and mceph/mceph (m/m) mice (B lower panel) double immunostained for N-terminal Kv1.1 (green, left) and insulin (red, middle); merged to the right.The N-terminal of Kv1.1 is present in beta cells of both genotypes.Note that Kv1.1 is up-regulated in mceph/mceph mice compared to wild type mice.Specificity of the N-terminal Kv1.1-LI was tested by determining the level of Kv1.2-LI.Strong N-terminal Kv1.1-LI and very weak Kv1.2-LI in an islet from mceph/mceph mice assured that the Kv1.1-LI reflected presence of Kv1.1 (C).doi:10.1371/journal.pone.0018213.g002

Figure 3 .
Figure 3. Normal islet structure in mceph/mceph and Kv 1.1 null mice compared to wild type mice.Medium-power photomicrographs,showing-all at the same magnification x400-serial sections of an islet of Langerhans in the pancreatic parenchyma from mice belonging to the three strains (A).The fact that the islet chosen for this particular set of photomicrographs happens to be larger from the Kv1.1 null mouse strain than those from the other two strains, is by pure incidence only, and does not, by no means, imply that the Kv1.1 null islets in general would be larger than those in a normal murine pancreas.Over-expression of truncated Kv1.1 protein in mceph/mceph does not affect the expression of Kv2.1 protein.Western blot with a polyclonal antibody against Kv2.1 protein in islets from mceph/mceph.Column diagram shows compiled data from 3 individual experiments (C), blots are from a typical experiment (B).doi:10.1371/journal.pone.0018213.g003 Figure 3. Normal islet structure in mceph/mceph and Kv 1.1 null mice compared to wild type mice.Medium-power photomicrographs,showing-all at the same magnification x400-serial sections of an islet of Langerhans in the pancreatic parenchyma from mice belonging to the three strains (A).The fact that the islet chosen for this particular set of photomicrographs happens to be larger from the Kv1.1 null mouse strain than those from the other two strains, is by pure incidence only, and does not, by no means, imply that the Kv1.1 null islets in general would be larger than those in a normal murine pancreas.Over-expression of truncated Kv1.1 protein in mceph/mceph does not affect the expression of Kv2.1 protein.Western blot with a polyclonal antibody against Kv2.1 protein in islets from mceph/mceph.Column diagram shows compiled data from 3 individual experiments (C), blots are from a typical experiment (B).doi:10.1371/journal.pone.0018213.g003

Figure 4 .
Figure 4. Increased glucose-induced insulin secretion in islets from mceph/mceph vs. wild type mice.Glucose-induced insulin secretion from batch-type incubations of islets of Langerhans from wild-type (+/+) and mceph/mceph (m/m) mice, n = 11 (A).*P,0.03 vs. wild type.Dendrotoxin-K increases insulin secretion in wild type but not in mceph/mceph and Kv1.1 null mouse islets.Incremental insulin release from batch-type incubations of islets from wild-type (wt) and mceph/mceph (m/m) mice in response to 10 nmol/l dendrotoxin-K (dtx), n = 7 (B).The difference in response to dendrotoxin-K for wild-type and mceph/mceph mice was significant, *P = 0.02 (C).Each batch-type experiment was performed in triplicate or quadruplicate with three equal-sized islets per tube.Equal amounts of islets from wild-type (D, n = 4), mutant mceph/mceph (E, n = 4) or Kv1.1 null (F, n = 3) mice were loaded into separate perifusion chambers.For typographical reasons only every fifth error bar is shown in (E and F).The response to 10 nmol/l dendrotoxin-K in wild-type islets was significant (*P = 0.05).For details of how the dendrotoxin-K-effect was calculated see results.(G) summarises average insulin response to glucose level elevation from 3.3 to 16.7 mmol/l as well as addition of dendrotoxin-K in the islet perifusion experiments.(The insulin response to an increase in glucose concentration from 3.3 to 16.7 mmol/l was significantly increased in islets from mceph/mceph and Kv1.1 null vs. wild-type mice (*P = 0.01).The response to dendrotoxin-K in wild-type islets was significant ({P = 0.05), while that was not the case in mceph/mceph and Kv1.1 null islets.Increased insulin response to dendrotoxin-K and TEA in dispersed islet cells from wild type mice.Perifusion with 200 nmol/l dendrotoxin-K or 15 mmol/l TEA.0.5 mmol/l IBMX was present from min 5 to 51.One typical experiment out of three is shown (H).doi:10.1371/journal.pone.0018213.g004

Figure 5 .Figure 6 .
Figure 5. Dendrotoxin-K significantly blocks relative K + currents in wild type vs. mceph/mceph b-cells.Whole-cell patch-clamp recordings from wild-type and mceph/mceph pancreatic b-cells.I-V relationship for wild-type (A) and mceph/mceph (B) pancreatic b-cells before, during and after exposure to 20 nmol/l dendrotoxin-K were obtained by clamping the cells at 280 mV and subsequently depolarized in steps of 20 mV for 100 ms to 80 mV.Example traces are recorded from wild-type b-cell in (C) at +80 mV, and the current measured at +80 mV is plotted every 20 s (D).In (E), to compensate for channel run-down, in each recording whole-cell K + current was normalized to the estimated K + current by extrapolating the current before and after the toxin administration.*P,0.05;n.s., not significant.doi:10.1371/journal.pone.0018213.g005