Polyamine Flux in Xenopus Oocytes Through Hemi-Gap Junctional Channels


  • D. Enkvetchakul,

    1. Division of Renal Medicine, Finch University of Health Sciences/The Chicago Medical School, North Chicago, IL 60064, USA
    Search for more papers by this author
  • L. Ebihara,

    1. Department of Physiology and Biophysics, Finch University of Health Sciences/The Chicago Medical School, North Chicago, IL 60064, USA
    Search for more papers by this author
  • C. G. Nichols

    Corresponding author
    1. Department of Cell Biology and Physiology and Washington University School of Medicine, 660 South Euclid Avenue, St Louis, MO 63110, USA
    • Corresponding author
      C. G. Nichols: Department of Cell Biology and Physiology and Washington University School of Medicine, 660 South Euclid Avenue, St Louis, MO 63110, USA. Email: cnichols@cellbio.wustl.edu

    Search for more papers by this author


Diverse polyamine transport systems have been described in different cells, but the molecular entities that mediate polyamine influx and efflux remain incompletely defined. We have previously demonstrated that spermidine efflux from oocytes is a simple electrodiffusive process, inhibitable by external Ca2+, consistent with permeation through a membrane cation channel. Hemi-gap junctional channels in Xenopus oocytes are formed from connexin 38 (Cx38), and produce a calcium-sensitive (Ic) current that is inhibited by external Ca2+. Spermidine efflux is also calcium sensitive, and removal of external calcium increases both Ic currents and spermidine efflux in Xenopus oocytes. Injection of Cx38 cRNA or Cx38 antisense oligonucleotides (to increase or decrease, respectively, Cx38 expression) also increases or decreases spermidine efflux in parallel. Spermidine efflux has a large voltage-dependent component, which is abolished with injection of Cx38 antisense oligonucleotides. In addition, spermidine uptake is significantly increased in Cx38 cRNA-injected oocytes in the absence of external calcium. The data indicate that hemi-gap junctional channels provide the Ca2+-inhibited pathway for electrodiffusive efflux of polyamines from oocytes, and it is likely that hemi-gap junctional channels provide Ca2+ and metabolism-sensitive polyamine permeation pathways in other cells.

Polyamines play a critical role in cell growth, proliferation, and differentiation. The intracellular polyamine pool is highly regulated, and influx and efflux of polyamines are thought to be controlled as part of the cell growth, with increased influx and decreased efflux during periods of rapid cell growth (Wallace & Keir, 1981, 1986). Polyamines also physiologically regulate various plasma membrane channels (Lopatin et al. 1994; Musa & Veenstra, 2003) (Kir, NMDA, connexin (Cx)) by functioning as pore blockers, and thereby play an important role in shaping membrane potential responses and in cell to cell signalling in the case of connexin. In Xenopus oocytes, efflux of polyamines in part follows a simple electrodiffusive mechanism (Sha et al. 1996) through an undefined endogenous pathway. This efflux of polyamines can result in significant accumulation of polyamines in the bathing solution, as shown by the ability of oocyte-conditioned solutions to induce strong inward rectification of inward rectifier K+ channels in excised membrane patches (Lopatin et al. 1994).

In this study, we have examined the possibility that connexin 38 (Cx38) hemi-gap junctions may mediate efflux and uptake of spermidine in Xenopus oocytes. Gap junctions are intercellular membrane channels formed from two hexameric complexes of connexin subunits, known as hemi-channels or connexons, which reside in the plasma membranes of closely opposed cells. Several studies suggest that unpaired hemi-channels are also present in single plasma membranes. These channels are normally closed by the blocking effects of external divalent cations and voltage. However, following removal of external calcium, they open, forming large, nonselective pores that could conceivably allow permeation of polyamines.


Overexpression and depletion of connexin 38 in Xenopus oocytes

Female Xenopus laevis were maintained in tanks of fresh water at 20 °C. For oocyte isolation, animals were anaesthetized with MS222 (2 g l−1, added to the water). The anaesthetized animal was laid on ice, one ovary was removed via a mini-laparotomy, the incision sutured and the animal allowed to recover at room temperature. Following the final collection of oocytes, the animals were humanely killed by MS222 overdose. Capped mRNA was prepared from linearized Cx38 cDNA (Ebihara, 1996) using Message Machine kits (Ambion Inc., Austin, TX, USA). Xenopus oocytes were isolated using conventional techniques, and were pressure injected with 50-100 nl of Cx38 mRNA, or Cx38 antisense oligonucleotide as indicated. Oocytes were maintained at room temperature in ND96 solution (see below) supplemented with 1 mm MgCl2, penicillin (100 u ml−1) and streptomycin (100 µg ml−1). Experiments were performed 1-4 days after injection.


Oocytes were voltage clamped at room temperature using a commercial voltage-clamp amplifier (Warner Instruments, Inc.) in a small chamber (volume 200 µl) mounted on the stage of a SMZ-1 microscope (Nikon Instruments). Electrodes were filled with 3 m KCl and had tip resistances of 1-5 MΩ. Pclamp software and a Labmaster TL125 D/A convertor (Axon Instruments Inc.) were used to generate voltage pulses. Data were normally filtered at 1 kHz, and signals were recorded onto videotape by digitizing at 22 kHz (Neurocorder, Neurodata, NY, USA), or directly recorded onto a computer through Pclamp6 and Axotape software (Axon Instruments).


ND96 solution contained (mm): NaCl 96, KCl 2, CaCl2 2, Hepes 5, pH 7.5. KD98 solution contained (mm): KCl 98, Hepes 5, pH 7.5. For solutions in which the [K+] was varied, a total of 98 mm KCl + NaCl was used, with 5 mm Hepes, pH 7.5. CaCl2 (2 mm) was added as indicated.

[3H]Spermidine efflux assays

Oocytes were injected with antisense oligonucleotides or with capped mRNA as indicated; uninjected oocytes are indicated as controls. Two hours before the assay, oocytes were further injected with ≈50 nl of [3H]spermidine (specific activity = 15 Ci mmol−1) in 10 mm unlabelled spermidine. Oocytes were washed twice in ≈0.3 ml experimental solution, and then incubated for the specified duration in 0.3 ml experimental solution, after which the solution was removed and kept for counting. Oocytes were then washed twice with 0.3 ml of the next experimental solution, the wash solution saved as part of the count for the previous period. Washing and incubation was repeated for each solution change as required. After the last experimental solution, oocytes were washed twice (final solution pooled with the last period count), then lysed in 0.5 ml of 1 % sodium dodecyl sulfate and counted for spermidine activity remaining in the oocyte. Radioactivity was measured in a scintillation spectrometer.

[3H]Spermidine uptake assays

Oocytes were injected with antisense oligonucletides or with capped mRNA as indicated, two days prior to the assay. Oocytes were washed twice with ≈1 ml of KD98 solution with or without calcium as indicated, and incubated in sets of three oocytes in 0.1 ml of 1.4 mm[3H]spermidine in KD98 solution, with or without calcium as indicated. After 1 h, the bathing solution was removed, oocytes were washed twice with the experimental solution, and the wash solution was pooled with the bathing solution and counted. The oocytes were then lysed in 1 ml of 1 % SDS solution, and counted for spermidine uptake.


A calcium-sensitive current can be induced or reduced by connexin overexpression or suppression

Figure 1 compares the effect of removing external calcium on membrane currents recorded from control oocytes, and oocytes pretreated with Cx38 antisense oligomers. In these experiments, the oocytes were held at -20 mV and membrane conductance was monitored by applying 0.5 s voltage clamp steps to -50 and +50 mV every 5 s. Figure 1A shows representative currents recorded from an uninjected, control oocyte. In the presence of 2 mm[Ca2+]o, the membrane conductance is small. However, following removal of external calcium, there is a large, reversible, increase in conductance. Previous studies have implicated Cx38 hemi-gap junctional channels (Ebihara, 1996; Zhang et al. 1998) in providing the underlying permeation pathway for this calcium-sensitive current, and as shown in Fig. 1B, depleting Cx38 protein by injecting oocytes with a Cx38 antisense oligonucleotide essentially abolishes the current. Figure 2 shows mean current-voltage relationships from uninjected (A), Cx38 antisense oligonucleotide- (B) and Cx38 cRNA- (C), injected oocytes in the presence of 2 mm Ca2+, and following Ca2+ removal. In uninjected, control oocytes, the membrane conductance increased ≈10-fold (from 38 ± 3 to 445 ± 9 µS, n= 6) (measured as chord conductance between -10mv and +10 mV). The increase in membrane conductance was less than 2-fold (from 52 ± 11 to 71 ± 14 µS, n= 6) in Cx38 antisense-injected oocytes. Conversely, in Cx38 cRNA-injected oocytes, there was ≈30-fold (from 47 ± 13 to 1500 ± 78 µS, n= 5) increase in membrane conductance when external Ca2+ was removed. In all cases, the increase in membrane conductance in the absence of calcium was reversible upon reintroduction of Ca2+.

Figure 1.

Xenopus oocytes express a large calcium-sensitive current which is dependent on connexin 38 expression

Representative recordings of current from A, control, and B, Cx38 antisense oligonucleotide-injected, Xenopus oocytes; 0.5 s voltage clamp steps to -50 and +50 mV were applied every 5 s from a holding potential of -20 mV. Bath solutions (see Methods) were changed as shown. Control (A) oocytes demonstrate a large induction of current upon removal of external calcium; this current is reversibly inhibited when Ca2+ is re-introduced. Oocytes injected with antisense oligonucleotide (B) demonstrate a markedly reduced calcium-sensitive current.

Figure 2.

The calcium-sensitive current in Xenopus oocytes is increased or decreased following injection of Cx38 cRNA or Cx38 antisense oligonucleotides, respectively

Oocytes were voltage clamped at a holding potential of -20 mV and stepped from -80 mV to +80 mV in 10 mV increments in KD98 bath solution, in the presence and absence of calcium. Instantaneous current-voltage relationships are plotted for control (A), Cx38 antisense RNA-injected (B) or Cx38 sense oligonucleotide-injected oocytes (C, mean ±s.e.m., n= 3-6 oocytes).

Spermidine efflux is altered by the same manoeuvres that modify connexin 38 calcium-inhibited current

We previously (Sha et al. 1996) demonstrated that unidirectional efflux of [3H]spermidine from Xenopus oocytes is strongly dependent on membrane potential and extracellular [Ca2+], leading to the conclusion that efflux occurs by electrodiffusion through an unknown ion channel pathway. In high [K+]o (i.e. depolarizing solution), oocytes demonstrate a significantly higher spermidine efflux in the absence, than the presence, of external Ca2+, and this is reversible upon reintroduction of Ca2+ (Fig. 3). In control oocytes, fractional spermidine efflux reversibly increased ≈15-fold following removal of external calcium. As with the Ca2+-sensitive conductance, injection of oocytes with Cx38 antisense oligonucleotides virtually abolished the Ca2+-sensitive spermidine efflux, whereas the Ca2+-sensitive spermidine efflux was enhanced in Cx38 cRNA-injected oocytes (Fig. 3). In all oocytes, the increase in spermidine efflux was reversible upon reintroduction of external calcium into the bath solution.

Figure 3.

Spermidine efflux is calcium sensitive and depends on Cx38 expression

Fractional spermidine efflux measured from control oocytes, or from oocytes injected, 2 days prior to assay, with Cx38 cRNA or Cx38 antisense oligonucleotides (see Methods). Oocytes were injected with ≈50 nl of 70 mm[3H]spermidine (specific activity = 15 Ci mmol−1) in 10 mm unlabelled spermidine, and incubated in KD98 solution, with or without Ca2+ for 1 h. Oocytes demonstrate a reversible increase in spermidine efflux in zero [Ca2+], which is diminished in Cx38 antisense-injected oocytes, and augmented in Cx38 sense-injected oocytes (mean +s.e.m., n= 7-10 oocytes).

Influx of spermidine is increased with Cx38 overexpression

The above experiments implicate endogenous Cx38 hemi-channels in providing the major pathway for spermidine efflux from oocytes (see Discussion). We further examined the relevance of this pathway for uptake of spermidine. Fractional uptake of spermidine from the bath solution into oocytes was measured by incubating oocytes in KD98 with tritiated spermidine. In the absence of calcium, spermidine influx was also significantly increased in oocytes injected with Cx38, and decreased in antisense- and sense-injected oocytes (Fig. 4).

Figure 4.

Spermidine uptake is sensitive to fractional uptake of [3H]spermidine is enhanced with Cx38 overexpression

[3H]spermidine uptake was measured in KD98 (see Methods) for control, Cx38 antisense-, and Cx38 sense cmRNA-injected oocytes, in KD98 solution with or withour Ca2+ (mean +s.e.m., n= 4-6 oocytes). A significant increase in spermidine uptake was seen in oocytes injected with Cx sense RNA, in zero [Ca2+]. Significant differences are indicated (*P < 0.05, **P < 0.01, unpaired t test).

External [K+] and voltage dependence of spermidine efflux

The above data indicate that spermidine efflux occurs through gap-junctional hemi-channels. Consistent with this conclusion, we previously showed that efflux is strongly dependent on membrane potential, as expected for an electrodiffusive process. In oocytes overexpressing inwardly rectifying K+ channels (Kir1.1 or Kir2.1), the membrane potential is essentially clamped at EK (Sha et al. 1996). Figure 5 shows efflux of tritiated spermidine from Kir1.1-expressing oocytes as a function of external [K+], with calculated EK indicated. In the absence of Ca2+, spermidine efflux was clearly dependent on external [K+]. Assuming the membrane potential follows EK, spermidine efflux is voltage dependent, with greater efflux occurring at depolarized potentials, consistent with positively charged spermidine (+3 charge under physiological conditions) (Seiler et al. 1996) as the main effluxing species. In oocytes injected with Cx38 antisense oligonucleotides (bottom panel), voltage-dependent efflux of spermidine is virtually abolished.

Figure 5.

Spermidine efflux through Cx38 is electrodiffusive

Oocytes were injected with Kir1.1 or Kir2.1 cRNA to induce a large K+ conductance, and Cx38 antisense oligonucleotide as indicated. Fractional efflux of [3H]spermidine was measured in bathing solutions containing [K+]+[Na+]= 98 mm, with [K+] indicated. The 150 mm K+ solution was made without Na+. Spermidine efflux was minimal at all external [K+] when bath solution contained Ca2+ (top panel). Spermidine efflux increased in zero [Ca2+] as a function of external [K+] (middle panel), consistent with a dependence on membrane potential (clamped at approximate voltages indicated due to high expressed K+ channel conductance (Sha et al. 1996). Spermidine efflux is reduced to basal in oocytes injected with antisense Cx38 oligonucleotide, and the dependence on external [K+] is abolished. Estimated membrane potentials are indicated in parentheses, and were calculated as the reversal potential of potassium, EK=RT/F× log([K+]in/[K+]out), where T= 298°K, [K+]in= 98 mm, and R and F have their usual values. ND, not done.


Membrane transport of polyamines

Polyamine transport into and out of cells has been studied for many years. There are clearly several processes involved in the uptake and release of polyamines in different tissues and organisms (Seiler et al. 1996), and specific transporter proteins have been identified in lower organisms (Kashiwagi et al. 1986; Kashiwagi et al. 1996; Sugiyama et al. 1996; Igarashi & Kashiwagi, 1999). However, molecular mechanisms remain poorly understood in higher organisms. Cellular uptake has received the most attention, and it is clear that there are high affinity uptake systems with saturation at low micromolar levels of polyamines in oocytes and other animal tissues (Kano & Oka, 1976; Saunders et al. 1989; Khan et al. 1990; Gilad & Gilad, 1991). There is some evidence for ‘trans-acceleration’ of polyamine flux, whereby raising the transmembrane polyamine concentration paradoxically increases flux from the cis-side of the membrane, indicative of futile cycling of an exchange process (Byers & Pegg, 1990; Mackarel & Wallace, 1994). One reasonably consistent finding is that depolarization, or manoeuvres likely to induce depolarization (e.g. increased extracellular [K+], stimulation of depolarizing ion currents, metabolic poisoning), tend to stimulate efflux, or to inhibit influx (Fage et al. 1992, 1993; Nicolas et al. 1994; Poulin et al. 1995). There is evidence for ion channel blockers inhibiting polyamine efflux from some cells (Tjandrawinata et al. 1994), and we previously demonstrated that spermidine efflux from Xenopus oocytes occurs by simple electrodiffusion (Sha et al. 1996), consistent with flux through an ion channel. In the absence of external divalent cations, spermidine efflux from oocytes is strongly dependent on the membrane potential (Em), and under voltage clamp, the effect of changing external [K+] is entirely a consequence of the change in Em, and not of changes in external [Na+], or [K+]per se. These results provide direct evidence against a Na+ or K+ cotransport or countertransport mechanism being involved in spermidine efflux under these conditions (Sha et al. 1996). The specific inhibition of this efflux by trans-Ba2+ or Ca2+ ions further argues that efflux occurs through a protein pore, rather than because of the appearance of ‘holes’ in the membrane itself.

Cx38 provides efflux and influx pathways for spermidine

Our previous study gave no indication of the channels that constitute the pathway for spermidine efflux. Polyamines act as pore blockers in several ion channels (Nichols & Lopatin, 1997; Williams, 1997; Kerschbaum et al. 2003), and cause physiological rectification in inward rectifier K+ channels (Bianchi et al. 1996; Nichols et al. 1996; Shyng et al. 1996). Various endogenous ion conductances have been described in Xenopus oocytes and we have previously shown that polyamines can efflux through a Na+/K+-ATPase γ-subunit-activated pathway (Sha et al. 2001), although this pathway is not constitutively activated. A weakly outward rectifying, Ca2+-inhibitable non-selective current is prominent in oocytes (Tupper & Maloff, 1973; Arellano et al. 1995), and importantly, this conductance is significantly reduced by injection with antisense oligonucleotides against the major endogenous connexin (Cx38) (Ebihara, 1996; Zhang et al. 1998). Connexins form wide pores permeable to large organic cations and anions including tetrapentylammonium, and polyamines are known to act as pore blockers in Cx40 gap junctional channels (Musa & Veenstra, 2003), indicating that they gain ready access to the pore. Depending on the detailed pore architecture, polyamines may be permeable through gap junctional channels composed of other connexins. In the present study, we considered the possibility that Cx38 gap junction hemi-channels may be the relevant pathway for constitutive electrodiffusive polyamine efflux from the oocyte. We have confirmed the major role of Cx38 in generating the Ca2+-inhibitable current, both by antisense depletion (Fig. 2B) and by enhancement of the current by Cx38 overexpression (Fig. 2C). Importantly, spermidine efflux is increased or decreased in direct parallel to the effects of these manipulations on membrane conductance. This unequivocally identifies Cx38 hemi-gap junctional channels as the pathway for the Ca2+-inhibitable spermidine flux.

In many cells, uptake of polyamines is dependent on external [Na+] (Seiler et al. 1996). This may also be consistent with electrodiffusion, since replacement of Na+ with K+ will generally depolarize cells and hence reduce electrodiffusive uptake. Here we have shown that hemi-channels underlie voltage-dependent efflux in oocytes. Similarly spermidine influx was increased by injection of Cx38 mRNA, and was again inhibited by external calcium. In antisense-injected oocytes, no increase in influx was seen in the absence of external calcium, consistent with a significantly diminished expression of Cx38. However, the lack of increased spermidine influx in the absence of external Ca2+, in control oocytes suggests that influx of spermidine may also involve alternative significant pathways (Seiler et al. 1996).

Relevance of gap junctional hemi-channel-dependent polyamine flux in other cells

The Cx38 hemi-channel represents a major flux pathway by which spermidine exits the oocyte. Hemi-gap junctional channels are not an artifact of oocyte isolation, since they are present in non-defolliculated oocytes (Li et al. 1996; Zhang et al. 1998). In many other cell types, hemi-gap junctional channels may also be present as a reservoir from which functional gap junctions can be rapidly formed (Li et al. 1996). In freshly isolated cardiac myocytes, removal of extracellular calcium, or metabolic inhibition, leads to a large increase in a non-selective membrane conductance that has been attributed to the opening of Cx43 hemi-channels (John et al. 1999; Kondo et al. 2000). As with oocytes, solutions conditioned by isolated cardiac myocytes and other cells are capable of restoring strong inward rectification to K+ channels in isolated membrane patches (Lopatin et al. 1994), again indicating that significant polyamine efflux can occur. Based on the above results, we suggest that, in addition to providing a non-selective conductance which will affect electrical activity directly, opening of hemi-gap junctional channels may play a significant role in controlling cellular function in a range of tissues, by controlling polyamine homeostasis, and by affecting cellular and membrane properties that depend on intracellular polyamine levels.


This work was supported by NIH Grants HL54171 (to C.G.N.), DK60086 (to D.E.) and EY10589 (to L.E.).