Synthetic bactericidal peptide based on CAP37: A 37 ... - Europe PMC

Proc. Natl. Acad. Sci. USA Vol. 90, pp. 4733-4737, May 1993 Immunology

Synthetic bactericidal peptide based on CAP37: A 37-kDa human neutrophil granule-associated cationic antimicrobial protein chemotactic for monocytes (host defense/serine esterases/lipopolysaccharide binding domain/Gram-negative bacteria)

H. ANNE PEREIRA*t, IMRE ERDEM*, JAN

POHLt, AND JOHN K. SPITZNAGEL*

*Department of Microbiology and Immunology, and *Microchemical Facility, Emory University School of Medicine, Atlanta, GA 30322

Communicated by Merrill W. Chase, February 12, 1993 (received for review October 2, 1992)

popolysaccharide (LPS) binding was tested against LPS and lipid A.

CAP37 (cationic antimicrobial protein of moABSTRACT lecular mass 37 kDa) is a multifunctional protein isolated from the granules of human neutrophils. It is antibiotic and chemotactic and binds lipopolysaccharide. A synthetic peptide, amino acid sequence NQGRHFCGGALIHARFVMTAASCFQ, based on residues 20-44 of CAP37 protein mimics its antibiotic and lipopolysaccharide binding action. Peptide 20-44, at the concentrations tested, has antibacterial activity against Salmonella typhimurium, Pseudomonas aeruginosa, Escherichia coli, Enterococcus faecalis, and Staphylococcus aureus. The bactericidal action of the peptide was pH dependent, with maximum activity at pH 5.0 and pH 5.5 and decreased activity at pH 7.0. Various truncations, substitutions, and other modifications in the sequence deteriorate its activity. Free sulfhydryl groups and/or disulfide bridge formation are required for optimum antibiotic activity, since substitution of serines for, or alkylation of, cysteine residues 26 and 42 eliminates bactericidal activity. Evidently amino acids 20-44 represent an important, perhaps principal, antibacterial domain of CAP37. This peptide should provide new insight into the mechanism of antimicrobial activity of CAP37 and may serve as a model for new, useful, synthetic antibiotics.

MATERIALS AND METHODS Bacterial Strains. The bacterial strains used in this study were: S. typhimurium LT2, S. typhimurium SH9178 (Rb2 chemotype) derived from SH5014 (17), its isogenic polymyxin B-resistant derivative strain SH7426 (Rb2 chemotype), and S. typhimurium SH7518 (Re chemotype) (18, 19); Pseudomonas aeruginosa (American Type Culture Collection ATCC no. 27853); E. coli (ATCC no. 25922); Staphylococcus aureus (502A, ATCC no. 27219); Enterococcus faecalis; Streptococcus pyogenes; Listeria monocytogenes; Proteus mirabilis; P. mirabilis Re45 (20); Proteus vulgaris; Morganella morganii; and six clinical isolates of S. typhimurium (C5500, C5549, C5590, C5597, C5608, and C5656). Synthesis of Peptides Based on CAP37. Peptides were synthesized by solid-phase synthesis (21) on an Applied Biosystems model 430A peptide synthesizer (0.1-mmol or 0.5-mmol scale) using the phenylacetamidomethyl copoly(styrene/divinylbenzene) resins and the tert-butyloxycarbonyl (Boc)-protected amino acids (Bachem). To incorporate the respective amino acids, we employed Boc-Arg(mesitylenesulfonyl), Boc-Asp(benzyl), Boc-Cys(4-methoxybenzyl) or Boc-Cys(acetamidomethyl), Boc-Glu(benzyl), BocHis(benzyloxycarbonyl) or Boc-His(2,4,-dinitrophenyl), Boc-Lys(2-chlorobenzyloxycarbonyl), Boc-Met, BocSer(benzyl), Boc-Thr(benzyl), and Boc-Tyr(2-bromobenzyloxycarbonyl). The manufacturer's coupling protocols (software version 1.40) were used throughout. Most of the Bocamino acids in all peptides were double-coupled to achieve coupling efficiencies typically higher than 99.5%, as determined for the peptidyl-resins by a preview analysis using solid-phase sequencing in an Applied Biosystems model 477A/120A peptide sequencer, according to the manufacturer's protocol. The peptides were deprotected and cleaved off the resin in liquid HF/p-cresol/dimethyl sulfide (10:1:0.5) or in liquid HF/anisole (10:1) for 90-120 min at -5°C. The resins were washed with cold diethyl ether and the peptides were extracted into 50%o (vol/vol) aqueous acetic acid and lyophilized. The peptides were purified by reverse-phase (RP)HPLC on an Aquapore OD-300 C18 or Aquapore RP-300 C8 silica column (1 x 10 cm, 20-,um particle size, 300-A pore size, Applied Biosystems) or on a MRPH-gel RP polystyrene

The oxygen-independent bactericidal action of human neutrophil polymorphonuclear leukocytes is due to a number of potent antimicrobial cationic proteins (1). Among these are the small (2-4 kDa) proteins, the defensins (2); the larger (24-29 kDa) serine proteases such as cathepsin G, elastase, and p29b (3-7); CAP37 (cationic antimicrobial protein) (8, 9), a 37-kDa protein also called azurocidin (7, 10) or BP30 (5); and the 55- to 57-kDa bactericidal permeability-increasing protein (BPI) (11) also called CAP57 (8). We isolated and purified CAP37 from human neutrophils in 1984 (8). It had very strong antibacterial activity against Gram-negative bacteria, particularly Salmonella typhimurium and Escherichia coli (8, 9). Recently, we have shown that, in addition, CAP37 is a potent and specific chemoattractant for monocytes (12). Furthermore, we reported the amino acid sequence of CAP37 (13) and cloned its gene (14). These results have been confirmed by others (15, 16). We now report synthesis of a peptide based on amino acids 20-44 of CAP37 that mimics the antibiotic action of CAP37 protein and extends its range. This domain was defined by synthesizing various overlapping peptides corresponding to the amino acid sequence of the mature protein and screening them in a bactericidal assay using S. typhimurium (SH9178) as the test organism. Once the active domain of CAP37 was identified, its specificity was tested against a number of different Gram-negative and Gram-positive bacteria. Its li-

column (1 x 10 cm, 5-,um particle size, 500-A pore size, The

Abbreviations: CAP37, cationic antimicrobial protein of 37 kDa; LPS, lipopolysaccharide; BPI, bactericidal permeability-increasing protein; ACM, Boc-Cys(acetamidomethyl); LB, Luria-Bertani. TTo whom reprint requests should be addressed at the present address: Department of Pathology, Room 434 Biomedical Sciences Building, The University of Oklahoma Health Sciences Center, P.O. Box 26901, Oklahoma City, OK 73190.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

4733

4734

Immunology: Pereira et al.

Proc. Natl. Acad. Sci. USA 90 (1993)

100

-+--

._

1-25 244

-U-

38-53

-0-

43.53 72.80

c

95.122

CL-

-U-

102-122

-a-D

113-122

-0-

13D-146 140-165

200

Peptide concentraton (lg/ml) FIG. 1. Antimicrobial activities of CAP37 peptides tested at 200, 100, 50, 25, and 12.5 pg/ml against S. typhimurium SH9178 at pH 5.5. The percent killed is determined as detailed above. Vertical bars represent the standard error of the mean.

Nest Group, Southboro, MA) using a linear gradient of acetonitrile in 0.1% aqueous trifluoroacetic acid and were stored at -20°C freeze dried in the form of trifluoroacetate salts. The purity and integrity of the crude and HPLC-purified peptides were confirmed by (i) amino acid compositional analysis using an Applied Biosystems model 420A/130A derivatizer/phenylthiocarbamyl amino acid-analyzer equipped with an on-line automatic HCI vapor hydrolysis unit (software version 3.43); (ii) automated sequencing (13) under the conditions of the solid-phase Edman degradation (to detect any incompletely deprotected and/or side-chain modified amino acid residues); and (iii) microbore (RP)-HPLC as described above (size, 1 x 250 mm) and an Applied Biosystems 140A micropurification system equipped with a 1000S diode-array UV/visible scanning detector and a 2.3-,ul microbore flow cell (13). The mass of peptide 20-44 was confirmed at 2722 by mass spectrometry. Synthesis of Elastase Peptide 20-44 and Cathepsn G Peptide 20-47. Peptides were synthesized corresponding to amino acid residues 20-44 of human neutrophil elastase and amino acid residues 20-47 of cathepsin G. To preserve maximum sequence similarity between the alignment of cathepsin G and CAP37, the cathepsin G sequence required the insertion of residues 22-24, thus making the resultant peptide three amino acids longer than the elastase and CAP37 peptides. These three peptides were tested against S. typhimurium SH9178 in the bactericidal assay. Bactericidal Assay. All Salmonella strains and E. coli, P. aeruginosa, and S. aureus were grown in Luria-Bertani (LB) broth and plated on LB agar plates (22). E. faecalis and S.

pyogenes were grown in brain heart infusion (BHI) broth (Difco) and plated on blood agar plates (Remel, Lenexa, KS). L. monocytogenes was grown in BHI broth and plated on BHI plates. Proteus strains and M. morganii were grown in LB broth and plated on nonswarming agar plates (23). Logarithmic-phase bacteria of the required strain were adjusted to a final concentration of 5 x 103 organisms per ml in tryptone saline at pH 5.5 (8). Stock solutions of peptides were made up at concentrations of 1 mg/ml in sterile endotoxin-free water. Subsequent dilutions were all made in tryptone saline at pH 5.5. Bactericidal assays were performed according to the technique of Shafer et al. (8). To 100 ,ul of peptide at the desired concentrations was added 100 ,ul of the bacterial suspension (-500 bacteria) in a 96-well sterile microtiter tray (Corning). The microtiter tray was incubated at 37°C for 1 hr, and 100 Al of the contents from each well was plated on LB agar plates. After incubation overnight at 37°C, the colony-forming units were counted. Controls were incubated for 60 min in the absence of peptide. Viable counts were done on similar control wells at the beginning of each experiment (i.e., time 0) to monitor bacterial growth during the 60-min incubation period. We have reported previously that growth of bacteria must occur for the antibiotic proteins to kill them. Bacterial killing is expressed as percent killed and calculated according to the following equation: [(control - test)/control] x 100 = % killed, where control = number of bacteria present in 100 A1 after 60 min of incubation in 200 ,l of tryptone saline in the absence of antibiotic protein or peptide and test = number of bacteria present after incubation in tryptone saline containing the peptide. Tests were performed in triplicate. The points on graphs and areas of histograms are the averages; vertical bars represent standard error of the mean. Each figure is representative of results obtained in at least three independent experiments. The inhibitory action of LPS (binding) for antibacterial action of peptide 20-44 was assayed in a similar fashion except that the peptide was used at 75 uM (95% killing). Graded concentrations of endotoxins were added to the peptide. Endotoxins were smooth S. typhimurium LPS, S. typhimurium lipid A (lipid AS), and S. minnesota lipid A (lipid AM). All were from Ribi Immunochem. Bactericidal Activity of CAP37 Peptides. Peptides comprising amino acids 1-25, 20-44, 38-53, 43-53, 72-80, 95-122, 102-122, 113-122, 130-146, and 140-165 of the CAP37 molecule (13) were assayed at concentrations of 200, 100, 50, 25, and 12.5 ,ug/ml for bactericidal activity against our standard test organism, S. typhimurium SH9178. To study bactericidal activity in the region of the first 44 amino acids at the amino terminus, additional peptides were synthesized with amino acids 1-44, 1-5, 6-25, 20-29, 29-44, 24-42, 24-44, 26-42, and 27-41 and were tested between 150

120

* juM [3125 yuM

100-

144

1-5

1

24

4

2

2

2741

FIG. 2. Bactericidal activity of 11 peptides tested at concentrations ranging from 150 ,uM to 12.5 ;M at pH 5.5 against S. typhimurium SH9178.

Immunology: Pereira et al. 100 -

80 -

0) .)


E2

125pM

19

100 PM

*

751M

60 -

25

=

O:

c

12.5

pM pM

0e 40 -

$

20 A

I

.

20-44 SER

20-44 ACM

20.44

FIG. 3. Effect of cysteine residues at positions 26 and 42 in the CAP37 peptide 20-44. The bactericidal activity of CAP37 peptide 20-44 was compared with the activities of peptide 20-44-ACM (cysteine blocked with ACM side chains) and peptide 20-44-Ser (cysteine residues replaced by serine) using S. typhimurium SH9178 as the test organism at pH 5.5.

AM and 6.2 ,uM. Molar concentrations of the freeze-dried peptides were calculated using the nominal molecular mass of the peptide. A Boc-Cys(acetamidomethyl)-peptide 20-44 (peptide 2044-ACM) was synthesized wherein the sulfhydryl groups of cysteine residues at positions 26 and 42 were permanently protected by the ACM side chain protecting group. Peptide 20-44-Ser was synthesized, serine residues replacing Cys-26 and Cys-42. The specificity of the bactericidal activity of peptide 20-44 was evaluated against a number of Gram-negative and Grampositive bacteria. To study the effect of pH on the antibiotic activity of peptide 20-44 the tryptone saline used in the experiments was adjusted to pH 5.0, 5.5, 6.0, 6.5, and 7.0 with HCI. Bactericidal assays were conducted as described above against our standard S. typhimurium. RESULTS Bactericidal Activity of CAP37 Peptides. Fig. 1 demonstrates the antimicrobial capacities of 10 synthetic peptides tested against S. typhimurium SH9178. Peptide 20-44 killed 99.3% of the bacteria at 200 ,g/ml (73 ,uM) and 90% at 100 ,g/ml, significantly more than the other 9. It is bactericidal rather than bacteriostatic since the viable counts at both of these concentrations were less than the starting inoculum. Optimal antibiotic action of peptide 20-44 was expressed between pH 5.0 and pH 5.5 (data not shown). At pH 6.0 and higher its bactericidal action failed to be expressed. This pH optimum resembles that of CAP37 (9). Minimal Number of Amino Acids Between Residues 1 and 44 Essential for Bactericidal Activity. The data in Fig. 1 indicated

4735

that peptide 20-44 had potent antibiotic activity. Consequently, it was important to determine whether shortening or lengthening the peptide could result in enhanced activity. Thus peptides from the region of 20-44 were synthesized. Their bactericidal activities are depicted in Fig. 2. Peptide 20-44 was by far the most active of any peptide we tested. Peptide 1-44, which incorporated peptides 1-25 and 20-44, was active between 150 uM and 100 AM but was ineffective at the lower concentrations, whereas peptide 20-44 had strong activity even at 50 ,uM. In addition, the activity of 20-44 was lost when the cysteine residues were separated in peptides 20-29 and 29-44. Shortening the 25-amino acid peptide 20-44 deteriorated the activity. Bangalore et al. (24) have shown that the amino-terminal pentapeptide IIGGR of cathepsin G has bactericidal activity for S. aureus and Neisseria gonorrhoeae at concentrations of 4.3 x 10-4 M. Therefore we prepared peptide 1-5 of CAP37, IVGGR, and tested its antibacterial activity (Fig. 2). This peptide at 75 ,M killed only 36% of SH9178, whereas peptide 20-44 killed 94%. Significance of Cysteines 26 and 42 and the Potential for Disulfide Bond Formation. Defensins (2) have six highly conserved cysteine residues with the potential to form intramolecular disulfide bonds conferring cyclic structure, postulated to be important for their antibiotic activity (25). According to Vaara (26), polymyxin B upon linearization of the cyclic peptide lost its antimicrobial action. Peptide 20-44 contains two cysteine residues, in positions 26 and 42, forming a disulfide bridge in CAP37 (13). Therefore, the importance of these two cysteine residues was assessed for activity of the peptide. Activity was lost when the two cysteines were either replaced by serine residues (20-44-Ser) or alkylated by the ACM group (20-44-ACM) and thus lacked the potential for disulfide bond formation (Fig. 3). As noted above, separation of the cysteines on peptides 20-29 and 29-44 resulted in inactive products (Fig. 2). Bactericidal Activity of Peptide 20-44 Against Various Bacterial Strains. Fig. 4 shows the effect of CAP37 peptide 20-44 on various strains of S. typhimurium in a range of LPS chemotypes and on six clinical isolates of S. typhimurium. Among the nonclinical isolates the peptide was most active against the deep rough Re mutant SH7518 and measurably less active against the smooth LT2 strain. The polymyxin B-resistant isogenic derivative of SH9178, SH7426, was least sensitive to the action of the peptide. The differences in activity of the peptide with regard to SH7518 and SH7426 paralleled the differences in activity of the mature protein against these strains. The six clinical isolates of S. typhimurium tested showed various levels of sensitivity. Of the other microorganisms tested (Fig. 5), P. aeruginosa and E. coli were very sensitive to the action of CAP37

100

so 00

OL

40 75

20

-lIM

50 pM

04

C5500

C5549

C5590

pM

C5597

C5608

C5658

LT2

* 25

pM

E

1iM

12.5

SH7426 SH7518

FIG. 4. Bactericidal activity of peptide 20-44 against Salmonella strains other than SH9178. The antimicrobial effect of CAP37 peptide 20-44 at 75, 50, 25, and 12.5 uM at pH 5.5 was tested against S. typhimurium clinical isolates C5500, C5549, C5590, C5597, C5608, and C5656 and

S. typhimurium LT2, SH7426, and SH7518.



4736

100

100 'a

E1D1

80

60

150 FM 125 pM i 10 gM

-

U75 ~iM

±M pM 12.5 RM 50 25

C

80

)

7

2 40

-

75M a

00

20

0)

40

0

Ecoli E.faecalis Ps.aerug S.aureus L.mono

Pr.vuig

P.mirab Re45

FIG. 5. Bactericidal activity of peptide 20-44 against various Gram-negative and Gram-positive microorganisms: E. coli, E. faecalis, P. aeruginosa (Ps.aerug), S. aureus, L. monocytogenes (L.mono), P. mirabilis (P.mirab), P.mirabilis Re45 (Re45), P. vulgaris (Pr.vulg), and M. morganii.

peptide. Surprisingly, the Gram-positive organisms S. aureus and E. faecalis that are resistant to CAP37 were as sensitive to the peptide as some Salmonella strains. However, two other Gram-positive organisms, S. pyogenes and L. monocytogenes, were resistant to the peptide and to CAP37. All strains of Proteus tested and M. morganji were resistant to peptide 20-44. Comparison of Bactericidal Activity of CAP37 Peptide 2044, Elastase Peptide 20-44, and Cathepsin G Peptide 20-47. The amino acid sequences of cathepsin G, elastase, and CAP37 are -40% homologous as previously reported (12, 27). CAP37 and elastase share 13 identical residues between amino acids 20 and 44 (Fig. 6), whereas CAP37 and cathepsin G have 12 identical residues. As seen in Fig. 7, the cathepsin G peptide at 75 ,uM killed only 23% of SH9178, whereas the peptide 20-44 killed 94%. The cathepsin G peptide was sparingly soluble in water and had to be solubilized in 10% dimethyl sulfoxide (DMSO). Controls indicated that 10% DMSO had no effect on killing of S. typhimurium SH9178. The elastase peptide 20-44 showed activity but clearly less than the CAP37 peptide. Activity of CAP37 and Peptide 20-44 Inhibited by Endotoxin and Lipid A. One LPS and two lipid A preparations (Fig. 8) inhibited (bound) the antibiotic action of peptide 20-44 in a dose-dependent manner. The reason for the somewhat less efficient binding due to lipid A is unclear.

DISCUSSION To establish a structural basis for the antibiotic action of CAP37, and a basis for possible synthesis of novel antibiotics, we synthesized 10 peptides based on the amino acid sequence of residues 1-165 of the 222-residue native protein. Peptide 20-44 had significant antimicrobial action, which, like CAP37, killed P. aeruginosa, S. typhimurium, and E. coli (Fig. 5). Proteus strains were resistant, as they are to CAP37. Unexpectedly, two Gram-positive organisms, S. aureus and E. faecalis, resistant in our hands to the entire CAP37 molecule, were sensitive to the peptide (20). In other respects the activity of the short synthetic peptide paralleled the activity of the native protein. In general, the smooth strains of Salmonella were more resistant and a very rough strain with Re LPS such as SH7518 (Fig. 4) was more sensitive to it. A polymyxin B-resistant (pmrA) strain of Salmonella (SH7426) was cross-resistant to peptide 20-44 as it was to CAP37. In general, the peptide qualitatively mimics the CAP37

20

-N-

CATG

20

-I

ELAS

20

-L-R

*'0-S-P-A,--

20

M.morgani

0T 20-44 CAP37

20-44 ELAS

20-47 CAT G

FIG. 7. Bactericidal activity of CAP peptide 20-44, elastase peptide 20-44 (ELAS), and cathepsin G peptide 20-47 (CAT G), tested at concentrations of 150, 125, 100, 75, 50, and 25 ,uM using S. typhimurium SH9178 as the test organism.

action of the native protein but has a broader antimicrobial spectrum including Gram-positive bacteria. Among the antibiotic, cationic proteins, CAP37 is unique because of its multifunctional nature: it is antibiotic (8) and chemotactic (12), binds heparin (15) and LPS, and has the structure of a serine proteinase. Although lacking enzymatic activity, it shares significant homology with numerous serine proteases that are important in inflammation, including human neutrophil elastase and cathepsin G (12). The lack of serine esterase activity is ascribable to replacement of two of the three active site residues (histidine and serine) (13). Interestingly, the ability of cathepsin G to kill bacteria also is independent of its enzymatic activity (4, 24). Bangalore et al. (24) and Shafer et al. (28) synthesized a pentapeptide, IIGGR, and a heptapeptide, HPQYNQR, based on the primary structure of cathepsin G. These short synthetic cathepsin G antimicrobial peptides are quite distinct from the CAP37 peptide 20-44, even though CAP37 and cathepsin G have strong sequence homology. These short synthetic structures had antibiotic activity at concentrations of 1.19 x 10-4 M to 5 x 10-5 M, whereas the cathepsin G protein had activity at 4 x 10-6 M, suggesting that other domains ofthe protein must contribute to its greater activity. Similar to the cathepsin G peptide, peptide 20-44, in losing so much protein structure, seems to have lost some antimicrobial activity. The peptide may, nevertheless, have advantages that offset this loss. It apparently reaches its lipid A target with less steric hindrance from LPS than CAP37 protein, roughly 10 times its size. This is suggested by the fact that its LD50, 5 x 10-5 M, is the same for LT2 (wild-type smooth LPS) as for strain SH9178 (rough Rb2 LPS), whereas the LD50 of CAP37 protein for the LT2 is 4.7 x 10-6 M and for the SH9178 is 6.6 x 10-7 M. The possible contribution of the carboxy-terminal 57 amino acid residues of CAP37 must be evaluated. Peptide 20-44 has cysteines at positions 26 and 42. These two cysteines are conserved in serine proteases and form one of the four intramolecular disulfide bonds in these enzymes (13). To determine if intramolecular disulfide bonding has a role in its action, two analogs of peptide 20-44 were synthesized. In one the sulfhydryl groups of the cysteines were protected by the ACM group; in the other, serine residues were substituted for the cysteines. The resulting linear pep-

Q G-

T-A

H-N

*V-A-

44

FIG. 6. Alignment of CAP37 peptide 20-44 with elastase peptide 20-44 (ELAS) and cathepsin G peptide 20-47 (CATG) showing sequence

identity.



co

4737

phosphate groups of lipid A. This would reduce the number of negative charges in the outer membrane and leave fewer putative binding sites for CAP37 (reviewed in ref. 1). We may speculate that the cationic peptide molecules bind with the anionic phosphate groups of the LPS and initiate their antimicrobial action. In addition, the active antimicrobial structure of the peptide evidently requires the potential to form an intramolecular disulfide bond. These properties-the size of the peptide, its basic charge, and hydrophobicity-all may contribute to its antibiotic behavior in ways yet to be determined. In any event, it appears to be a promising candidate for antibiotic drug development.

60

~40LU

X 20 0

30

50 100 ENDOTOXIN jg/mi

200

FIG. 8. Inhibition (binding) of 75 ,uM peptide 20-44 due to smooth LPS (*), lipid AS (o), and lipid AM (o) at 25, 50, 75, 100, and 200 jug/ml. When tested against SH9178, all three gave dosedependent inhibition of killing. Logarithmic regression lines (FreeLance Software, Lotus Corp., Cambridge, MA) gave the best fit as shown.

tides lacked antibiotic activity. Thus the potential for disulfide bond formation is necessary for the peptide's maximum antibiotic activity. Potential for disulfide bond formation evidently is necessary but is not sufficient for maximal antibiotic activity, since our cathepsin G peptide 20-47 also had two cysteines in identical positions to the CAP37 peptide but had meager antibiotic activity (see above and Fig. 7), although there was 50%o sequence identity. Although its amino acid sequence is quite different from theirs, CAP37 peptide 20-44 resembles other known naturally occurring antibiotic peptides such as the mammalian and insect defensins (2), which tend to be low molecular mass polypeptides, are characteristically cationic and/or hydrophobic, and have distinctly different antibiotic spectrums. BPI, a very large antimicrobial protein, also bears some similarities to CAP37 in that both are cationic and hydrophobic and both affect Gram-negative bacteria. Yet there is no sequence homology between the principal antibiotic domain of BPI and of peptide 20-44. The bactericidal domain of BPI is confined to a peptide, obtained by limited proteolysis of the purified native protein, and comprising its 199 amino-terminal residues (29). The entire CAP37 protein is strongly basic, having 23 positively charged residues and 15 acidic residues (13). What accounts for the antibiotic activity of peptide 20-44? The net charge on peptide 20-44 is +2. However, at pH < 6 the charge on the peptide is more likely to be +4, owing to the presence of two histidine residues. Charge alone, however, fails to explain its activity, since peptide 130-146 (Fig. 1) has a +4 charge but insignificant activity. Perhaps peptide 20-44 contributes to the overall antibiotic action of CAP37 by combination of its hydrophobicity (56% of its amino acid residues are hydrophobic), a basic charge, and relatively small size. Such features would facilitate its interaction with lipid A on the surface of Gram-negative organisms. LPS, and two lipid A preparations, inhibited the antibiotic action of peptide 20-44 (Fig. 8) and CAP37 (not shown), strongly suggesting that the bactericidal and lipid A binding domains of CAP37 are the same. An S. typhimurium strain, SH7426, resistant to polymyxin B was cross-resistant to CAP37 (9). Indeed, SH7426 is also cross-resistant to peptide 20-44. This bacterial phenotype may be due to substitution of 4-aminopentose on the acidic

We thank Drs. H. K. Ziegler, W. M. Shafer, P. Fields, J. R. Scott, and Y. Y. Yu for providing us with the various bacterial strains used in this study, Dr. G. Churchward for performing the data base searches, and Dr. K. W. Jackson, University of Oklahoma, Oklahoma City, for the mass spectrometric assay of peptide 20-44. The technical help of Larry Martin, Nancy Moore-Martin, Debra Houry, Beverly Boltin, and Lance Wells is gratefuily acknowledged. This research was supported by research grants A128018 (H.A.P.) and AI26589 (J.K.S.) from the U.S. Public Health Service of the United States and a research grant award from the American Lung Association (H.A.P.). I.E. is supported by scholarships from the Council of Higher Education, Government of Turkey, Ankara, and Ege University, Izmir, Turkey. 1. 2. 3. 4.

Spitznagel, J. K. (1990) J. Clin. Invest. 86, 1381-1386. Lehrer, R. I., Ganz, T. & Selsted, M. E. (1991) Cell 64, 229-230. Odeberg, H. & Olsson, I. (1975) J. Clin. Invest. 56, 1118-1124. Shafer, W. M., Onunka, V. & Martin, L. E. (1986) Infect. Immun. 54,

184-188. 5. Wasiluk, K. R., Skubitz, K. M. & Gray, B. H. (1991) Infect. Immun. 59, 4193-4200. 6. Thorne, K. J. I., Oliver, R. C. & Barrett, A. J. (1976) Infect. Immun. 14, 555-563. 7. Campanelli, D., Detmers, P. A., Nathan, C. F. & Gabay, J. E. (1990) J. Clin. Invest. 85, 904-915. 8. Shafer, W. M., Martin, L. E. & Spitznagel, J. K. (1984) Infect. Immun. 45, 29-35. 9. Shafer, W. M., Martin, L. E. & Spitznagel, J. K. (1986) Infect. Immun. 53, 651-655. 10. Gabay, J. E., Scott, R. W., Campanelli, D., Griffith, J., Wilde, C., Marra, M. N., Seeger, M. & Nathan, C. F. (1989) Proc. Natd. Acad. Sci. USA 86, 5610-5614. 11. Weiss, J., Elsbach, P., Olsson, I. & Odeberg, H. (1978) J. Biol. Chem. 253, 2664-2672. 12. Pereira, H. A., Shafer, W. M., Pohl, J., Martin, L. E. & Spitznagel, J. K. (1990) J. Clin. Invest. 85, 1468-1476. 13. Pohl, J., Pereira, H. A., Martin, N. M. & Spitznagel, J. K. (1990) FEBS Lett. 272, 200-204. 14. Morgan, J. G., Sukiennicki, T., Pereira, H. A., Spitznagel, J. K., Guerra, M. E. & Larrick, J. W. (1991) J. Immunol. 147, 3210-3214. 15. Flodgaard, H., Ostergaard, E., Bayne, S., Svendsen, A., Thomsen, J., Engels, M. & Wollmer, A. (1991) Eur. J. Biochem. 197, 535-547. 16. Almeida, R. P., Melchior, M., Campanclli, D., Nathan, C. F. & Gabay, J. E. (1991) Biochem. Biophys. Res. Commun. 177, 688-695. 17. Sukupolvi, S., Vaara, M., Helander, I. M., Viljanen, P. & Makela, P. H. (1984) J. Bacteriol. 159, 704-712. 18. Shafer, W. M., Casey, S. G. & Spitznagel, J. K. (1984) Infect. Immun. 43, 834-838. 19. Farley, M. M., Shafer, W. M. & Spitznagel, J. K. (1987) Infect. Immun. 55, 1536-1539. 20. Shafer, W. M., Engle, S. A., Martin, L. E. & Spitznagel, J. K. (1988) Infect. Immun. 56, 51-53. 21. Merrifeld, R. B. (1986) Science 232, 341-347. 22. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) in Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Lab., Plainview, NY), 2nd Ed., pp. 55-73. 23. Belas, R., Erskine, D. & Flaherty, D. (1991)J. Bacteriol. 173, 6289-6293. 24. Bangalore, N., Travis, J., Onunka, V. C., Pohl, J. & Shafer, W. M.

(1990) J. Biol. Chem. 265, 13584-13588.

25. Selsted, M. E. & Harwig, S. S. L. (1989) J. Biol. Chem. 264, 4003-4007. 26. Vaara, M. (1981) Drugs Exp. Clin. Res. 17, 437-444. 27. Pereira, H. A., Spitznagel, J. K., Pohl, J., Wilson, D. E., Morgan, J. G., PaLings, I. & Larrick, J. W. (1990) Life Sci. 46, 189-196. 28. Shafer, W. M., Pohl, J., Onunka, V. C., Bangalore, N. & Travis, J. (1991) J. Biol. Chem. 266, 112-116. 29. Ooi, C. E., Weiss, J., Doerfler, M. E. & Elsbach, P. (1991) J. Exp. Med. 174, 649-655.