Histone-induced damage of a mammalian epithelium: the role of protein and membrane structure TERI J. KLEINE,1 PETER N. LEWIS,2 AND SIMON A. LEWIS1 of Physiology and Biophysics, University of Texas Medical Branch, Galveston, Texas 77555; and 2Department of Biochemistry, University of Toronto, Toronto, Canada M5S 1A8 1Department Kleine, Teri J., Peter N. Lewis, and Simon A. Lewis. Histone-induced damage of a mammalian epithelium: the role of protein and membrane structure. Am. J. Physiol. 273 (Cell Physiol. 42): C1925–C1936, 1997.—In a previous report [T. J. Kleine, A. Gladfelter, P. N. Lewis, and S. A. Lewis. Am. J. Physiol. 268 (Cell Physiol. 37): C1114–C1125, 1995], we found that the cationic DNA-binding proteins histones H4, H1, and H5 caused a voltage-dependent increase in the transepithelial conductance in rabbit urinary bladder epithelium. In this study, results from lipid bilayer experiments suggest that histones H5-H1 and H4 form variably sized conductive units. Purified fragments of histones H4 and H5 were used to determine the role of histone tertiary structure in inducing conductance. Isolated COOH- and NH2-terminal tails of histone H4, which are random coils, were inactive, whereas the central a-helical domain induced a conductance increase. Although the activities of the central fragment and intact histone H4 were in many ways similar, the doseresponse relationships suggest that the isolated central domain was much less potent than intact histone H4. This suggests than the NH2- and COOH-terminal tails are also important for histone H4 activity. For histone H5, the isolated globular central domain was inactive. Thus the random-coil NH2- and COOH-terminal tails are important for H5 activity as well. These results indicate that histone molecules interact directly with membrane phospholipids to form a channel and that protein tertiary structure and the degree of positive charge play an important role in this activity. tight epithelium; mammalian bladder; toxicity; ion permeability CATIONIC PROTEINS (proteins with a net positive charge) are known to be cytotoxic to eukaryotic cells as well as to possess antimicrobial properties. Examples of such cationic proteins are numerous and include protamine sulfate, histones, major basic protein (MBP), and eosinophil peroxidase (EPO) (6, 7, 13, 19). Although histones are normally contained within the nucleus of the cell, conditions such as cell death and lysis cause the release of histones from the cell. Purified histones have been found to increase cell membrane permeability to small monovalent cations and anions, and it has been proposed that this increase in membrane permeability leads to cell swelling and ultimately cell lysis (7). Thus the release of histones may be pathologically important in conditions of significant cell death, such as that which occurs during the breakdown of sperm. In this regard, Mendizabal and Naftalin (12) demonstrated that human semen was toxic to rat colonic mucosa, resulting in a focal loss of epithelial cells (i.e., a loss of the local barrier function of the colon). In addition, some patients with diabetes mellitus suffer from retrograde ejaculation (a painful disorder), which results in the delivery of semen into the lumen of the bladder. Given the cytotoxic nature of histones on the bladder epithelium, the pain resulting from retrograde ejaculation might be caused by the loss of bladder barrier function, allowing ready access of urine to underlying sensory neurons. The membrane conductances induced by histone share a number of properties with protamine, MBP (unpublished observations), and EPO (unpublished observations), including voltage dependence and reversal by calcium, suggesting a common mechanism among these proteins. Thus the effects of histone on membrane permeability are of interest not only because of a potential pathological role of histones on disruption of the barrier function of colonic and bladder epithelia but also as a model for the mechanism of action of other cationic proteins. In a previous report (7), it was demonstrated that histones H1, H4, and H5 increase apical membrane permeability in rabbit urinary bladder epithelium, which ultimately led to cytotoxicity. The permeabilities were characterized in terms of the dose-response relationship, voltage dependence, ion selectivity, and reversibility. However, it was unclear whether histone was forming a channel in the cell membrane or was instead increasing activity of a native membrane channel. Other questions posed by this earlier study involve defining the structures of the histone molecule that participate in activity. Histones H4 and H5 are two similar yet structurally distinct cationic DNA-binding proteins (Table 1). Both have two random-coil tails that flank a central domain. This central domain of H4 contains an a-helical region that spans amino acids 55–67 (3). A second span from 70–90 has been demonstrated to be a-helical when histone H4 is associated with the nucleosome but not when purified histone H4 polymerizes in solution (8). In contrast, the central domain of histone H5 is globular (2). Another structural difference between histones H4 and H5 is that the COOH-terminal tails of histone H4 can associate into b-sheets to form high-molecular-weight aggregates (8). Although both histone H5 and the isolated globular domain have been shown to self-associate under certain conditions (11), significant aggregation of intact H5 in solution does not occur. In this report, the following questions are addressed. 1) Do histones induce the formation of a channel? and 2) What are the active domain(s) of the histone molecule? The data presented in this study suggest that histones interact with phospholipids to induce the formation of a channel. In addition, the central fragment of histone H4 (amino acids 25–67) is important for channel formation, whereas the random-coil tails are important in potentiating the activity of the channel- 0363-6143/97 $5.00 Copyright r 1997 the American Physiological Society Downloaded from journals.physiology.org/journal/ajpcell (052.090.062.141) on January 12, 2022. C1925 C1926 HISTONE DAMAGES A MAMMALIAN EPITHELIUM Table 1. Characteristics of histones H4 and H5 and their fragments Protein Amino Acids Basic Amino Acids Positive Charge, % Acidic Amino Acids Mol Wt Histone H4 NH2-tail a-Helix COOH-tail 1 COOH-tail 2 Histone H5 Globular H5 1–102 1–23 25–67 69–102 86–102 1–190 22–102 23 8 10 6 3 66 16 22.5 33.3 23.3 17.6 17.6 34.6 19.7 7 0 3 2 0 5 4 11,300 2,366 4,696 4,003 1,886 20,900 8,900 No. of positive charges includes all arginine and unmethylated lysine residues. Negatively charged amino acids include glutamic acid and aspartic acid. Data on histone H4 fragments are from a previously published report by Lewis et al. (8). with nitrogen gas, and the lipids were resuspended in decane. Bilayer formation has been described previously (5). The lipids were painted across a 100–200 µM aperture of a Delrin cup until a bilayer with a capacitance of 200–500 pF was formed. Experiments were performed in symmetric 150 mM KCl in twice-distilled water at room temperature. Histone was added to the solution in the cis-chamber, and the trans-chamber was defined as a ground. The solutions in both chambers were stirred with magnetic stirring bars. Voltage was passed, and the resulting current (I) was measured via Ag-AgCl electrodes connected to a bilayer voltage clamp (5). These were continuously monitored on an oscilloscope as well as passed through a pulse-code modulator (Sony) and recorded on videotape. The data were subsequently digitized and analyzed using pClamp 6.0 (Axon Instruments). Transepithelial Voltage Clamping Experiments forming domains. For histone H5, the isolated central globular domain was inactive, demonstrating the importance of the random-coil tails for the conductive activity of histone H5. MATERIALS AND METHODS Purification of Histones and Histone Fragments Purification of histones H4 (mol wt 11,294) and H5 (mol wt 20,900) has been described previously (8, 9). In brief, histones were separated from chicken erythrocytes by gel filtration using a Bio-Gel P-10 column (150 3 2.7 cm) eluted with 0.02 M HCl. Fractions contained purified H4 and an H5-H1 (4:1) mixture. H5 was separated from H1 by ion-exchange chromatography. Histones were recovered from concentrated column fractions by precipitation with acidified acetone (0.1% HCl) and then washed with acetone and dried under a vacuum. Production of histone H4 fragments has been described previously (8). Intact H4 was cleaved at three aspartic acid residues (amino acids 24, 68, and 85) as follows. Briefly, H4 was dissolved in 0.25 M acetic acid and heated for 6 h at 105°C. The mixture was then fractionated on a Sephadex G-50 column (150 3 2.7 cm) and eluted with 0.02 M HCl. Eluted fractions were monitored at 200 nm and were appropriately pooled. The pooled fractions were then dialyzed against absolute ethanol, and 10 volumes of acidified acetone (0.1% HCl) were then added to precipitate the peptides. The peptides were washed with dry acetone and dried under a vacuum. Fragments 1–23 and 69–102 were separated by preparative electrophoresis (20) using a Sephadex G-10 column (50 3 1.1 cm) with 0.05 M acetic acid buffer at 500 V (1.4 mA) for 3.5 h. Purification of the globular domain of H5 has been previously described (2). First, histone H5 (20 mg/ml) was dissolved in 0.2 M K2SO4 and 50 mM tris(hydroxymethyl)aminomethane · HCl buffer, at pH 8. Next, trypsin was added at an enzyme-to-substrate ratio of 1:1,000 at 20°C for 2 h. The digestion was then quenched using 0.02% 1-chloro-3-tosylamido-7-amino-heptanone, and the globular H5 was isolated using a Sephadex G-50 F column. The sample was then dialyzed against 20 mM HCl and recovered by acetone precipitation. All histones were dissolved in distilled, deionized water to make concentrated stock solutions, which were stored at 0°C. Bilayer Experiments All phospholipids were purchased in chloroform from Avanti Polar Lipids (Alabaster, AL). The chloroform was evaporated Tissue preparation. Urinary bladders were excised from 3-kg male New Zealand White rabbits and were washed in NaCl Ringer (see Solutions below). The smooth muscle was dissected away, and the epithelium was mounted on a ring of 2 cm2 exposed area. The ring was transferred to a temperaturecontrolled, modified Ussing chamber (10) where the serosal side of the epithelium was held against a nylon mesh by a slight excess of solution in the mucosal chamber. Both the mucosal and serosal chambers initially held a bathing solution of NaCl Ringer and were aerated with 95% O2-5% CO2 while integral water jackets maintained the temperature of the bathing solution at 37°C. The mucosal chamber was modified to reduce the volume to 4.5 ml (the serosal chamber volume was 15 ml). This was done to minimize the amount of protein used in the experiments. The serosal chamber was stirred by a magnetic spin bar at the bottom of the chamber while the mucosal chamber was stirred by adding the 95% O2-5% CO2 at the bottom of the chamber and allowing it to bubble upward. Solutions. NaCl Ringer contains (in mM) 111.2 NaCl, 25 NaHCO3, 10 glucose, 5.8 KCl, 2.0 CaCl2, 1.2 KH2PO4, and 1.2 MgSO4. In KCl Ringer, all Na1 salts were substituted with the appropriate K1 salts. Unless otherwise noted, all experiments with histones and histone fragments were performed using KCl Ringer as the mucosal bathing solution. Histones, histone fragments, and amino acid heteropolymers were suspended in distilled H2O as a stock solution that was added in microliter quantities to the mucosal solution. Poly(Lys-Ala) 1:1 and poly(D-Glu-D-Lys) 6:4 were purchased from Sigma (St. Louis, MO). Transepithelial Electrophysiological Methods Electrical measurements. All electrical measurements were made under voltage clamp conditions unless otherwise noted. The transepithelial voltage (Vt ) was measured with Ag-AgCl wires placed adjacent to either side of the epithelium (serosal solution ground) while I was passed from Ag-AgCl electrodes placed in the rear of each hemichamber. Both sets of electrodes were connected to an automatic voltage clamp (Warner Instruments). Transepithelial resistance and its inverse, transepithelial conductance (Gt ), were calculated using Ohm’s law from I required to clamp the epithelium 10 mV from the holding voltage under voltage clamp conditions. Data acquisition. I and voltage outputs of the voltage clamp were connected to an analog-to-digital converter (Axon Instruments) interfaced with a computer that calculated values for resistance and short-circuit current (Isc ). Vt and I were continuously monitored on an oscilloscope. All data were Downloaded from journals.physiology.org/journal/ajpcell (052.090.062.141) on January 12, 2022. HISTONE DAMAGES A MAMMALIAN EPITHELIUM C1927 This plot is linear if the added protein changes only the cell resistance when Vt is clamped to 0 mV. The y-intercept of the line will be equal to the junctional conductance (Gj ), and the slope will be the inverse of the cellular electromotive force (Ec ), which is the sum of the apical and basolateral membrane equivalent batteries. Current-voltage relationship. The steady-state difference current-voltage (I-V) relationship of the protein-induced conductance was calculated using the method of Tzan et al. (19). This method involves measuring the transepithelial I-V relationships in both the absence and the presence of added protein; the difference between these two relationships is the voltage dependence of I flowing through the protein-induced conductance. First, the tissue was voltage clamped to Vt 5 0 mV, and the transepithelial I responses to computergenerated voltage pulses 30 ms long and of increasing magnitude and alternating polarity were measured. Next, the transepithelial potential was voltage clamped to 270 mV, protein was added to the mucosal solution and equilibrated for 3 min, and then the transepithelial potential was clamped to 0 mV. The conductance was allowed to reach a steady state before the I-V relationship was again measured. The difference between the I-V relationships in the presence and absence of added protein was then fit by the constant-field equation to determine the relative ionic permeabilities of the protein-induced conductance. Data analysis and statistics. Curve fitting was done on an IBM-AT using NFIT (Island Products, Galveston, TX). Statistics were calculated using INSTAT (GraphPAD Software, San Diego, CA). Data are shown as means 6 SE. of phospholipid bilayers composed of a 5:3:2 ratio of phosphatidylserine (PS), phosphatidylethanolamine (PE), and phosphatidylcholine (PC) (wt/wt/wt). Voltage was first held at 0 mV (trans-side ground) during an equilibration period of several minutes and then was clamped to either 2100 mV or 100 mV. At either voltage, channels of variable sizes appeared (Fig. 1A). When the voltage polarity was reversed, channels were still evident. The all points histogram of the I tracing demonstrates the variability in the magnitude of the channels (Fig. 1B). In eight bilayers, the channel conductances ranged in size from 4 to 20 pS. H5-H1 (185 nM) was also tested on bilayers composed of only PE. Channel activity was observed at both 1100 mV and 2100 mV (data not shown). However, in contrast to the PS:PE:PC bilayers, only single-channel events with a low probability of opening (0.036 6 0.005, n 5 4) were observed in PE bilayers. This suggests that histones have a stronger affinity for negatively-charged phospholipids, which results in increased channel activity. Histone H4 (89 nM) was also added to bilayers composed of a 5:3:2 ratio of PS:PE:PC. Channels of variable sizes appeared at both 100 mV and 2100 mV (Fig. 2A). In addition to distinct channel openings and closings, there were sporadic increases in noise that could have been the result of channels flickering open and closed. Because of this noisy channel activity, the peaks in the all points histograms were broad and not well resolved (Fig. 2B). Therefore, I amplitudes were determined by measuring each distinct individual opening. The distinct channels were variable in size, ranging from 2 to 15 pS (53 distinct openings, n 5 2). In another bilayer, after histone addition, several large spikes (100–160 pS) appeared followed by breakdown of the bilayer (not shown). These results suggest that histone H4, like histone H5-H1, is capable of forming channels of variable sizes in phospholipid bilayers. RESULTS Activity of the Histone Fragments printed out with the time of data acquisition and were additionally stored on hard disk. Equivalent circuit analysis. The method of Yonath and Civan (21) was used to differentiate between an increase in the conductance of the cell membrane or tight junctions. Gt (µS/cm2 ) was plotted as a function of Isc (in µA/cm2 ) when Vt 5 0 mV in the presence of added protein. This plot was then fit by the equation Gt 5 (Isc /Ec) 1 Gj (1) In this section, we first report the effects of purified histones on phospholipid bilayers. Then the fragments of histones H4 and H5 and their ability to induce an increase in Gt in rabbit urinary bladder epithelium are compared. Finally, the synthetic proteins poly(Lys-Ala) and poly(Glu-Lys) are tested to determine the role of negatively-charged amino acids. Histones Induce a Conductance in Bilayers In previous reports of the conductive effect of histones on rabbit urinary bladder epithelium, it was unclear whether histones were directly inducing channel formation or whether they were indirectly affecting epithelial conductance, for example, by activating a native membrane channel or a second messenger system which would in turn cause an increase in conductance (7). Therefore, purified histones were tested on phospholipid bilayers. First, a 4:1 mixture of histones H5 and H1 (H5-H1) was tested on phospholipid bilayers. Histone H5-H1 (185 nM) was added to the solution bathing the cis-side The activity of whole histones has previously been described in rabbit urinary bladder epithelium (7). In this report, the effects of the histone fragments are tested to identify the active domains. The activity of the fragments is characterized in terms of time course, dose response, voltage sensitivity, ion selectivity, and site of action. Differences between the activity of the histone fragments and the intact histones are also described. Unless otherwise noted, the mucosal solution was a KCl Ringer in all histone and histone fragment experiments. The effect of histone fragments on Gt of rabbit urinary bladder epithelium. A previous report suggested that histones induce a voltage-sensitive conductance in the apical membrane of rabbit urinary bladder (7). In urinary bladder epithelium, the apical membrane has a high resistance to ion flux, whereas the basolateral membrane is quite permeable to K1 and Cl2. Therefore, the voltage across the apical membrane can be controlled by the transepithelial potential. The apical membrane voltage (Va ) is calculated as the difference between Vt and the basolateral membrane voltage (Vb ), Downloaded from journals.physiology.org/journal/ajpcell (052.090.062.141) on January 12, 2022. C1928 HISTONE DAMAGES A MAMMALIAN EPITHELIUM Fig. 1. Histone-induced variably sized channels in phospholipid bilayers. A: current (I) trace demonstrating channel activity of histone H5-H1 in phospholipid bilayer. Bilayer was composed of 5:3:2 ratio of phosphatidylserine (PS), phosphatidylethanolamine (PE), and phosphatidylcholine (PC). Histone H5-H1 (186 nM) was added to cis-chamber, with bilayer voltage clamped to 0 mV (trans-side ground), and then solution was stirred (not shown). Next voltage was clamped to 100 mV. Solution was symmetric 150 mM KCl. Data were filtered at 30 Hz, amplified by a factor of 10, and sampled at 120 Hz. Solid horizontal line indicates I at 0 mV. B: all points histogram for I trace. Data have been corrected for baseline I. Tallest peak corresponds to baseline. Histogram is shown fitted using Marquardt’s least squares fitting algorithm. Best fit peaks were at DI of 0, 0.4, 0.6, and 1.0 pA. which is relatively constant at 255 mV [as previously determined using microelectrodes (10)]. For example, when Vt is clamped at 270 mV (serosa ground), Vb is 255 mV, and therefore Va is 115 mV relative to the mucosal solution. If Vt is clamped to 0 mV, the apical membrane cell potential is 255 mV. Under normal conditions, Gt of the rabbit urinary bladder epithelium displayed little or no voltage sensitivity. However, when either histone H1, H4, or H5 was added to the mucosal solution, there was a rapid, voltage-sensitive increase in Gt when Vt was clamped from 270 mV to 0 mV (7). All three of the histones induced an increase in Gt only when the voltage gradient across the apical membrane was cell interior negative. To identify the histone domains responsible for inducing this voltage-sensitive increase in Gt, five fragments of histones H4 and H5 were tested. The four histone H4 fragments were as follows: a portion of the NH2terminal tail (amino acids 1–23), the remainder of the NH2-terminal tail connected with an a-helical portion of the central domain (25–67), and two fragments from the COOH-terminal tail (69–102 and 86–102). The fifth fragment that was tested was the central globular domain of histone H5-(22–102). A number of the physical characteristics of these fragments is summarized in Table 1. These fragments were tested in the following manner: first, protein was added to the mucosal bathing solution while Vt was clamped at 270 mV, allowed to equilibrate for 3 min, and then clamped to 0 mV. The isolated NH2- and COOH-terminal tails of H4 (fragments 1–23, 69–102, and 86–102) and the globular domain of H5 (fragment 22–102) failed to induce a conductance over a wide range of concentrations when Vt was clamped to 0 mV (time course data not shown; see Dose-response relationship). In contrast, the H4 a-helix fragment (25–67) induced an increase in Gt at Vt 5 0 mV or when the apical membrane potential was cell interior negative. The time course of the response for the 850 nM histone fragment H4-(25–67) is shown in Fig. 3. Data are shown fitted by the equation DGt(t) 5 Gh(1 2 e2k h t) (2) where DGt(t) is the change in the transepithelial conductance as a function of time, normalized to the concentration of histone (µS · cm22 · µM21 ); Gh is the maximal histone-induced conductance, normalized for concentration; and kh (unit of inverse seconds) is the rate constant of the conductance change. This equation has been previously demonstrated to describe the time Downloaded from journals.physiology.org/journal/ajpcell (052.090.062.141) on January 12, 2022. C1929 HISTONE DAMAGES A MAMMALIAN EPITHELIUM Fig. 3. Time course of conductance change in rabbit urinary bladder epithelium induced by H4 fragment 25–67. Histone H4 fragment 25–67 (850 nM) was added to mucosal solution at transepithelial voltage (Vt ) of 270 mV and equilibrated for 3 min. Voltage was clamped to 0 mV at time 0. Smooth curve is best fit of Eq. 2 to data. Best fit value for fragment-induced conductance (Gh ) was 66 µS · cm22 · µM21 and rate constant (k) was 0.09 s21. Mucosal solution was KCl Ringer, and serosal solution was NaCl Ringer. Gt, transepithelial conductance. Fig. 2. Histone H4-induced channels in bilayers. A: histone H4 (89 nM) was added to cis-chamber while voltage was held at 0 mV (trans-side ground), solution was stirred, and then voltage was clamped to indicated voltages. At both 100 mV and 2100 mV, 2 different activities were evident. One was a sporadic increase in noise, which may be a channel flickering open and closed. Also evident were longer openings of various magnitudes. Data were filtered at 100 Hz, amplified by a factor of 10, and sampled at 500 Hz. I at 0 mV is indicated by solid horizontal line. Bilayer was 5:3:2 ratio of PS, PE, and PC. Solution was symmetric 150 mM KCl. B: all points histogram of H4-induced channel activity at 100 mV. Data have been corrected for baseline I. Histogram was generated from I trace at 100 mV using Fetchan (pClamp 6.0). course of the conductance change induced by intact histones H1, H4, and H5 (7). Thus the H4-(25–67) fragment induces a voltage-sensitive increase in Gt that follows a time course similar to intact histone H4. To determine if loss of the NH2- and COOH-terminal tails resulted in a change in either the potency or the speed of the conductance increase, the magnitude (Gh ) and the rate constant (k) of the induced conductances for H4-(25–67) and intact histone H4 were compared. As shown in Table 2, the H4-(25–67)-induced conductance was significantly smaller than that of intact H4 on a micromolar basis. The rate constant was also markedly slower for H4-(25–67). Thus the loss of the NH2- and COOH-terminal tails resulted in a reduction of histone H4 activity. Dose-response relationship. The dose-response relationship also indicated that the histone fragment H4(25–67) was much less potent than intact H4. The magnitude of the induced conductance was determined as a function of concentration for both the fragments and intact histone H4 (Fig. 4). The H4 fragments (1–23), (69–102), and (86–102) did not alter Gt at any of the tested concentrations. The globular domain of H5 also did not elicit a response for concentrations ranging from 102 to 3,700 nM (3 tissues, data not shown). In contrast, the magnitude of the conductance increased as a function of protein concentration for both H4-(25– 67) and intact histone H4. The shape of the doseresponse relationship for the H4-(25–67) fragment is sigmoidal. Such a sigmoidal relationship possibly suggests that several molecules of the fragment combine to induce a conductance. This is in contrast to the dose response of intact histone H4, which is hyperbolic. The dose-response relationships were normalized to minimize tissue variability. For the fragment, conductance was normalized to the conductance induced by 426 nM fragment (21 6 5 µS/cm2, n 5 9). For H4, conductance was first normalized to the conductance Table 2. Comparison of histone H4 and H4(25–67)-induced conductance Protein n kh , s21 Gh , µS · cm22 · µM21 H4-(25–67) Histone H4 7 31 0.03 6 0.007 0.21 6 0.03 71 6 20 1,270 6 150 Values are means 6 SE; n 5 no. of experiments. Comparison of rate constant (kh ) and magnitude (Gh ) of conductance induced by intact histone H4 and H4 fragment (25–67). Both kh and the Gh induced by H4-(25–67) are significantly different from that of intact H4 (P , 0.0001, Welch’s t-test). Values were calculated by fitting time courses of conductance induced by 89 nM intact histone H4 or 426 nM H4-(25–67) by Eq. 2. Downloaded from journals.physiology.org/journal/ajpcell (052.090.062.141) on January 12, 2022. C1930 HISTONE DAMAGES A MAMMALIAN EPITHELIUM voltage than intact histone H4. After protein addition and equilibration, the tissue was clamped to various Vt and the magnitude and rate constant were measured at each voltage. As shown in Fig. 5, A and B, both the magnitude and the rate constant were exponential functions of the transepithelial (and apical) membrane voltage. Values were normalized to the conductance change or rate constant measured at 0 mV to correct for tissue variability. To determine the steepness of the voltage sensitivity for both the magnitude and the rate constant, the data were fit by the following equation Fig. 4. Dose-response relationships for histone H4 fragments and intact histone H4. Varying amounts of protein were added to mucosal solution at Vt 5 270 mV and equilibrated, and then Vt was clamped to 0 mV. Time courses of conductance responses were then fitted by Eq. 2 to determine values for Gh. There was no change of conductance for fragments H4-(1–23) (3 tissues, s), H4-(69–102) (4 tissues, k), or H4-(86–102) (1 tissue, q). For fragment H4-(25–67) (n) and intact histone H4 (r), magnitude of conductance change increased as function of protein concentration, although fragment was much less potent than intact histone H4. For fragment, data from 7 tissues were normalized to magnitude of conductance change elicited by 426 nM H4-(25–67) (21 6 5 µS/cm2, n 5 9). For intact histone H4, data from 14 tissues were normalized to change at 89 nM H4 (96 6 9 µS/cm2, n 5 32) and then again normalized to ratio of maximum increase in transepithelial conductance (Gmax ) for intact histone H4 to Gmax for fragment. When fit by Hill equation, best fit values for H4-(25–67) were: Gmax, 3.47 times conductance change at 426 nM, or 73 µS/cm2; Km, 567 nM; and Hill coefficient (nH ), 2.73 (n 5 9). For intact histone H4, Gmax was 3.9 times the conductance induced by 89 nM H4, or 374 µS/cm2, and Km was 348 nM (n 5 32). induced by 89 nM H4 (96 6 9 µS/cm2, n 5 32) to minimize tissue variability. The conductances for H4 were then normalized to the ratio of the best fit values for maximum conductance change (Gmax ) for intact H4 and H4-(25–67) so that the conductances induced by these proteins could be compared. When fit by the Hill equation, the Gmax for the fragment was 73 µS/cm2 (or 3.47 times greater than the conductance change induced by 426 nM of the fragment), the Michaelis constant (Km ) was 567 nM, and the Hill coefficient (nH ) was 2.73, suggesting that three fragments are necessary to form a conductive unit (n 5 9). For intact H4, Gmax was 374 µS/cm2 (or 3.9 times the conductance induced by 89 nM H4), Km was 348 nM (n 5 32), and nH was 1. The large decrease in both Gmax and Km for the fragment in comparison with intact H4 suggests a decrease in the number of binding sites and in the binding affinity of the fragment. The rate constant of the fragment-induced conductance change was independent of the bath concentration of the fragment. When the fragment concentration was doubled from 426 nM to 852 nM, there was no significant change in the rate constant (P 5 0.9, n 5 4, data not shown). In contrast, for intact histone H4, it has been shown that the rate constant was weakly dependent on histone concentration (7). Voltage sensitivity. Both the magnitude and the rate constant of the H4-(25–67)-induced conductance increased as a function of the applied voltage. In addition, the H4 fragment was more sensitive to the applied Gt(V) 5 G(0)eeNVt /kT (3) where G(0) is the conductance change (µS/cm2 ) at 0 mV (or alternatively, the rate constant at 0 mV), Gt(V) (µS/cm2 ) is the total conductance change (or rate constant) at a particular voltage, N is an empirical constant that indicates the degree of voltage sensitivity, Vt is the transepithelial voltage (mV), e is the electron charge (1.602 3 10219 C), k is the Boltzmann constant (1.38 3 10223 J/K), and T is temperature (310 K). For the magnitude of the conductance change, N was 1.5 for the histone fragment and 0.99 for intact histone H4. This suggests that the magnitude of the conductance induced by the H4-(25–67) fragment was more steeply a function of membrane voltage than that induced by the intact H4 molecule. The rate constant for H4-(25–67) was also voltage sensitive, although it was not as strongly affected by voltage as the magnitude of the conductance change. The best-fit value for N for H4-(25–67) was 0.85 (n 5 7, 4 tissues). This is in contrast to intact histone H4, for which the rate constant was shown to be independent of voltage (7). Site of action. Previously, it has been shown that intact histone H4 predominantly affects the apical, rather than junctional or basolateral, membrane permeability. To determine if the central helical domain has the same site of action, the method of Yonath and Civan (21) was used (see MATERIALS AND METHODS ). While Vt was clamped to 0 mV, the changes in conductance (Gt ) and Isc were monitored. Changes in Gt can occur at three sites: the apical membrane, basolateral membrane, or tight junctions. If the protein-induced conductance increases as a linear function of Isc (see Eq. 1), this suggests that only the cellular conductance (apical and/or basolateral membrane conductance) was increased by histones. The best fit values to Eq. 1 for the fragment are Ec 5 260 6 3 mV and Gj5 28 6 6 µS/cm2 (n 5 6). As shown in Fig. 6, the relationship between the change in Gt and the change in Isc was linear for the a-helical fragment, with a slope similar to that of intact histone H4. This suggests that the fragment acts on the apical membrane and that removal of the tails does not alter the specificity of the site of action. Ion selectivity. The H4-(25–67)-induced conductance was nonselective for K1 and Cl2. The ion selectivity was determined from the steady-state difference I-V relationship (see MATERIALS AND METHODS ). An example of this Downloaded from journals.physiology.org/journal/ajpcell (052.090.062.141) on January 12, 2022. HISTONE DAMAGES A MAMMALIAN EPITHELIUM C1931 Fig. 6. Histone H4 and isolated central domain have same site of action. This was determined from a plot of Gt vs. short-circuit current (Isc ). Linear relationships suggest an increase in conductance of cell membrane rather than tight junction. Data are shown fitted by Eq. 1. Best fit values for intact histone H4 (r): junctional conductance (Gj ), 25 µS/cm2; cellular electromotive force (Ec ), 265 mV. Best fit values for H4-(25–67) fragment (n): Gj, 61 µS/cm2; Ec, 256 mV. relationship, which was linear, is shown in Fig. 7. The data were fit by the constant-field equation to determine the K1 and Cl2 permeabilities. For 426 nM H4-(25–67), the best fit value for the K1 permeability was 3.6 6 1.2 3 1028 cm/s (n 5 4), and the ratio of Cl2 to K1 permeability (PCl/PK ) was 0.8 6 0.2 (n 5 4), indicating that the fragment-induced conductance was nonselective for these two ions. This is in agreement with the I-V relationship for intact histone H4 and suggests that the ion specificity of the induced conductance is conferred by the central helical domain. Fig. 5. Voltage sensitivity of H4-(25–67) fragment. Protein was added to mucosal solution at Vt 5 270 mV and allowed to equilibrate, and then Vt was clamped to more positive potentials. Gh and k were then determined using Eq. 2. Data were normalized to Gh and k at 0 mV. A: magnitude of H4-(25–67)-induced conductance was more voltage sensitive than that of intact histone H4. Data from 4 tissues for the fragment (n) and 13 tissues for H4 (r) were fitted by Eq. 3. Best fit values for empirical constant indicating degree of voltage sensitivity (N) for fragment and H4 were 1.5 and 0.99, respectively, indicating that fragment has a steeper voltage dependence. Conductance change at 0 mV was 58 6 15 µS · cm22 · µM21 (n 5 11) for 426 nM fragment and 1,210 6 153 µS · cm22 · µM21 (n 5 21) for 89 nM histone H4. B: rate constant was voltage sensitive for H4-(25–67). Data from 4 tissues are shown fitted by Eq. 3; best fit value for N was 0.85. Rate constant at 0 mV was 0.03 6 0.007 (n 5 7). Fig. 7. Current-voltage (I-V) relationship of H4-(25–67) fragment. This I-V relationship was generated by subtracting control I-V relationship at 0 mV from I-V curve measured for steady-state conductance induced by 640 nM histone fragment at 0 mV in KCl Ringer. Data were then fitted by constant-field equation to determine K1 and Cl2 permeabilities (see MATERIALS AND METHODS ). Downloaded from journals.physiology.org/journal/ajpcell (052.090.062.141) on January 12, 2022. C1932 HISTONE DAMAGES A MAMMALIAN EPITHELIUM the time for the induced conductance to reverse 90% was measured. For the fragment, the average reversal time was 147 6 14 s (n 5 10). This value is very similar to the 90% reversal time of intact H4, which was reported to be 145 s (7). Thus loss of the tails did not affect the rate of reversal of the induced conductance at 270 mV. Role of Acidic Amino Acids Fig. 8. Voltage-induced reversal of H4-(25–67) conductance. After fragment-induced increase in Gt had reached a plateau at Vt 5 0 mV, voltage was reversed to 270 mV (time 0). This resulted in a large, rapid increase in Gt followed by a biphasic decrease. Smooth curve through data is best fitted by a double exponential equation modeling reversal as 2 conductances in parallel. Best fit values: Gr, 18 µS/cm2; kr, 0.08 s21; Gs, 34 µS/cm2; ks, 0.01 s21 where subscripts r and s are rapid and slow, respectively. Reversibility of the induced conductance. Previously, it has been demonstrated that the intact histone H4induced conductance can be reversed be changing Vt from 0 mV back to 270 mV (7). To determine if the loss of the NH2- and COOH-terminal tails affected voltagedependent reversal, Vt was clamped back to 270 mV after the fragment-induced conductance reached a plateau at 0 mV. The time course of the conductance reversal is shown in Fig. 8. Note that the conductance increased first before decreasing. This response was observed in all of the voltage reversal experiments (n 5 20). The average increase was 31 6 3 µS/cm2. In contrast, only 44% of the intact histone H4 voltage reversals were reported to show this response (7). This initial jump was hypothesized to result from the intact histone H4 partitioning through the cell membrane into the cytoplasm, where it would induce a conductance when the voltage gradient across the apical membrane was cell interior positive. This was further supported by the serosal addition of histone resulting in an increase in apical membrane conductance, with the opposite voltage polarity as mucosal histone, suggesting that histone entered the cell through the basolateral membrane rather than by crossing the tight junctions and entering into the mucosal solution (7). Because the initial jump was more frequent with the H4-(25–67) than with intact histone H4, this suggests that the fragment may partition through the membrane at Vt 5 0 mV more easily than intact histone H4. After the initial jump, when clamping from 0 mV to 270 mV, the reversal of the H4-(25–67)-induced conductance followed the form of a double exponential, consistent with intact histone H4. To determine if the loss of the COOH- and NH2terminal tails affected the rate of reversal at 270 mV, As the fragment experiments demonstrate, the tertiary structure of histones is important for their conductive activity. It has been demonstrated that positively charged amino acids located within the protein molecule are important as well (18). However, the role of negatively charged amino acids within the protein molecule has not been characterized. To determine the role of acidic amino acids within the protein, two synthetic molecules were tested: poly(Lys-Ala) 1:1 and poly(D-Glu-D-Lys) 5.7:4.3. Both of these proteins had a molecular weight that ranged from 20,000 to 50,000; the average molecular weight was 41,600 for poly(LysAla) and 23,000 for poly(Glu-Lys), as determined by viscosity measurements (Sigma). These proteins are similar in that approximately one-half of each protein is composed of cationic amino acids. However, they differ in their net charge density. The charge density is defined as the percentage of the total protein molecule that is composed of similarly charged amino acids (18). The net charge density is the difference between the positive and negative charge densities of the protein. Poly(Lys-Ala) had a net charge density of 50% positive charge, whereas poly(Glu-Lys) had a net 14% negative charge density. The effect of poly(Lys-Ala) and poly(Glu-Lys) on Gt. Each synthetic protein was tested on the urinary bladder epithelium in the same manner as the histone fragments; protein was added to the mucosal solution while Vt was clamped at 270 mV and equilibrated for 3 min. Then Vt was clamped to 0 mV. Typical time courses for each protein are shown in Fig. 9. Note that the concentration of poly(Glu-Lys) was 24-fold higher than poly(Lys-Ala). The time course data for poly(Lys-Ala) were fitted by a double exponential equation that assumes two conductive states in parallel DGt(t) 5 Gp1(1 2 e2k1t) 1 Gp2(1 2 e2k2t) (4) where DGt(t) is the time-dependent change in the transepithelial conductance, normalized to the concentration of added protein (µS · cm22 · µM21 ); Gp1 and Gp2 are the two protein-induced conductances, normalized for concentration (µS · cm22 · µM21 ); t is time; and k1 and k2 are the rate constants (inverse seconds) for entering the respective conductance states. Best-fit values for 72 nM poly(Lys-Ala) are Gp1 5 700 6 180 µS · cm22 · µM21, k1 5 1.0 6 0.2 s21, Gp2 5 4,100 6 750 µS · cm22 · µM21, and k2 5 0.02 6 0.005 s21 (n 5 4). In contrast, the time course for poly(Glu-Lys) was fit by Eq. 2, which is a single exponential equation. Downloaded from journals.physiology.org/journal/ajpcell (052.090.062.141) on January 12, 2022. HISTONE DAMAGES A MAMMALIAN EPITHELIUM C1933 phospholipid bilayers. The results suggest that histone is capable of inducing channel formation and that this activity is enhanced by the presence of anionic phospholipids in the membrane. The importance of histone tertiary structure and the effect of negatively charged amino acids within the protein molecule were explored in an effort to elucidate the structural features of the protein that contribute to activity. Histone Channels and Membrane-Binding Sites Fig. 9. Time course of effects of poly(Lys-Ala) (s) and poly(Glu-Lys) (r) on Gt. Poly(Lys-Ala) (72 nM) or poly(Glu-Lys) (1,740 nM) was added to mucosal solution at Vt 5 270 mV and allowed to equilibrate, and at time 0, Vt was clamped to 0 mV. Smooth curves are best fits of data by Eq. 2 for poly(Glu-Lys) and Eq. 4 for poly(Lys-Ala). Best fit values for poly(Glu-Lys): Gh, 16 µS · cm22 · µM21; k, 0.15 s21. Best fit values for poly(Lys-Ala): Gp1, 820 µS · cm22 · µM21; k1, 1.1 s21; Gp2, 4,900 µS · cm22 · µM21; k2, 0.01 s21; where Gp1 and Gp2 and k1 and k2 are protein-induced conductances and rate constants, respectively, for 2 conductance states in parallel. Data were normalized for protein concentration. Mucosal solution was KCl Ringer, and serosal solution was NaCl Ringer. The results of the phospholipid bilayer experiments are important from two perspectives: first, because channel activity occurred in the presence of histone, this indicates that histone is capable of forming channels; second, the channel activity was increased in anionic phospholipid-containing bilayers compared with bilayers composed only of neutral phospholipids. This suggests that anionic membrane-binding sites, although not required for histone activity, enhance channel formation. This is perhaps a result of the electrostatic interaction between histones and the anionic phospholipid, resulting in an increased amount of histone binding or greater stability of the histonephospholipid interaction. Histones have been reported to preferentially bind to anionic phospholipids (16). Purified histones were demonstrated to bind to the anionic phospholipids cardiolipin and PS with high avidity but not to the zwitterionic phospholipid PC. Histone Fragments Best-fit values for 1,740 nM poly(Glu-Lys) are Gh 5 18 6 6 µS · cm22 · µM21 and k 5 0.17 6 0.03 s21 (n 5 5). The dose-response relationships. Poly(Lys-Ala) was much more potent than poly(Glu-Lys), as indicated by the dose-response relationship. The relationship between the magnitude of the conductance change and the concentration of added protein was determined in the same manner as described for the histone fragment (see above). For poly(Lys-Ala), the time courses were fit by Eq. 4, and Gp1 and Gp2 were added to determine the total conductance change. As shown in Fig. 10, the magnitude of the conductance induced by either protein increased as a function of protein concentration, but poly(Lys-Ala) was much more potent than poly(GluLys). There are several possibilities to explain the reduced activity by poly(Glu-Lys). One possibility is that the net negative charge repels poly(Glu-Lys) from an anionic binding site for cationic proteins. Another explanation is that, by interacting electrostatically, the acidic residue glutamic acid neutralizes the basic amino acid lysine, which could result in loss of the voltage sensitivity of the peptide. An alternative explanation is that the poly(Glu-Lys) molecules electrostatically interact to form aggregates that are inactive. The 25–67 fragment of histone H4 was capable of increasing Gt in a manner similar to intact histone H4. Both the fragment and H4 conductances displayed comparable time courses, sites of action, voltage dependence, and ion selectivity. Because of the limited quantity of fragment available, the long-term toxicity was DISCUSSION In this paper, the components of both the cellular membrane and the protein that are important for conductive activity were examined. First, the identity of the membrane-binding site and the ability of histone to induce ion channel formation were explored using Fig. 10. Dose-response relationships for both poly(Lys-Ala) and poly(Glu-Lys). Magnitudes of conductance induced by proteins was determined for a wide range of protein concentrations. For both proteins, magnitude of conductance increased as a function of protein concentration. Poly(Lys-Ala) (s) is much more potent than poly(GluLys) (r). Downloaded from journals.physiology.org/journal/ajpcell (052.090.062.141) on January 12, 2022. C1934 HISTONE DAMAGES A MAMMALIAN EPITHELIUM not determined. Therefore, it is yet to be determined if the fragment is able to induce the same degree of toxicity as intact histone H4. The conductive activity of the H4-(25–67) fragment suggests that this fragment is important for channel formation and may contain the channel-forming domain. Both the structure of this fragment and of intact histone H4 has been described in detail (3, 8). Comparison of the structures may help explain the similarities and differences in the activities of these two proteins. With the use of fragments of histone H4, it has been determined that the stretch of amino acids from 50 to 67 is critically involved in the formation of both the a-helix and multimeric aggregates. The a-helix is composed of two sections, residues 55–67 and 70–90. The helical wheel diagram of the helical portions of histone H4 suggests that the stretch of amino acids involved in helix formation and aggregation (amino acids 50–67) is somewhat amphipathic in character (Fig. 11). The hydrophilic amino acids are located on one side of the helix, opposite to the majority of the hydrophobic amino acids, making one side of the helix much more polar than the rest of the helix. This suggests that histone H4 [and the H4-(25–67) fragment] may belong to a family of amphipathic a-helical proteins that increase membrane permeability (4, 14, 15, 22). These proteins are believed to aggregate into barrel-like structures, with their outward-facing hydrophobic sides interacting with the phospholipid bilayer, while the hydrophilic faces line the channel. Further investigation is necessary to determine if histone H4 behaves similarly. When the H4-(25–67) fragment aggregates in solution, the portion of the H4-(25–67) fragment that is not helical (residues 25–54) is incorporated into the aggregates rather than being free in solution. In contrast, histone H4 has an additional stretch of NH2-terminal tail (residues 1–24) that is highly cationic and is a random coil in solution. The carboxyl tail of histone H4 is also a random coil but is not quite as cationic as the amino tail; the COOH-terminal tail is important for the formation of high-molecular weight aggregates by histone H4 (23). Both histone H4 and the 25–67 fragment are each long enough to span the cell membrane (.20 amino acids); therefore, individual protein molecules may form the conductive unit. In addition, both of these proteins aggregate, and therefore channels might also be formed by the polymerized proteins. Previous reports by Lewis et al. (8) indicate that these two proteins aggregate differently. Both intact histone H4 and H4(25–67) rapidly aggregate along the a-helical residues in a parallel fashion. Intact histone H4 then forms high-molecular weight aggregates by slowly forming b-sheets along the COOH-terminal tails. Fragment H4-(25–67), in contrast, quickly forms b-sheets at the NH2-terminal of the fragment (amino acids 25–34). The fragment, however, is not able to form high-molecular weight aggregates, because the COOH-terminal is required for this process (23). There are a number of possible explanations for the lower potency of the H4-(25–67) fragment. 1) The Fig. 11. Helical wheel diagrams of a-helical portions of histone H4. Hydrophobic residues are underlined. Charged residues are indicated in bold and marked with appropriate charge. a-helical region from amino acids 70–90, which is found in intact histone H4 but not in the fragment, may contribute to the formation of a more stable membrane channel. Studies suggest that increasing the length of an amphipathic a-helix increases both channel formation in bilayers and toxicity in bacteria (1). 2) The NH2and/or the COOH-terminal tails of histone H4 are important for membrane binding or for stabilizing the conductance in the membrane. Loss of the tails would then either make it more difficult for the fragment to associate with the membrane or result in the fragment dissociating from the membrane more easily. 3) The b-structure of the fragment aggregates may impede their ability to form a conductance. 4) The aggregation of the COOH-terminal tails of intact histone H4 into high-molecular-weight aggregates could lead to the formation of larger channels than H4-(25–67), which does not form high-molecular weight aggregates. 5) The Downloaded from journals.physiology.org/journal/ajpcell (052.090.062.141) on January 12, 2022. HISTONE DAMAGES A MAMMALIAN EPITHELIUM fragment has a weaker voltage sensitivity than intact H4. Preliminary analysis of washout experiments suggests that the rate constants at which both histone H4 and the fragment become nonconductive at 0 mV are similar, whereas the rate constant of formation of the conductance is much slower for the fragment compared with intact H4. This suggests that the difference in the dose-response relationships (i.e., the maximum conductance and binding affinity) is because of the fragment’s slower rate constant for the formation of a conductance than H4. Interestingly, the central domain of histone H5 did not induce a conductance. One possible explanation is that the positively charged amino acids within this central domain are not accessible to the membrane surface and thus cannot associate with the cell membrane binding site. The three-dimensional structures of both histone H5 and the isolated globular domain have been described previously (2, 17). The central domain is compact and is composed of three a-helical regions and an anti-parallel b-sheet. This complex structure may not be able to interact with the cell membrane because of steric hindrance. The COOH-terminal of histone H5 is a long and highly charged random-coil tail. This tail is composed of amino acids 103–190 and is 50% cationic. In contrast, the NH2-tail is small; it spans residues 1–21 and is 29% cationic. These random-coil tails, which have been removed from the purified globular domain, may be necessary for membrane binding of histone H5. Both histone H5 and its globular domain have been reported to form high-molecular weight aggregates (11). For globular H5, multimers ranged from 2 to 14 monomers. It is as yet unclear whether individual protein molecules or aggregates are responsible for the conductive activity of histone H5. It has been reported that histone H1 also increases membrane permeability in rabbit urinary bladder epithelium (7). Histone H1 is structurally very similar to histone H5, and histone H5 is regarded as an ‘‘extreme variant’’ of histone H1 (17). They have similar molecular weights (21,000 for H5, 24,000 for H1) and similar structures. Both have a central globular domain flanked by random-coil tails (2). The major structural differences between these two proteins are that histone H5 contains more arginine and serine than histone H1 and that histone H5 has a shorter NH2-terminal tail (21 amino acids compared with 35 amino acids for H1). Histones H1 and H5 display another interesting difference; the globular domain of histone H1 does not form aggregates as readily as the globular domain of histone H5 (11). One might predict that, because of the structural similarities between H1 and H5, the isolated central domain of histone H1 would be inactive. Predicted Activity of the Histone Fragments Based on Positive Charge The model developed by Tzan et al. (18) is useful in predicting the relationship between cationic charge and induced conductance. The conductance induced by cationic proteins increased as the square of both the total number of cationic residues and the density of the C1935 cationic charge within the protein molecule. This model was developed using synthetic proteins that are composed of cationic and neutral amino acids and are random coils in solution. With the use of this model and histone H1 as a reference molecule, the magnitude of the conductance change that would be induced solely on the basis of positive charge can be predicted for the histone fragments. The calculated values for the fragments indicate that they would be predicted to induce an appreciable change in conductance at the concentrations used in the experiments. There are a number of possible explanations for the lack of activity that was demonstrated by these molecules. Some of the fragments may be too small to be active. For example, the smallest fragment was 86–102, which is only 16 amino acids long and therefore is not long enough to span the cell membrane. The fragment 1–23 is predicted to be the most active of the fragments based on charge but is barely long enough to span the membrane and may not be able to form a stable conductance. Acidic Amino Acids Addition of negatively charged molecules such as DNA or pentosan polysulfate has been demonstrated to decrease the conductive effect of histone (7), most likely by an electrostatic interaction that neutralizes the cationic charge of the histone. These data suggest that negative charges within the protein molecule also are inhibitory, although the mechanism of inhibition is unclear. The differences in the magnitude of the conductance changes induced by poly(Lys-Ala) and poly(Glu-Lys) are not a result of the difference in positive charge between the two molecules. The model developed by Tzan et al. (18) was also used to predict the magnitude of the conductance change that would be induced by both poly(Lys-Ala) and poly(Glu-Lys). Poly(Lys-Ala) is about as active as predicted, whereas poly(Glu-Lys) is much less active than predicted on the basis of its positive charge. This deviation by poly(Glu-Lys) suggests that the negative charge is reducing the ability to induce a conductance. However, the ability to induce a conductance was not entirely abolished by the net negative charge of the molecule. One possible explanation is that because the synthetic proteins are made by a random distribution of amino acids, they are therefore a heterogeneous mixture of proteins with an average ratio of 6:4 glutamic acid to lysine. A certain percentage of the poly(Glu-Lys) will have a higher concentration of positive charge (and an equivalent amount will have a lower proportion of positive charge). The portion with the higher degree of cationic charge might have a sufficient amount of positive charge so that the net charge on the protein molecule is positive. A net positive charge might then result in the protein being able to induce a conductance. Summary These results indicate that histone is capable of forming a conductive unit and identify the central Downloaded from journals.physiology.org/journal/ajpcell (052.090.062.141) on January 12, 2022. C1936 HISTONE DAMAGES A MAMMALIAN EPITHELIUM portion of histone H4 as the domain that is responsible for many of the conductive properties of this molecule. The fragment of histone H4 that spans amino acids 25–67 forms a voltage-dependent, non-ion-selective conductance in the apical membrane of rabbit urinary bladder epithelium. This region is predominantly ahelical in structure. This suggests that histone H4 behaves similarly to a number of channel-forming, amphipathic, a-helical proteins. Further studies are needed to more definitively describe the mechanism by which these proteins increase membrane permeability. This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-51382 to S. A. Lewis, Medical Research Council of Canada Grant MT-5453 to P. N. Lewis, and James W. McLaughlin Fellowship Fund to T. J. Kleine. Address for reprint requests: S. A. Lewis, Dept. of Physiology and Biophysics, Univ. of Texas Medical Branch, 301 Univ. Blvd., Galveston, TX, 77555-0641. Received 20 June 1997; accepted in final form 19 August 1997. REFERENCES 1. Agawa, Y., S. Lee, S. Ono, H. Aoyagi, M. Ohno, T. Taniguchi, K. Anzai, and Y. Kirino. Interaction with phospholipid bilayers, ion channel formation, and antimicrobial activity of basic amphipathic a-helical model peptides of various chain lengths. J. Biol. Chem. 266: 20218–20222, 1991. 2. Aviles, F. J., G. E. Chapman, G. G. Kneale, C. CraneRobinson, and E. M. Bradbury. The conformation of histone H5: isolation and characterisation of the globular segment. Eur. J. Biochem. 88: 363–371, 1978. 3. Crane-Robinson, C., H. Hayashi, P. D. Cary, G. Briand, P. Sautiere, D. Krieger, G. Vidali, P. N. Lewis, and J. TomKun. The location of secondary structure in histone H4. Eur. J. Biochem. 79: 535–548, 1977. 4. Cruciani, R. A., J. L. Barker, S. R. Durell, G. Raghunathan, H. R. Guy, M. Zasloff, and E. F. Stanley. Magainin 2, a natural antibiotic from frog skin, forms ion channels in lipid bilayer membranes. Eur. J. Pharmacol. 226: 287–296, 1992. 5. Hamilton, S. L., L. Mejia-Alvarez, M. Fill, M. J. Hawkes, K. L. Brush, W. P. Schilling, and E. Stefani. [3H]-PN200–110 and [3H]-ryanodine binding and reconstitution of ion channel activity with skeletal muscle membranes. Anal. Biochem. 183: 31–41, 1989. 6. Hirsch, J. G. Bactericidal action of histone. J. Exp. Med. 108: 925–944, 1958. 7. Kleine, T. J., A. Gladfelter, P. N. Lewis, and S. A. Lewis. Histone-induced damage of a mammalian epithelium. Am. J. Physiol. 268 (Cell Physiol. 37): C1114–C1125, 1995. 8. Lewis, P. N., E. M. Bradbury, and C. Crane-Robinson. Ionic strength induced structure in histone H4 and its fragments. Biochemistry 14: 3391–3400, 1975. 9. Lewis, P. N., J. G. Guillemette, and S. Chan. Histone accessibility determined by lysine-specific acetylation on chicken erythrocyte nuclei. Eur. J. Biochem. 172: 135–145, 1988. 10. Lewis, S. A., D. C. Eaton, C. Clausen, and J. M. Diamond. Nystatin as a probe for investigating the electrical properties of a tight epithelium. J. Gen. Physiol. 70: 427–440, 1977. 11. Maman, J. D., T. D. Yager, and J. Allan. Self-association of the globular domain of histone H5. Biochemistry 33: 1300–1310, 1994. 12. Mendizabal, M. V., and R. J. Naftalin. In vivo exposure of rat colonic mucosa to human semen induces mucosal cytolysis, abolishes fluid absorption and raises paracellular permeability (Abstract). J. Physiol. (Lond.) 446: 411P, 1992. 13. Motojima, S., E. Frigas, D. A. Loegering, and G. J. Gleich. Toxicity of eosinophil cationic proteins for guinea pig tracheal epithelium in vitro. Am. Rev. Respir. Dis. 139: 801–805, 1989. 14. Oka, T., Y. Murata, T. Nakanishi, H. Yoshizumi, H. Hayashida, Y. Ohtsuki, K. Toyoshima, and A. Hakura. Similarity, in molecular structure and function, between the plant toxin purothionin and the mammalian pore-forming proteins. Mol. Biol. Evol. 9: 707–715, 1992. 15. Peitsch, M. C., and J. Tschopp. Assembly of macromolecular pores by immune defense systems. Curr. Opin. Cell Biol. 3: 710–716, 1991. 16. Periera, L. F., F. M. Marco, R. Boimorto, A. Caturla, A. Bustos, E. G. de la Concha, and J. L. Subiza. Histones interact with anionic phospholipids with high avidity; its relevance for the binding of histone-antihistone immune complexes. Clin. Exp. Immunol. 97: 175–180, 1994. 17. Ramakrishnan, V., J. T. Finch, V. Graziano, P. L. Lee, and R. M. Sweet. Crystal structure of globular domain of histone H5 and its implications for nucleosome binding. Nature 362: 219– 223, 1993. 18. Tzan, C. J., J. R. Berg, and S. A. Lewis. Modification of epithelial permeability by cationic polypeptides. Am. J. Physiol. 265 (Cell Physiol. 34): C1637–C1647, 1993. 19. Tzan, C., J. Berg, and S. A. Lewis. Effect of protamine sulfate on the permeability properties of the mammalian urinary bladder. J. Membr. Biol. 133: 227–242, 1993. 20. Whitehead, J. S., E. Kay, J. Y. Lew, and L. M. Shannon. A preparative column electrophoresis apparatus using Sephadex G-25. Anal. Biochem. 40: 287–291, 1971. 21. Yonath, J., and M. M. Civan. Determination of the driving force of the Na1 pump in toad bladder by means of vasopressin. J. Membr. Biol. 5: 366–385, 1971. 22. Young, J. D., and C. C. Liu. Multiple mechanisms of lymphocytemediated killing. Immunol. Today 9: 140–144, 1988. 23. Ziccardi, R., and V. Schumaker. Interaction of histone f2al with T7 deoxyribonucleic acid. Cooperativity of histone binding. Biochemistry 12: 3231–3235, 1973. Downloaded from journals.physiology.org/journal/ajpcell (052.090.062.141) on January 12, 2022.