Non-muscle myosin II and myosin light chain kinase are downstream targets for vasopressin signaling in the renal collecting duct.

We have previously demonstrated that vasopressin increases the water permeability of the inner medullary collecting duct (IMCD) by inducing trafficking of aquaporin-2 to the apical plasma membrane and that this response is dependent on intracellular calcium mobilization and calmodulin activation. Here, we address the hypothesis that this water permeability response is mediated in part through activation of the calcium/calmodulin-dependent myosin light chain kinase (MLCK) and regulation of non-muscle myosin II. Immunoblotting and immunocytochemistry demonstrated the presence of MLCK, the myosin regulatory light chain (MLC), and the IIA and IIB isoforms of the non-muscle myosin heavy chain in rat IMCD cells. Two-dimensional electrophoresis and matrix-assisted laser desorption ionization time-of-flight mass spectrometry identified two isoforms of MLC, both of which also exist in phosphorylated and non-phosphorylated forms. 32P incubation of the inner medulla followed by autoradiography of two-dimensional gels demonstrated increased 32P labeling of both isoforms in response to the V2 receptor agonist [deamino-Cys1,D-Arg8]vasopressin (DDAVP). Time course studies of MLC phosphorylation in IMCD suspensions (using immunoblotting with anti-phospho-MLC antibodies) showed that the increase in phosphorylation could be detected as early as 30 s after exposure to vasopressin. The MLCK inhibitor ML-7 blocked the DDAVP-induced MLC phosphorylation and substantially reduced [Arg8]vasopressin (AVP)-stimulated water permeability. AVP-induced MLC phosphorylation was associated with a rearrangement of actin filaments (Alexa Fluor 568-phalloidin) in primary cultures of IMCD cells. These results demonstrate that MLC phosphorylation by MLCK represents a downstream effect of AVP-activated calcium/calmodulin signaling in IMCD cells and point to a role for non-muscle myosin II in regulation of water permeability by vasopressin.

Vasopressin regulates the water permeability of the renal collecting duct epithelium in part by inducing translocation of aquaporin-2-containing intracellular vesicles to the apical region of the principal cells, where they fuse with the apical plasma membrane (1). The resulting increase in water permeability accelerates water reabsorption from the tubule lumen to the blood. Regulation of the water permeability of the collecting duct is a key element of the homeostatic process whereby vasopressin regulates body water balance.
Vasopressin acts in collecting duct principal cells by binding to the V 2 receptor located in the basolateral membrane. Ligand binding activates adenylyl cyclase VI (2) via the heterotrimeric G-protein G s , resulting in an increase in cAMP production. In addition, V 2 receptor occupation by vasopressin elicits an increase in intracellular calcium (3)(4)(5)(6)(7). The increase in intracellular calcium is seen at physiological concentrations of vasopressin and can be elicited in response to exposure of the cells to cell-permeant forms of cAMP (7), suggesting that vasopressin-induced calcium mobilization is triggered by cAMP, presumably via protein kinase A activation. Recent confocal imaging studies of the inner medullary collecting duct (IMCD) 1 have shown that the vasopressin-induced increase in intracellular calcium is oscillatory in nature (8). The calcium release is mediated by ryanodine-sensitive intracellular calcium channels in the collecting duct cells (7).
Isolated perfused tubule studies have demonstrated that the ability of vasopressin to increase water permeability in the IMCD is blocked by intracellular Ca 2ϩ chelation with BAPTA or by blockade of ryanodine-sensitive Ca 2ϩ channels by ryanodine, indicating that vasopressin-induced Ca 2ϩ mobilization is important for translocation of aquaporin-2 and the water permeability response in the IMCD (7). Furthermore, inhibitors of calmodulin, viz. W-7 and trifluoroperazine, cause a reversible inhibition of vasopressin-dependent water permeability and aquaporin-2 trafficking, implicating calmodulin in the vasopressin response (7). Calmodulin inhibitors also interfere with the ability of vasopressin to increase water permeability in toad urinary bladder, an amphibian analog of the renal collecting duct (9).
Calmodulin is a ubiquitous calcium-dependent regulatory protein with multiple regulatory targets (10). One group of well characterized regulatory targets is the calmodulin-regulated kinases (11). We hypothesize that one of these kinases, myosin light chain kinase (MLCK), is a regulatory target for calmodulin in the IMCD and plays a role in mediating the water permeability response to vasopressin. MLCK regulates actin filament organization in many cell types by phosphorylating the regulatory light chain of myosin II (11). Myosins constitute a large superfamily of ATP-dependent molecular motor proteins that interact with actin microfilaments (12). A role for the actin cytoskeleton in the vasopressin action to increase water permeability has been proposed based on studies demonstrating that F-actin-depolymerizing agents, viz. cytochalasins, decrease the osmotic water permeability response in toad urinary bladder to vasopressin (13)(14)(15)(16). Studies by Hays and co-workers have demonstrated that vasopressin and cAMP depolymerize F-actin in the apical region of toad urinary bladder (17)(18)(19) and IMCD cells (20). They have proposed from these studies that a dense cortical matrix of microfilaments in the subapical region of collecting duct cells constitutes a barrier blocking the access of water channel vesicles to the apical plasma membrane and that vasopressin stimulates fusion of these vesicles with the apical plasma membrane by depolymerizing this matrix and eliminating the barrier. These observations present a conundrum: if vasopressin acts to increase water permeability by depolymerizing subapical F-actin, then why does cytochalasin, which acts by depolymerizing actin, decrease rather than increase osmotic water permeability? The answer proposed by Hays and co-workers (16) is that actin microfilaments have at least two roles in collecting duct principal cells corresponding to two or more hypothetical pools of F-actin. 1) One pool is proposed to be involved in the subapical barrier function discussed above. 2) The other pools are proposed to be involved in migration of water channel vesicles from diverse regions of the cytoplasm to the subapical region, a process that undoubtedly depends on microtubules as well. Depolymerization of the former pool by cytochalasins would tend to enhance water permeability; depolymerization of the latter pools would tend to decrease water permeability. Because these actions are in series, overall inhibition might dominate if the long-range migration of water channel vesicles to the subapical region is rate-limiting. It is in the context of these considerations that we contemplate the potential role of actinbased motors, the myosins, in regulation of aquaporin-2 (AQP2) trafficking.
Myosin II is one class of myosins that is likely to be involved in regulation of AQP2 trafficking. Myosin II proteins are multimeric, with two heavy chains of 170 -240 kDa and two pairs of light chains (16 -23 kDa) called the essential and regulatory light chains (12). In cardiac and skeletal muscle, myosin II assembles into stable myofibrils involved in muscle contraction. In non-muscle cells, myosin II isoforms form transient filaments in different locations in the cell, including contractile bundles in stress fibers, gel-like networks in the cortical web, and tight bundles in filopodia and microvilli (21). They are involved in diverse processes in the cell, including cytokinesis, cell motility, and cell shape maintenance. Myosin II is regulated chiefly by phosphorylation of its regulatory light chain by MLCK. Phosphorylation of Ser 19 at the N terminus of the myosin regulatory light chain (MLC) has been shown to increase the ability of myosin II to assemble into bipolar filaments in part by increasing actin-activated Mg 2ϩ -ATPase activity, resulting in an enhancement of myosin-actin filament interaction and the formation of actin fibers responsible for maintaining the cell architecture. Here, we use immunochemical methods, mass spectrometry-based proteomics methods, phospho-MLC-specific antibodies, and water permeability measurements in isolated perfused tubules to investigate the potential role of myosin II, MLC, and MLCK in the water permeability response to vasopressin in the rat IMCD.

EXPERIMENTAL PROCEDURES
Animals-Pathogen-free male Sprague-Dawley rats (80 -120 g; Taconic Farms Inc., Germantown, NY) were used for osmotic water permeability measurements in isolated perfused tubule experiments and in immunoblot experiments for detection of myosin II isoforms, MLCK, and MLC in the IMCD. Vasopressin-deficient Brattleboro rats (180 -220 g; Harlan Sprague Dawley Inc., Indianapolis, IN) were used in experiments for detection of vasopressin-stimulated MLC phosphorylation using phospho-specific antibodies. Male Hannover-Wistar rats (Møllegaard) were used for immunohistochemistry and studies of MLC phosphorylation using 32 P autoradiography. All rats had ad libitum access to water and regular rat chow before experiments.
Perfusion of Isolated Renal Tubules-IMCD segments were microdissected from kidney inner medullas with Dumont No. 5 forceps without enzymatic pretreatment. The dissection solution was the same as the perfusate solution and contained 120 mM NaCl, 2 mM K 2 HPO 4 , 5 mM KCl, 25 mM NaHCO 3 , 2 mM CaCl 2 , 1.2 mM MgSO 4 , and 5.5 mM glucose (290 mosM). The tubules were placed on miniature glass pipettes and perfused in vitro at 37°C following the method described originally by Burg et al. (26). Osmotic water permeability was determined by measuring the transepithelial flux of water resulting from an imposed transepithelial bath-to-lumen gradient of 200 mosM. The lumen was perfused with the physiological perfusate solution described above, and the peritubular bath solution was the same as the perfusate solution except for the addition of 111 mM NaCl, raising the osmolality to 490 mosM. 1 mM fluorescein sulfonate (Molecular Probes, Inc., Eugene, OR) was added to the luminal perfusate as an impermeant luminal marker that will be concentrated upon movement of water from the lumen to the peritubular bath. Fluorescein sulfonate concentrations in the perfusate and collected fluid were measured by continuous flow fluorometry (27), allowing calculation of transepithelial water flux and the osmotic water permeability coefficient (P f ) according to the equation of Al Zahid et al. (28).
Preparation of IMCD Suspensions-IMCD suspensions were prepared as described previously (29). Inner medullas were dissected from rat kidney, minced, and digested by incubation in dissection fluid containing collagenase B (2 mg/ml; Roche Applied Science) and hyaluronidase (600 units/ml; Worthington) with continuous mixing. One-third of the inner medullary suspension was collected without fractionation ("whole inner medullary"). The remaining two-thirds of inner medullary suspension was subjected to three low speed centrifugations (each at 80 ϫ g for 30 s) to separate IMCD tubule fragments in the pellets ("IMCD-enriched") from the lighter non-IMCD structures in the supernatants ("non-IMCD"). Samples were homogenized and solubilized in Laemmli buffer (10 mM Tris, pH 6.8, 1.5% SDS, 6% glycerol, 0.05% bromphenol blue, and 40 mM dithiothreitol) before loading for SDS-PAGE.
Immunoblotting-Proteins were resolved by SDS-PAGE on polyacrylamide gels (Bio-Rad) and transferred electrophoretically to nitrocellulose membranes. Blots were blocked for 30 min with 5% nonfat dry milk in wash buffer (42 mM Na 2 HPO 4 , 8 mM NaH 2 PO 4 , 150 mM NaCl, and 0.05% Tween 20, pH 7.5), rinsed, and probed with the respective primary antibodies overnight at 4°C. The immune complexes were detected with horseradish peroxidase-conjugated immunoglobulin G (1:5000 dilution; Pierce). Sites of antibody-antigen reaction were detected by enhanced chemiluminescence (Kirkegaard & Perry Laboratories, Gaithersburg, MD) with exposure to light-sensitive film (XAR-2, Eastman Kodak Co.). Band density was measured using a laser densitometer (Amersham Biosciences).
Two-dimensional Gel Electrophoresis for Fluorescent Labeling of Proteins-IMCD suspensions were solubilized in lysis buffer containing 7 M urea, 2 M thiourea, 4% CHAPS, and 30 mM Tris, pH 8.8. After shearing the DNA by passage through a 21-gauge needle, the lysate was centrifuged at 14,000 ϫ g for 15 min to pellet insoluble material. The sample was then loaded on a single pH 3-10 immobilized pH gradient strip (24 cm; IPGphor, Amersham Biosciences) for isoelectric focusing. After isoelectric focusing, the strips were equilibrated for 15 min in a solution containing 6 M urea, 30% glycerol, 2% SDS, 50 mM Tris, pH 8.8, 0.002% bromphenol blue, and 10 mg/ml dithiothreitol (DTT), followed by a second 15-min equilibration with iodoacetamide (25 mg/ml) instead of DTT. Strips were briefly rinsed with 1ϫ SDS-PAGE buffer and applied to a 12.5% polyacrylamide gel for electrophoresis at a constant 1-2 watts/gel overnight.
The two-dimensional gels were stained sequentially with Pro-Q Diamond and SYPRO Ruby (both from Molecular Probes, Inc.) following the manufacturer's recommended procedures. Spots of interest were picked and digested with trypsin using an automated system (Ettan robotics system, Amersham Biosciences). Proteins were analyzed on a MALDI-TOFpro mass spectrometer operating in positive ion reflectron mode at a 20-kV accelerating potential with 8-shot pulsed extraction enabled. Trypsin autodigestion peaks were used as internal calibrants. Peptide masses were searched against the NCBI non-redundant rat data base using a proprietary implementation of ProFound, a program that calculates the likelihood of the correct identification based on the theoretical number of peptides in a trypsin digest of a given protein and the number of peptides matched. Alkylation of all cysteines and oxidation of some methionines were assumed.
Two-dimensional Gel Electrophoresis and 32 P Autoradiography-Inner medullas dissected from normal rat kidney were preincubated with Krebs bicarbonate buffer (124 mM NaCl, 4 mM KCl, 26 mM NaHCO 3 , 10 mM glucose, 1.5 mM MgSO 4 , 1.5 mM CaCl 2 , 0.25 mM KH 2 PO 4 , and 1 mM sodium butyrate) at 30°C for 15 min, followed by a 15-min incubation with 1 mCi of [ 32 P]orthophosphoric acid in the same solution. After washing, the kidney inner medullas were incubated with or without DDAVP (10 Ϫ7 M) at 30°C for 10 min. 95% O 2 and 5% CO 2 were supplied to the solution during all incubation periods. The inner medullas were then rapidly frozen to Ϫ20°C until homogenization for two-dimensional gel analysis.
Inner medullary samples were homogenized in lysis buffer containing 8.9 M urea, 2% (v/v) Triton X-100, 2% (v/v) immobilized pH gradient buffer, pH 3-10 non-linear, 0.13 M DTT, and 8 mM phenylmethylsulfonyl fluoride. First dimension isoelectric focusing was performed using nonlinear pH 3-10 immobilized pH gradient strips. The immobilized pH gradient strip was rehydrated for 20 h at room temperature in 400 l of protein sample containing ϳ200 g of protein.
Isoelectric focusing was carried out on a Multiphor II electrophoresis unit (Amersham Biosciences) at 500 V for 0.01 h, 500 V for 3 h, 3500 V for 5 h, and 3500 V for 20 h in gradient mode at 17°C using a MultiTemp III thermostatic circulator (Amersham Biosciences). Prior to second dimension SDS-PAGE, the immobilized pH gradient strip was equilibrated twice for 15 min with gentle agitation in 10 ml of equilibration solution (0.1% Tris-HCl, pH 6.8, 5.5 M urea, 0.3% glycerol, 0.035 M SDS, and 0.065 M DTT). DTT was replaced with iodoacetamide (14 mM), and traces of bromphenol blue dye were added to make the second equilibration solution. For the second dimension, SDS-PAGE was carried out using 5% polyacrylamide in the stacking gel and 12% polyacrylamide in the resolving gel. Electrophoresis for the second dimension was performed vertically at a maximum voltage of 50 V and a current of 5 mA for ϳ20 h. Gels were visualized by silver staining using the protocol of Mortz et al. (30) optimized for high sensitivity protein identification by mass spectrometry, following which they were dried and exposed to lightsensitive film. The dry silver-stained gels and the corresponding films were scanned using a Bio-Rad GS-710 calibrated imaging densitometer. The spots were analyzed using Bio-Rad Multi-Analyst software (Version 1.02), which designates a volume (spot area ϫ density) to each spot proportional to the amount of protein. To correct for a slightly higher amount of loaded protein on the control gel compared with the DDAVP gel, a correction factor (N f ) was calculated from x well defined spots (internal standards; x ϭ 13 in Experiment 1 and x ϭ 22 in Experiment 2) on the two gels: N f ϭ 3(spot i(control) /spot i(DDAVP) )/x. This correction factor was applied as a multiplier for the volume of the analyzed spots on the film from the DDAVP-treated sample.
For identification of MLC by mass spectrometry, gels containing protein spots selected for identification were re-swelled in water. The protein spots in question were excised from the gels and subjected to trypsin digestion. Identification was performed by MALDI-TOF mass fingerprinting and post-source decay identification of specific peptides performed as a service by Protagen AG (Bochum, Germany).
Immunocytochemistry in Rat Kidney Tissue-Kidneys from normal male Hannover-Wistar rats were fixed by retrograde perfusion via the aorta with 3% paraformaldehyde in 0.1 M cacodylate buffer, pH 7.4, and post-fixed for 1 h in the same fixative. Kidney slices containing all kidney zones were dehydrated, embedded in paraffin, and subsequently cut in 2-m sections on a Leica rotary microtome. The sections were dewaxed in xylene, followed by rehydration using 99 and 96% ethanol. At this point, the sections were incubated in 0.3% H 2 O 2 in methanol to block endogenous peroxidase activity. After a rinse with 96% ethanol, the sections were rehydrated using 70% ethanol and finally water. To reveal antigens, the sections were placed in 10 mM Tris buffer, pH 9.0, supplemented with 0.5 mM EGTA and heated in a microwave oven for 10 min. Nonspecific binding of immunoglobulin was prevented by incubating the sections in 50 mM NH 4 Cl for 30 min, followed by blocking in phosphate-buffered saline (PBS) supplemented with 1% bovine serum albumin, 0.05% saponin, and 0.2% gelatin. The sections were incubated overnight at 4°C with primary antibodies diluted in 10 mM PBS, pH 7.4, containing 0.1% Triton X-100 and 0.1% bovine serum albumin. Subsequently, the sections were incubated with horseradish peroxidase-linked goat anti-rabbit secondary antibodies (P448, Dako Corp., Glostrup, Denmark). Labeling was visualized by the 3,3Ј-diaminobenzidine technique, and the sections were counterstained using Mayer's hematoxylin.
Cells in each chamber were fixed with 500 l of 4% paraformaldehyde in Dulbecco's PBS for 15 min at room temperature. Cells were washed three times (5 min each) with Tris-buffered saline containing 0.1% Triton X-100 at room temperature and incubated in 3% bovine serum albumin (fraction V) in the same buffer for 1 h. The cells were stained with Alexa Fluor 568-phalloidin (1:200; Molecular Probes, Inc.) for 20 min. After washing, the cells were coated with SlowFade antifade solution (Molecular Probes, Inc.) and covered with a coverslip. The IMCD cells were analyzed on the stage of a Zeiss Model 410 inverted confocal microscope equipped with differential interference contrast optics and a ϫ40 objective lens.

Effects of Latrunculin B and Jasplakinolide on Osmotic
Water Permeability of the IMCD-Prior studies implicating the actin cytoskeleton in regulation of water permeability by vasopressin have been done in toad urinary bladder and cell culture models (see the Introduction). To address whether the actin cytoskeleton plays a clear-cut role in regulation of water permeability in the renal collecting duct, we perfused isolated IMCD segments in vitro and measured the osmotic water permeability coefficient (P f ) using a non-absorbable fluorescent marker, fluorescein sulfonate. Fig. 1A shows the effect of latrunculin B, which inhibits the polymerization of actin by sequestering G-actin monomers (31). 50 M latrunculin B caused a rapid decrease in vasopressin-dependent osmotic water permeability (basal, 127 Ϯ 26 m/s; AVP, 793 Ϯ 103 m/s; and AVP ϩ latrunculin B, 187 Ϯ 34 m/s (n ϭ 6; p Ͻ 0.01)). Fig. 1B shows the effect of jasplakinolide on P f . Jasplakinolide stabilizes F-actin (32) and induces condensation of actin filaments in the cortical region of cells. 1 M jasplakinolide caused a significant 66% reduction in vasopressin-stimulated water permeability (basal, 199 Ϯ 27 m/s; AVP, 636 Ϯ 67 m/s; and AVP ϩ jasplakinolide, 216 Ϯ 54 m/s (n ϭ 4; p Ͻ 0.02)). These results demonstrate that the vasopressin-dependent water permeability in the IMCD is dependent on the state of organization of actin microfilaments in the collecting duct cells.
Effects of Vasopressin, Latrunculin B, and Jasplakinolide on F-actin Organization in IMCD Cells- Fig. 2 shows x-z reconstructions of confocal images of Alexa Fluor 568-phalloidinlabeled primary cultures of IMCD cells exposed to no agent (control), 10 Ϫ9 M AVP, 50 M latrunculin B, or 1 M jasplakinolide. AVP caused a redistribution of F-actin in the cells with a reduction in labeling in the subapical region and near the base of the cells, but Alexa Fluor 568-phalloidin labeling was sustained along the lateral aspect of the cells. The reduction in subapical labeling is consistent with the conclusion of Hays and co-workers (33) that vasopressin causes depolymerization of the actin network in the subapical cortex of IMCD cells. In addition, AVP caused significant cell shape changes. Cell height increased (control, 12.3 Ϯ 0.3 m; and AVP, 15.3 Ϯ 0.6 m (n ϭ 6; p Ͻ 0.001), and there was a change in the configuration of the intercellular space. The latter was highlighted by the actin labeling, which outlined a V-like pattern in the controls (Fig. 2, arrowheads) that closed after vasopressin. These cell shape changes are consistent with the view that vasopressin causes rapid remodeling of the collecting duct cell, associated with reorganization of the actin cytoskeleton. Fig. 2 also demonstrates that latrunculin B and jasplakinolide, at the same concentrations used for the osmotic water permeability measurements in Fig. 1, had the expected effects on the organization of actin filaments. Latrunculin B treatment resulted in an irregular pattern of Alexa Fluor 568-phalloidin labeling consistent with marked disruption of the organization of actin filaments especially in the apical and lateral cell cor-tices. Interpretation of the jasplakinolide labeling is somewhat clouded by the fact that jasplakinolide and phalloidin bind at the same site on F-actin, such that jasplakinolide may be expected to competitively reduce Alexa Fluor 568-phalloidin binding. Nevertheless, it appeared that jasplakinolide caused clear qualitative changes in Alexa Fluor 568-phalloidin labeling, including a loss of the sharp pattern of labeling at the lateral margins of the cell as well as increased labeling throughout the cytoplasm.
Localization of Myosin II Isoforms in the Inner Medulla-To test whether either non-muscle myosin IIA or IIB is expressed in the IMCD, we carried out immunoblotting of IMCD-enriched cell suspensions purified from whole inner medulla by low speed centrifugation (Fig. 3A). The degree of enrichment of this fraction relative to the whole inner medulla and the remnant non-IMCD fraction is illustrated by immunoblots for aquaporin-1 (Henle's loop and vasa recta marker) and AQP2 (collecting duct marker). Immunoreactive myosins IIA and IIB were both readily detectable in the IMCD-enriched fraction. Myosin IIB appears to be selectively expressed in IMCD cells versus other inner medullary cell types, whereas myosin IIA is expressed in both IMCD and non-IMCD cells. Immunocytochemical localization (Fig. 3B) confirmed expression of both myosins IIA and IIB in the IMCD. In contrast to myosin IIA, which was found in both the IMCD and thin limbs, immunoreactivity for myosin IIB in the inner medulla was restricted to the collecting ducts.
MLCK Expression in the IMCD-The smooth muscle MLCK gene produces several transcripts via alternative promotors and alternative splicing (34,35). Two major smooth muscle MLCK proteins have been described, a "short form" (992 amino acids) and a "long form" (1914 amino acids) (34). We used an antibody raised against turkey gizzard MLCK (25) to determine whether smooth muscle MLCK is expressed in the IMCD (Fig. 4). These immunoblots revealed three bands in the IMCDenriched cell fraction. A lower band at ϳ130 kDa had the same mobility as the MLCK detected in skeletal muscle and is con- sistent with the predicted size of the short isoform. The doublet at 202-220 kDa was expressed predominantly in the IMCDenriched fraction of the inner medulla and is consistent with the predicted size of the long isoform. Presumably, the doublet is due to one of the described splicing variants (35). All three bands were ablated after the antibody was pre-adsorbed by turkey gizzard MLCK. Thus, both the short and long forms of smooth muscle MLCK are present in IMCD cells.
Identification of MLC Proteins in the IMCD- Fig. 5A shows an immunoblot for MLC in renal IMCD cells isolated from the inner medulla. A doublet was demonstrated at ϳ20 kDa, suggesting the presence of two distinct MLC proteins, both of which are markedly enriched in the IMCD relative to whole inner medullary or non-IMCD cells.
Detection of Phospho-MLC in the IMCD-To investigate whether phosphorylated isoforms of MLC are present in the IMCD, we carried out two-dimensional electrophoresis with sequential staining using a fluorescent dye (Pro-Q Diamond) that recognizes phosphoproteins (36), followed by a general protein stain (SYPRO Ruby). The superimposed images are shown in Fig. 5B, false-colored with green to represent Pro-Q Diamond staining (phosphoproteins) and with red to represent SYPRO Ruby staining (total proteins). As highlighted in the inset, there was a constellation of four spots, all of which were identified as MLC by MALDI-TOF mass spectrometry (see also the analysis of silver-stained gels discussed below). This constellation was centered at 19 -20 kDa and pI 4.7-4.9, consistent with the values expected for MLC gene products. Two of the spots (with coordinates A-0 and B-0) (Fig. 5B, inset) were labeled with SYPRO Ruby but not with Pro-Q Diamond, indicating non-phosphorylated forms. Spots A-1 and B-1 are putatively phosphorylated forms of spots A-0 and B-0, respectively. The additional faint bands labeled A-2 and B-2 are putatively doubly phosphorylated MLC.
Regulation of MLC Phosphorylation by Vasopressin-To test whether stimulation of the V 2 receptor increases MLC phosphorylation, inner medullary tissue from normal male Hannover-Wistar rats was exposed to DDAVP (10 Ϫ7 M) or vehicle for 10 min in the presence of 32 P, homogenized, and subjected to two-dimensional gel electrophoresis, followed by silver staining (Fig. 6A) and autoradiography (Fig. 6B). MALDI-TOF mass spectrometry identified spots 1 and 2 as MLC. 32 P labeling corresponding to these two spots was increased after DDAVP exposure (Fig. 6C).
Part A in Table I shows the peptides that were identified by MALDI-TOF corresponding to spots 1 and 2 in Fig. 6A. Part B shows the amino acid sequences of three smooth muscle MLC isoforms in rat, identified by BLAST analysis using the B isoform of MLC as the reference sequence. The peptides identified on the silver-stained gel are shown in color (both spots 1 and 2 in red, spot 1 alone in green, and spot 2 alone in blue). The The regulatory light chain of myosin is phosphorylated by MLCK sequentially at two sites, Ser 19 and Thr 18 . Monophosphorylation of MLC at Ser 19 or diphosphorylation at Ser 19 / Thr 18 by MLCK is associated with an increase in the actinactivated Mg 2ϩ -ATPase activity of the heavy chain. To test whether the DDAVP-induced MLC phosphorylation seen in kidney inner medulla is a result of phosphorylation of these two amino acids by MLCK, we used phosphorylation-specific antibodies against monophospho-Ser 19 MLC and diphospho-Ser 19 / Thr 18 MLC for immunoblotting of IMCD proteins. Fig. 7A shows the changes in monophospho-MLC and diphospho-MLC abundance following exposure of IMCD suspensions to 10 Ϫ9 M DDAVP. DDAVP increased monophosphorylation of MLC within 30 s of exposure, and this effect was sustained for up to 2 min after DDAVP stimulation. DDAVP increased diphosphorylation of MLC at later time points, and labeling with the diphospho-specific antibody was significantly increased at 5 min after DDAVP stimulation. Equal loading among lanes was confirmed by densitometry of Coomassie Blue-stained SDSpolyacrylamide gels loaded with the same samples. DDAVP also caused a sustained 2-3-fold increase in phosphorylation of AQP2 at Ser 256 throughout the 30-min experimental period. AQP2, indicating that it does not inhibit protein kinase A at the concentration used. These results thus confirm the specificity of the drug effect and demonstrate that phosphorylation of MLC induced by DDAVP is mediated by MLCK.
Effect of MLCK Inhibitors on the Osmotic Water Permeability Coefficient in IMCD Tubules-To seek functional evidence for a role for MLCK in vasopressin signaling in the IMCD, we tested the effects of two MLCK inhibitors on osmotic water permeability in isolated perfused IMCD segments. Fig. 8 shows that, in control tubules, 10 Ϫ10 M vasopressin induced a 6-fold increase in P f from 104 Ϯ 15 to 643 Ϯ 65 m/s. In contrast, in tubules pretreated with the MLCK inhibitor ML-7 at 25 M (Fig. 8,) vasopressin increased P f only to 254 Ϯ 30 m/s, a significant reduction of 60% (n ϭ 4; p Ͻ 0.01) compared with the tubules exposed to vasopressin alone. This concentration of ML-7 was shown in Fig. 7 to have no effect on protein kinase A-mediated phosphorylation of AQP2; and therefore, the inhibition of P f in these experiments is unlikely to be due to an effect of ML-7 on protein kinase A activity in the IMCD. Furthermore, in tubules preincubated with another MLCK inhibitor, ML-9 at 50 M, the P f increase was also inhibited. It increased only to 285 Ϯ 53 m/s, a reduction of 56% (n ϭ 4; p Ͻ 0.02) compared with tubules exposed to vasopressin alone (Fig.  8). The inhibitory effect of ML-7 on P f appeared to be dose-dependent. ML-7 at a lower concentration (5 M) inhibited AVPstimulated P f by 34% (basal, 235 Ϯ 99 m/s; AVP, 698 Ϯ 30 m/s; and AVP ϩ ML-7, 458 Ϯ 32m/s (n ϭ 3; p Ͻ 0.01 versus AVP alone)). These inhibitory effects of MLCK inhibitors demonstrate that the water permeability response to vasopressin is dependent in part on phosphorylation of the myosin light chain via MLCK. DISCUSSION Intracellular transport of organelles is a fundamental process that is essential to many cellular functions. Most organelle movements are driven along microtubules and actin filaments by motor proteins. In general, it is believed that long-distance transport across cells is mediated by microtubule-based motors (dyneins and kinesins), whereas local transport is mediated by actin-based motors (myosins) (37,38). The movement of AQP2containing vesicles from the cell interior to the apical region is an example of organelle transport that undoubtedly involves both types of motors. Already, AQP2 vesicles from collecting duct cells have been shown to associate with the motor protein dynein (39), which moves vesicular cargo in the direction of the minus ends of microtubules, located in epithelial cells near cortical actin network beneath the plasma membrane. Vasopressin causes long-distance transport of AQP2 vesicles from throughout the collecting duct cell to the apical region of the cell (1,40,41), and this translocation is microtubule-dependent (41). However, in other cell types manifesting long-distance translocation of vesicles, e.g. translocation of melanosomes in Xenopus melanophores, both microtubule-based motors and actin-based motors are necessary for the overall translocation process (42,43). Presumably, the actin cytoskeleton and associated myosins are needed to deliver the translocating vesicles to microtubules at the proximal end of the translocation proc- ess and then to move the delivered vesicles from the distal end of the microtubules to the target region of the cell.
Although the myosin family is very large, the various myosin isoforms can be subclassified into two general types (12): 1) unconventional myosins, including myosins I, V, and VI, which can function to translocate vesicles along an actin filament, and 2) conventional myosins (both muscle-specific myosin II and non-muscle myosin II isoforms), which associate to form myosin filaments and function to move and organize actin filaments. There is already ample evidence for the involvement of several unconventional myosins in vesicular trafficking (38,43). In this study, we have focused on the potential role of conventional myosins, specifically non-muscle myosins IIA and IIB, in collecting duct cells as mediators of the water permeability response to vasopressin, a response that occurs as a result of cAMP-and Ca 2ϩ /calmodulin-dependent trafficking of AQP2 (7,44).
The immunochemical data presented in this work show that myosin IIA and IIB heavy chains are expressed in IMCD cells (Fig. 3) and that at least two regulatory light chain gene products are expressed as well (Fig. 5 and Table I). The use of phosphoprotein-specific fluorescent labeling of two-dimensional gels of IMCD cell proteins showed that phosphorylated forms of MLC proteins were present in the IMCD (Fig. 5), and 32 P autoradiography of two-dimensional gels of proteins incubated with 32 P showed that the extent of MLC phosphorylation was increased by DDAVP in IMCD cells (Fig. 6). Immunoblotting with antibodies specific for the Ser 19 -and Ser 19 /Thr 18phosphorylated forms of MLC demonstrated that the vasopressin-induced phosphorylation occurred at sites known to be phosphorylated by MLCK, a Ca 2ϩ /calmodulin-regulated enzyme (Fig. 7). The presence of both the long and short isoforms of smooth muscle MLCK was demonstrated by immunoblotting (Fig. 4). Isolated perfused tubule experiments demonstrated that two selective MLCK inhibitors, ML-7 and ML-9, markedly decreased the water permeability of the collecting ducts (Fig. 8) without a change in the amount of Ser 256 -phosphorylated AQP2, i.e. without inhibiting protein kinase A (Fig. 7B). Thus, we conclude from these studies that non-muscle myosin II (either myosin IIA or IIB or both) plays an important role in the vasopressin-mediated water permeability increase in the IMCD. Furthermore, we conclude that MLCK phosphorylates two MLC isoforms and is an important downstream component of vasopressin signaling in IMCD cells.
These results add to findings from previous studies supporting a role for calcium and calmodulin in the collecting duct response to vasopressin. Vasopressin, working through the V 2 receptor, causes intracellular calcium oscillations (8) in the IMCD, which are dependent on calcium release via the ryanodine-sensitive calcium release channel (7). Both chelation of intracellular calcium and calmodulin inhibitors interfere with the water permeability response and AQP2 trafficking in the IMCD (7). Our study implicates a calcium/calmodulin-dependent enzyme, MLCK, in the overall water permeability response to vasopressin in the IMCD.
Our conclusions appear to contrast with those of Lorenz et al. (45), who concluded from studies in primary cultures of IMCD cells that calcium mobilization plays no role in fusion of AQP2 vesicles with the apical plasma membrane. We propose that culturing IMCD cells may induce changes in the expression of critical proteins involved in trafficking, resulting in phenotypic differences relative to the native IMCD cells used in our study. For example, AQP2 trafficking in primary cultures of IMCD cells is predominantly directed to the basolateral plasma membrane rather than the apical plasma membrane as seen in native IMCD cells. Of course, calcium could play important roles other than regulation of MLCK activity in the IMCD. These may include both calmodulin-dependent and calmodulin-independent processes. One process that is important to water channel trafficking in the IMCD is depolymerization of the cortical actin network (20), which may be dependent in part on the calcium-dependent actin-severing enzymes adseverin and gelsolin (46). These proteins function to cleave actin filaments and to cap the exposed plus end to prevent repolymerization. Actin polymerization in the collecting duct is also controlled by small GTPases of the Rho family (47,48).
What roles might myosins IIA and IIB play in regulation of water permeability in the collecting duct? The increase in water permeability mediated by vasopressin results from regulated trafficking of intracellular AQP2 vesicles to the apical plasma membrane (1). One demonstrated role of non-muscle myosin II in protein trafficking is in the process of vesicle budding from the trans-Golgi network (49,50). Therefore, inhibition of the function of myosin II via MLCK inhibitors or calmodulin inhibitors may reduce the rate of AQP2 transfer from the trans-Golgi network into the endosomal pathway and thereby reduce AQP2 delivery to the plasma membrane. In-deed, both calmodulin inhibitors and ryanodine (which inhibits vasopressin-induced intracellular calcium oscillations) result in accumulation of AQP2 in the perinuclear region of IMCD cells, consistent with a Golgi localization (7). Inhibition of cell metabolism in AQP2-transfected LLC-PK1 cells by lowering the temperature to 20°C also results in AQP2 accumulation in the trans-Golgi network, consistent with the view that budding and translocation of vesicles from the trans-Golgi network may be the rate-limiting step in AQP2 trafficking (51). Another possible role of non-muscle myosin II in AQP2 trafficking is in the apical region of the cell, where myosin II might be involved in organizing actin filaments in a manner conducive to delivery of AQP2 vesicles to the plasma membrane, similar to what has been proposed in the exocytosis of synaptic vesicles (52). A third possibility involves a newly described role of non-muscle myosin II in the structure of the so-called "membrane skeleton" (53,54), which underlies lipid rafts in plasma membranes and has been proposed to participate in protein sorting at the plasma membrane level (55). These regions are the sites of formation of caveolae and internalization of caveolae to form caveosomes (56). A subset of apical membrane proteins are transferred to the apical plasma membrane indirectly through a two-step process involving the initial transfer to the basolateral plasma membrane via the exosome complex, followed by transcytosis initiated by internalization of caveolae from basolateral membrane raft regions (57). Whether AQP2 is routed to the apical plasma membrane via this process has not yet been investigated. If so, the transcytosis component may be dependent on myosin II at the basolateral protein sorting step for internalization of caveolae or for translocation of caveosomes. Furthermore, this indirect targeting model may provide an explanation for the finding in the IMCD (58,59) and in earlier parts of the collecting duct system (60) that a substantial fraction of total cellular AQP2 is "mistargeted" to the basolateral plasma membrane.
Dynamics of MLC Phosphorylation-Studies of the kinetics of the water permeability increase in response to vasopressin in isolated perfused IMCD cells show that the earliest water permeability increase is detectable 30 -40 s after addition of vasopressin (27), which is in accord with the period of rapid MLC phosphorylation detected in our study (Fig. 7A). These response times are consistent with our conclusion that MLC phosphorylation may play a role in the collecting duct water permeability response. However, one dilemma is how the seemingly transient increase in phosphorylation of MLC can explain the prolonged water permeability increase induced by vasopressin. A simple explanation is that transient MLC phosphorylation may serve as a switch to initiate the water permeability response but may not be necessary to sustain the response. This sort of mechanism was inferred from studies of contraction of bovine tracheal smooth muscle in response to neural stimulation (61), where it was shown that, although the monophosphorylation of MLC undergoes a transient increase, the force response is sustained. An alternative explanation is that MLC phosphorylation and dephosphorylation may be a highly dynamic phenomenon with time constants on the same order as those of calcium dynamics in the cell with oscillations occurring with a period of 10 -30 s (8). The oscillatory period of the calcium spikes appears to vary from cell to cell so that the calcium spikes, although initially synchronized after vasopressin exposure, are soon entirely out of phase. Thus, if MLC phosphorylation and dephosphorylation occur in lock step with the calcium oscillations, one would expect that MLC phosphorylation would be initially synchronized but soon reach a desynchronized state so that, at any given time, MLC would be phosphorylated in a minority of cells. Because immunoblotting FIG. 8. Effect of MLCK inhibitors on P f of isolated perfused IMCD segments. After cannulation of isolated IMCD segments, tubules were warmed to 37°C for 40 min before 10 Ϫ10 M AVP was added to the peritubular bath. Osmotic water permeability was measured fluorometrically using a fluorescent luminal marker, fluorescein sulfonate. In all groups, vehicle (0.1% Me 2 SO (DMSO)) or kinase inhibitor was added to the peritubular bath during a 30-min equilibration period and was present in the AVP period throughout the experiment. The three curves show time courses of AVP stimulation of P f in control tubules with vehicle alone, in tubules pretreated with the MLCK inhibitor ML-7 (25 M), and in tubules pretreated with the MLCK inhibitor ML-9 (50 M).
would reflect a weighted average of all cells, increased phosphorylation of MLC would only be detectable early in the response to vasopressin when the cells were synchronized. Thus, it is possible that MLC phosphorylation is oscillatory and sustained, rather than transient.
Vasopressin-induced Cell Shape Changes-The results presented in this study also provide further insight into the possible causes of cell shape changes induced by vasopressin. Previous ultrastructural studies of isolated perfused collecting ducts have consistently demonstrated an increase in cell height in response to vasopressin (62)(63)(64). Because these structural changes were dependent on imposition of an osmotic gradient across the cells, it has been believed that the increase in cell height merely reflects cell swelling due to the rapid entry of water across the apical plasma membrane. However, recent studies of epithelial cell structure have made it clear that there are multiple interactions between the cytoskeleton and the plasma membrane, including focal adhesions, zonula adherens, tight junctions, and the rigid cortical actin network, that resist cell shape changes. For example, the stiffness of the cell cortex of cultured MDCK or NIH-3T3 cells is markedly increased by increases in phospho-MLC through expression of a constitutively active form of MLCK or application of the phosphatase inhibitor okadaic acid; in contrast, cytochalasin D markedly decreases cell cortical stiffness (65,66). Furthermore, expression of the constitutively active form of MLCK decreases the cell volume of NIH-3T3 cells and decreases the cell volume response to hypotonicity (66). Thus, it is likely that vasopressin-induced cell shape changes involve alterations of the cytoskeleton. In this study, analysis of confocal immunofluorescence images of IMCD cells in primary culture revealed a significant increase in cell height in response to vasopressin (Fig. 2). This increase in cell height was associated with a marked rearrangement of actin filaments, as determined by Alexa 568phalloidin staining, from a relatively asymmetrical distribution to a more organized distribution with relocation to the peripheral regions of the cell. Further studies to address the role of the actin cytoskeleton and associated myosins in vasopressin-induced cell shape changes appear to be warranted.