BKIP-1, an auxiliary subunit critical to SLO-1 function, inhibits SLO-2 potassium channel in vivo

Auxiliary subunits are often needed to tailor K+ channel functional properties and expression levels. Many auxiliary subunits have been identified for mammalian Slo1, a high-conductance K+ channel gated by voltage and Ca2+. Experiments with heterologous expression systems show that some of the identified Slo1 auxiliary subunits can also regulate other Slo K+ channels. However, it is unclear whether a single auxiliary subunit may regulate more than one Slo channel in native tissues. BKIP-1, an auxiliary subunit of C. elegans SLO-1, facilitates SLO-1 membrane trafficking and regulates SLO-1 function in neurons and muscle cells. Here we show that BKIP-1 also serves as an auxiliary subunit of C. elegans SLO-2, a high-conductance K+ channel gated by membrane voltage and cytosolic Cl− and Ca2+. Comparisons of whole-cell and single-channel SLO-2 currents in native neurons and muscle cells between worm strains with and without BKIP-1 suggest that BKIP-1 reduces chloride sensitivity, activation rate, and single-channel open probability of SLO-2. Bimolecular fluorescence complementation assays indicate that BKIP-1 interacts with SLO-2 carboxyl terminal. Thus, BKIP-1 may serve as an auxiliary subunit of SLO-2. BKIP-1 appears to be the first example that a single auxiliary subunit exerts opposite effects on evolutionarily related channels in the same cells.

BKIP-1 (BK channel interacting protein-1), a single pass membrane protein, is an auxiliary subunit of SLO-1 in C. elegans 24 . It elevates the half-maximal voltage for activation (V 50 ) and slows activation rate of SLO-1 at lower [Ca 2+ ] but reduces the V 50 and shows no obvious effect on activation rate at higher [Ca 2+ ] 24 . These effects of BKIP-1 are similar to those of β4 subunit on mammalian Slo1 6 . In addition, BKIP-1 plays an important role in SLO-1 surface expression 24 . Interestingly, a recent study shows that BKIP-1 genetically interacts with both SLO-1 and SLO-2 to control terminal differentiation of a pair of C. elegans olfactory neurons 28 . However, it remains to be determined how BKIP-1 may affect SLO-2 function. Here, through analyzing the effects of bkip-1 loss-of-function mutation on SLO-2 function in C. elegans neurons and muscle cells, we show that BKIP-1 is an inhibitory auxiliary subunit of SLO-2. BKIP-1 causes significant decreases in SLO-2 apparent Cl − and voltage sensitivities, activation rate, and single-channel open probability. These effects of BKIP-1 on SLO-2 are in contrast to those of BKIP-1 on SLO-1 24 , suggesting that a single auxiliary subunit may disparately regulate two different Slo channels within the same cells. The identification of BKIP-1 as regulators of both SLO-1 and SLO-2 reveals a new aspect of versatility of auxiliary/regulatory subunits in Slo channel functions.
BKIP-1 reduces SLO-2 Cl − but not Ca 2+ sensitivity. Given that SLO-2 is activated by Cl − and Ca 2+ on the cytosolic side 21,29 , BKIP-1 might regulate SLO-2 function through altering its Cl − and/or Ca 2+ sensitivities. We explored these possibilities by analyzing the effects of BKIP-1 on SLO-2 single-channel activity in inside-out patches of body-wall muscle cells. To identify SLO-2 single-channel events, we obtained membrane patches from mutants of several K + channels expressed in muscle cells using pipettes with a tip resistance of ~20 MΩ, and recorded single-channel activities using a bath solution containing 100 μM Ca 2+ and 50 mM Cl − . Single-channel activities were always observed in patches from shk-1(ok1581) but never in patches from either slo-2(nf101) or slo-2(nf101);shk-1(ok1581) (Fig. 2a), suggesting that SLO-2 is the only contributor to the single-channel activities under our experimental conditions. Because an earlier study reported the observation of SLO-1 single-channel activities in muscle cells under comparable experimental conditions (100 μM Ca 2+ and 150 mM Cl − in bath solution) 31 , we tried to identify any SLO-1 single-channel activity by using recording pipettes of a much larger tip size (resistance ~2 MΩ). Single-channel events were observed in only a small percentage (~20%) of patches from slo-2(nf101) but never in patches from slo-2(nf101);slo-1(md1745) (Fig. 2b), suggesting that they resulted from SLO-1. SLO-1 displayed flickery openings with sub-conductance states (Fig. 2b), which is very distinct from SLO-2. Density of SLO-1 in the muscle cell membrane appears to be extremely low; we never had a patch showing more than one active SLO-1 channel even though we spent a lot of time trying to get patches from various areas of the muscle cells, including dense body areas where SLO-1 is enriched 19 .
SCieNTiFiC RepoRts | (2017) 7:17843 | DOI:10.1038/s41598-017-18052-z BKIP-1 shapes muscle action potentials. SLO-2 plays important roles in setting the resting membrane potential and shaping action potentials in C. elegans body-wall muscle cells 22 . We therefore determined whether BKIP-1 regulates muscle resting membrane potential and action potentials. BKIP-1 does not affect the resting membrane potential but increases the amplitude of action potentials, the number of spikes per train, and inter-spike intervals (Fig. 5a). In addition, BKIP-1 increases the rise time of action potentials without altering their decay time and mid-peak width (Fig. 5b). BKIP-1 also has no effect on the afterhyperpolarization (Fig. 5b). These effects of BKIP-1 are conceivably due to the regulation of SLO-2 function.

BKIP-1 inhibits SLO-2 in motor neurons.
Among motor neurons important to C. elegans locomotion are three types: A, B, and D. The A and B types mediate backward and forward movements, respectively, and contract muscles by releasing acetylcholine, whereas the D type relaxes muscle by releasing GABA (γ-aminobutyric acid) 32 . Our previous study showed that SLO-2 is an important conductor of delayed outward currents in VA5, VB6 and VD5, which are representatives of A, B, and D type motor neurons, respectively 21 . To determine whether BKIP-1 regulates SLO-2 in these motor neurons, we compared whole-cell currents between SLO-2 and SLO-2 + BKIP-1. BKIP-1 causes a significant decrease in delayed outward currents of VA5 and VB6 but not VD5 (Fig. 6).
Because SLO-2 contributes a much smaller proportion of the total delayed outward currents in VD5 (33%) than either VA5 (80%) or VB6 (67%) 21 and the inhibitory effects of BKIP-1 on delayed outward currents are small even in VA5 and VB6, we wondered whether SLO-2 in VD5 is also regulated by BKIP-1 but this effect was not detected in the recorded whole-cell currents. To address this possibility, we analyzed the effects of BKIP-1 on SLO-2 single-channel properties in patches that apparently contained only one SLO-2 channel. Our analyses show that BKIP-1 inhibited SLO-2 P o by ~50% in all three neurons without altering the single-channel current amplitude (Fig. 7a,b). Analyses of single-channel open and closed times suggest that there are at least 2 different open states and 3 different closed states, and that the inhibitory effect of BKIP-1 on SLO-2 P o mainly results from decreased open durations of the two open states (Fig. 7c,d). Taken together, our analyses of the effects of BKIP-1 on SLO-2 single-channel activity suggest that BKIP-1 inhibits SLO-2 in both cholinergic and GABAergic motor neurons.

Discussion
This study establishes BKIP-1 as a novel auxiliary subunit of SLO-2 through experiments with native cells. Our results suggest that BKIP-1 modulates SLO-2 through several mechanisms, including reducing apparent Cl − and voltage sensitivities, slowing activation rate, and altering either the duration or proportion of open or closed states. These diverse effects of BKIP-1 on SLO-2 are reminiscent of those of mammalian β subunits on Slo1. For example, β subunits may alter Slo1 apparent voltage and Ca 2+ sensitivities, slow Slo1 activation, and regulate the duration and frequency of Slo1 single channel openings (see references 13,14 for reviews). The fact that BKIP-1 SLO-2 is expressed in body-wall muscle cells and many neurons, including ventral cord motor neurons 29,34 . In body-wall muscle cells, SLO-2 plays major roles in setting the resting membrane potential, repolarizing action potentials, and producing afterhyperpolarization 22 . In motor neurons, SLO-2 sets the resting membrane potential, and inhibits neurotransmitter release through shortening the duration and reducing charge transfer rate of spontaneous postsynaptic current bursts 21 , which are the electrical signals used by motor neurons to instruct muscle activity 35 . The present study shows that the presence of BKIP-1 causes significant changes in biophysical properties of SLO-2 in both muscle cells and motor neurons, and in muscle action potential properties, suggesting that BKIP-1 is important to the function of SLO-2 in many cells. Although BKIP-1 inhibits SLO-2 P o in all the cells examined, its effects on channel open and closed states are somewhat variable. C. elegans has at least 8 different isoforms of SLO-2 and 2 different isoforms of BKIP-1 due to alternative splicing (www.wormbase.org). Conceivably, the variable effects of BKIP-1 on SLO-2 open and closed states among the muscle cells and neurons BKIP-1 was initially identified as an auxiliary subunit of SLO-1 through a genetic screen for suppressors of a sluggish phenotype caused by expressing a hyperactive SLO-1 in C. elegans 24 . SLO-1 is a prominent K + channel in both neurons and body-wall muscle cells 19,20,36 . In neurons, SLO-1 colocalizes with presynaptic markers 20,37 , and serves as a potent negative regulator of neurotransmitter release 19 . In body-wall muscle cells, SLO-1 colocalizes with EGL-19 (Ca V 1/L-type voltage-gated Ca 2+ channel) 27 , and plays an important role in regulating Ca 2+ mobilization 20 . bkip-1 and slo-1 mutants are indistinguishable in phenotypes 24 , suggesting that BKIP-1 is indispensable for SLO-1 function in vivo. The finding of BKIP-1 as a regulator of both SLO-1 and SLO-2 demonstrates that a single auxiliary subunit may differentially regulate two members of the Slo family within the same cells.
BKIP-1 was first implicated in SLO-2 function in a recent study to determine the roles of SLO-1 and SLO-2 in asymmetric differentiation of a pair of olfactory sensory neurons (AWC) in C. elegans 28 . While single loss-of-function mutants of slo-1, slo-2 and bkip-1 show normal AWC differentiation, a combination of any two mutants of the three genes disrupts the asymmetry of AWC neurons. These observations led to the suggestion that BKIP-1 is required for the function of both SLO-1 and SLO-2 in AWC neurons. The apparent difference in BKIP-1 effects on SLO-2 between AWC neurons and motor neurons might be due to the usage of different isoforms of SLO-2 and/or BKIP-1 in these cells.
Although SLO-1 and SLO-2 are both expressed in muscle cells and neurons, single-channel events in inside-out patches obtained with the smaller pipette tip size appeared to be entirely due to SLO-2. SLO-1 single-channel events could only be observed in patches of muscle cells using pipettes with a much larger tip size, and even so only ~20% of the patches displayed SLO-1 activity. These observations are very different from those of an earlier study, which reported that single-channel events of SLO-1 are frequently observed and those of SLO-2 run down quickly in inside-out patches of body-wall muscle cells 31 . We do not know whether the apparent differences in results between these two studies are due to differences in experimental conditions or other factors. Because all the experiments of the previous study were performed with worms of wild-type slo-2 genetic background, it is unclear whether the results or interpretations have been complicated by SLO-2.
The modulatory effects of BKIP-1 on SLO-2 are relatively weak compared with those of many mammalian Slo1 regulatory proteins. For example, the γ1 subunit (LRRC26) and β1 subunit shift the V 50 of Slo1 by over 100 mV 2,11 , and the β2 and β3 subunits cause great inactivation of Slo1 8,10 . Among the possible causes for the relatively weak regulatory effect of BKIP-1 on SLO-2 are a much narrower physiological range of the membrane potential of C. elegans muscle cells and neurons compared with that of mammalian neurons and muscle cells. In C. elegans, muscle cells have a resting membrane potential of approximately −27 mV, and depolarize to approximately +20 mV at the peak of action potentials 22 ; and motor neurons have a resting membrane potential of −46 to −72 Figure 6. BKIP-1 inhibits SLO-2 whole-cell currents in VA5 and VB6 cholinergic motor neurons but not in VD5 GABAergic motor neuron. Shown are sample current traces and the current-voltage relationship of SLO-2 (VA5, n = 5; VB6, n = 6; VD5, n = 5) and SLO-2 + BKIP-1 (VA5, n = 12; VB6, n = 12; VD5, n = 9), which represent shk-1(ok1581);bkip-1(zw10) and shk-1(ok1581) strains, respectively. All values are shown as mean ± SE. The double asterisk (**) indicates a statistically significant difference at p < 0.01 levels (two-way ANOVA).
SCieNTiFiC RepoRts | (2017) 7:17843 | DOI:10.1038/s41598-017-18052-z mV depending on the types of neurons, and can depolarize to about −20 mV. In contrast, mammalian muscle cells and neurons generally have a resting membrane potential of −70 to −90 mV, and reaches as high as +50 mV at the peak of action potentials. Consistent with this speculation, the effect of BKIP-1 on SLO-1 V 50 is also small (~20 mV). It thus appears that nature has exquisitely tuned the properties of BKIP-1 to allow the proper functions of both SLO-2 and SLO-1 in vivo.

Electrophysiology.
Electrophysiological experiments were performed with adult hermaphrodites. An animal was immobilized on a glass coverslip by applying Vetbond TM Tissue Adhesive (3 M Company, St. Paul, MN). Application of the glue was generally restricted to the dorsal anterior portion of the animal, allowing the tail to sway freely during the experiment. A short longitudinal incision was made along the glued region. After clearing the viscera by suction through a glass pipette, the cuticle flap was folded back and glued to the coverslip, exposing several ventral body-wall muscle cells and a small number of motor neurons anterior to the vulva. The dissected worm preparation was treated with collagenase A (Roche Applied Science, catalogue number 10103578001, 0.5 mg/ml) for 10-15 sec and perfused with the extracellular solution for 5 to 10-fold of bath volume. Borosilicate glass pipettes were used as electrodes for voltage-clamp whole-cell or single-channel recordings. In whole-cell recordings, pipette tip resistance for recording from body-wall muscle cells was 3-5 MΩ whereas that for recording from motor neurons was ~20 MΩ. Classical whole-cell configuration was obtained by applying a negative pressure to the recording pipette. Series resistance was compensated to ~70%. In single-channel recordings with inside-out patches, pipette tip resistance was 20-30 MΩ for both muscle cells and motor neurons. The specific  ] was calculated using online software (http://web.stanford.edu/~cpatton/web-maxcS.htm). The 100-μM Ca 2+ solution was also used in experiments analyzing SLO-2 single-channel biophysical properties. In the experiments analyzing the effects of Cl − or Ca 2+ on SLO-2 activity, the recording pipette was inserted into the opening of a Perfusion Pencil TM (Automate Scientific, Inc., Berkeley, CA, USA) through which various solutions were perfused using a perfusion controller (VALVELINK8.2, Automate Scientific, Inc.).

Data analyses.
Clampfit (Molecular Devices) was used for the quantification of most electrophysiological data. The amplitude of SLO-2 or SHK-1 whole-cell current was quantified from the mean amplitude during the last 100 ms of each voltage step. The SLO-2 whole-cell current data were then converted to conductance (G) and fitted to the Boltzmann function G/G max = 1/{1 + exp[(V 50 − V)/k]}, where G max is the fitted value for maximal conductance, V 50 is the voltage of half maximal activation of conductance, and k is the term for the voltage dependence of activation in units of mV. The activation rate of SLO-2 whole-cell current was quantified by fitting the initial 500-ms current trace of each voltage step to two exponentials. The open probability SLO-2 in patches used to determine the effects of varying Cl − or Ca 2+ concentration was quantified from the current integral above the baseline, and the concentration-open probability curves were fitted with the equations y = Bottom + (Top-Bottom)/(1 + 10^((EC50-X)*slope) (for Cl − response) and y = Bottom + (Top-Bottom)/ (1 + 10^((LogEC50-X)*Slope)) (for Ca 2+ response), where Top and Bottom are the maximum and minimum responses, X is the test concentration, and Slope is the Hill slope. Patches containing only one channel were used for quantifying SLO-2 opening frequency, single-channel current amplitude distribution, and open and closed time analyses. The QuB software (https://qub.mandelics.com/) was used to fit open and closed times to exponentials, and to quantify the τ values and relative areas of the fitted components, which were automatically determined by the software. The entire recording of each experiment (30 sec duration) was used in such analyses. Statistical comparisons were performed with OriginPro (version 9, OriginLab, Northampton, MA, USA) using either unpaired t-test or one-way analysis of variance (ANOVA), as specified in figure legends. All values are shown as mean ± s.e. p < 0.05 is considered to be statistically significant. The sample size (n) equals to either the number of cells (Figs 1, 5 and 6) or membrane patches (Figs 2, 3, 4 and 7). Data graphs were made with OriginPro. Data availability. All data generated or analyzed in this study are included in this published article.