Overlapping pharmacology of Ca2+-activated Cl− and K+ channels

Opinion

TRENDS in Pharmacological Sciences

Vol.28 No.1

Overlapping pharmacology of Ca2+-activated ClS and K+ channels Iain A. Greenwood1 and Normand Leblanc2 1

Ion Channels and Cell Signalling Research Centre, Division of Basic Medical Sciences, St George’s, University of London, London SW17 0RE, UK 2 Department of Pharmacology/MS 318, Center of Biomedical Research Excellence (COBRE), University of Nevada School of Medicine, Reno, NV 89557-0270, USA

Research into Ca2+-activated ClS channels is hampered by the inability to decipher their molecular identity and the fact that all extant ClS channel blockers have effects on other ion channels. Most notably, ClS channel blockers such as the fenamates (e.g. niflumic acid and flufenamic acid) activate Ca2+-dependent K+ channels, although other pharmacological overlaps have been discovered. In this article, we highlight the complex pharmacology of Ca2+-activated ClS channels and the caveats associated with using these blockers – a necessary requirement because many researchers use ClS channel blockers as probes for ClS channel activity. Moreover, we discuss the argument for a common structural motif between Ca2+-activated ClS channels and Ca2+-dependent K+ channels, which has led to the possibility that the molecular identity of ClS channels will be revealed by research in this new direction, in addition to the use of existing candidates such as the CLCA, Bestrophin and tweety genes. Introduction Compared with the wealth of information about cation channels, the understanding of the properties of anion channels is markedly more fragmentary. Anion channels can be subdivided into five major categories [1]: (i) Cl channels that are activated by an increase in concentration of the intracellular second messenger signal cAMP and encoded by the cystic fibrosis transmembrane regulator (CFTR) gene; (ii) voltage-sensitive channels that are encoded by the CLC family of genes; (iii) channels activated by inhibitory neurotransmitters in the CNS (termed ‘ligand-gated’) that bind to a specific receptor (e.g. GABAA or glycine) that is an integral part of the channel protein; (iv) channels that are activated by changes in cell volume (so-called ‘swelling-activated’ or ‘volume-sensitive’); and (v) channels that are activated by an increase in intracellular Ca2+ concentration ([Ca2+]i), the so-called Ca2+-activated Cl channels (ClCa). Considerable equivocation exists about the molecular identity of the last two types of anion channel. The latter type has been recorded from a wide range of cell types, including cardiac myocytes, neurons, Xenopus oocytes, epithelial cells, secretory cells, endothelial cells and smooth muscle cells [2–4]. In smooth muscle cells, the activation of ClCa is believed to be an excitatory Corresponding author: Greenwood, I.A. ([email protected]). Available online 5 December 2006. www.sciencedirect.com

depolarizing mechanism because Cl ions are actively accumulated so that the activation of the channel by an increase in [Ca2+]i leads to Cl ion efflux [4]. The current evoked by these channels (IClCa) can be subdivided into two main classes [2]. One class, exemplified by epithelial IClCa, lacks time-dependent kinetics and exhibits a low Ca2+ sensitivity (Kd > 1 mM) that is increased by phosphorylation mediated by Ca2+/calmodulin-dependent protein kinase II (CaMKII) [1,4,5]. The other class of IClCa, which has been recorded in Xenopus oocytes, smooth muscle cells, endothelial cells and parotid acinar cells, exhibits a high Ca2+ sensitivity (Kd < 500 nM), outward rectification, distinctive voltage-dependent activation and deactivation kinetics, a small unitary conductance (1–3 pS) and modulation by Ca2+-dependent phosphorylation [4]. These characteristics of IClCa in different cell types have been summarized in several reviews [2–4], and the goal of this article is to highlight the recent pharmacological findings regarding these channels and to propose that a structural commonality exists between Ca2+-activated K+ and Cl channels. Existing molecular candidates for Ca2+-activated ClS channels Unlike the majority of ion channels, the molecular identity of the channels that underlie IClCa in every cell type is unknown. The lack of structural information about these channels has hampered the identification of their role in physiological processes. It is speculated that the IClCa in epithelial cells, the activity of which is determined by Ca2+-dependent phosphorylation, is encoded by the long isoform of the human voltage-gated Cl channel gene CLC-3, the activity of which is increased by CaMKIIinduced phosphorylation [5]. Two major gene families, those encoding CLCA and bestrophins, have been proposed as molecular correlates for the ClCa channels found in the majority of excitable cells, which are thought to be opened by direct Ca2+ binding [4–6]. There is now strong evidence that bestrophin proteins, which are encoded by Best1–Best4, form ion channels [3,7–10] because: (i) the protein sequence predicts four–six distinct transmembrane domains; (ii) current activation occurs at physiological [Ca2+]i; (iii) mutations of specific amino acids within the hypothesized pore region alter anion conductance and selectivity [3,7,8,10]; and (iv) downregulating the expression of genes encoding bestrophin by RNA

0165-6147/$ – see front matter ß 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.tips.2006.11.004

2

Opinion

TRENDS in Pharmacological Sciences

Vol.28 No.1

Table 1. Effects of different ClS channel blockers on ion channelsa,b ClS channel blocker Niflumic acid c

Flufenamic acid c Mefenamic acid c

Meclofenamic acid c NPPB

A-9-C IAA94

Ethacrynic acid Tamoxifen

Ionic current ICl(swell) IK(BKCa)

Effect Inhibit Stimulate

VDCC If IKv4 IClCa IKs IK(BKCa) IK(BKCa) IKS IK(dr) IK(ATP) IKCNQ2/3 ICl(swell) IK(IKCa) ICFTR IK(BKCa) IClCa VDCC IK(BKCa) IK(ATP) IK(BKCa) IK(ATP) ICl(swell) IK(BKCa)

Decrease Decrease Decrease Stimulate Increase Stimulate Stimulate Increase Increase Increase Increase Inhibit Inhibit Activate Stimulate Stimulate Decrease Stimulate Stimulate Stimulate Stimulate Inhibit Stimulate

Cell type Portal vein SMCs Lipid bilayer, portal vein SMCs, urethra SMCs Cerebral artery SMCs, urethra SMCs Cardiac myocytes Xenopus oocytes VSMCs Xenopus oocytes SMCs, lipid bilayer Lipid bilayer, portal vein SMCs Xenopus oocytes Canine jejunum SMCs Urethra SMCs CHO cells NIH-3T3 cells, 8226 myeloma cells Glioblastoma cells HEK-293 cells Portal vein SMCs PA SMCs Cerebral artery SMCs Portal vein SMCs Portal vein SMCs Portal vein SMCs Portal vein SMCs Portal vein SMCs Colonic SMCs

Refs [29] [17,20,30] [31,32] [33] [34] [21,22] [35] [17,20] [17,20] [35] [36] [30] [37] [38] [39] [40] [18] [23] [39] [18] [18] [18] [18] [29] [19]

a

All agents shown are used as blockers of IClCa in various cell types. Abbreviations: ICFTR, Cl current generated by the gene encoding the cystic fibrosis transmembrane regulator CFTR; ICl(swell), swelling-activated Cl current; If, hyperpolarization-activated pacemaker current; IK(ATP), ATP-sensitive K+ channel current; IK(BKCa), Ca2+-activated K+ channel current; IK(dr), delayed rectifier K+ current; IK(IKCa), K+ current generated by the gene encoding the intermediate conductance Ca2+-activated K+ channel; IKs, slow delayed rectifier K+ current generated by the coassociation of proteins encoded by KCNQ1 and KCNE1 (minK); IKv4, transient outward K+ current belonging to the Kv4 family of voltage-dependent K+ channel genes; NPPB, 5nitro-2-(3-phenylpropylamino)benzoic acid; NSC, non-selective cation current; PA, pulmonary artery; SMC, smooth muscle cell; VDCC, dihydropyridine-sensitive voltagedependent Ca2+ channel; VSMC, vascular smooth muscle cell. c These agents are structurally similar. b

silencing reduces the native IClCa in a Drosophila S2 cell line [9]. However, the heterologous expression of genes encoding bestrophin does not recapitulate all the characteristics of native IClCa. By contrast, there is only weak evidence of CLCA gene products forming ClCa channels. The biophysical properties of CLCA-generated channels – including mCLCA1, which is expressed in smooth muscle – do not correlate with those of native IClCa recorded in the same cell type [11]. Moreover, the predicted topology of proteins encoded by CLCA genes (i.e. between one and five transmembrane domains, depending on the hydropathy programme) is inconsistent with these proteins forming ion channels [12–14]. It is now thought that these proteins are secreted by cells [13,14], and they have been implicated as cell adhesion molecules in the development of various tumours and asthma [1,6]. Alternatively CLCA proteins might function as signalling mediators because they augment the activity of cAMP-dependent Cl channels [6,14,15] and undergo numerous protein–protein interactions [8,15]: for example, with an ancillary b subunit associated with Ca2+-dependent K+ channels [16]. It is now speculated that CLCA gene products are regulators of channel expression and function [6,13–15]. Pharmacology of Ca2+-activated ClS channels in smooth muscle cells The inability of researchers to determine the identity of the protein underlying IClCa is compounded by the fact that there is no effective blocker to act as a probe for this www.sciencedirect.com

conductance, although a range of structurally different agents has been identified as Cl channel blockers [2–4] (Table 1). The pharmacology of Cl channels is plagued by low potency, low selectivity (Table 1) and tissue variability, but researchers still use Cl channel blockers as probes for ClCa in functional studies without due consideration of these caveats. Do these pharmacological observations reflect nonspecific cross-reactivity or is there a common structural basis for these observations? In smooth muscle cells, several Cl channel blockers – including fenamates, anthracene-9carboxylic acid (A-9-C), indanyloxyacetic acid (IAA94) and ethacrynic acid – activate large-conductance Ca2+-activated K+ channels (BKCa) (Box 1 and Table 1) at similar concentrations to those required to block IClCa [17,18]. Tamoxifen, which inhibits IClCa in endothelial cells, also activates BKCa in excised patches from various smooth muscle cells via an effect on the auxiliary b subunit [19]. Fenamates such as niflumic, flufenamic and mefenamic acids also activate BKCa in lipid bilayers [20]. Interestingly, recent studies of IClCa in vascular myocytes showed that Cl channel blockers such as niflumic acid (NFA) and A-9-C stimulate persistently activated ClCa, resulting in augmented IClCa [21–23]. Although these effects have not been reported for other cell types, the similarity among native IClCa in smooth muscle cells, secretory cells and endothelial cells indicates that this is a real phenomenon. The presence of stimulatory sites on ClCa [22] and the relative predominance of this modulation over the inhibitory effect of NFA could explain why the efficacy of Cl channel blockers differs between different smooth muscle cell types.

Opinion

TRENDS in Pharmacological Sciences

Vol.28 No.1

3

Box 1. Functional and molecular properties of BKCa channels BKCa denotes ‘big K+ channel activated by internal Ca2+’. This widely used acronym is based on the extremely high flux of K+ ions through this channel when it opens; with [K+] at 150 mM on both sides of the membrane, the channel exhibits a unitary conductance of 200 pS [26]. These voltage- and time-dependent K+ channels are activated by membrane depolarization and are modulated by intracellular Ca2+, which sensitizes the channels to voltage by shifting their open probability towards potentials that are more negative. Physiologically, the opening of these channels by Ca2+ influx across the plasma membrane, by Ca2+-induced Ca2+ release (CICR) or by spontaneously occurring Ca2+ sparks mediated by Ca2+ release through ryanodine receptors in the endoplasmic reticulum (ER) or sarcoplasmic reticulum (SR) acts as a negative-feedback mechanism that promotes membrane repolarization in action-potential-generating cells or hyperpolarization in cells responding to tonic changes in membrane potential. The pore-forming a subunit of BKCa channels

is encoded by a single gene (Slo1) that encodes a protein with seven transmembrane-spanning regions (S0–S6) (Figure I). The channel protein is characterized by a short extracellular N-terminus, a voltage-sensing domain (S4), a pore (or P-loop) domain located between segments S5 and S6, and a long intracellular C-terminus that comprises four relatively hydrophobic domains (S7–S10) and a stretch of acidic amino acid residues near the S10 segment that is speculated to be the Ca2+ sensor (‘Ca2+ bowl’ in Figure I); in addition, numerous regions are involved in regulation by phosphorylation. The a subunit is regulated by one of four distinct regulatory b subunits (b1–b4), which contain two transmembrane-spanning segments (T1 and T2 in Figure I) that are encoded by separate genes and that increase the sensitivity of the channel to internal Ca2+. The channel comprises four a and four b subunits, with each b subunit located in one of four grooves formed by the homotetrameric poreforming a subunits (top view in Figure I).

Figure I. Functional and molecular properties of BKCa channels.

The overlap in BKCa and ClCa pharmacology is supported by recent findings that the structurally dissimilar agents 1,3-dihydro-1-[2-hydroxy-5-(trifluoromethyl)phenyl]-5-trifluoromethyl-2H-benzimidazol-2-one (NS1619) and isopimarene, which are considered to be selective activators of BKCa, also enhance IClCa in vascular myocytes [24] at concentrations similar to those affecting BKCa [25,26]. These agents stimulate BKCa via an interaction with the pore-forming a subunit encoded by the Slo1 gene [25], not through an effect on the b subunit. Moreover, the effects of NS1619 on IClCa [24] are reversed by the specific BKCa blockers paxilline and penitrem A [26], which also inhibit control IClCa and abrogate the stimulatory effect of the Cl channel blocker NFA on this conductance [24]. These experiments do not demonstrate conclusively that there is a structural similarity between ClCa and BKCa www.sciencedirect.com

but, in association with the observation that Cl channel blockers activate BKCa, indicate that a homologous motif could exist. Intriguingly, regarding the probability that CLCA gene products augment an endogenous Cl channel with pore properties similar to native IClCa, it has been shown that mCLCA1 expression products physically interact with the b1 subunit, which is normally associated with BKCa [16] (Box 1). Molecular identity revisited Although the possibility of nonspecific interactions cannot be ruled out, the pharmacological profile generated over the past decade, including the most recent data, indicates that part of that channel complex could be similar to BKCa (Figure 1) or that BKCa itself might act not only as a cation channel but also as a regulator of physically associated

4

Opinion

TRENDS in Pharmacological Sciences

Vol.28 No.1

Figure 1. Summary of the pharmacological data regarding native IClCa and a putative structural model for the native channel. (a) Representations of native Ca2+-activated Cl channel that is activated either by a transient increment in [Ca2+] (i) or by a sustained increase in [Ca2+] (ii). An example of currents evoked by this technique is shown to the left of each model (scalar = 100 ms in each case). These illustrations highlight the dual effects of Cl channel blockers (CCBs) and that the stimulatory effect of these agents becomes more predominant with greater stimulation of the Cl channel. Sustained IClCa is augmented by the BKCa activators NS1619 and isopimaric acid (IpA) (ii). (b) BKCa channels encoded by the Slo1 gene are modulated positively by CCBs, in addition to NS1619 and IpA. (c) Theoretical scheme for the putative structure of the channel underlying native IClCa. Colours relate to the schemes shown in (a,b), where the native Cl channel is represented in blue and the native BKCa channel is represented in brown. The two models are based on the accrued pharmacological data that indicate that the native ClCa channel is formed from a distinct protein with a region that has a high similarity to the BKCa channel (left-hand model) or comprises a BKCa protein background but with a Cl -permeable pore (right-hand model).

ClCa. An amino acid sequence alignment of mouse Slo1 (mSlo1) with all mouse members of the CLCA or Bestrophin family of genes revealed no homology. The recently identified tweety gene encodes a Cl channel protein with homology to BKCa [27,28]. However, the Cl channel generated by the expression of tweety exhibits a unitary conductance 50–100-fold larger than smooth muscle IClCa and lacks the characteristic time-dependent kinetics [27,28]. Another possibility is that the mouse gene encoding BKCa, mSlo1, also encodes ClCa. Although mammalian Slo1 has numerous splicing sites [26], none of the splice variants identified experimentally correlates with a Cl sensitive pore, indicating that products of this gene cannot underlie IClCa. There are still considerable challenges ahead for those researchers who have been beguiled by the elusive ClCa. Because most ion channels seem to be the product of one or more pore-forming proteins associated with ancillary subunits, it is likely that the channel underlying IClCa is a heteromultimer. Could the identity of native ClCa lie in a www.sciencedirect.com

combination of an anion pore in a K+ channel background (Figure 1)? Future lines of investigation that would consolidate the pharmacological findings so far include experiments with mSlo1-deficient transgenic mice, determination of whether the IClCa generated by the expression of CLCA, Bestrophin or tweety genes is modified by the application of BKCa blockers and activators, and whether native IClCa can be recreated by the co-expression of these genes with Slo1. There are few clues about the identity of ClCa. It is unknown whether the native channel is multimeric or formed from one protein. Moreover, it is unknown whether ancillary subunits are associated, as is the case with native BKCa. However, the pharmacological discoveries made during the past ten years indicate that the Cl channel might be formed by a protein containing domains that are similar to that encoded by Slo1 but with a Cl -permeable pore or that it might be formed from a distinct protein that forms an anion-permeable pore but which has a region with a high homology to BKCa protein (Figure 1).

Opinion

TRENDS in Pharmacological Sciences

Acknowledgements This work was supported by grants to N.L. from the NIH (HL 1 R01 and HL075477–01) and to I.A.G. by the British Heart Foundation and The Wellcome Trust. This publication was also made possible by grant NCRR 5P20 RR15581 (to N.L.) from the National Center for Research Resources (NCRR), a component of the NIH that supports COBRE at the University of Nevada School of Medicine. The contents of the article are solely the responsibility of the authors and do not necessarily represent the official views of the NCRR or the NIH.

References 1 Jentsch, T.J. et al. (2002) Molecular structure and physiological function of chloride channels. Physiol. Rev. 82, 503–568 2 Frings, S. et al. (2000) Neuronal Ca2+-activated Cl channels – homing in on an elusive channel species. Prog. Neurobiol. 60, 247–289 3 Hartzell, C. et al. (2005) Calcium-activated chloride channels. Annu. Rev. Physiol. 67, 719–758 4 Leblanc, N. et al. (2005) Regulation of calcium-activated chloride channels – a complex picture emerges. Can. J. Physiol. Pharmacol. 83, 541–556 5 Huang, P. et al. (2001) Regulation of human CLC-3 channels by multifunctional Ca2+/calmodulin-dependent protein kinase. J. Biol. Chem. 276, 20093–20100 6 Loewen, M.E. and Forsyth, G.W. (2005) Structure and function of CLCA proteins. Physiol. Rev. 85, 1061–1092 7 Tsunenari, T. et al. (2003) Structure–function analysis of the bestrophin family of anion channels. J. Biol. Chem. 278, 41114–41125 8 Qu, Z.Q. et al. (2004) Mouse bestrophin-2 is a bona fide Cl channel: identification of a residue important in anion binding and conduction. J. Gen. Physiol. 123, 327–340 9 Chien, L.T. et al. (2006) Single Cl channels activated by Ca2+ in Drosophila S2 cells are mediated by bestrophins. J. Gen. Physiol. 128, 247–259 10 Qu, Z. et al. (2006) The anion-selective pore of the bestrophins, a family of chloride channels associated with retinal degeneration. J. Neurosci. 26, 5411–5419 11 Britton, F. et al. (2002) Comparison of the properties of CLCA1 generated currents and ICl(Ca) in murine portal vein smooth muscle cells. J. Physiol. 539, 107–117 12 Elble, R.C. et al. (2006) The putative chloride channel hCLCA2 has a single C-terminal transmembrane segment. J. Biol. Chem. 281, 29448– 29454 13 Gibson, A. et al. (2005) HCLCA1 and MCLCA3 are secreted non-integral membrane proteins and therefore are not ion channels. J. Biol. Chem. 280, 27205–27212 14 Mundhenk, L. et al. (2006) Both cleavage products of the mCLCA3 protein are secreted soluble proteins. J. Biol. Chem. 281, 30072–30080 15 Loewen, M.E. et al. (2004) pCLCA1 lacks inherent chloride channel activity in an epithelial colon carcinoma cell line. Am. J. Physiol. 287, G33–G41 16 Greenwood, I.A. et al. (2002) The large conductance potassium channel b-subunit can interact with and modulate the functional properties of a calcium-activated chloride channel, CLCA1. J. Biol. Chem. 277, 22119– 22122 17 Greenwood, I.A. and Large, W.A. (1995) Comparison of the effects of the fenamates on Ca-activated chloride and potassium currents in rabbit portal vein smooth muscle cells. Br. J. Pharmacol. 116, 2939– 2948 18 Toma, C. et al. (1996) Activation of potassium currents by inhibitors of calcium-activated chloride conductance in rabbit portal vein smooth muscle cells. Br. J. Pharmacol. 118, 513–520

www.sciencedirect.com

Vol.28 No.1

5

19 Dick, G.M. et al. (2001) Tamoxifen activates smooth muscle BK channels through the regulatory b1 subunit. J. Biol. Chem. 276, 34594–34599 20 Ottolia, M. and Toro, L. (1994) Potentiation of large conductance KCa channels by niflumic, flufenamic, and mefenamic acids. Biophys. J. 67, 2272–2279 21 Piper, A.S. et al. (2002) Dual effect of blocking agents on Ca2+-activated Cl currents in rabbit pulmonary artery smooth muscle cells. J. Physiol. 593, 117–131 22 Ledoux, J. et al. (2005) Dynamics of Ca2+-dependent Cl channel modulation by niflumic acid in rabbit coronary arterial myocytes. Mol. Pharmacol. 67, 163–173 23 Piper, A.S. and Greenwood, I.A. (2003) Anomalous effect of anthracene9-carboxylic acid on calcium-activated chloride currents in rabbit pulmonary artery smooth muscle cells. Br. J. Pharmacol. 138, 31–38 24 Leblanc, N. et al. (2006) BKCa channel modulators influence Ca2+activated Cl channels and tone in vascular smooth muscle cells. Biophys. J. 90, 537a 25 Imaizumi, Y. et al. (2002) Molecular basis of pimarane compounds as novel activators of large-conductance Ca2+-activated K+ channel asubunit. Mol. Pharmacol. 62, 836–846 26 Ghatta, S. et al. (2006) Large-conductance, calcium-activated potassium channels: structural and functional implications. Pharmacol. Ther. 110, 103–116 27 Suzuki, M. and Mizuno, A. (2004) A novel human Cl channel family related to Drosophila flightless locus. J. Biol. Chem. 279, 22461–22468 28 Suzuki, M. (2006) The Drosophila tweety family: molecular candidates for large-conductance Ca2+-activated Cl channels. Exp. Physiol. 91, 141–147 29 Greenwood, I.A. and Large, W.A. (1998) Properties of a chloride current activated by cell swelling in rabbit portal vein vascular smooth muscle cells. Am. J. Physiol. 275, H1524–H1532 30 Teramoto, N. et al. (2003) Multiple effects of mefenamic acid on K+ currents in smooth muscle cells from pig proximal urethra. Br. J. Pharmacol. 140, 1341–1350 31 Doughty, J.M. et al. (1998) Non-specificity of chloride channel blockers in rat cerebral arteries: block of the L-type calcium channel. J. Physiol. 507, 433–439 32 Toland, H.M. et al. (2000) Ca2+-activated Cl current in sheep lymphatic smooth muscle. Am. J. Physiol. 279, C1327–C1335 33 Accili, E.A. and DiFrancesco, D. (1996) Inhibition of the hyperpolarization-activated current (if) of rabbit SA node myocytes by niflumic acid. Pflugers Arch. 431, 757–762 34 Wang, H.S. et al. (1997) Unexpected and differential effects of Cl channel blockers on the Kv4.3 and Kv4. 2 K+ channels. Implications for the study of the Ito2 current. Circ. Res. 81, 711–718 35 Busch, A.E. et al. (1994) Positive regulation by chloride channel blockers of IsK channels expressed in Xenopus oocytes. Mol. Pharmacol. 46, 750–753 36 Farrugia, G. et al. (1993) Potassium current in circular smooth muscle of human jejunum activated by fenamates. Am. J. Physiol. 265, G873– G879 37 Peretz, A. et al. (2005) Meclofenamic acid and diclofenac, novel templates of KCNQ2/Q3 potassium channel openers, depress cortical neuron activity and exhibit anticonvulsant properties. Mol. Pharmacol. 67, 1053–1066 38 Ehring, G.R. et al. (1994) Swelling-activated chloride channels in multidrug-sensitive and -resistant cells. J. Gen. Physiol. 104, 1129–1161 39 Fioretti, B. et al. (2004) NPPB block of the intermediate-conductance Ca2+-activated K+ channel. Eur. J. Pharmacol. 497, 1–6 40 Wang, W. et al. (2005) Activating cystic fibrosis transmembrane conductance regulator channels with pore blocker analogs. J. Biol. Chem. 280, 23622–23630