Idea Transcript
INFECTION AND IMMUNITY, Sept. 1984, p. 692-696 0019-9567/84/090692-05$02.00/0 Copyright © 1984, American Society for Microbiology
Vol. 45, No. 3
Effect of Zinc and Phosphate on an Antibacterial Peptide Isolated from Lung Lavage F. MARC LAFORCEl* AND DOROTHY S. BOOSE2 Medical Service, Veterans Administration Medical Center,1 and Division of Infectious Diseases, University of Colorado School of Medicine,2 Denver, Colorado 80220 Received 19 March 1984/Accepted 6 June 1984
We have recently described an antibacterial system that was found in normal rabbit lung lavage and that damages Escherichia coli cells (6). In our previous study, fractionation of lung lavage led to the identification of a lowmolecular-weight peptide which mediated this antibacterial activity. Peptide activity was stable at acid and alkaline pH values and at 95°C but was reversed with 10-4 M EDTA. Cell damage was characterized by the failure of lavageexposed bacteria to multiply on deoxycholate agar, and electron photomicrographs confirmed bacterial damage (6). The purpose of this paper is to further characterize this peptide and to show that peptide damage to bacteria is dependent on zinc and is associated with release of phosphorus from E. coli and that phosphate can reverse antibacterial
column with either 0.01 M ammonium acetate or 0.01 M acetic acid. Individual peaks were lyophilized and repassed on a Sephadex G-15 column with 0.01 M acetic acid. Protein determinations were carried out by the biuret method of Mokrasch and McGilvery (7). Polyacrylamide gel separation. Gel analysis of the peptide was done with a 20% acrylamide separating gel and a 3% acrylamide stacking gel (2). Protein samples (200 ,ug) contained 0.025 M Tris-hydrochloride (pH 6.8), 2% sodium dodecyl sulfate, 2% beta-mercaptoethanol, 10% glycerol, 0.002% pyronin Y, and 0.02% bromphenol blue. The gel was run at a constant voltage of 50 V for about 16 h and at 100 V for an additional 5 h. Peptide stained best with 0.1% bromphenol blue-50% methanol-10% acetic acid. Protein standards (3,000 to 43,000 daltons) were run at the same time. Antibacterial studies. Peptide peaks were tested for their antibacterial activity by techniques previously described (6). Briefly, 104 stationary-phase E. coli cells (serotype 026) were incubated with 10 [Lg of peptide in physiological saline. Quantitative cultures were taken initially and 30 min later with Trypticase soy agar (BBL Microbiology Systems, Cockeysville, Md.) supplemented with 0.085% sodium deoxycholate (Difco Laboratories, Detroit, Mich.). Organisms suspended in physiological saline served as controls. Human insulin (10 Fg/ml) in physiological saline was also tested as a peptide control and was not active. Organisms were counted the following day, and results were expressed as the percentage of microorganisms remaining at 30 min. The effect of various anions and cations on the test system were evaluated. Solutions tested included Hanks balanced salt solution (pH 7.4), 0.01 M ammonium acetate (pH 6.9), 0.05 M barbital buffer (pH 7.3), 0.01 M phosphate-buffered saline (pH 7.0), and 0.05 M each sodium and potassium phosphate buffers (pH 7.5). To determine whether phosphate loading could affect antibacterial activity of peptide, we incubated washed E. coli cells in either 0.05 M sodium phosphate buffer (pH 7.5) or physiological saline for 30 min. Bacteria were then washed twice in physiological saline. Phosphate-loaded and salineincubated bacteria were then challenged with 10 ,ug of peptide, and antibacterial activity was determined as previously described.
activity. MATERIALS AND METHODS Isolation of peptide from rabbit lung lavage. Techniques used to harvest and process rabbit lung lavage have been described previously (6). Briefly, New Zealand rabbits weighing 2 to 4 kg each were sacrificed, and their lungs were lavaged with three 50-ml volumes of physiological saline. Macrophages and the surfactant pellet were removed by centrifugation, and the supernatant was passed through a PM 10 membrane filter (Amicon Corp., Lexington, Mass.). The ultrafiltrate was lyophilized and separated on a Sephadex G-15 column (Pharmacia Fine Chemicals, Inc., Piscataway, N.J.). Our initial column separations were done in distilled water or 0.01 M ammonium acetate (6), although more recently we used 0.01 M acetic acid. We currently pool and lyophilize ultrafiltrate from three rabbits and pass the material on a Sephadex G-15 column (2.5 by 50 cm) with 0.01 M acetic acid at an elution rate of 36 ml/h. Individual 3-ml fractions are read at 220 nm with a Beckman model 25 spectrophotometer. The initial separation yielded a sharp, early peak, a small second peak, and a later, broad peak (Fig. 1). The second and third peaks were heavily contaminated with salt. The third peak was desalted by repassing it on a Sephadex G-15 *
Corresponding author. 692
Downloaded from http://iai.asm.org/ on February 14, 2019 by guest
Incubation of Escherichia coli (104 organisms per ml) in cell-free rabbit lung lavage for 30 min at 37°C resulted in a 70% reduction in colony counts on deoxycholate agar. A low-molecular-weight peptide (about 3,400 daltons), with zinc as a cofactor, was responsible for this activity. The peptide was isolated by Sephadex G-15 separation of lyophilized rabbit lung lavage which had been centrifuged to remove macrophages and suspended phosphlipids and passed through a 10,000-dalton (pore size) membrane filter. Peptide activity against E. coli was inhibited by phosphate buffer but not by borate, Tris, or barbital buffer. Bacteria incubated in phosphate buffer and then washed in saline were resistant to peptide activity. Antibacterial activity was also inhibited when peptide-exposed bacteria were incubated in phosphate buffer before deoxycholate treatment. 32P-radiolabeled E. coli cells lost about 20% of their radiolabel after 15 min of incubation with peptide.
LUNG ANTIBACTERIAL PEPTIDE
VOL. 45, 1984
E
O
0.1
I.a./A0-P3 G-15 .01 M z
radiolabel counts in the total and supernatant samples were determined in a scintillation counter (Nuclear Liquimat; Picker Corp., White Plains, N.Y.). Data were expressed as the percentage of total 32p in the supernatant after incubation with peptide. Organisms suspended in physiological saline served as controls. Viability studies were often done as part of these studies. Zinc studies. EDTA was known to reverse eluate activity (6), a finding which suggested that a metal cofactor was necessary for activity. To better define this question, we incubated active ultrafiltrates with graded concentrations of EDTA (10-s, 10-6, and 10-7 M) to determine at which concentration reversal of antibacterial activity occurred. For most ultrafiltrates, this concentration proved to be 10-5 M EDTA. Zinc, calcium, or magnesium at a concentration of 10-5 M was added singly to eluates with EDTA, and antibacterial activity against E. coli was measured as described above. The pH was kept constant at 6.8 with barbital buffer. To ensure that controls for these experiments reflected as closely as possible the major cation composition of ultrafiltrate samples, we measured eight ultrafiltrates for calcium, magnesium, and zinc by use of a Perkin-Elmer 305 B atomic absorption spectrophotometer. Calcium and magnesium concentrations averaged about 4 x 10-5 and 1 x 10-5 M, respectively, whereas zinc levels were usually less than 1 ,ug/ml. Consequently, 0.15 M saline with 4 x 10-5 M calcium and 1 x 10-4 M magnesium, termed artificial ultrafiltrate, was used as the control solution for these studies. Statistical analysis. Arithmetic means and standard deviations were calculated, and comparisons were made by the Student t test. RESULTS Results of Sephadex G-15 separation of lyophilized ultrafiltrate are shown in Fig. 1. Original separation yielded three peaks which were designated peak 1 (P1), peak 2 (P2), and peak 3 (P3). Protein measurements showed that about 75% of biuret protein was located in P1. Pooled ultrafiltrate from three rabbits yielded 300 to 600 ,ug of P1. In about half the runs, only P1 and P3 were seen. When the broad P3 was desalted by repassing on Sephadex G-15 with 10 mM ammonium acetate, three peaks were recovered and, when P1 from this second pass was then passed on 10 mM acetic acid, it eluted at the same position as did the original P1. ConTABLE 1. Effect of various buffers on activity of peptide (P1) against E. colia
Pl~
% Bacteria
~ P3/ ~
Solutionb
Molarity
pH
remaining at 30 minc
G- 15 .0 M
G-15
/ 30
\
40
FRACTION FIG.
1.
.OIM ACETIC ACID
60
70
80
NUMBER
Separation of lyophilized ultrafiltrate of G-15 Sephadex
showing initial separation into P1, P2, and P3. Desalting P3 led to
further recovery of P1.
Saline Ammonium acetate Barbital HBSS HBSS without Ca or Mg
0.15 0.01 0.05 0.15 0.15
5.2 6.9 7.3 7.3 7.3
1 5 4 96 115
7.0 PBS 0.01 120 Sodium + potassium 0.05 7.2 82 phosphate buffer 7.2 0.05 Potassium phosphate 17 7.2 Sodium phosphate 0.05 63 a Each test included buffer, 10 jLg of peptide, and 104 E. coli organisms per in ml incubated for 30 min at 37°C and cultured on 0.085% deoxycholate agar. b HBSS, Hanks balanced salt solution; PBS, phosphate-buffered saline. c
Results are the means of three studies.
Downloaded from http://iai.asm.org/ on February 14, 2019 by guest
The rates at which various solutions reversed peptide activity were determined as follows. E. coli cells were incubated with 10 ,ug of peptide for 30 min, at which time antibacterial activity was determined with deoxycholate agar as described. Samples of 0.1 ml of peptide-damaged E. coli cells were also taken and placed in 9.9 ml of 0.00001 to 0.1 M phosphate buffer, Hanks balanced salt solution, physiological saline, 0.0015 M saline, or 0.001 M calcium chloride. Samples were taken at 5, 10, and 20 min and plated on deoxycholate agar. Phosphate was noted to reverse peptide activity, and the effect of temperature on this reversal was measured by repeating the studies described above at 37 and 40C. Radiolabel studies. E. coli organisms were radiolabeled with 32p (as 32pj; New England Nuclear Corp., Boston, Mass.) and washed twice with 0.15 M saline, and 108 organisms per ml were incubated with peptide. At 15 or 30 min later, a 0.2-ml sample was placed in a scintillating vial. A second 0.2-ml sample was added to 1.8 ml of saline and centrifuged at 3,000 x g, and 1 ml of the supernatant was placed in a scintillation vial. Bray solution was added, and
693
694
INFECT.1IMMUN.
LAFORCE AND BOOSE
z z
cr
TABLE 2. Effect of temperature on reversal of peptide damage by phosphate % Bacteria
Conditions
remaining
Peptide exposure E. coli and peptide in 0.15 M saline ................. ............ E. coli in 0.15 M saline (control) .......
5 110
Reversal studies'
250C ................. Phosphate (0.001 M) ........... Phosphate (0.001 M) and calcium (0.001 M) .......
40C
Phosphate (0.001 M) ............................ Phosphate (0.001 M) and calcium (0.001 M) .......
33 91
16 21
a In reversal studies, peptide-exposed E. coli organisms were held in phosphate buffers at 25 or 4°C for 15 min before being plated on deoxycholate agar.
not at 4°C (Table 2). Bacteria previously held in phosphate buffer and then washed in saline (phosphate-loaded) were more resistant to the antibacterial activity of peptide than were bacteria held in saline (Table 3). Zinc was an important cofactor for peptide antibacterial activity. EDTA at 10-5 M always reversed ultrafiltrate activity, and activity in EDTA-treated ultrafiltrate could be reconstituted with the addition of 10-5 M zinc but not with equimolar concentrations of calcium or magnesium (Table 4). E. coli organisms labeled with 32p lost radiolabel when exposed to peptide (Fig. 3), and about one-third of the radiolabel was recovered in supernatant after 30 min of incubation with peptide. Release of radiolabel into the supernatant was associated with bacterial damage, as evidenced by the failure of peptide-damaged bacteria to grow on deoxycholate agar. Recovery of bacterial radiolabel was dose related. An increase in the concentration of peptide from 5 to 20 ,ug/ml led to progressive increases in supernatant radiolabel after 30 min of incubation. DISCUSSION Pulmonary antibacterial defenses depend on integrated activity of respiratory reflexes, mucociliary clearance, lung phagocytes, and the immune status of the lung (3, 8). The role of lung secretions other than immunoglobulins in overall lung antibacterial defenses has been a subject of controversy. Early studies have suggested that surfactant does not have direct antibacterial activity, since growth of E. coli, Streptococcus viridans, and Staphylococcus aureus in dog
w
TABLE 3. Effect of phosphate loading of E. coli on peptide activity z
% Bacteria remaining at 30 min'
Treatment
w
0 w
a.
Phosphate loaded
Saline Sa i......91 n 9 Saline + peptide ................................. 50 .
TIME (minutes) FIG. 2. Reversal of peptide activity by phosphate and calcium.
..................................
.
Not phosphate loaded Saline ......................................... Saline + peptide ................................. a Results are the means of three experiments.
102 11
Downloaded from http://iai.asm.org/ on February 14, 2019 by guest
versely, P1 was broken into separate peaks when an ammonium acetate buffer was used. We felt that these results probably indicated that the peptide underwent aggregation in an acid pH. Polyacrylamide gel electrophoresis yielded a spot (estimated molecular weight, 3,400) which rapidly destained and could not be permanently fixed to the gel. E. coli organisms incubated with peptide in physiological saline were killed when cultured on deoxycholate agar. In general, peptide recovered from 0.01 M acetic acid separation was more active than that recovered from 0.01 M ammonium acetate. Usually, 10 ,ug of peptide resulted in about 5 to 20% of the original bacterial inoculum remaining at 30 min. As previously mentioned, a culture of peptideexposed E. coli cells on a noninhibitory medium such as Trypticase soy or blood agar showed no evidence of bacterial kill. The buffers chosen for activity studies proved to be important. All buffers which contained phosphate interfered with peptide activity against E. coli, whereas barbital, Tris, and acetate buffers all supported peptide activity against E. coli (Table 1). The type of phosphate buffer also seemed to be important; sodium phosphate was more active in reversing peptide activity than was potassium phosphate, suggesting that the cation present could also influence the rate of reversal of peptide activity. Addition of glucose did not enhance reversal. Calcium enhanced the reversal of peptide damage by phosphate (Fig. 2). More than 90% of the E. coli organisms incubated with 10 jig of peptide in saline for 30 min were unable to grow on deoxycholate agar, but this damage could be reversed. For example, holding the sample in 0.0015 M saline did not result in the reversal of peptide damage, whereas holding it in 0.15 M saline resulted in some recovery. Calcium at 0.001 M had little effect. Phosphate alone at 0.001 M could reverse the defect, but the rate of reversal was markedly accelerated by the addition of 0.001 M calcium. Temperature was an important variable. Peptide-induced damage to E. coli organisms was completely reversed by phosphate and calcium solutions at room temperature but
LUNG ANTIBACTERIAL PEPTIDE
VOL. 45, 1984
TABLE 4. Importance of zinc as a cofactor in ultrafiltrate activity against E. coli Treatment
(molarity)a
% Bacteria
remaining'
UF Alone ....................................... 30.7 + EDTA (10-6) ............................... 56.1 + + + +
EDTA EDTA EDTA EDTA
18.3 31.3 .................... 111.8 + 11.9 (10-5) (10-5) + Ca (10-5) .................... 121.7 ± 15.0 (10-5) + Mg (10-5) ................... 113.7 ± 9.5 (10-5) + Zn (10-5) .................... 25.0 ± 10.4c
Artificial UF Alone ........................................ 81.7 + EDTA (10-5) ............................... 95.7 + EDTA (10-5) + Ca (10-5) .................... 93.1 + EDTA (10-5) + Mg (10-5) ................... 116.8 + EDTA (10-5) + Zn (10--5) .................... 83.3
+
+
15.1
+
± 34.4
+ 10.4 +22.8 ± 9.6'
a UF, Ultrafiltrate. Artificial ultrafiltrate is 0.15 M saline with Ca (4 M) and Mg (1 x 10-5 M). b Results are the (± standard deviation) of seven experiments. C P < 0.001.
x
10-5
100-
_
_
_
_
-A _A _
-
-100
X
m
-
PEPTIDE ---SALINE
80-
en 60-
-80 rn
z w -60 o -g
w
-J Ict 40-
-40
z
mP1 P1 i;
K
w
m
0~
-20 z c)
U0
I
I
15
30
MINUTES FIG. 3. Release of 32P radiolabel from peptide-exposed bacteria. Viability data are also shown. Symbols: 0, percent release of 32p; A, percent bacteria remaining.
was removed after absorption with bentonite (12, 13). Moredetailed biochemical studies identified the bacterial inhibitor as a low-molecular-weight (630) peptide which required the metal cation zinc. The peptide was inactivated by digestion with carboxypeptidase, and the amino acid composition of the peptide was 3-glutamine-glutamic acid, 1 glycine and 1 lysine (13). Our experimental data are quite similar to those reported in amniotic fluid studies and suggest that lowmolecular-weight antibacterial peptides may be broadly distributed. Phosphate reversal of antibacterial activity was impressive, although not surprising, since several antibacterial peptides have been shown to be sensitive to phosphate (10, 13). Phosphate alone at concentrations greater than 0.05 M sufficed to reverse peptide damage. Addition of calcium, however, markedly decreased the level of phosphate necessary to reverse peptide activity. The mechanism involved in this reversal of activity is not shown. Failure to protect at 40C suggests that the effect is not simply one of cation absorption to E. coli cell walls or membranes. Leakage of 32p from peptide-treated E. coli organisms is consistent with membrane damage, and electron microscopy studies of peptide-treated E. coli organisms have shown evidence consistent with leakage of intracellular contents (6). To date, we have been unable to produce a specific antibody to this peptide. Until such an antibody is produced, it will be difficult to quantitate the role that this antibacterial system plays in in situ antibacterial activity in the lung. ACKNOWLEDGMENTS We gratefully acknowledge the advice and support of Stewart Gordon and the technical help of Nancy Miller. This work was supported by the Veterans Administration Research Service. LITERATURE CITED 1. Coonrod, J. S., and K. Yoneda. 1983. Detection and partial characterization of antibacterial factor(s) in alveolar lining material in rats. J. Clin. Invest. 71:129-141. 2. Gordon, A. H. 1975. Electrophoresis of proteins in polyacrylamide and starch gels, p. 73-132. In T. S. Work and E. Work
Downloaded from http://iai.asm.org/ on February 14, 2019 by guest
or rabbit surfactant is enhanced as compared with controls (4, 5). These concepts are now being reevaluated as a result of new and exciting studies by Coonrod and Yoneda (1), who have shown that rat surfactant has profound pneumococcidal activity by stimulating pneumococcal autolysin. This activity is somewhat labile, which probably explains why this antimicrobial action was missed in earlier studies. These investigators have also shown that rat surfactant also has important antibacterial activity against several nonpneumococcal, gram-positive bacteria. Their studies have also demonstrated that a lysophospholipid, palmitoyl lysophosphatidyl choline, has many of the properties seen in the studies on raw surfactants. These data have reopened the proposition that pulmonary lysophospholipids have important antibacterial activity. Lehrer and co-workers have recently characterized two microbicidal cationic proteins from rabbit alveolar macrophages. Alveolar macrophages were mechanically disrupted by centrifugation at 27,000 x g, and a supernatant fraction was extracted with citric acid, which yielded two cationic protein bands. These proteins had maximum antimicrobial activity against Candida spp. and intermediate activity against gram-positive organisms but were inactive against gram-negative rods (9). These peptides have been purified, and amino acid analysis has shown them to be rich in arginine and half-cystine (10). It is likely that these proteins play an important role in the intracellular killing of Candida spp. Whether these peptides are secreted and exist freely in the acellular fraction of lung lavage is currently not known. Our studies have shown that normal rabbit lavage contains a complex antibacterial system which damages E. coli organisms. Fractionation of lung lavage yielded a low-molecularweight protein which had antibacterial activity and needed zinc as a cofactor. Evidence for the importance of zinc stemmed from EDTA studies of ultrafiltrate, in which it was shown that zinc could reconstitute activity in EDTA-treated specimens, a phenomenon not seen with calcium or magnesium. The similarity between this peptide and an antimicrobial system in amniotic fluid is striking (13). Amniotic fluid samples collected from women at 36 to 40 weeks of gestation were all bactericidal or bacteriostatic against E. coli (11). The inhibitor was heat stable and sensitive to phosphate and
695
696
3.
4. 5. 6. 7.
(ed.), Laboratory and techniques in biochemistry and molecular biology. Worth-Holland Publishing Co., Amsterdam. Green, G. M., G. J. Jakab, R. B. Low, and G. S. Davis. 1977. Defense mechanisms of the respiratory membrane. Am. Rev. Respir. Dis. 115:479-514. Jalowayski, A. A., and S. T. Giammona. 1972. The interaction of bacteria with pulmonary surfactant. Am. Rev. Respir. Dis. 105:236-241. LaForce, F. M. 1976. Effect of alveolar lining material on phagocytic and bactericidal activity of lung macrophages against Staphylococcus aureus. J. Lab. Clin. Med. 88:691-699. LaForce, F. M., and D. S. Boose. 1981. Sublethal damage of Escherichia coli by lung lavage. Am. Rev. Respir. Dis. 124:733737. Mokrasch, L. C., and R. W. McGilvery. 1956. Purification and properties of fructose-1,6-diphosphatase. J. Biol. Chem. 221:909-927. Newhouse, M., J. Sanchis, and J. Bienenstock. 1976. Lung
INFECT. IMMUN.
defense mechanisms. N. Engl. J. Med. 295:990-998, 1045-1052. 9. Patterson-Delafield, J., R. J. Martinez, and R. I. Lehrer. 1980. Microbicidal cationic proteins in rabbit alveolar macrophages: a potential host defense mechanism. Infect. Immun. 30:180-192. 10. Patterson-Delafield, J., D. Szklarek, R. J. Martinez, and R. I. Lehrer. 1981. Microbicidal cationic proteins of rabbit alveolar macrophages: amino acid composition and functional attributes. Infect. Immun. 31:723-731. 11. Sachs, B. P. 1979. Activity and characterization of a low molecular fraction present in human amniotic fluid with broad spectrum antibacterial activity. Br. J. Obstet. Gynecol. 86:8186. 12. Schlievert, B. S., W. Johnson, and R. P. Galask. 1976. Bacterial growth inhibition by amniotic fluid. Am. J. Obstet. Gynecol. 125:899-905. 13. Schlievert, P., W. Johnson, and R. P. Galask. 1976. Isolation of a low-molecular-weight antibacterial system from human amniotic fluid. Infect. Immun. 14:1156-1166.
Downloaded from http://iai.asm.org/ on February 14, 2019 by guest
8.
LAFORCE AND BOOSE