Advertisement

hSK4/hIK1, a Calmodulin-binding KCa Channel in Human T Lymphocytes

ROLES IN PROLIFERATION AND VOLUME REGULATION*
Open AccessPublished:May 21, 1999DOI:https://doi.org/10.1074/jbc.274.21.14838
      Human T lymphocytes express a Ca2+-activated K+ current (IK), whose roles and regulation are poorly understood. We amplified hSK4 cDNA from human T lymphoblasts, and we showed that its biophysical and pharmacological properties when stably expressed in Chinese hamster ovary cells were essentially identical to the native IK current. In activated lymphoblasts, hSK4 mRNA increased 14.6-fold (Kv1.3 mRNA increased 1.3-fold), with functional consequences. Proliferation was inhibited when Kv1.3 and IK were blocked in naive T cells, but IK block alone inhibited re-stimulated lymphoblasts. IK and Kv1.3 were involved in volume regulation, but IK was more important, particularly in lymphoblasts. hSK4 lacks known Ca2+-binding sites; however, we mapped a Ca2+-dependent calmodulin (CaM)-binding site to the proximal C terminus (Ct1) of hSK4. Full-length hSK4 produced a highly negative membrane potential (V m) in Chinese hamster ovary cells, whereas the channels did not function when either Ct1 or the distal C terminus was deleted (V m ∼0 mV). Native IK (but not expressed hSK4) current was inhibited by CaM and CaM kinase antagonists at physiological V m values, suggesting modulation by an accessory molecule in native cells. Our results provide evidence for increased roles for IK/hSK4 in activated T cell functions; thus hSK4 may be a promising therapeutic target for disorders involving the secondary immune response.
      Both voltage-gated and Ca2+-activated K+(KCa)
      The abbreviations used are: KCa, Ca2+-activated K+; AgTx-2, agitoxin-2; bp, base pairs; CaM, calmodulin; CHO, Chinese hamster ovary; ChTx, charybdotoxin; HEK, human embryonic kidney; IK, intermediate-conductance KCa channel in T lymphocytes; kb, kilobase (pairs); MgTx, margatoxin; PCR, polymerase chain reaction; RVD, regulatory volume decrease; SK, small-conductance KCachannels; TFP, trifluoperazine; Ca2+i, free Ca2+ concentration; BAPTA, 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid; PHA, phytohemagglutinin; pS, picosiemens.
      1The abbreviations used are: KCa, Ca2+-activated K+; AgTx-2, agitoxin-2; bp, base pairs; CaM, calmodulin; CHO, Chinese hamster ovary; ChTx, charybdotoxin; HEK, human embryonic kidney; IK, intermediate-conductance KCa channel in T lymphocytes; kb, kilobase (pairs); MgTx, margatoxin; PCR, polymerase chain reaction; RVD, regulatory volume decrease; SK, small-conductance KCachannels; TFP, trifluoperazine; Ca2+i, free Ca2+ concentration; BAPTA, 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid; PHA, phytohemagglutinin; pS, picosiemens.
      channels are widely expressed in immune cells, including human T lymphocytes. Drugs that block voltage-gated Kv1.3 channels inhibit T lymphocyte activation and proliferation, volume regulation, and cell-mediated cytotoxicity (for review, see Ref.
      • Lewis R.S.
      • Cahalan M.D.
      ). Inasmuch as these functions involve Ca2+ influx through channels activated by depletion of Ca2+ stores, one widely proposed role for K+channels is to maintain a negative membrane potential and large driving force for Ca2+ entry. However, the relative roles of KCa versus Kv1.3 channels in these cell functions are not known, partly owing to the previous lack of potent KCa blockers that do not also block Kv1.3 channels.
      Two KCa channels have been found in lymphocytes and lymphocytic cell lines. They differ in biophysical and pharmacological properties (
      • Mahaut-Smith M.
      • Schlichter L.C.
      ,
      • Mahaut-Smith M.
      • Schlichter L.C.
      ,
      • Grissmer S.
      • Lewis R.S.
      • Cahalan M.D.
      ,
      • Grissmer S.
      • Nguyen A.N.
      • Cahalan M.D.
      ,
      • Schlichter L.C.
      • Pahapill P.
      • Schumacher P.A.
      ). An apamin-sensitive, small conductance channel (7–8 pS) is the prevalent KCa channel in the commonly used Jurkat T cell line (
      • Grissmer S.
      • Lewis R.S.
      • Cahalan M.D.
      ) and is also present in rat T and human B lymphocytes (
      • Mahaut-Smith M.
      • Schlichter L.C.
      ,
      • Schlichter L.C.
      • Pahapill P.
      • Schumacher P.A.
      ). However, a corresponding apamin-sensitive whole-cell current has not been identified in normal human T cells, perhaps a result of channel rundown we observed after cell disruption (
      • Mahaut-Smith M.
      • Schlichter L.C.
      ). Instead, a KCa channel we first described (
      • Mahaut-Smith M.
      • Schlichter L.C.
      ,
      • Mahaut-Smith M.
      • Schlichter L.C.
      ) is the prevalent KCa channel in resting and activated human T lymphocytes. It is a charybdotoxin-sensitive, inwardly rectifying channel (15–35 pS in symmetrical K+ solutions (
      • Grissmer S.
      • Nguyen A.N.
      • Cahalan M.D.
      ,
      • Schlichter L.C.
      • Pahapill P.
      • Schumacher P.A.
      )) that is commonly called “IK,” for intermediate conductance KCa. Recently, a molecular candidate for IK was cloned from a human placental cDNA library (hSK4 (
      • Joiner W.J.
      • Wang L.-Y.
      • Tang M.D.
      • Kaczmarek L.K.
      )) and subsequently from human pancreas (hIK1 (
      • Ishii T.M.
      • Silvia C.
      • Hirschberg B.
      • Bond C.T.
      • Adelman J.P.
      • Maylie J.
      )) and a human lymph node library (hKCa4 (
      • Logsdon N.J.
      • Kang J.
      • Togo J.A.
      • Christian E.P.
      • Aiyar J.
      )).
      IK current increases in the 3–4 days following activation of human T cells (
      • Grissmer S.
      • Nguyen A.N.
      • Cahalan M.D.
      ). Thus, it is anticipated that this KCa current will be especially important for secondary immune responses of activated T cells (lymphoblasts), including proliferation and volume regulation. The regulatory volume decrease (RVD) that follows T cell swelling is known to depend on K+ (and Cl) channels (
      • Grinstein S.
      • Foskett J.K.
      ,
      • Sarkadi B.
      • Parker J.C.
      ,
      • Lang F.
      • Busch G.L.
      • Volkl H.
      ). Although Kv1.3 is involved in RVD in some resting T cells (
      • Deutsch C.
      • Chen L.-Q.
      ), the relative contribution of IK versus Kv1.3 channels is not known, either in resting human T cells or in lymphoblasts. We previously reported that intracellular Ca2+ rises immediately after human T cells are exposed to a hypotonic shock (
      • Schlichter L.C.
      • Sakellaropoulos G.
      ); thus, we predicted that KCa currents would also subserve RVD.
      In the present study we cloned hSK4 from human T lymphoblasts, expressed the channels stably in CHO cells, and compared the salient biophysical and pharmacological properties of the native and cloned channels. All intrinsic properties examined were indistinguishable, supporting the view that hSK4 homotetramer forms the α subunit of the IK channel of lymphoblasts. We found that hSK4 mRNA expression is strongly up-regulated after T cell activation; thus we predicted (and observed) an increased role for IK current in lymphoblasts compared with resting T cells. Although hSK4 is functionally a KCachannel, that is activated by a rise in intracellular Ca2+, it was recently reported that brain SK channels are not gated directly by Ca2+ but rather by Ca2+ interacting with calmodulin that is irreversibly bound to the channel (
      • Xia X.-M.
      • Fakler B.
      • Rivard A.
      • Wayman G.
      • Johnson-Pais T.
      • Keen J.E.
      • Ishii T.
      • Hirschberg B.
      • Bond C.T.
      • Lutsenko S.
      • Maylie J.
      • Adelman J.P.
      ). We have presented preliminary data showing that the lymphoblast IK current is inhibited by antagonists of calmodulin and CaM kinase (
      • Chang M.C.
      • Schlichter L.C.
      ,
      • Khanna R.
      • Chang M.C.
      • Joiner W.
      • Kaczmarek L.K.
      • Schlichter L.C.
      ). We now show details of this inhibition of native IK current and that calmodulin binds directly to the hSK4 channel protein in a Ca2+-dependent manner. The CaM binding domain resides in the proximal part of the C terminus, since binding to this region occurs in the absence of flanking sequence and is eliminated in constructs lacking this region. Unlike the study of heterologously expressed brain SK channels (
      • Xia X.-M.
      • Fakler B.
      • Rivard A.
      • Wayman G.
      • Johnson-Pais T.
      • Keen J.E.
      • Ishii T.
      • Hirschberg B.
      • Bond C.T.
      • Lutsenko S.
      • Maylie J.
      • Adelman J.P.
      ), we provide evidence for additional modulation of hSK4 in lymphocytes, results that have important implications for the existence of accessory molecules and cell-specific KCa channel regulation.

      DISCUSSION

      Comparison of the Cloned hSK4 with IK Current in Activated T Cells

      The present results are entirely consistent with the IK current in T lymphocytes being the product of the hSK4/hIK1/hKCa4 gene, which was recently cloned from cDNA libraries from human placenta (
      • Joiner W.J.
      • Wang L.-Y.
      • Tang M.D.
      • Kaczmarek L.K.
      ), pancreas (
      • Ishii T.M.
      • Silvia C.
      • Hirschberg B.
      • Bond C.T.
      • Adelman J.P.
      • Maylie J.
      ), and lymph node (
      • Logsdon N.J.
      • Kang J.
      • Togo J.A.
      • Christian E.P.
      • Aiyar J.
      ). Since the product we cloned is 100% identical to hSK4/hIK1/hKCa4, differences in properties of the native lymphocyte IK and exogenously expressed hSK4 channels are not expected unless such properties are determined by something other than the α subunit of the channel. In principle, differences could arise if the channel forms heteromultimers with another protein, if alternative splice variants exist, or if the channel interacts with accessory molecules. It is intriguing that multiple transcript sizes are commonly seen for this channel,i.e. 2.6 and 3.8 kb (
      • Joiner W.J.
      • Wang L.-Y.
      • Tang M.D.
      • Kaczmarek L.K.
      ), ∼2.1 kb, and at least one larger band (
      • Ishii T.M.
      • Silvia C.
      • Hirschberg B.
      • Bond C.T.
      • Adelman J.P.
      • Maylie J.
      ), 2.2 kb, with two larger bands (
      • Logsdon N.J.
      • Kang J.
      • Togo J.A.
      • Christian E.P.
      • Aiyar J.
      ), and a prominent 2.2-kb band with a weaker 2.6-kb band (present study, data not shown).
      Biophysical properties of the lymphocyte IK current have been described at the single channel (
      • Mahaut-Smith M.
      • Schlichter L.C.
      ,
      • Grissmer S.
      • Nguyen A.N.
      • Cahalan M.D.
      ,
      • Verheugen J.A.
      • Vijverberg H.P.M.
      • Oortgiesen M.
      • Cahalan M.D.
      ) and whole-cell level (
      • Grissmer S.
      • Nguyen A.N.
      • Cahalan M.D.
      ). Whereas channel gating is independent of voltage, it is highly sensitive to intracellular free Ca2+, activating at <200 nmin T and B cells, reaching half-maximal activation at about 450 nm, and maximal activation at ∼1 μm (
      • Mahaut-Smith M.
      • Schlichter L.C.
      ,
      • Grissmer S.
      • Nguyen A.N.
      • Cahalan M.D.
      ). Both the single channel and whole-cell current versusvoltage (I-V) relations are inwardly rectifying with symmetrical K+ concentrations on both sides of the membrane (
      • Grissmer S.
      • Nguyen A.N.
      • Cahalan M.D.
      ,
      • Schlichter L.C.
      • Pahapill P.
      • Schumacher P.A.
      ,
      • Verheugen J.A.
      • Vijverberg H.P.M.
      • Oortgiesen M.
      • Cahalan M.D.
      ). The single channel I-V relation is linear under physiological Na+/K+ gradients, which, together with the voltage-independent gating, results in a whole-cell current that is nearly linear (
      • Grissmer S.
      • Nguyen A.N.
      • Cahalan M.D.
      ,
      • Verheugen J.A.
      • Vijverberg H.P.M.
      • Oortgiesen M.
      • Cahalan M.D.
      ). Expressed hSK4/hIK1/hKCa4 currents (Refs.
      • Joiner W.J.
      • Wang L.-Y.
      • Tang M.D.
      • Kaczmarek L.K.
      ,
      • Ishii T.M.
      • Silvia C.
      • Hirschberg B.
      • Bond C.T.
      • Adelman J.P.
      • Maylie J.
      ,
      • Logsdon N.J.
      • Kang J.
      • Togo J.A.
      • Christian E.P.
      • Aiyar J.
      and present study) have the following features in common with the lymphocyte IK current (Refs.
      • Mahaut-Smith M.
      • Schlichter L.C.
      ,
      • Mahaut-Smith M.
      • Schlichter L.C.
      ,
      • Grissmer S.
      • Nguyen A.N.
      • Cahalan M.D.
      ,
      • Joiner W.J.
      • Wang L.-Y.
      • Tang M.D.
      • Kaczmarek L.K.
      , and
      • Verheugen J.A.
      • Vijverberg H.P.M.
      • Oortgiesen M.
      • Cahalan M.D.
      and present study); activation by sub-micromolar free Ca2+, time- and voltage-independent gating, inwardly rectified single channel I-V relations in symmetrical K+ (10–35 pS), and nearly linear I-Vs in physiological Na+/K+ gradients (∼10 pS).
      The pharmacological profiles of native IK and hSK4/hIK1/hKCa4 are also similar. Native IK in lymphoblasts and hSK4 expressed in CHO cells were blocked by ChTx (IC50 2–10 nm) but very poorly by iberiotoxin (IC50 >200 nm), margatoxin (IC50 >100 nm), or tetraethylammonium (IC50 30–40 mm). Clotrimazole showed a similar potency for blocking IK in lymphoblasts (present study) and for heterologously expressed hSK4 (IC50 25–60 nm(Ref.
      • Ishii T.M.
      • Silvia C.
      • Hirschberg B.
      • Bond C.T.
      • Adelman J.P.
      • Maylie J.
      and present study)). The hSK4/hIK1/hKCa4 channel is expected to be insensitive to both apamin and d-tubocurarine since it lacks two necessary amino acids in the putative pore (
      • Ishii T.M.
      • Maylie J.
      • Adelman J.P.
      ) and, as expected, neither the lymphoblast IK nor the expressed hSK4 current were significantly inhibited by apamin (IC50 >100 nm (Refs.
      • Grissmer S.
      • Nguyen A.N.
      • Cahalan M.D.
      and
      • Joiner W.J.
      • Wang L.-Y.
      • Tang M.D.
      • Kaczmarek L.K.
      ,
      • Ishii T.M.
      • Silvia C.
      • Hirschberg B.
      • Bond C.T.
      • Adelman J.P.
      • Maylie J.
      ,
      • Logsdon N.J.
      • Kang J.
      • Togo J.A.
      • Christian E.P.
      • Aiyar J.
      and present study) ord-tubocurarine (IC50 >250 μm; present study).

      Increased Role for hSK4 in Lymphoblast Proliferation

      K+ channel activity is important during the early activation phase of naive T cells, especially for maintaining a hyperpolarized membrane potential, promoting a rise in intracellular Ca2+, and permitting a cascade of events that culminates in interleukin-2 production (
      • Lewis R.S.
      • Cahalan M.D.
      ,
      • Lin C.S.
      • Boltz R.C.
      • Blake J.T.
      • Nguyen M.
      • Talento A.
      • Fischer P.A.
      • Springer M.S.
      • Sigal N.H.
      • Slaughter R.S.
      • Garcia M.L.
      • Kaczorowski G.J.
      • Koo G.C.
      ,
      • Price M.
      • Lee S.C.
      • Deutsch C.
      ). In the first few hours after mitogenic stimulation, precisely when Ca2+ elevation is necessary (
      • Crabtree G.R.
      • Clipstone N.A.
      ,
      • Timmerman L.A.
      • Clipstone N.A.
      • Ho S.N.
      • Northrop J.P.
      • Crabtree G.R.
      ), K+ channel blockers, or other means of depolarizing T cells (high external K+, voltage clamp), inhibit T cell activation (
      • Cheung R.K.
      • Grinstein S.
      • Gelfand E.W.
      ) by compromising Ca2+ influx and the resulting rise in Ca2+. Early studies using non-selective K+ channel blockers (e.g.quinidine, 4-aminopyridine) were later substantiated by more selective peptide toxins including charybdotoxin, which blocks both IK (K d ∼2–6 nm (Refs.
      • Mahaut-Smith M.
      • Schlichter L.C.
      and
      • Grissmer S.
      • Nguyen A.N.
      • Cahalan M.D.
      and present study)) and Kv1.3 channels (K d ∼1 nm (
      • Lewis R.S.
      • Cahalan M.D.
      ,
      • Grissmer S.
      • Nguyen A.N.
      • Cahalan M.D.
      ,
      • Chang M.C.
      • Schlichter L.C.
      ,
      • Khanna R.
      • Chang M.C.
      • Joiner W.
      • Kaczmarek L.K.
      • Schlichter L.C.
      ,
      • Verheugen J.A.
      • Vijverberg H.P.M.
      • Oortgiesen M.
      • Cahalan M.D.
      )), and margatoxin or noxiustoxin which block Kv1.3 but not IK channels (
      • Lewis R.S.
      • Cahalan M.D.
      ,
      • Chang M.C.
      • Schlichter L.C.
      ,
      • Khanna R.
      • Chang M.C.
      • Joiner W.
      • Kaczmarek L.K.
      • Schlichter L.C.
      ,
      • Koo G.C.
      • Blake J.T.
      • Talento A.
      • Nguyen M.
      • Lin S.
      • Sirotina A.
      • Shah K.
      • Mulvany K.
      • Hora D.
      • Cunningham P.
      • Wunderler D.L.
      • McManus O.B.
      • Slaughter R.
      • Bugianesi R.
      • Felix J.
      • Garcia M.
      • Williamson J.
      • Kaczorowski G.J.
      • Sigal N.H.
      • Springer M.S.
      • Feeney W.
      ,
      • Schlichter L.C.
      • Chung I.
      • Chang M.C.
      ). From the limited functional studies using blockers that discriminate between Kv1.3 and other K+ channels, Kv1.3 appears to be important for activation of naive T cells through pathways that are Ca2+-dependent (
      • Lewis R.S.
      • Cahalan M.D.
      ,
      • Lin C.S.
      • Boltz R.C.
      • Blake J.T.
      • Nguyen M.
      • Talento A.
      • Fischer P.A.
      • Springer M.S.
      • Sigal N.H.
      • Slaughter R.S.
      • Garcia M.L.
      • Kaczorowski G.J.
      • Koo G.C.
      ).
      The contribution of KCa channels to T cell activation and proliferation is still poorly understood. Since we previously found that mitogens activate KCa channels in cell-attached patches from naive human T cells, we proposed that they also play a role in T cell activation (
      • Schlichter L.C.
      • Sakellaropoulos G.
      ). In the present study we observed a 14.6-fold increase in hSK4 transcripts by 3–4 days after mitogenic stimulation. During the same period, mRNA levels for Kv1.3 increased only 1.3-fold. These changes are consistent with previous patch clamp studies showing an approximate doubling in Kv1.3 current (see Ref.
      • Lewis R.S.
      • Cahalan M.D.
      ), a 30-fold increase in the ChTx-sensitive KCacurrent (
      • Grissmer S.
      • Nguyen A.N.
      • Cahalan M.D.
      ) and an increase in hKCa4 mRNA (
      • Logsdon N.J.
      • Kang J.
      • Togo J.A.
      • Christian E.P.
      • Aiyar J.
      ). The prediction that IK will be increasingly important for the secondary immune response (e.g. proliferation of previously activated lymphoblasts) is supported by the present results. There is also a recent report that Ca2+ signaling and proliferation were more strongly inhibited by ChTx in lymphoblasts than in naive T cells (
      • Verheugen J.A.H.
      ); however, ChTx does not discriminate between IK and Kv1.3 channels. To separate better the contributions of Kv1.3 and IK to T cell function, we used AgTx-2 to block Kv1.3 and clotrimazole to block IK. Consistent with our expectations, IK block more effectively inhibited proliferation of lymphoblasts than naive T cells. Furthermore, despite the greater Kv1.3 channel block (AgTx-2 at ∼25 K d) than IK block (clotrimazole at ∼6 K d), IK block was more effective in inhibiting lymphoblast proliferation, i.e. by 65.0% compared with 18.4% for Kv1.3 block. Blocking both channels (AgTx-2 + clotrimazole) was approximately additive, reducing lymphoblast proliferation by 86.8%. Proliferation of naive T cells was also sensitive to blocking both channels (36.5% inhibition), consistent with previous reports of reduced proliferation when IK + Kv1.3 were blocked with ChTx (
      • Lewis R.S.
      • Cahalan M.D.
      ,
      • Lin C.S.
      • Boltz R.C.
      • Blake J.T.
      • Nguyen M.
      • Talento A.
      • Fischer P.A.
      • Springer M.S.
      • Sigal N.H.
      • Slaughter R.S.
      • Garcia M.L.
      • Kaczorowski G.J.
      • Koo G.C.
      ,
      • Price M.
      • Lee S.C.
      • Deutsch C.
      ).
      How might both Kv1.3 and IK channels contribute to T cell proliferation? Within seconds after stimulating the T cell receptor, tyrosine-kinase mediated activation of phospholipase C produces inositol 1,4,5-trisphosphate and quickly triggers Ca2+release from internal stores. A plasma membrane channel (the Ca2+ release-activated Ca2+ channel) then opens to allow Ca2+ influx, which is required for several hours. Ca2+ release-activated Ca2+ channel opening is not voltage-dependent, but Ca2+ influx is strongly driven by the membrane potential. Thus, any means of increasing the K+ conductance and hyperpolarizing the cell will facilitate Ca2+ entry. Kv1.3 is voltage-gated, activated by depolarization, and its steady-state activity is maximal between −50 and −30 mV in resting T cells, depending on post-translational modulation (
      • Schlichter L.C.
      • Chung I.
      • Chang M.C.
      ). Hence it is likely to play a role only when the membrane is moderately depolarized. In contrast, gating of the IK/hSK4 channel is voltage-independent but exquisitely sensitive to internal Ca2+; thus, it is well designed to open whenever Ca2+ rises marginally above the resting level. Rather than Ca2+ changing in a sustained manner after T cell receptor stimulation, Ca2+ and membrane potential can oscillate (
      • Lewis R.S.
      • Cahalan M.D.
      ,
      • Verheugen J.A.H.
      ,
      • Verheugen J.A.H.
      • Vijverberg H.P.M.
      ). Collectively, these complementary properties would allow the cell alternately to use IK channels when Ca2+ is high, even if the membrane is hyperpolarized, and Kv1.3 channels during periods of low Ca2+ and/or depolarization.

      Role of hSK4 and Kv1.3 in Volume Regulation

      Volume regulation in leukocytes and other mammalian cells has been extensively reviewed (
      • Grinstein S.
      • Foskett J.K.
      ,
      • Sarkadi B.
      • Parker J.C.
      ,
      • Lang F.
      • Busch G.L.
      • Volkl H.
      ). The RVD involves ion efflux through separate K+ and anion channels. Its rate and extent depend on the combined ion conductances; thus if either the K+ or Cl current is small (or blocked pharmacologically) it will limit volume recovery. Identifying the particular K+channel(s) that underlie RVD has been problematic, largely due to the lack of selective blockers and uncertainty over whether swelling evokes a rise in intracellular Ca2+. For instance, early evidence of a role for Kv1.3 was not convincing since it first relied on nonspecific K+ channel blockers and then on ChTx (
      • Grinstein S.
      • Foskett J.K.
      ,
      • Sarkadi B.
      • Parker J.C.
      ,
      • Grinstein S.
      • Dupre A.
      • Rothstein A.
      ) which we now know blocks both Kv1.3 and IK. Kv1.3 can confer RVD when transfected into a mouse T cell line that lacks voltage-gated K+ currents (
      • Deutsch C.
      • Chen L.-Q.
      ); however, that study did not address the presence or role of KCa channels. To determine whether intracellular Ca2+ is elevated in human T cells during RVD, we previously designed a perfused cuvette system for fluorometric measurements (
      • Schlichter L.C.
      • Pennefather P.S.
      • van Staden C.J.
      • Valentine E.A.
      ). We found that hypotonic shock elicits a rapid, biphasic rise in Ca2+ (
      • Schlichter L.C.
      • Sakellaropoulos G.
      ) which comprises release from internal stores and influx across the plasma membrane. Ca2+reached a peak in 1–2 min and remained elevated for at least 10 min, and the entire pattern was indistinguishable from Ca2+signaling during T cell activation. Thus, we proposed (
      • Schlichter L.C.
      • Sakellaropoulos G.
      ) that both KCa and Kv1.3 channels will contribute to RVD, with Ca2+ and voltage oscillations alternately opening each type of K+ channel.
      Previous studies of RVD in T cells have been restricted to resting cells and show a stereotypical response (
      • Grinstein S.
      • Foskett J.K.
      ,
      • Sarkadi B.
      • Parker J.C.
      ,
      • Deutsch C.
      • Chen L.-Q.
      ); within 1–2 min after a hypotonic shock, T cells swelled to ∼120% of their original volume and then returned to their original volume within 5–15 min. Our results, using right angle light scattering to measure RVD, are in excellent agreement both in the extent of swelling (∼120%) and in the time course, with maximal swelling within 1–2 min and nearly full recovery within 6 min for both resting T cells and lymphoblasts. We have now assessed the relative contributions of IK and Kv1.3 channels to RVD in resting T cells compared with activated lymphoblasts. Owing to the dramatic increase in IK and small increase in Kv1.3 in lymphoblasts, we expected RVD to depend more on IK than Kv1.3 channels in lymphoblasts.
      We examined the role of each channel type in the initial swelling to maximal volume and in the degree of recovery by 6 min after a 56% hypotonic shock. For resting T cells, blocking Kv1.3 or IK, or both channels simultaneously, increased the maximal volume, implying that K+ efflux through both channels occurs even during the initial swelling phase. Consistent with this conclusion, in previous studies T cells swelled less than predicted for a passive osmometer (
      • Grinstein S.
      • Foskett J.K.
      ,
      • Sarkadi B.
      • Parker J.C.
      ,
      • Lang F.
      • Busch G.L.
      • Volkl H.
      ). We had anticipated a role for IK channels in resting cells, since IK is expressed (Refs.
      • Schlichter L.C.
      • Pahapill P.
      • Schumacher P.A.
      and
      • Verheugen J.A.H.
      • Vijverberg H.P.M.
      and present study), and intracellular Ca2+ rises after a hypotonic shock (
      • Schlichter L.C.
      • Sakellaropoulos G.
      ). For lymphoblasts, IK block was very effective in increasing the maximal volume, whereas Kv1.3 block had no effect. Thus, in lymphoblasts volume regulation also proceeds during the swelling phase but IK plays a much greater role than Kv1.3 current at this time.
      The extent of recovery after the maximal volume is reached reflects both the K+ and Cl conductances during the RVD phase. Substantial recovery occurred within 6 min in both cell types, and the slightly greater recovery in lymphoblasts is consistent with up-regulation of IK/hSK4 expression. RVD was significantly inhibited by blocking Kv1.3 channels in both resting T cells and lymphoblasts but was more effective in resting cells. For both cell types IK block was more effective than Kv1.3 block; moreover, inhibition of RVD was greater in lymphoblasts. Thus, both K+ channels contribute to RVD, but their relative importance is opposite as follows: Kv1.3 plays a greater role in resting cells, and IK is more important in lymphoblasts. Not only are these results predicted from the up-regulated expression of IK/hSK4 in lymphoblasts but IK activation implies that hypotonic shock elicits an early and sustained rise in Ca2+ in activated lymphoblasts, as we have previously shown for resting T cells (
      • Schlichter L.C.
      • Sakellaropoulos G.
      ).
      Although there is insufficient information to calculate K+fluxes through Kv1.3 and IK channels during RVD, some predictions can be made by considering their expression and biophysical properties. Flux through each channel type is proportional to the number of channels (n), their open probability (P o), their single channel conductance (γ), and the driving force, which is the same at a given voltage. For resting T cells, the number of Kv1.3 channels per cell is much larger than IK/hSK4 channels, and γ is similar (∼10 pS in a normal Na/K gradient). So, for IK channels to play a substantial role in RVD, theirP o must be much larger than theP o of Kv1.3 channels in resting cells. Kv1.3 contribution will be controlled by the membrane potential since these channels require moderate depolarization to be tonically active (
      • Schlichter L.C.
      • Chung I.
      • Chang M.C.
      ,
      • Pahapill P.A.
      • Schlichter L.C.
      ,
      • Chung I.
      • Schlichter L.C.
      ). A simple model is that Ca2+ remains elevated thereby activating IK, and the membrane potential remains hyperpolarized, thereby reducing the opening of Kv1.3 channels. IK/hSK4 expression is much higher in lymphoblasts than in resting T cells, and since IK gating is voltage-independent, this channel is expected to contribute more whether or not the membrane potential fluctuates, provided Ca2+ remains modestly elevated.

      Calmodulin-dependent Modulation, Evidence for More Than One Mechanism

      Our electrophysiological results implicate calmodulin (CaM) in regulating native IK channels in lymphoblasts. In principle, CaM antagonists could act by interfering with interactions between CaM and the channel protein, by interacting with accessory CaM-binding molecules (e.g. CaM kinases or channel β subunits, if they exist), or by directly interfering with the channel protein. Some predictions can be made from the mechanism by which the antagonists inhibit CaM (
      • Hait W.N.
      • Lazo J.S.
      ,
      • Braun A.P.
      • Schulman H.
      ,
      • Tokumitsu H.
      • Chijiwa T.
      • Hagiwara M.
      • Mizutani A.
      • Teresawa M.
      • Hidaka H.
      ). In cell-free systems, CaM changes conformation when at least two of its four Ca2+-binding domains are saturated (K d ∼2.4 μmCa2+). CaM antagonists can then bind reversibly to a newly exposed hydrophobic site (
      • Weiss B.
      • Prozialeck W.
      • Cimino M.
      • Barnette M.S.
      • Wallace T.L.
      ) thereby preventing interactions between CaM and target proteins. Thus, it is expected that excess CaM will competitively reduce inhibition by titrating the amount of drug available for inhibition.
      Direct interactions of some CaM antagonists in the pore of some K+ channels have been proposed when their potency for CaM inhibition differed from that of channel inhibition (
      • McCann J.D.
      • Welsh M.J.
      ,
      • Klockner U.
      • Isenberg G.
      ,
      • Kihira M.
      • Matsuzawa K.
      • Tokuno H.
      • Tomita T.
      ), or the drugs were effective even when Ca2+ was not elevated (
      • Kihira M.
      • Matsuzawa K.
      • Tokuno H.
      • Tomita T.
      ), or exogenously added CaM did not compete with the antagonists (
      • Klockner U.
      • Isenberg G.
      ). We found that trifluoperazine and W-7 produced time- and voltage-dependent decreases in both native IK and expressed hSK4 current at positive potentials, which may reflect a direct drug interaction with the channel protein. Of greater physiological relevance is the inhibition of native IK channels we observed at negative membrane potentials. In this case the potency of inhibition by W-7 and TFP was consistent with effects on CaM, and as expected for competitive drug binding, excess internal CaM significantly relieved the inhibition by W-7. This result also rules out significant channel block by W-7 from the outside at negative potentials. CaM antagonists may affect the lymphoblast IK current through interactions between CaM or other CaM-binding molecules and the channel protein. Such interactions must differ for hSK4 channels stably expressed in CHO cells since, at negative potentials, these currents were not inhibited by TFP or W-7, and calmidazolium was less effective than on IK currents. As discussed below, differences in actions on native IK and hSK4 channels may reflect multiple sites of action.

      Direct Interactions between CaM and IK/hSK4 Channels

      Despite the exquisite Ca2+ sensitivity of IK/hSK4 gating, the primary amino acid sequence of the α subunit contains no known Ca2+-binding sites, that is no E-F hands, C2 domains (
      • Shao X.
      • Davletov B.A.
      • Sutton R.B.
      • Sudhof T.C.
      • Rizo J.
      ), or Ca2+ “bowls” (
      • Schreiber M.
      • Salkoff L.
      ). We found that channels made from wild-type, full-length hSK4 α subunits bind to calmodulin. Although binding was greatly inhibited when Ca2+ was chelated, some binding remained. Deletion mutants of several cytoplasmic regions that are relatively conserved between hSK4 and brain SK channels showed that CaM binding was restricted to the proximal C-terminal tail of hSK4 (a region we call “Ct1,” see Fig. 7). Deleting the Ct1 region prevented the expression of functional Ca2+-gated hSK4 channels. When wild-type hSK4 was expressed and whole-cell membrane potential (V m) recordings were made with micromolar intracellular Ca2+ to maximally activate hSK4,V m became highly negative owing to the hyperpolarizing K+ conductance. In contrast, the Ct1-deleted channel failed to produce a hyperpolarizing K+conductance, and V m remained essentially at zero. There are two possibilities as follows: without CaM binding the channels did not open in response to high Ca2+ (as is the case for SK2 channels (
      • Xia X.-M.
      • Fakler B.
      • Rivard A.
      • Wayman G.
      • Johnson-Pais T.
      • Keen J.E.
      • Ishii T.
      • Hirschberg B.
      • Bond C.T.
      • Lutsenko S.
      • Maylie J.
      • Adelman J.P.
      )) or the mutant channels did not assemble properly in the cell membrane. In the future it will be useful to examine the assembly and trafficking of mutant hSK4 channels, particularly since the Ct2-deletion mutant (lacking the leucine zipper region) also failed to produce a hyperpolarizing K+conductance. The α subunits of brain SK channels (SK1–3) also bind to CaM in the proximal part of the cytoplasmic C terminus, and CaM apparently serves as the Ca2+-binding gate (
      • Xia X.-M.
      • Fakler B.
      • Rivard A.
      • Wayman G.
      • Johnson-Pais T.
      • Keen J.E.
      • Ishii T.
      • Hirschberg B.
      • Bond C.T.
      • Lutsenko S.
      • Maylie J.
      • Adelman J.P.
      ). Most of the C terminus of SK2 channels (4 α helices, A–D) bound to CaM, whereas a “post-D” C-terminal tail (corresponding to our Ct2, leucine-zipper region) did not. If helices A–D were all present, binding was independent of Ca2+, whereas helices B and C and B–D conferred Ca2+-dependent binding to CaM.
      Our results on hSK4 share several features with SK2 (
      • Xia X.-M.
      • Fakler B.
      • Rivard A.
      • Wayman G.
      • Johnson-Pais T.
      • Keen J.E.
      • Ishii T.
      • Hirschberg B.
      • Bond C.T.
      • Lutsenko S.
      • Maylie J.
      • Adelman J.P.
      ), but they also differ in ways that are consistent with additional sites of interaction or modulation of the native channel in lymphocytes. Although brain SK channels have only modest overall homology (∼40%) to hSK4 (
      • Joiner W.J.
      • Wang L.-Y.
      • Tang M.D.
      • Kaczmarek L.K.
      ), some regions are more homologous. The 95-amino acid CaM-binding domain (Ct1) that we identified in hSK4 is the same channel region as helices A–C in SK2. Helix A in SK2 is highly homologous to the corresponding region of hSK4 (79% identical), whereas regions B and C have much lower homology (20% identical). For SK2, the Ca2+-independent CaM binding and patch clamp studies in which calmidazolium failed to inhibit the expressed channels were taken as evidence that CaM binds constitutively and irreversibly to brain SK channels (
      • Xia X.-M.
      • Fakler B.
      • Rivard A.
      • Wayman G.
      • Johnson-Pais T.
      • Keen J.E.
      • Ishii T.
      • Hirschberg B.
      • Bond C.T.
      • Lutsenko S.
      • Maylie J.
      • Adelman J.P.
      ). We found that most of the CaM binding to hSK4 was Ca2+-dependent; however, as explained earlier, the binding assay we used would not detect channels irreversibly bound to CaM. Several possible explanations for these differences will require further study. For instance, there may be more than one CaM-binding site with different affinities, as is the case for cation channels in retinal rods (
      • Grunwald M.E.
      • Yu W.-P.
      • Yu H.-H.
      • Yau K.-W.
      ), with a lower affinity site that is reversible and Ca2+-dependent, perhaps in Ct1 of hSK4 (helices B–D in SK2). CaM antagonists inhibited the native current in lymphoblasts, with little or no inhibition of expressed hSK4 currents, a result that is consistent with the failure of 1000 nm calmidazolium to prevent CaM binding to the expressed hSK4 protein. A further possibility is that weaker CaM channel binding in lymphoblasts, perhaps a result of other protein-channel interactions (see below), allows more effective competition by CaM antagonists.

      Evidence for Accessory Molecules in Lymphocytes

      The striking differences in inhibition at negative potentials of native IKversus expressed hSK4 channels provide the first evidence that accessory molecules (other than CaM) modulate a member of the SK channel family. Candidate molecules include CaMK, calcineurin, and β subunits analogous to those interacting with voltage-gated K+ channels. Although β subunits have not been identified for SK channels, there is evidence that apamin-sensitive SK channels form hetero-oligomers. In a variety of cells expressing SK channels, apamin binds to both high (59 or 86 kDa) and low (30 or 33 kDa) molecular mass polypeptides that are integral membrane proteins (
      • Wadsworth J.D.F.
      • Torelli S.
      • Doorty K.B.
      • Strong P.N.
      ). It is unlikely that such accessory molecules are essential for channel activity since all known members of the SK family, with the exception of rSK1 (
      • Joiner W.J.
      • Wang L.-Y.
      • Tang M.D.
      • Kaczmarek L.K.
      ), are functional in expression systems, includingXenopus oocytes, HEK, and CHO cells. This observation also implies that the Ca2+-binding site, which is thought to be CaM bound to the channel (
      • Xia X.-M.
      • Fakler B.
      • Rivard A.
      • Wayman G.
      • Johnson-Pais T.
      • Keen J.E.
      • Ishii T.
      • Hirschberg B.
      • Bond C.T.
      • Lutsenko S.
      • Maylie J.
      • Adelman J.P.
      ), functions normally in these cells.
      Since the CaM kinase antagonist, KN-62, inhibited native IK current but had no effect on hSK4, it is necessary to consider how CaM kinase might selectively modulate the current in lymphocytes. hSK4 contains a potential phosphorylation site in the C-terminal domain that should accommodate either CaM kinase II (or protein kinase A) (
      • Braun A.P.
      • Schulman H.
      ); however, this site is not necessarily phosphorylated. It may be that CHO cells have insufficient CaM kinase (T cells express high levels of CaM kinase II and IV (
      • Hanissian S.H.
      • Frangakis M.
      • Bland M.M.
      • Jawahar S.
      • Chatila T.A.
      )) or that the site of phosphorylation is not on the channel protein itself, but rather on a β subunit or other unknown accessory molecule. Interestingly, in lymphoblasts the increased inhibition by W-7 in the presence of KN-62 is consistent with dual modulation by CaM binding and CaM kinase.
      Potential modulation of lymphocyte IK/hSK4 channels by CaM and CaM kinase is of broader importance. Early in T cell activation or lymphoblast re-activation, there is a rise in intracellular Ca2+ that activates CaM-dependent enzymes. These include CaM kinases II and IV (
      • Hanissian S.H.
      • Frangakis M.
      • Bland M.M.
      • Jawahar S.
      • Chatila T.A.
      ) and calcineurin (protein phosphatase 2B), which is highly expressed in lymphocytes and crucial for T cell proliferation (
      • Crabtree G.R.
      • Clipstone N.A.
      ). CaM antagonists can inhibit some lymphocyte functions that either trigger conductive K+fluxes or are sensitive to the membrane potential of the cell (which depends on K+ channels). T cell activation (
      • Nakabayashi H.
      • Komada H.
      • Yoshida T.
      • Takanari H.
      • Izutsu K.
      ), cell-mediated cytotoxicity (
      • Rees R.C.
      • Parker S.
      • Platts A.
      • Blackburn M.G.
      • MacNeil S.
      ), and volume regulation (
      • Grinstein S.
      • Dupre A.
      • Rothstein A.
      ) are inhibited both by K+ channel blockers and by CaM antagonists. Our present results provide new evidence that a specific K+ channel (IK/hSK4) that is important for at least two of these functions (proliferation and volume regulation) is susceptible to CaM antagonists, thus providing a link between CaM and K+channels in regulating lymphocyte function.

      Acknowledgments

      We are very grateful to Dr. E. F. Stanley (NINDS, National Institutes of Health, Bethesda) for helpful discussions and comments on the manuscript, Dr. J. R. G. Challis (University of Toronto) for human placental tissue, and Dr. O. T. Jones for the use of equipment.

      REFERENCES

        • Lewis R.S.
        • Cahalan M.D.
        Annu. Rev. Immunol. 1995; 13: 623-653
        • Mahaut-Smith M.
        • Schlichter L.C.
        J. Physiol. (Lond.). 1989; 415: 69-83
        • Mahaut-Smith M.
        • Schlichter L.C.
        Pfluegers Arch. 1989; 414: S164-S165
        • Grissmer S.
        • Lewis R.S.
        • Cahalan M.D.
        J. Gen. Physiol. 1992; 99: 63-84
        • Grissmer S.
        • Nguyen A.N.
        • Cahalan M.D.
        J. Gen. Physiol. 1993; 102: 601-630
        • Schlichter L.C.
        • Pahapill P.
        • Schumacher P.A.
        Receptors Channels. 1993; 1: 201-215
        • Joiner W.J.
        • Wang L.-Y.
        • Tang M.D.
        • Kaczmarek L.K.
        Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 11013-11018
        • Ishii T.M.
        • Silvia C.
        • Hirschberg B.
        • Bond C.T.
        • Adelman J.P.
        • Maylie J.
        Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 11651-11656
        • Logsdon N.J.
        • Kang J.
        • Togo J.A.
        • Christian E.P.
        • Aiyar J.
        J. Biol. Chem. 1997; 272: 32723-32726
        • Grinstein S.
        • Foskett J.K.
        Annu. Rev. Physiol. 1990; 52: 399-414
        • Sarkadi B.
        • Parker J.C.
        Biochim. Biophys. Acta. 1991; 1071: 407-427
        • Lang F.
        • Busch G.L.
        • Volkl H.
        Cell. Physiol. Biochem. 1998; 8: 1-45
        • Deutsch C.
        • Chen L.-Q.
        Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 10036-10040
        • Schlichter L.C.
        • Sakellaropoulos G.
        Exp. Cell Res. 1994; 215: 211-222
        • Xia X.-M.
        • Fakler B.
        • Rivard A.
        • Wayman G.
        • Johnson-Pais T.
        • Keen J.E.
        • Ishii T.
        • Hirschberg B.
        • Bond C.T.
        • Lutsenko S.
        • Maylie J.
        • Adelman J.P.
        Nature. 1998; 395: 503-507
        • Chang M.C.
        • Schlichter L.C.
        Soc. Neurosci. Abstr. 1996; 22: 1444
        • Khanna R.
        • Chang M.C.
        • Joiner W.
        • Kaczmarek L.K.
        • Schlichter L.C.
        Biophys. J. 1998; 74: 35
        • Chomczynski P.
        • Sacchi N.
        Anal. Biochem. 1987; 162: 156-159
        • Chapin S.J.
        • Enrich C.
        • Aroeti B.
        • Havel R.J.
        • Mostov K.E.
        J. Biol. Chem. 1996; 271: 1336-1342
        • Schumacher P.A.
        • Sakellaropoulos G.
        • Phipps D.J.
        • Schlichter L.C.
        J. Membr. Biol. 1995; 145: 217-232
        • Weinstein H.
        • Mehler E.L.
        Annu. Rev. Physiol. 1994; 56: 213-236
        • Linse S.
        • Helmersson A.
        • Forsen S.
        J. Biol. Chem. 1991; 266: 8050-8054
        • McGann L.E.
        • Walterson M.L.
        • Hogg L.M.
        Cytometry. 1988; 9: 33-38
        • Verheugen J.A.
        • Vijverberg H.P.M.
        • Oortgiesen M.
        • Cahalan M.D.
        J. Gen. Physiol. 1995; 105: 765-794
        • Alvarez J.
        • Montero M.
        • Garcia-Sancho J.
        J. Biol. Chem. 1992; 267: 11789-11793
        • Kohler M.
        • Hirschberg B.
        • Bond C.T.
        • Kinzie J.M.
        • Marrion N.V.
        • Maylie J.
        • Adelman J.P.
        Science. 1996; 273: 1709-1713
        • Ishii T.M.
        • Maylie J.
        • Adelman J.P.
        J. Biol. Chem. 1997; 272: 23195-23200
        • Koo G.C.
        • Blake J.T.
        • Talento A.
        • Nguyen M.
        • Lin S.
        • Sirotina A.
        • Shah K.
        • Mulvany K.
        • Hora D.
        • Cunningham P.
        • Wunderler D.L.
        • McManus O.B.
        • Slaughter R.
        • Bugianesi R.
        • Felix J.
        • Garcia M.
        • Williamson J.
        • Kaczorowski G.J.
        • Sigal N.H.
        • Springer M.S.
        • Feeney W.
        J. Immunol. 1997; 158: 5120-5128
        • Walsh M.P.
        Rev. Clin. Basic Pharmacol. 1985; 5: 35-69
        • Tanaka T.
        • Ohmura T.
        • Yamakado T.
        • Hidaka H.
        Mol. Pharmacol. 1982; 22: 408-412
        • Gietzen K.
        • Wuthrich A.
        • Bader H.
        Biochem. Biophys. Res. Commun. 1981; 101: 418-425
        • Lin C.S.
        • Boltz R.C.
        • Blake J.T.
        • Nguyen M.
        • Talento A.
        • Fischer P.A.
        • Springer M.S.
        • Sigal N.H.
        • Slaughter R.S.
        • Garcia M.L.
        • Kaczorowski G.J.
        • Koo G.C.
        J. Exp. Med. 1993; 177: 637-645
        • Price M.
        • Lee S.C.
        • Deutsch C.
        Proc. Natl. Acad. Sci. U. S. A. 1989; 86: 10171-10175
        • Crabtree G.R.
        • Clipstone N.A.
        Annu. Rev. Biochem. 1994; 63: 1045-1083
        • Timmerman L.A.
        • Clipstone N.A.
        • Ho S.N.
        • Northrop J.P.
        • Crabtree G.R.
        Nature. 1996; 383: 837-840
        • Cheung R.K.
        • Grinstein S.
        • Gelfand E.W.
        J. Immunol. 1983; 131: 2291-2295
        • Schlichter L.C.
        • Chung I.
        • Chang M.C.
        Cell. Physiol. Biochem. 1997; 7: 159-171
        • Verheugen J.A.H.
        Cell. Physiol. Biochem. 1997; 7: 188-202
        • Verheugen J.A.H.
        • Vijverberg H.P.M.
        Cell Calcium. 1995; 17: 287-300
        • Grinstein S.
        • Dupre A.
        • Rothstein A.
        J. Gen. Physiol. 1982; 79: 849-868
        • Schlichter L.C.
        • Pennefather P.S.
        • van Staden C.J.
        • Valentine E.A.
        Pfluegers Arch. 1992; 421: 400-402
        • Pahapill P.A.
        • Schlichter L.C.
        J. Physiol. (Lond.). 1990; 422: 103-126
        • Chung I.
        • Schlichter L.C.
        J. Membr. Biol. 1997; 156: 73-85
        • Hait W.N.
        • Lazo J.S.
        J. Clin. Oncol. 1986; 4: 994-1012
        • Braun A.P.
        • Schulman H.
        Annu. Rev. Physiol. 1995; 57: 415-445
        • Tokumitsu H.
        • Chijiwa T.
        • Hagiwara M.
        • Mizutani A.
        • Teresawa M.
        • Hidaka H.
        J. Biol. Chem. 1990; 265: 4315-4320
        • Weiss B.
        • Prozialeck W.
        • Cimino M.
        • Barnette M.S.
        • Wallace T.L.
        Ann . N. Y. Acad. Sci. 1980; 356: 319-345
        • McCann J.D.
        • Welsh M.J.
        J. Gen. Physiol. 1987; 89: 339-352
        • Klockner U.
        • Isenberg G.
        Am. J. Physiol. 1987; 253: H1601-H1611
        • Kihira M.
        • Matsuzawa K.
        • Tokuno H.
        • Tomita T.
        Br. J. Pharmacol. 1990; 100: 353-359
        • Shao X.
        • Davletov B.A.
        • Sutton R.B.
        • Sudhof T.C.
        • Rizo J.
        Science. 1996; 273: 248-251
        • Schreiber M.
        • Salkoff L.
        Biophys. J. 1997; 73: 1355-1363
        • Grunwald M.E.
        • Yu W.-P.
        • Yu H.-H.
        • Yau K.-W.
        J. Biol. Chem. 1998; 273: 9148-9157
        • Wadsworth J.D.F.
        • Torelli S.
        • Doorty K.B.
        • Strong P.N.
        Arch. Biochem. Biophys. 1997; 346: 151-160
        • Hanissian S.H.
        • Frangakis M.
        • Bland M.M.
        • Jawahar S.
        • Chatila T.A.
        J. Biol. Chem. 1993; 268: 20055-20063
        • Nakabayashi H.
        • Komada H.
        • Yoshida T.
        • Takanari H.
        • Izutsu K.
        Biol. Cell. 1992; 75: 55-59
        • Rees R.C.
        • Parker S.
        • Platts A.
        • Blackburn M.G.
        • MacNeil S.
        Biosci. Rep. 1987; 7: 771-775