Identification of CRAMP, a Cathelin-related Antimicrobial Peptide Expressed in the Embryonic and Adult Mouse*

Cathelicidins are the precursors of potent antimicrobial peptides that have been identified in several mammalian species. Prior work has suggested that members of this gene family can participate in host defense through their antimicrobial effects and activate mesenchymal cells during wound repair. To permit further study of these proteins a reverse transcriptase-polymerase chain reaction approach was used to identify potential mouse homologs. A full-length 562-base pair cDNA clone was obtained encoding an NH2-terminal prepro domain homologous to other cathelicidins and a unique COOH-terminal peptide. This gene, namedCramp for cathelin-relatedantimicrobial peptide, was mapped to chromosome 9 at a region of conserved synteny to which genes for cathelicidins have been mapped in pig and man. Northern blot analysis detected a 1-kilobase transcript that was expressed in adult bone marrow and during embryogenesis as early as E12, the earliest stage of blood development. Reverse transcriptase-polymerase chain reaction also detected CRAMP expression in adult testis, spleen, stomach, and intestine but not in brain, liver, heart, or skeletal muscle. To evaluate further the expression and function of CRAMP, a peptide corresponding to the predicted COOH-terminal region was synthesized. CD spectral analysis showed that CRAMP will form an amphipathic α-helix similar to other antimicrobial peptides. Functional studies showed CRAMP to be a potent antibiotic against Gram-negative bacteria by inhibiting growth of a variety of bacterial strains (minimum inhibitory concentrations 0.5–8.0 μm) and by permeabilizing the inner membrane of Escherichia colidirectly at 1 μm. Antiserum against CRAMP revealed abundant expression in myeloid precursors and neutrophils. Thus, CRAMP represents the first antibiotic peptide found in cells of myeloid lineage in the mouse. These data suggest that inflammatory cells in the mouse can use a nonoxidative mechanism for microbial killing and permit use of the mouse to study the role such peptides play in host defense and wound repair.

Endogenous antimicrobial peptides play an important role in innate immunity (1)(2)(3). More than 100 microbicidal peptides have been isolated from plants and animals (4). The role of these defense peptides in mammals has been inferred from their expression in neutrophil granules and at sites exposed to multiple microbes such as the skin and gastrointestinal tract. To exert their antimicrobial effect these peptides adhere to and permeabilize the surface membranes of potential pathogens. This activity is a consequence of several common features such as a high content of basic residues and the tendency of some to adopt an amphipathic conformation. However, antimicrobial peptides show marked diversity in structure and antimicrobial spectrum. One class of antimicrobial peptides, the cathelicidinderived peptides, contains a highly conserved prepro region that is homologous to cathelin, a putative cysteine-proteinase inhibitor originally isolated from pig leukocytes (5). Cathelicidins have been identified in several species including pig, cow, sheep, rabbit, and man (6 -13). The high degree of conservation of their cathelin domain suggests that the members of this family evolved from a common ancestor gene through duplication and modification (7). In general, the precursors of these peptides are stored in neutrophil granules. Upon stimulation, the cathelin domain is cleaved proteolytically to allow the mature COOH-terminal antimicrobial peptide to be released.
The antimicrobial portion of the cathelicidin-derived gene family is highly diverse in terms of structure and function. In the pig, the cathelicidin PR-39 has been found to have potent activity against Gram-negative bacteria and also to function as a stimulator of syndecan-1 and -4 expression on fibroblasts and endothelia (14). Observations of functions beyond antimicrobial activity have also been made for other host defense peptides (15) and suggest the need for further study of the role these peptides play in vivo.
The mouse is a highly useful animal model to study the function of the immune system in events such as wound repair and cutaneous inflammation. Surprisingly, despite the well characterized immune system in the mouse, antimicrobial peptides have only been identified in the mouse intestine (16,17) and appear to be absent from neutrophils (18). Thus, in this investigation we sought to identify a mouse member of the cathelicidin gene family and describe its expression and antimicrobial function. We report a full-length cDNA sequence derived from mouse marrow which is a member of the cathelicidin gene family. This gene, named Cramp, for cathelin-related antimicrobial peptide, mapped to a single region on murine chromosome 9, homologous to the map locations of cathelicidins in man and pig. Transcripts for Cramp were expressed in multiple mature tissues and during embryogenesis. Finally, CRAMP protein was identified by immunostaining in murine bone marrow cells and neutrophils and behaved structurally and functionally as a potent antimicrobial agent.
cDNA Cloning-Total RNA was extracted from C57BL/6 mouse femoral marrow cells with guanidinium thiocyanate (19). To isolate potential murine cDNA homologs to the cathelin-related antimicrobial peptides, a 5Ј and 3Ј RACE strategy was applied (Life Technologies, Inc.; 3Ј and 5Ј RACE systems, Gaithersburg, MD). For 3Ј RACE, cDNA synthesis was carried out using adapter primer 5Ј-GGCCACGCGTCGACTAG-TAC(T) 17 -3Ј. Amplification toward the 3Ј end was done first with the universal amplification primer 5Ј-(CUA) 4 GGCCACGCGTCGACTAGT-AC-3Ј and with a cathelin-specific primer-1, 5Ј-TCGGAAGCTAATCT-CTAC-3-3Ј, which was designed based on a base pairs 165-182 sequence of porcine prepro-PR-39. A second nested amplification was then done with a cathelin-specific primer-2, 5Ј-(CAU) 4 CTGGACCAGCCGC-CCAAG-3Ј designed based on base pairs 195-212 of prepro-PR-39 and the universal amplification primer. For amplification toward the 5Ј end, the gene-specific primer-1, 5Ј-TTTGCGGAGAAGTCCAGC-3Ј, based on a sequence derived from 3Ј RACE, was used for cDNA synthesis. Amplification was done with an anchor primer, provided by Life Technologies, Inc., and with a gene-specific primer-2, 5Ј-(CAU) 4 GAAATT-TTCTTGAACCG-3Ј. Products of 3Ј and 5Ј RACE were cloned into pAMP1 for sequencing by automated sequencer (model 373A, Perkin-Elmer) in both directions using primers against M13 and T7. Three independent clones were sequenced in both directions.
Northern Blot Analysis-Northern blot analysis of total RNA was performed as described previously (14). Approximately 10 g of total RNA was extracted from whole C57BL/6 mouse embryos at different gestational ages and separated by electrophoresis through a 1% agarose/formaldehyde gel. RNA was transferred to a GeneScreen Plus membrane (DuPont NEN). Hybridization was carried out at 65°C in QuikHyb Solution (Stratagene) and probed using [ 32 P]dCTP random primer-labeled cDNA, corresponding to base pairs 79 -249 of murine CRAMP cDNA. Filters were washed twice for 15 min in 2 ϫ SSPE (0.18 M NaCl, 0.01 M Na 2 H 2 PO 4 , 1 mM EDTA, pH 7.7), 0.1% sodium dodecyl sulfate at room temperature and then twice in 0.2 ϫ SSPE, 0.1% sodium dodecyl sulfate at 55°C.
Revere Transcriptase-PCR-Total RNA from marrow, testis, stomach, small intestine, liver, lung, skeletal muscle, brain, heart, spleen, kidney, and large intestine of 12-week-old C57BL/6 mice was prepared with guanidinium thiocyanate (19). Approximately 5 g of DNasetreated total RNA from each tissue was annealed at 42°C with random hexamers in a solution containing 10 mM Tris-HCl, pH 8.3, 50 mM KCl, and 5 mM MgCl 2 . The annealed hexamers were then extended by adding dNTP at a final concentration of 0.1 mM, 1.5 units of RNasin (Promega, Madison), and 1 unit of Superscript reverse transcriptase (Life Technologies, Inc.) at 42°C for 90 min. The resulting cDNA was then amplified with the specific primers 5Ј-GCTGATGTCAAAAGAAT-CAGCG-3Ј and 5Ј-TCCCTCTGGAACTGCATGGTTCC-3Ј, based on base pairs 10 -32 and 357-378, respectively, of the CRAMP cDNA sequence. These primers were used with the following thermal cycle profile: 95°C for 5 min; 20 cycles of 94°C for 1 min, 55°C for 1 min, and 72°C for 1.5 min; and a final extension step of 72°C for 7 min. A second round of PCR was performed with nested primers 5Ј-CTGTGGCGGT-CACTATCACT-3Ј and 5Ј-GTTCCTTGAAGGCACATTGC-3Ј based on base pairs 49 -68 and 291-310, respectively, of the CRAMP cDNA sequence. Amplification conditions were as above, but a higher annealing temperature of 58°C was used. Products were separated on 3% agarose gel and photographed using Eagle Eye apparatus (Stratagene).
Analytical Assays-The peptide concentration was measured by the absorbance of Phe at 257 nm and using a molar extinction coefficient of 195 (20). Mass determinations were performed with an API I ionspray mass spectrometer (PE SCIEX, Toronto, Canada). Circular dichroism spectra were recorded at room temperature on a Jasco J-600 spectropolarimeter (Jasco Corp., Tokyo, Japan). Peptide samples (10 -25 M) were dissolved in 5 mM sodium phosphate buffer, pH 7.0, in the absence or presence of 15, 30, and 45% (v/v) trifluoroethanol. The ␣-helical content was estimated by the method of Chen et al. (21) from the [] 222 measurements, using Ϫ1,500 and Ϫ39,500 (1-2.5/n) deg-cm 2 dmol Ϫ1 , as the values for 0% and 100% of helix, respectively, with n indicating the number of residues.
Bacterial Growth Suppression and Membrane Permeabilization-Antibacterial and antifungal minimum inhibitory concentrations (MIC) were determined by a microdilution susceptibility test described previously (9). The following bacterial strains were tested: Escherichia coli ATCC 25922, ML-35 and D21, Salmonella typhimurium ATCC 14028, Pseuodomonas aeruginosa ATCC 27853, Serratia marcescens ATCC 1800, Proteus vulgaris ATCC 13315, Staphylococcus aureus ATCC 25923, Cowan 1, and two methicillin-resistant clinical isolates carrying the mecA gene (provided by L. Dolzani, Department of Biomedical Sciences, University of Trieste, Italy), Staphylococcus epidermidis ATCC 12228, and Bacillus megaterium BM 11. The MIC values were determined after incubation at 37°C for 18 h. The antifungal activity was determined using clinical isolates of Candida albicans and Cryptococcus neoformans. The assay conditions were similar to those used for bacteria, except that the fungal species were grown and tested in Sabouraud liquid medium, and the MIC was determined after incubation at 30°C for 36 -48 h.
The bactericidal activity of CRAMP-1 and CRAMP-2 against midlog phase cultures of E. coli ATCC 25922, P. aeruginosa ATCC 27853, and S. aureus ATCC 25923 was tested in low ionic strength buffer (10 mM sodium phosphate buffer, pH 7.4). Bacteria (0.4 -0.6 ϫ 10 6 colonyforming units/ml) were incubated in the absence (controls) or in the presence of different amounts of peptides in a 96-well microtiter plate (final volume of 150 l). After a 1-h incubation at 37°C, samples were serially diluted with sterile PBS, plated in duplicate on Mueller-Hinton agar, and incubated for 16 -18 h to allow colony counts (9). Bacterial inner membrane permeabilization was evaluated with the E. coli ML-35 strain as described previously (22).
Generation of Primary Antibody-Antibodies to synthetic CRAMP-1 were raised in New Zealand White rabbits by an initial intramuscular injection of 150 g of CRAMP-1 in complete Freund's adjuvant followed by five booster injections of 150 g of CRAMP in incomplete Freund's at 3-week intervals. Serum was collected 10 days after boosting.
Immunohistochemistry-4 ϫ 10 5 Swiss-Webster mouse bone marrow cells in 100 l of PBS, 1% bovine serum albumin, or 5 ϫ 10 6 C57BL/six peripheral blood cells in Tris-buffered saline, 5 mM EDTA, were attached to glass slides by cytospin at 400 ϫ g for 5 min in a Shandon Cytospin 3 cytocentrifuge, fixed in methanol, then blocked for 1 h with a solution of 5% goat serum in PBS, 1% bovine serum albumin. Slides were then incubated for 1 h with a 1:200 dilution of rabbit anti-CRAMP antiserum or nonimmune rabbit serum. To confirm specificity of antibody binding, parallel slides were treated identically with rabbit-anti CRAMP serum that had been preincubated for 1 h at room temperature with 20 g/ml synthetic CRAMP-1. Following primary antibody incu-bation slides were washed three times for 5 min in PBS and then incubated for 1 h at room temperature with a 1:250 dilution of fluorescein-conjugated goat anti-rabbit IgG (Cappel Research Products, Durham, NC). Each slide was washed three times for 5 min with Tris-buffered saline. Coverslips were mounted with Prolong 228 Antifade mounting medium (Molecular Probes, Eugene, OR) and cells photographed with Kodak Elite II Ektachrome ASA400 film on a Zeiss Axiophot microscope.

Molecular Cloning of a Cathelin-related cDNA in the
Mouse-To isolate potential murine homologs to the family of antimicrobial peptides related by a cathelin-like domain, primers were designed using sequence information from the conserved cathelin-like domain of the porcine gene PR-39. A nested RACE PCR strategy was used, capitalizing on the highly invariant cathelin-like prepro domain from all species studied to date. Three independently derived clones were sequenced in both directions to give the cDNA sequence shown in Fig. 1. No other related cDNAs were identified by this approach despite screening more than 500 clones for isolates related by the 5Ј cathelin domain but distinct in the 3Ј region.
The murine cathelin-related cDNA contains an open reading frame of 516 base pairs and encodes a predicted translation product of 172 amino acids. Sequence comparisons between the prepro regions of the predicted murine (CRAMP) pig (PR-39) and human (FALL-39) cathelicidins identify significant primary sequence similarities; the murine protein maintains 52% identity with PR-39 and FALL-39 in the cathelin-like domain and 80% identity with either individually (Fig. 2). This sequence similarity includes the conservation of cysteines at positions 83, 94, 105, 122 and potential processing sites at the carboxyl terminus of the cathelin portion of the peptide. The previously reported cDNA sequences of human FALL-39 and porcine PR-39 encode different peptides in the 3Ј region. Similarly, the predicted translation product of CRAMP is distinct in this region.
Cramp mRNA Is Expressed during Embryogenesis and in Adult Tissues-To evaluate the expression of Cramp mRNA in the mouse, RNA was extracted from whole murine embryos isolated at various gestational ages and from whole adult mouse tissues. Northern blot of total RNA from whole embryos demonstrated abundant 1-kilobase message (Fig. 3, A and B) mRNA was detectable as early as gestational days 12 and 13 and increased relative to ribosomal RNA during development. In adult tissues, this transcript was detectable by Northern blot analysis only in marrow total RNA (data not shown). To increase the sensitivity for detection in adult tissues, a reverse transcriptase-PCR strategy was employed. Total RNA from brain, heart, spleen, kidney, testis, colon, liver, marrow, stomach, small intestine, lung, and skeletal muscle was prepared, and reverse transcriptase-PCR was performed using nested primers to specifically amplify cDNA. As a control for all reverse transcriptase-PCR experiments, primers selected for ␤-actin were chosen, and reactions were included lacking either RNA or reverse transcriptase. Cramp transcripts were detectable in spleen, testis, colon, marrow, stomach, and small intestine (Fig. 3C). Faint bands corresponding to Cramp were also seen in heart, lung, and skeletal muscle but were not detectable in brain, kidney, and liver. All samples were positive with actin-specific primers and negative when performed in the absence of RNA or reverse transcriptase.
Cramp Is Located on Mouse Chromosome 9 -Using the probe spanning base pairs 79 -249 of Cramp, we identified ScaI fragments of 10.2 kilobases in parental M. musculus DNA, 11.7 in NFS/N and C58/J, and 14.2 in M. spretus. Inheritance of the variant fragments was typed in the progeny of two sets of genetic crosses and compared with inheritance of more than 1000 genetic markers previously typed and mapped in these mice. CRAMP showed linkage to markers on distal Chr 9 (Fig.  4) and was positioned just distal to Gnat1. This region of mouse Chr 9 is homologous to human 3p23-21, consistent with the location of the human homolog of this gene, Fall-39.
CRAMP Can Form an Amphipathic ␣-Helix-The translation products of the cathelicidins are processed to release carboxyl-terminal peptides with antimicrobial activity. To investigate the properties of the carboxyl-terminal peptide predicted by the CRAMP cDNA, two peptides were synthesized. The first FIG. 1. Nucleotide and predicted amino acid sequence of CRAMP. The nucleotide sequence is numbered on the left, and the amino acid sequence is numbered from the first methionine on the right. The sequence of the predicted mature protein is underlined. The stop codon is marked with an asterisk, and the cysteine residues are in boldface type. peptide, CRAMP-1, was prepared based on the predicted elastase cleavage site at position 138 -139 and is a 33-amino acid peptide. The second, CRAMP-2, was prepared based on potential processing at the dibasic site at position 133-134 and is a 38-amino acid peptide. Helical wheel projections of both CRAMP-1 and CRAMP-2 predict that these peptides will form an amphipathic helix similar to that for FALL-39 (13) and non-cathelin-related antimicrobial peptides such as the maganins and cecropins. To evaluate this directly, circular dichroism (CD) spectral analysis of CRAMP-1 and CRAMP-2 was done (Fig. 5). The spectra of 10 -25 M synthetic CRAMP-1 in buffer suggest that the peptide is in a random coil configuration and that it will assume a helical structure in the presence of 15-45% (v/v) trifluoroethanol. The CD spectra of CRAMP-2 (not shown) were similar to CRAMP-1.
CRAMP Is a Potent Antimicrobial Peptide-The capacity of CRAMP-1 and CRAMP-2 to function as antimicrobial agents was evaluated directly on a variety of microbes (Table I). Both peptides demonstrated potent antimicrobial activity against Gram-negative bacteria with MIC values in the range of 0.5-8.0 M. CRAMP peptides were less active against Gram-positive strains (MIC 32-64 M) and not active at the concentrations tested against P. vulgaris, C. albicans, and selected strains of S. aureus. The CRAMP peptides were also found to have potent direct bactericidal activity when assayed in low ionic strength buffer (Table II). Under these conditions even the Gram-positive bacterium (S. aureus) was killed at concentrations as low as 1 M.
The kinetics of E. coli inner membrane permeabilization was evaluated for both peptides. The reaction rate for permeabilization with CRAMP-2 was faster than for CRAMP-1 (Fig. 6). At a concentration of 2.5 M, CRAMP-2 has the capacity to permeabilize bacteria and thus unmask cytoplasmic ␤-galactosidase activity at a rate that is 93% of that achieved by sonication. Steady-state bacterial permeabilization was reached after 4 min. CRAMP-1 at the same concentration induced about 60% permeabilization with steady state achieved by 8 min. Despite the potent ability to lyse bacterial inner membranes, the CRAMP peptides (50 M) did not lyse human or sheep eryth- FIG. 3. CRAMP is expressed during embryogenesis and in adult mouse tissues. Panel A, Northern analysis of total RNA from mouse embryos at indicated gestational age. kb, kilobases. Panel B, graph quanitifying the relative abundance of CRAMP mRNA/28 S RNA. Panel C, photograph of reverse transcriptase-PCR products from mRNA isolated from adult mouse tissues. Lanes: 1, molecular weight standards; 2, brain; 3, heart; 4, spleen; 5, kidney; 6, testis; 7, colon; 8, liver; 9, bone marrow; 10, stomach; 11, small intestine; 12, lung; 13, skeletal muscle.

TABLE I
Antimicrobial activity of CRAMP-1 and CRAMP-2 MIC was defined as the lowest concentration of peptide preventing visible growth after incubation with bacteria for 18 h or with fungi for 36 -48 h. All strains tested were grown in Mueller-Hinton broth, except B. megaterium, which was grown in LB medium, and C. albicans and C. neoformans, which were grown in Sabouraud liquid medium. Results were determined using approximately 1.5 ϫ 10 5 (bacteria) and 0.5 ϫ 10 5 (fungi) colony-forming units/ml and are the mean of at least three independent determinations with a divergence of not more than one MIC value with respect to those reported here.  16 16 rocyte membranes (data not shown).

Immunolocalization of CRAMP in Granulocytes-
The expression of CRAMP protein in vivo was evaluated with a polyclonal antibody prepared against CRAMP-1. Prominent immunostaining was detectable in myeloid cells present within mouse femoral marrow (Fig. 7A). This staining was specific for CRAMP since addition of excess synthetic peptide abolished staining (Fig. 7B). In peripheral leukocytes, CRAMP was detected in abundance in neutrophil granules (Fig. 7C) but was not seen in monocytes, lymphocytes, or eosinophils. Thus, CRAMP protein is expressed in vivo in a location similar to that shown for cathelin-related antimicrobial peptides in other species.

DISCUSSION
In this study we identify, molecularly clone, and characterize a novel antimicrobial peptide in the mouse. This molecule is a new member of a large family of antimicrobial peptide precursors found in human (13,28), pig (6,8,14,29,30), cow (9, 31), rabbit (12), and sheep (10,11). The homology among the precursors of these antimicrobial peptides lies in a common NH 2terminal region of the unprocessed precursor known as cathelin and has led to use of the term cathelicidins to describe the gene family. The mouse gene described here shares a high degree of homology with the cathelin domain contained within all members of this family; it maps to a region of mouse distal chromosome 9 homologous to the chromosomal location of cathelicidins in pig and human, it is antimicrobial, and it is expressed primarily in neutrophil granules. Therefore, this gene, named Cramp, encodes a new endogenous antimicrobial peptide and identifies a novel host defense molecule in mouse for further study of mechanisms of innate immunity.
The cathelicidins exert their antimicrobial effect based on the structurally diverse COOH-terminal domain of these peptides (7). To identify candidate cathelicidins in mouse, primers were designed against various regions within the highly evolutionarily conserved NH 2 -terminal cathelin domain of these gene products. This approach has been successful previously in identifying multiple cathelin-related genes in several species (7). Only a single cathelicidin was identified by this approach in mouse despite the use of multiple primers and tissue sources for RNA and large scale screening of more than 500 cathelinrelated clones. An additional myeloid gene product, B9, has been described in mouse which contains a cathelin-like domain yet is not known to be antimicrobial (32). The inability to detect this transcript may reflect limitations inherent in a reverse transcriptase-PCR approach such as difficulties in detecting transcripts in low abundance. Thus, although no additional antimicrobial cathelicidin has been described in mice, the current work cannot rule out the presence of multiple cathelinrelated antimicrobial peptides such as found in other species.
Identification of a murine cathelin-related gene enabled us to use the mouse for study of tissue-specific and developmental expression. We found that the expression of Cramp transcripts in the mouse is regulated in a tissue-specific and temporal developmental pattern. In the adult, transcript expression was most abundant in marrow. The finding was consistent with our observation that the protein was found abundantly in granulocytes and in marrow cells of myeloid lineage and agrees with the sites of expression of cathelicidins in man and pig (14,28,33). However, CRAMP mRNA was also detectable at low levels in mouse testis, the gastrointestinal tract, spleen, heart, lung, and skeletal muscle. It is possible that the presence of CRAMP transcripts in a variety of tissues is due to the presence of small numbers of myeloid precursors in those organs. However, a lack of CRAMP transcripts in liver, kidney, and brain, tissues that are likely to contain significant amounts of blood cell contaminants, argues against this. Furthermore, pig cathelicidins have been found in the gastrointestinal tract, and the human cathelicidin FALL-39 has been found in testis (13). Our findings are therefore supported by the tissue-specific expres-  sion of cathelicidins in other species.
Analysis of whole embryo RNA during development demonstrated that CRAMP transcripts are also expressed during embryogenesis. Expression was detected as early as embryonic day 12 in the mouse and increased subsequently. The timing of CRAMP expression coincides with the earliest colonization of the fetal murine liver with hematopoietic stem cells but precedes the establishment of granulocytic lineages (34). Thus, as in adult tissues, CRAMP RNA may be expressed by nonimmune cells. The significance of these observations is 2-fold. First, finding constitutive expression of antimicrobials in a variety of tissues suggests that these tissues may have an innate ability to resist microbial infection before or without stimulating an inflammatory cell influx. Secondly, certain antimicrobial peptides such as PR-39 are known to possess additional biologic activity including the ability to function as stimulators of cell surface syndecan expression (14). Expression of CRAMP in a variety of tissues and during embryogenesis therefore is consistent with functions in addition to antimicrobial activity. Further work to evaluate cell type-specific expression and biological function of CRAMP is needed to address these issues directly.
The 3Ј region of CRAMP predicts a translation product that is an effective antimicrobial peptide. Analysis of the processing of the cathelicidins suggested that CRAMP may be cleaved at multiple sites (28,35). Helical wheel projections of two potential cleavage products predicted that the mature peptides would acquire an ␣-helical conformation with an amphipathic arrangement. This structural prediction is similar to a subset of cathelicidins which includes the human FALL-37, rabbit CAP 18, and pig PMAP-37, but not the cathelicidins characterized by proline and arginine-rich or tryptophan-rich mature peptides. Direct analysis of both synthetically produced CRAMP peptides confirmed this structural prediction and established the antimicrobial function of CRAMP. CD spectral analysis demonstrated that these synthetic peptides can assume an ␣-helical structure. Antimicrobial assays done against a variety of microbes demonstrated potent activity against Gram-negative bacteria and rapid permeabilization of the inner membrane of E. coli. These results demonstrate that CRAMP is a potent, membrane-active antibacterial peptide. Recently, a similar cDNA from the mouse was described (36). The corresponding synthetic peptide lacked five NH 2 -terminal amino acids (GLLRK) used here in CRAMP-1 and was not antimicrobial. This deletion is likely to be responsible for the lack of antimicrobial activity because structural analysis done in other cathelicidin-derived peptides has shown that the NH 2terminal domain of the mature peptide is critical for activity (37). Purification of the native peptide from the mouse will be necessary to establish its native sequence.
In summary, identification of CRAMP and characterization of its expression and antimicrobial activity demonstrate that the murine host defense system includes members of the cathelicidin family. This information permits many powerful exper-imental tools available in the mouse to be applied to the study of these antimicrobial peptides. Such work may lead to greater insight into the function of these peptides in protecting organisms from infection and influencing cell behaviors.