Developmental regulation of voltage-dependent delayed rectifier potassium current (I Kv) ofXenopus primary spinal neurons regulates the waveform of the action potential. I Kv undergoes a tripling in density and acceleration of it activation kinetics during the initial day of its appearance. Another voltage-dependent potassium current, the A current, is acquired during the subsequent day and contributes to further shortening of the impulse duration. To decipher the molecular mechanisms underlying this functional differentiation, we are identifying potassium channel genes expressed in the embryonic amphibian nervous system. Potassium channels consist of pore-forming (α) as well as auxiliary (β) subunits. Here, we report the primary sequence, developmental localization, and functional properties of twoXenopus Kvβ genes. On the basis of primary sequence, one of these (xKvβ2) is highly conserved with Kvβ2 genes identified in other species, whereas the other (xKvβ4) appears to identify a new member of the Kvβ family. Both are expressed in developing spinal neurons during the period of impulse maturation but in different neuronal populations. In a heterologous system, coexpression of xKvβ subunits modulates properties of potassium current that are developmentally regulated in the endogenousI Kv. Consistent with xKvβ4's unique primary sequence, the repertoire of functional effects it has on coexpressed Kv1α subunits is novel. Taken together, the results implicate auxiliary subunits in regulation of potassium current function and action potential waveforms in subpopulations of embryonic primary spinal neurons.
- potassium channels
- Kv1α subunits
- auxiliary Kvβ subunits
- Xenopus embryos
- electrical excitability
- spinal cord neurons
In several systems, voltage-dependent potassium current (IKv) undergoes developmental regulation and dynamically regulates emerging patterns of neuronal excitability (for review, see Spitzer and Ribera, 1998). In embryonic amphibian spinal neurons, maturation ofI Kv converts long-duration action potentials to brief sodium-dependent spikes (Barish, 1986; O'Dowd et al., 1988; Lockery and Spitzer, 1992). Maturation ofI Kv involves a tripling in its density and acceleration of activation and inactivation kinetics (Barish, 1986;O'Dowd et al., 1988; Ribera and Spitzer, 1990). Definition of the molecular mechanisms that underlie development of excitability requires identification of the potassium channels expressed in embryonicXenopus spinal neurons.
Voltage-dependent potassium channels are oligomeric proteins composed of pore-forming (Kvα) and auxiliary subunits. Kvα subunit genes belong to one of several related subfamilies (Kv1–Kv9). Kv1α subunits are expressed and contribute to potassium current maturation in the embryonic Xenopus spinal cord (Ribera and Nguyen, 1993; Jones and Ribera, 1994; Ribera, 1996). Auxiliary cytoplasmic Kvβ subunits tightly associate with Kv1α subunits (Rehm and Lazdunski, 1988; Parcej et al., 1992; Rettig et al., 1994; Scott et al., 1994; Rhodes et al., 1995, 1996; Nakahira et al., 1996; Sewing et al., 1996). The majority of studies of vertebrate Kvβ gene function have examined channels expressed in heterologous systems, and consequently their in vivo function is not yet defined. In heterologous systems, the effects of Kvβ subunits are diverse and range from increasing surface expression/current density (Shi et al., 1996; Accili et al., 1997a) to accelerating activation and inactivation kinetics (Rettig et al., 1994; England et al., 1995a,b;Majumder et al., 1995; Heinemann et al., 1996; Morales et al., 1996). In addition, Kvβ subunits share sequence and structural similarity with aldo-keto reductases, raising the possibility that they function as enzymes with as yet unspecified substrates (McCormack and McCormack, 1994; Gulbis et al., 1999).
Although little is known regarding the in vivo function of Kvβ subunits, less is known about their contribution to emerging properties of neuronal excitability [but see, Butler et al. (1998)]. We sought to determine whether embryonic spinal neurons express Kvβ subunits, and if so, whether their mRNAs are detectable in spinal neurons that express xKv1α subunits (Ribera, 1990; Ribera and Nguyen, 1993). Here we report that Xenopus embryonic spinal neurons express Kvβ subunit genes during the period of maturation ofI Kv. One of these (xKvβ2) appears to be the Xenopus ortholog of mammalian Kvβ2 on the basis of high amino acid identity (97%). The other gene shares less identity (<73%) with previously identified Kvβ genes. We propose that it is a new member of the Kvβ family, xKvβ4. Heterologous coexpression of xKvβs with xKv1α subunits leads to modulation of properties of potassium current that are developmentally regulated in the endogenousI Kv: current density, activation kinetics, and extent of inactivation. In situ hybridization indicates that xKvβ2 and xKvβ4 are expressed contemporaneously with previously studied Kv1α genes in subpopulations of spinal neurons. Taken together, the results implicate xKvβ subunits in regulation ofI Kv and emerging excitability of embryonic spinal neurons.
MATERIALS AND METHODS
Animals. Xenopus embryos were produced byin vitro fertilization and staged according to Nieuwkoop and Faber (1967).
Isolation of Xenopus Kvβ cDNAs, DNA sequencing, and analysis. 32P-labeled probes directed against the entire coding regions of rat brain Kvβ1 and Kvβ2 genes (kindly provided by Drs. Ken Nakahira and James Trimmer, Department of Biochemistry, SUNY Stony Brook, Stony Brook, NY) were used to screen aXenopus tadpole brain cDNA library (gift of Dr. Nicholas C. Spitzer, Department of Biology, University of California San Diego, La Jolla, CA) at reduced stringency. One full-length clone (xKvβ-A) and two partial-length clones (xKvβ-B1p, xKvβ-B2p) were obtained. The missing 5′ ends of the partial clones were obtained by further screening of cDNA libraries using either conventional methods or PCR. All isolated clones were sequenced over both strands. DNA sequences were read and entered into a computer using a GEL READER interface and software (CBS Scientific, Del Mar, CA) and analyzed using DNASTAR software (Madison, WI).
Whole-mount in situ hybridization. The nonradioactive whole-mount detection method (Harland, 1991) was used with minor modifications as described previously (Ferreiro et al., 1993; Burger and Ribera, 1996). cRNA sense and antisense probes corresponding to the 3′ untranslated region (UTR) and 3′ coding region of xKvβ-A, xKvβ-B1, and xKvβ-B2 were used. For xKvβ-A, the probe contained 700 bp of coding region and 150 bp of 3′ UTR; for both xKvβ-B1 and xKvβ-B2, the probes contained 500 bp of coding region and 700 bp of 3′ UTR. cRNA probes were synthesized in the presence of digoxigenin-labeled UTP (Boehringer Mannheim, Indianapolis, IN) and hybridized to whole-mount albino embryos. After removal of probe, embryos were incubated with alkaline phosphatase-conjugated anti-digoxigenin antibody (Boehringer Mannheim). The alkaline phosphatase reaction product was developed in the presence of chromogenic substrate (Boehringer Mannheim). Whole-mount embryos were either cleared in Murray's solution (2:1 benzyl benzoate/benzyl alcohol) for photography or embedded in plastic (JB-4 embedding kit; Polysciences, Warrington, PA). Photography of whole mounts and sections was performed with Kodak Ektachrome 160T film using appropriate color filters. Photographic images were digitized (Nikon Cool Scanner), and composite figures were constructed using Adobe Photoshop Software.
Oocyte recording. The entire coding regions of xKv1.1α and xKv1.2 α potassium channel genes were previously cloned into the pSP64T expression vector, and cRNA was synthesized as described (Ribera and Nguyen, 1993). An EcoRI–XhoI fragment containing the entire coding region of xKvβ4 along with 671 bp of 3′ UTR was cloned into the pCS2+ expression vector (Rupp et al., 1994;Turner and Weintraub, 1994); the xKvβ2-pCS2+ expression construct was constructed by insertion of a PCR-generated cDNA encoding xKvβ2 into pCS2+. Synthesis of capped cRNA was achieved by linearization of a xKvβ-pCS2+ expression vector with NotI followed byin vitro transcription with SP6 RNA polymerase (Promega, Madison, WI) in the presence of ribonucleotide triphosphates (Pharmacia Biotech, Piscataway, NJ) and cap analog (Boehringer Mannheim). RNA concentrations were determined spectrophotometrically and confirmed by electrophoresis in agarose-formaldehyde gels. Stage VI oocytes were removed and defolliculated as described previously (Ribera, 1990;Ribera and Nguyen, 1993; Burger and Ribera, 1996). xKvβ and xKv1α cRNAs were either injected alone or coinjected at ratios of 1–30:1 (xKvβ/ xKv1α). For two-electrode voltage-clamp (TEVC) experiments, oocytes were injected with 50 nl of RNA solution (Kv1 RNA, 2.5–5 μg/ml; Kvβ RNA, 75–150 μg/ml). Higher expression was required for macropatch experiments, and oocytes were injected with 100 nl of a more concentrated RNA solution (Kv1 RNA, 5–21 μg/ml; Kvβ RNA, 150–630 μg/ml). Oocytes were incubated at 18°C in Barth's solution containing the following (in mm): 88 NaCl, 1 KCl, 0.41 CaCl2, 0.33 Ca(NO3)2, 2.4 NaHCO3, 0.82 MgSO4 and 5 Na HEPES, pH 7.4.
TEVC methods were used [Axoclamp 2A amplifier (Axon Instruments, Foster City, CA) or Oocyte Clamp OC-725C (Warner Instruments, Hamden, CT)] to record voltage-activated outward currents induced in oocytes 24–48 hr after injection of RNA. For detailed analysis of activation kinetics, data were obtained 48–72 hr after injection using the cell-attached macropatch configuration and a patch-clamp amplifier (Axopatch 200B amplifier, Axon Instruments) to avoid the constraints imposed by the large capacitative transients of the whole oocyte. Data acquisition and analysis were accomplished with the pCLAMP suite of programs and AxoGraph software (Axon Instruments). For TEVC recordings, currents were sampled at 100–200 μsec and filtered at 5–10 kHz. For macropatch recordings, currents were sampled at 30–100 μsec and filtered at 5 kHz, and data from 10 runs were averaged to improve the signal-to-noise ratio. The leak and capacitative transient currents were subtracted using the P/4 protocol of the Clampex program (pCLAMP). For TEVC recordings, the bath consisted of Barth's solution and the electrode solution consisted of 3 m KCl and 10 mm HEPES, pH 7.4. Electrode resistances ranged between 0.1 and 0.7 MΩ. Oocytes were not used if the holding current exceeded −200 nA at a potential of −80 mV or if the resting membrane potential was positive to −40 mV. For macropatch recordings, the bath consisted of 115 mm KCl, 1.8 mm EGTA, and 10 mm HEPES, pH 7.4, and electrodes were filled with Barth's solution. Electrode resistances ranged between 1.5 and 2.5 MΩ.
Current amplitudes were measured 50–55 msec into a 60 msec pulse (steady-state level), except for data obtained on coexpression of xKvβ4. In the latter case, peak values were measured to avoid the effects produced by inactivation; induction of inactivation was analyzed separately (see Fig. 7). Conductance–voltage (G–V) relationships were determined for recordings from oocytes that did not saturate the amplifier at the most depolarized voltage step (+100 mV). Conductance was determined by dividing current amplitudes by the driving force assuming an internal K+ concentration of 100 mm and E K = −116 mV; recordings that did not achieve a plateau value of conductance (G max; determined from Boltzmann fits) were excluded. Normalized G–Vplots were obtained by dividing values of G by maximumG, G max, as determined from a fit to the Boltzmann equation. The Boltzmann parameters,V 1/2 and k, were obtained from fits of the Boltzmann equation to G–Vplots. Time to half-maximum (t 1/2) was determined as the time required to achieve half-maximum amplitude during a 60 msec voltage step. Current densities were determined by dividing current amplitudes by the oocyte surface area, assuming a 1 mm diameter (∼106μm2). Levels of statistical significance were assessed using the nonparametric Mann–Whitney test; pvalues <0.05 were indicative of statistical significance.
Identification of Xenopus Kvβ Potassium Channel Subunits
Given the high identity among previously identified Kvβ subunits, we began our search for Xenopus Kvβ clones by screening cDNA libraries with probes derived from rat (kindly provided by Drs. K. Nakahira and J. S. Trimmer, Department of Biochemistry, SUNY, Stony Brook, NY). One full-length (xKvβ-A) and two partial-length cDNA (xKvβ-B1p and xKvβ-B2p) clones were isolated. The missing 5′ ends of the partial clones were obtained through a combination of further library screening and PCR-based approaches (see Materials and Methods) to yield cDNA clones containing complete open reading frames (xKvβ-B1 and xKvβ-B2, respectively).
The xKvβ-A coding region contains 1101 base pairs (bp) and predicts a protein of 367 amino acids and a molecular weight (MW) of 40 kDa (Fig.1). Alignment of the predicted amino acid sequence of Xenopus Kvβ-A with mammalian Kvβ sequences reveals >97% identity with Kvβ2 protein (Table1). In contrast, only 76–77 and 62–67% exists with Kvβ1 and Kvβ3 subfamily members. On the basis of these comparisons, we consider Kvβ-A to be the Xenopus ortholog of Kvβ2 and refer to it as xKvβ2.
xKvβ-B1 and xKvβ-B2 both contain open reading frames that are 1203 bp in length. They share 93 and 94% overall identity at both the nucleotide and predicted amino acid levels over the coding region. Kvβ subunits belong to the superfamily of aldo-keto reductases; for this superfamily, Jez et al. (1997) proposed that ≥97% amino acid identity is indicative of allelic forms. Given their 94% amino acid identity and 94 and 89% nucleotide identity of coding regions and 3′ UTR, respectively, xKvβ-B1 and xKvβ-B2 sequences are likely to be allelic variants. Consistent with this, both genes have similar mRNA expression patterns (data not shown). Only results for the xKvβ-B1 allelic form are presented below.
The coding sequence of xKvβ-B1 predicts a protein containing 401 amino acids (44 kDa MW) (Fig. 1). Alignment of the predicted amino acid sequence of xKvβ-B1 with mammalian Kvβ subfamily members reveals 69–74, 70–73, and 70–74% identity with mammalian Kvβ1, Kvβ2, and Kvβ3 subfamily members (Table 1). In contrast, xKvβ2 is ≥97% identical to Kvβ2 genes previously cloned in other vertebrate species. Within a species (e.g., human), the different Kvβ genes (i.e., Kvβ1, Kvβ2, Kvβ3) share ∼70–75% identity (Table 1), which is the degree of identity found between xKvβ-B1 and previously identified Kvβ genes. On this basis, xKvβ-B1 appears to identify a new Kvβ subfamily, and we refer to xKvβ-B1 as xKvβ4.
Kvβ genes share sequence and structural similarities with the NAD(P)H-dependent oxidoreductase superfamily (McCormack and McCormack, 1994; Gulbis et al., 1999). The xKvβ proteins show the same extent of sequence identity (13–49%) with these enzymes as do mammalian Kvβ subunits. Although aspects of the xKvβ4 sequence are novel, motifs that define characteristic β-α-β barrel structures and critical active site residues are conserved (D119, Y124, and K154 of xKvβ4) (Fig. 1), indicating that xKvβ4 belongs to the β subunit cluster of aldo-keto reductases (Jez et al., 1997; Gulbis et al., 1999).
xKvβ2 and xKvβ4 transcripts are present in excitable tissues of the embryo
Whole-mount in situ hybridization was performed to determine the expression patterns of xKvβ2 and xKvβ4 genes inXenopus embryos. Embryos ranging in age between 1 and 2 d [Stage (St) 8–36] were examined, because maturation ofI Kv in spinal neurons occurs during this period. Action potentials are first detected in sensory spinal neurons at 20 hr (St 18) (Baccaglini and Spitzer, 1977). At this time, potassium currents are of small density (Barish, 1986; O'Dowd et al., 1988). During the subsequent day, there is a progressive increase inI Kv density and consequent shortening of the action-potential duration. Appearance ofI Kv in myocytes (somitic tissue) is delayed with respect to that in neurons but also undergoes functional upregulation once it is present (25 hr, St 23) (Ribera and Spitzer, 1990).
In the developing embryo, excitable tissues express xKvβ2 in a spatially and temporally dynamic pattern (Fig.2). xKvβ2 transcripts are first detected in somites (26 hr, St 24). Within 9 hr, (St 29, 35 hr), Kvβ2 is detected in the spinal cord as well. One hour later (St 30, 36 hr), xKvβ2 mRNA expression is downregulated in somites and upregulated in the spinal cord. By St 35–36 (50 hr), xKvβ2 expression is limited to the spinal cord and no longer present in the somites. In transverse sections of 2 d embryos (St 35), xKvβ2 mRNA expression is detected only in dorsal spinal neurons, including the mechanosensory Rohon-Beard cells.
xKvβ4 expression is limited to the nervous system in a spatially and temporally dynamic pattern. Kvβ4 transcripts are first detected in the brain and only the most anterior regions of the spinal cord (St 24, 26 hr). Nine hours later (St 29), Kvβ4 mRNA is detected in posterior regions of the spinal cord as well. During the subsequent day (Fig.2 I), the spatial pattern of expression remains constant, although the intensity of staining obtained is stronger, suggesting upregulation of gene expression during this period.
xKvβ2, but not xKvβ4, increases functional expression of xKv1α channels
Coexpression of mammalian Kvβ with Kv1 subunits leads to increases in current density and acceleration of activation and inactivation kinetics (Rettig et al., 1994; Majumder et al., 1995;McCormack et al., 1995; Heinemann et al., 1995, 1996; Morales et al., 1996; Shi et al., 1996; Uebele et al., 1996; Accili et al., 1997a,b). These properties of potassium current are of particular interest, because they are developmentally regulated in the endogenousI Kv. We coexpressed xKvβ with xKv1.1α and Kv1.2α subunits, because the developing spinal cord expresses the mRNAs for these α-subunits (Ribera, 1990; Ribera and Nguyen, 1993).
xKv1α and xKvβ mRNAs were coinjected into Xenopusoocytes at several different ratios ranging between 1 and 30:1 (xKvβ/ xKv1α). For ratios of 1–4:1, no effects are observed. At ratios of 15–30:1, similar effects are observed, which are statistically significant at a ratio of 30:1 (xKvβ/xKv1α). Thus, only data obtained using the 30:1 ratio are presented.
Coexpression of xKvβ2 with either xKv1.1α or xKv1.2α subunits increases current amplitude 1.6- and 2.0-fold, respectively (+20 mV) (Fig. 3 A). In contrast, coexpression of xKvβ4 with either xKv1α subunit has no effect on current amplitude. One possible explanation for the lack of effect of xKvβ4 coexpression is that this auxiliary subunit was not expressed at a sufficiently high level. However, xKvβ4 coexpression does have functional effects on kinetic properties (see below). On this basis, lack of expression of xKvβ4 subunits at the 30:1 ratio (xKvβ/xKv1α) is unlikely to account for the absence of effect on current density.
Current densities for channels expressed in oocytes ranged between 1 and 10 pA/μm2, which includes the range found in developing spinal neurons (0.5–1.5 pA/μm2 at +30 mV) (O'Dowd et al., 1988;Jones and Ribera, 1994). On coexpression of xKvβ2, current densities increased two- to threefold, just asI Kv density does in developing spinal neurons.
xKvβ1 and xKvβ2 modify functional properties of xKv1α channels
The increase in current densities observed on coexpression of xKvβ2 subunits may reflect changes in the conductance–voltage relationship of the current, the number of functional channels present in the plasma membrane, or both. To distinguish between these possibilities, we examined conductance–voltage relationships of currents carried by the various subunit combinations. We determined the maximum value of the conductance-(G max) and voltage-dependent parameters derived from fits to the Boltzmann equation (V 1/2 and k).
For Kv1.1α and Kv1.2α subunit-containing channels, coexpression of xKvβ2 subunits leads to 1.6- and 1.8-fold increases inG max (Fig.4 A,B; Table 2). These data indicate that an increase in the number of functional channels contributes to the increased current amplitudes obtained on xKvβ coexpression (Fig. 3). The increase in G max found on xKvβ2 coexpression with xKv1.1α subunits accounts for the increase in current density. For Kv1.2α,xKvβ2 coexpression has a slightly smaller effect on G max than it did on current amplitudes. If the single-channel conductance is unaltered, the result suggests that there are about twice as many functional channels in the plasma membrane on coexpression of xKvβ2 with either xKv1.1α or xKv1.2α subunits. In contrast, xKvβ4 coexpression does not alter either G max values (Fig.4 A,B; Table 2) or current amplitudes (Fig. 3).
Next, we examined the possibility that coexpression of either xKvβ2 or xKvβ4 affects voltage-dependent properties of steady-state activation. In fact, coexpression of xKβ4, but not xKvβ2, with xKv1.1α shifts the activation curve to more positive voltages (Fig.4 C); V 1/2 is increased by ∼5 mV (Table 2). Such a shift inV 1/2 on Kvβ coexpression would lead to a reduction in current amplitude in the range of −20 to +40 mV. However, current amplitudes are unaffected by xKvβ4 coexpression (Fig. 3).
Similar analyses were performed for Kv1.2α-containing channels. Coexpression of either xKvβ2 or xKvβ4 leads to small depolarizing shifts in the G–V relationship (Fig.4 D; Table 2). Such shifts would lead to a decrease in current amplitude in the range of +10 to +50 mV. However, current amplitude is unaffected by xKvβ4 coexpression and increased by xKvβ2 coexpression. On the basis of these data, the increased current amplitudes observed on xKvβ2 coexpression are most likely caused by an increase in channel number rather than changes in steady-state voltage-dependent properties of the expressed channels.
xKvβ subunits affect activation kinetics of Kv1 currents
The endogenous I Kv undergoes a developmentally regulated acceleration of activation kinetics (O'Dowd et al., 1988). Mammalian Kvβ subunits accelerate activation kinetics of Kv1 currents (Heinemann et al., 1995, 1996; Uebele et al., 1996). In whole-oocyte recordings, xKv1 currents activate more rapidly on coexpression of either xKvβ2 or xKvβ4 (Fig.5). The time to half-maximum current (t 1/2) decreases, indicating that activation of the current is accelerated. Larger effects are observed when xKvβ subunits are coexpressed with xKv1.2α versus xKv1.1α. The effects of xKvβ coexpression are voltage dependent and less pronounced with increasing depolarization. Overall, coexpression of an xKvβ subunit can lead to a 25–50% decrease int 1/2, which is reminiscent of the 50% decrease in t 1/2 observed for the endogenous I Kv during maturation of the action potential.
To avoid the large, long-lasting capacitative transients obtained in whole-oocyte recordings that compromise analysis of kinetics of current activation, we also examined kinetics of activation in recordings from oocyte macropatches (Fig. 6). The values of t 1/2 obtained from the macropatch data are smaller than those obtained for whole-oocyte recordings (Fig.5 vs Fig. 6), which most likely reflects the better temporal resolution of the macropatch configuration. Nonetheless, as found for whole-oocyte recordings, coexpression of xKvβ2 or xKvβ4 with xKv1.1α leads to a decrease in the activation t 1/2. xKvβ coexpression effects are voltage-dependent and greater at less depolarized potentials. At voltages positive to +60 and +70 mV (xKvβ2 and xKvβ4, respectively), the effects are no longer significant.
Coexpression of xKvβ subunits with xKv1.2 α subunits also accelerates current activation (Fig. 6 A,C).t 1/2 values are reduced by as much as a factor of 2. Effects are voltage dependent, being more pronounced at less depolarized membrane potentials. In sum, Kvβ coexpression accelerates the kinetics of activation, and the specific effects depend on the particular combination of α- and β-subunits.
xKvβ4 subunits increase and accelerate inactivation of xKv1α currents
The endogenous I Kv undergoes a developmentally regulated change in inactivation properties (Ribera and Spitzer, 1990). Because mammalian Kvβ1 and Kvβ3 subunits enhance inactivation of Kv1 currents (Rettig et al., 1994; Heinemann et al., 1995, 1996; Majumder et al., 1995; McCormack et al., 1995; Morales et al., 1996; Accili et al., 1997a; Leicher et al., 1998), we also determined whether xKvβ subunits are capable of modulating inactivation of channels containing either xKv1.1α or xKv1.2α subunits. Inactivation was evaluated by calculating the ratio of the steady-state versus peak-current amplitudes achieved during a pulse (I ss/I pk). This ratio has a value close to 1 for sustained currents and decreases when inactivation is apparent. Typically, xKv1.1α and xKv1.2α currents have I ss/I pkvalues of ≥0.98 (n = 74 and 29, respectively; +100 mV), as expected for sustained currents.
Consistent with results found for Kvβ2 of other species, coexpression of xKvβ2 has no effect on inactivation of either Kv1.1α or Kv1.2α currents (Fig. 7 A). However, coexpression of xKvβ4 with either xKv1.1α or xKv1.2α subunits induces inactivation (i.e.,I ss/I pk ≤ 0.9) in 79 and 100%, respectively, of oocytes. For those oocytes havingI ss/I pk values ≤0.90 (+100 mV), inactivation was examined over a more extended range of voltages (Fig. 7 B). xKvβ4 increased the extent of inactivation of both xKv1.1α and xKv1.2α currents in a voltage-dependent manner, with more inactivation observed at more depolarized voltages. These effects of xKvβ4 coexpression resemble those reported for mammalian Kvβ1 and Kvβ3 subunits (Rettig et al., 1994; England et al., 1995a,b; Heinemann et al., 1995, 1996; Majumder et al., 1995; Morales et al., 1996; Accili et al., 1997a).
Our interest in Kvβ subunits derives from the fact that they modulate properties of voltage-dependent potassium current that are developmentally regulated in the I Kvof Xenopus spinal neurons. Although it is well known thatI Kv undergoes changes in density and kinetics of activation and inactivation, the underlying molecular mechanisms have yet to be defined. We find that xKvβ genes are expressed in spinal neurons of Xenopus embryos during the period of action-potential maturation. Although one xKvβ gene (xKvβ2) is highly conserved with mammalian forms, the other (xKvβ4) is unique with respect to both its primary sequence and functional effects. Coexpression of xKvβ with xKv1α subunits modifies those properties of potassium current that are required for maturation of the impulse and excitability.
xKvβ2 shares >97% identity with mammalian Kvβ2. In contrast, xKvβ4 shares only 69–74% identity with mammalian Kvβ1, Kvβ2, and Kvβ3. The level of identity found between the different Kvβ families within a species (e.g., human Kvβ1 vs Kvβ3) is only 70–75%. On this basis, xKvβ4 appears to identify a new Kvβ family. This conclusion is further supported by xKvβ4's novel functional effects, as discussed below. Although we screened several embryonic cDNA libraries and did extensive RT-PCR using embryonic mRNA, we have isolated only xKvβ2 and xKvβ4. It is possible that at later stages of development Xenopus expresses genes that are orthologous to mammalian Kvβ1 or Kvβ3. Kvβ1 is not a mammalian-specific gene, because Rajeevan et al. (1999) recently isolated a chick Kvβ1.1 ortholog that is 95% identical to mammalian Kvβ1.1. Previous Kvβ cloning efforts have focused on adult stages. Thus, it is possible that xKvβ4 reveals a Kvβ subfamily that is preferentially expressed during embryonic development.
Consistent with xKvβ2's high primary sequence identity to mammalian forms, the functional effects of xKvβ2 coexpression are similar to those reported for mammalian Kvβ2. Mammalian Kvβ2 coexpression increases the amplitude of Kv1.2α and Kv1.4α currents (McCormack et al., 1995; Accili et al., 1997a) and surface expression of Kv1.1α, Kv1.2α, and Kv1.6α subunits (Shi et al., 1996); mammalian Kvβ2 also accelerates Kv1.5α activation kinetics (Heinemann et al., 1996;Uebele et al., 1996). Similarly, xKvβ2 increases current density andG max and also accelerates activation kinetics when coexpressed heterologously with xKv1.1α and xKv1.2α subunits.
The most obvious functional effect of xKvβ4 coexpression is induction of inactivation of Kv1.1α and Kv1.2α channels (Fig. 7). In this respect, xKvβ4 is similar to mammalian Kvβ1 and Kvβ3, which have been shown to increase inactivation of coexpressed Kv1α subunits (Rettig et al., 1994; Heinemann et al., 1995, 1996; Majumder et al., 1995; McCormack et al., 1995; Morales et al., 1996; Accili et al., 1997b; Leicher et al., 1998). Induction of inactivation by Kvβ1 subunits is thought to involve an N-terminal ball domain and critically depends on residues that are sensitive to oxidation (Rettig et al., 1994; Wang et al., 1996). In mammalian Kvβ3 as well as in the newly reported xKvβ4, similar residues (cysteine and serine preceding several positively charged amino acids) exist, indicating that induction of inactivation by Kvβ1, Kvβ3, and xKvβ4 may share a common mechanism of action. However, only Kvβ1.1, a Kvβ1 splice variant, has been shown to induce inactivation of both Kv1.1α and Kv1.2α channels as does xKvβ4 (Heinemann et al., 1996). However, coexpression of xKvβ4 also accelerates activation kinetics, an effect not observed on coexpression of either mammalian Kvβ1 or Kvβ3 with Kv1.1α or Kv1.2α channels (Heinemann et al., 1995, 1996; Leicher et al., 1998). Furthermore, xKvβ4 coexpression leads to depolarizing shifts in the activation curve (Fig. 4). In contrast, coexpression of Kvβ1 or Kvβ3 subunits typically leads to hyperpolarizing shifts, if any (Pongs et al., 1999). Taken together with xKvβ4's unique primary sequence, these results support the designation of xKvβ4 as a new Kvβ family member.
The two xKvβ genes have different expression patterns, suggesting that they operate in a nonoverlapping population of neurons. The dorsally restricted expression of xKvβ2 within the spinal cord (e.g., Rohon-Beard cells) is reminiscent of the expression pattern found previously for xKv1.1 (Ribera and Nguyen, 1993), suggesting that xKv1.1α and xKvβ2 subunits form Kv complexesin vivo. The earliest that xKv1.1α mRNA is detected in Rohon-Beard cells is St 26, several hours after xKvβ2 mRNA is detected. xKvβ2 is transiently detected in somitic tissue between St 24 and 30 (26–35 hr). The earliest thatI Kv can be consistently recorded from developing myocytes is St 24 (26 hr). Thus, just as in neurons, Kvβ2 is expressed as soon as would be required to contribute to formation of the first functional potassium channels.
xKvβ4's expression pattern differs both spatially and temporally from that of xKvβ2. Spatially, xKvβ4's expression domain is limited to the nervous system but extends more ventrally within the developing spinal cord. Temporally, xKvβ4 has a pronounced anterior–posterior gradient in its expression pattern, which is less obvious for xKvβ2. The expression pattern of xKvβ4 is similar to that of a recently identified xKv1α gene, xKv1.3α (A. Hoffman and A. Ribera, unpublished data), suggesting that xKv1.3α and xKvβ4 form Kv complexes in vivo. However, coexpression of subunit mRNAs or proteins is not a reliable indicator of the final channel stoichiometry, because examination of mammalian brain Kv channel complexes indicates that only a subset of possible channel combinations actually exists (Rhodes et al., 1997; Shamotienko et al., 1997).
In Xenopus spinal neurons, developmental regulation ofI Kv is synchronized, despite the heterogeneity in neuronal identities as well as patterns of potassium channel gene expression (Ribera, 1996). Accordingly, each neuronal subpopulation must upregulate I Kv via molecular mechanisms that are specific to the potassium channel isoforms expressed by that particular neuronal subtype. However, the different neuronal populations must also temporally coordinate their developmental increases in I Kvdensity. The results presented here indicate that in those neuronal subpopulations using xKv1α channels, xKvβ mRNAs are present and their coexpression is able to induce relevant developmental changes inI Kv. Because xKvβ genes are expressed contemporaneously with Kv1α subunits, their potential contribution to functional upregulation ofI Kv would be regulated at a level subsequent to transcription. In this regard, the sequence and structural similarities between Kvβ subunits and aldo-keto reductase enzymes raise the possibility that developmental changes in either cellular redox potentials or concentration of substrate drive Kvβ modulation of potassium channel function (McCormack and McCormack, 1994; Gulbis et al., 1999). It is possible to overexpress wild-type and mutant forms of ion channel subunits in the Xenopus embryo (Jones and Ribera, 1994; Ribera 1996) and thus test experimentally how Kvβ subunits function in vivo.
This work was supported by National Institutes of Health (NIH) Grant F32 NS10250 and American Heart Association Grant CWFW-07-98 (M.A.L.), and NIH Grants T32 NS07083 and RO1-NS25217 (A.B.R.). We thank Dr. Kurt Svoboda for generous assistance in preparation of Figure 2, and Drs. Ken Nakahira and Jim Trimmer for providing the rat Kvβ1 and Kvβ2 cDNAs.
Correspondence should be addressed to Angeles B. Ribera, Department of Physiology and Biophysics, C-240, University of Colorado Health Sciences Center, Denver, CO 80262. E-mail:.