The Greater Celandine: Identification and Characterization of an Antimicrobial Peptide from Chelidonium majus

Abstract

Chelidonium majus, known as Greater Celandine, is a latex bearing plant that has been leveraged for its anticancer and antimicrobial properties. Herein, C. majus aerial tissue is mined for the presence of antimicrobial peptides. A highly abundant, cysteine-rich peptide with a length of 25 amino acids, deemed CM-AMP1, is characterized through multiple mass spectrometric approaches. Electron-activated dissociation is leveraged to differentiate between isoleucine and leucine residues, and complement conventional collision-induced dissociation to gain full sequence coverage of the full-length peptide. CM-AMP1 shares little sequence similarity to any proteins in publicly available databases, highlighting the novelty of its cysteine landscape and core motif. The presence of three disulfide bonds in the native peptide confers proteolytic stability, and antimicrobial activity is greatly decreased upon alkylation of the cysteine residues. Synthetic variants of CM-AMP1 are used to confirm activity of the full-length sequence and the core motif. To assess the biological impact, E. coli was grown in a sub-lethal concentration of CM-AMP1 and quantitative proteomics was used to identify proteins produced by the bacteria under stress, ultimately suggesting a membrane lytic antimicrobial mechanism of action. This study integrates multiple analytical methods for molecular and biological characterization of a unique antimicrobial peptide identified from C. majus.

Graphical Abstract

Antimicrobial peptides (AMPs) are produced ubiquitously across all kingdoms as a part of the innate immune response and have been shown effective against a broad spectrum of pathogens, including fungi, insects, and bacteria.^1–3 AMPs’ mechanism of action typically involves insertion into the target cell membrane which results in lysis of the pathogen.¹ The broad-spectrum nature of AMPs can be attributed to several common physiochemical properties. AMPs are typically less than 100 amino acids long, and are known for being highly basic, amphipathic, and cysteine rich.¹ Plant AMPs are classified based on their cysteine motifs, and are broadly sorted into seven major families: thionins, defensins, hevein-like, knottins, α-hairpinins, lipid transfer proteins, and snakins.^4–6 These cysteine motifs form disulfide bonds across the peptide and help confer structure and stability, both of which are crucial components to the overall activity against potential pathogens.⁷ To date over 3,500 naturally occurring AMPs have been discovered, with nearly 75% of them originating from animals.⁸ Plant AMPs represent a mere 10% of the antimicrobial peptide database,⁸ despite the longstanding prevalent use of plant extracts for medicinal use.^9–11

Mass spectrometry (MS) is an invaluable tool for discovery and molecular characterization of AMPs.¹² MS-based strategies can be used to determine primary peptide sequence and post-translational modifications with relatively low amounts of starting material compared to other methods, such as NMR or Edman degradation. Collision-induced activation (CID) has long been used for peptide / protein sequencing; however, evolving instrument capabilities have expanded the types of MS-based experiments possible. Modern instruments can use MSⁿ and electron-based fragmentation modes to obtain orthogonal fragment information, which ultimately aids in sequence characterization and reduces the need for alternative experiments.¹² Additionally, chemical derivatization followed by MS can readily aid in determining amino acid stoichiometry and structural motifs, such as disulfide bonds.¹² Increased availability of genomic information and the advancement of bioinformatic pipelines have expedited the discovery of AMPs. Programs such as Cysmotif searcher¹³ and SignalP¹⁴ have rapidly evolved with this expanded data availability, and increased the capacity for researchers to predict AMP candidates from sequence information. Collectively, advancements in MS, chemical derivatization strategies, and bioinformatics have catapulted researchers to explore a wide breadth of plants with historic medicinal applications.

Latex-bearing plants have been a compelling source of bioactive compounds for the past few decades, drawing attention due to their antimicrobial to anticancer activities.^15–17 It is a very diverse group of more than 20,000 plant species across over 40 families of angiosperm plants.¹⁸ Typical families include Euphorbiaceae, Asteraceae, Apocynaceae and Papaveraceae.^{19, 20} Laticiferous species are more frequently observed in tropical than in temperate regions.¹⁸ The metabolite composition across different latex-bearing plants varies, with a range of diverse chemical compounds, like alkaloids (Papaveraceae, e.g. Macleaya cordata, Papaver somniferum, Chelidonium majus), terpenoids (Euphorbiaceae, e.g. Hevea brasiliensis), phenolic glucosides (Cannabaceae, e.g. Cannabis sativa) or cardenolides (Asclepiadaceae, e.g. Asclepias curassavica).²¹ Some compounds are of pharmaceutical and economic importance, like the narcotic drug morphine from P. somniferum or the industrial raw material natural rubber (cis-1, 4-polyisoprene) from H. brasiliensis.²²

One pharmaceutically important species belonging to Papaveraceae is Chelidonium majus, or Greater Celandine, which is a common herbaceous plant found worldwide, mostly across Europe, East Asia and introduced to North and South America, Australia, and New Zealand. Its metabolite profile includes compounds like benzylisoquinoline alkaloids (e.g. chelidonine, sanguarine, coptisine, berberine, protopine), flavonoids, phenolic acids, unsaturated fatty acids, carotenoids, saponins, and defense-related proteins.^{15, 16, 23–25} Extracts have been used against skin ailments, like eczema, corns, Tinea infections, and tumors of the skin, but also for liver and biliary disorders and as a chalogogue.²⁶ Moreover, it possesses antiviral properties utilized in traditional folk medicine to remove papillae or warts from the skin, which result from human papilloma virus (HPV) infection.^{26, 27} Alkaloids and flavonoids produced by this plant have shown strong antimicrobial activity;¹⁷ however, antimicrobial peptides from this plant have remained largely unexplored.

Herein, we examined Chelidonium majus aerial tissue for the presence of AMPs using PepSAVI-MS,²⁸ and discovered a unique broad-spectrum antimicrobial peptide (deemed CM-AMP1). Multiple MS techniques were implemented for sequence determination, isoleucine / leucine differentiation, and disulfide bond connectivity. Synthetic analogs of the peptide were used to characterize the structural importance of the disulfide bonds and their contribution to the bioactivity of the peptide. The microbiological impact of this potent antimicrobial peptide is explored by measuring changes in E. coli protein abundance in the presence of sub-inhibitory concentrations, which ultimately informs the mechanism of action for this new peptide.

RESULTS AND DISCUSSION

Bioactivity and LC-MS Analysis of Chelidonium majus.

C. majus aerial tissue extract was enriched for positively charged species through strong cation exchange (SCX) chromatography. The retained fraction from SCX was then subjected to separation via reversed-phase chromatography, where fractions were collected every minute to facilitate creation of a peptide library for antibiotic activity testing. The peptide library was screened against the laboratory strain E. coli ATCC 25922 and the clinically relevant S. aureus USA300 LAC, a well-characterized strain of methicillin-resistant S. aureus (MRSA). % activity for each fraction was calculated relative to the positive controls (100 μg/mL ampicillin or erythromycin for E. coli and S. aureus, respectively). A robust region of activity was present against both pathogens, with activity reaching nearly 100% activity against the strain of MRSA relative to the positive control (Figure 1A).

Figure 1.

Bioactivity profile of C. majus peptide library. A) Antibiotic activity of C. majus reversed-phase fractions (n = 3) assayed against E. coli ATCC 25922 and S. aureus LAC. Activity was calculated relative to ampicillin and erythromycin controls for E. coli and S. aureus, respectively. Error bars represent +/− 1 standard deviation away from the mean values. B) Total ion chromatogram (TIC) of fraction 24. Inset represents the MS1 level of the most abundant peak at t = 12.72 min (red dashed line).

Fractions 11–32 were subject to LC-MS analysis and resulted in detection of 1,400 features. The most active fraction, #24, primarily consisted of a single compound represented by multiple charge states (Figure 1B). The highly abundant mass (2923.31 Da) is within the expected range of known AMPs (20–100 AAs), and the presence of a high charge state (m/z = 488.22, (+6)) is indicative of multiple signature basic residues, a common characteristic for AMPs. This peptide was deemed CM-AMP1.

Full Sequence Determination of CM-AMP1.

CM-AMP1 disulfide bonds were reduced with dithiothreitol and cysteine residues were alkylated with iodoacetamide to reduce fragmentation complexity and facilitate de novo sequencing. Sequencing efforts relied on two different fragmentation techniques, collision-induced and electron activated dissociation (CID and EAD, respectively). Spectra across all present charge states (+3 – +6) were used to facilitate residue identifications in both fragmentation strategies (Figure S1). Representative fragment spectra from the +5 charge state for CID and EAD experiments are shown in Figure 2A. Cumulative analysis of all fragment data resulted in a sequence assignment of CGCYCKSVDKKRRFF(I/L)PTC(I/L)RSCCN with 100% coverage across the peptide backbone.

Figure 2.

Sequence elucidation of CM-AMP1 enabled by complementary fragmentation strategies. A) Comparison of fragmentation spectra achieved by collision-induced dissociation (CID, top) and electron-activated dissociation (EAD, bottom) of the Cys-alkylated CM-AMP1 precursor (m/z = 655.31, M = 3271.51). B) Sequence coverage of the intact CM-AMP1 peptide (top). Cumulative observed fragments across precursor charge states +3 – +6 were counted at each position for both fragmentation strategies (bottom).

Observed fragment masses across all precursor charge states were tallied to identify trends specific to a given fragmentation technique. While both strategies produced robust coverage across the peptide (23/24 bonds for CID, 20/24 for EAD) (Figure 2B), there were notable differences for each technique. Specifically, CID fragments (b/y-type ions) showed improved coverage near both termini, while EAD fragments (c/z-type ions) showed improved coverage near the interior of the peptide (Figure 2B). Previous literature has shown that CID-type activation tends to cleave at the most labile bonds, which are typically near the termini of peptides and proteins.²⁹ While each technique was able to obtain relatively good coverage across the assigned sequence, both were necessary to obtain 100% fragmentation coverage across the entire peptide sequence.

Isoleucine / Leucine Differentiation Using Electron Activated Dissociation.

CM-AMP1 sequencing revealed the presence of two amino acid locations containing either a leucine or isoleucine residue; CGCYCKSVDKKRRFF(I/L)PTC(I/L)RSCCN. Traditional collision-induced fragmentation cannot differentiate between the two isomers; however, EAD can result in the unambiguous assignment of an I/L based on distinct neutral losses in the gas phase (Figure 3A).^30–32 To achieve improved coverage across the region in CM-AMP1 containing the two potential I/L positions, a digestion with LysC was used. The C-terminal fragment resulting from this digestion was analyzed for the presence of the neutral losses associated with each residue. Identification of a Leu neutral loss at the z₆ ion and an Ile neutral loss at the z₁₀ ion confirmed the presence of these two residues at positions 16 and 20 in the full-length peptide (Figure 3B).

Figure 3.

Leucine / Isoleucine determination in CM-AMP1. A) Overview of neutral losses occurring from electron-activated dissociation (EAD) of Leu and Ile amino acids. B) EAD fragmentation spectrum of LysC digested CM-AMP1, and resulting sequence assignment.

Disulfide bonds enumeration of CM-AMP1.

Cysteine residue stoichiometry can readily be identified through carbidomethylation mass shift. Chemical derivatization of oxidized cysteine residues with iodoacetamide (IAA) will result in a shift of +58.02, while reduced cysteine residues will result in a shift of +57.02. This experiment provides keen insight into the presence and number of disulfide bonds within a sequence. When fraction 24 was subjected to reduction and alkylation, a characteristic shift of 348.18 Da was observed, corresponding to the presence of six Cys residues and three disulfide bonds (Figure S2).

Disulfide bond connectivity can be mapped using modern MS techniques. Enzymatic cleavage or partial acid hydrolysis can deconvolute the peptide into fragments connected by a single disulfide.³³ Additionally, in-source fragmentation can be leveraged to preferentially break disulfide bonds, aiding in the identification of connectivity.³⁴ Establishing the natural connectivity of cysteines within the disulfide bonds is non-trivial for this peptide sequence, as many of the cysteine residues have very few, or no residues in between (CGCYCKSVDKKRRFFIPTCLRSCCN). Efforts to map disulfides were largely unsuccessful due to the proximity of the Cys residues and lack of enzymatically accessible cleavage sites to differentiate connectivity. Additionally, non-specific acid hydrolysis produces peptides that are not readily analyzed by MS due to short length, or limited ability to carry charge. While specific connectivity was not solved, the placement of three cysteines near each termini facilitate the peptide forming a compact secondary structure regardless of disulfide pattern.

CM-AMP1 Sequence Similarity Search.

The uniqueness of the CM-AMP1 sequence can be attributed to both the cysteine motif, as well as the highly basic “core” of the sequence that falls between the two innermost cysteine residues (CKSVDKKRRFFIPTC). A BLAST search of the non-redundant sequences on NCBI reveals very little sequence homology to any known protein, with the closest match scoring an Expected (E) value of 8.7 (Table S1), suggesting that many of the hits are insignificant. While many proteins contained partial hits to the core motif, the majority of the hits only map a few consecutive residues. This highlights the uniqueness of the CM-AMP1 sequence not only in the antimicrobial peptide space, but also among all proteins deposited in NCBI.

An additional BLAST was run against the UniProtKB reference proteomes + SwissProt database, resulting in a weak alignment to several cysteine-rich domains of serine protease inhibitors (Pfam PF01826). Alignment to this domain was primarily anchored on cysteine residues, however a few auxiliary residues matched those of the protease inhibitor domain (Figure S3A). This family of protease inhibitors typically have 10 Cys residues with conserved disulfide-bond connectivity (1–7, 2–6, 3–5, 4–10 and 8–9). Commonly, conserved disulfides in known AMP families can be used to aid in the elucidation of disulfide connectivity in new AMPs. In the alignment of CM-AMP1 against this family of protease inhibitors, conserved disulfide bonds cannot be used to help infer disulfide connectivity because many of the Cys positions within the confines of the CM-AMP1 alignment are connected to Cys residues outside of the CM-AMP1 alignment (ex. 1–7).

Additionally, CM-AMP1 sequence was compared to other known classes of antimicrobial peptides containing six cysteines (Figure S3B). The cysteine residues in CM-AMP1 most closely align to the residues from jS1,³⁵ an AMP from Jasminum sambac. However the length of sequence between C3-C4 in CM-AMP1 is much longer than that of jS1, which means that structural motifs may be very different in the two species.

C. majus SignalP and Cysmotif Peptide Identification.

The CM-AMP1 peptide is not a part of any predicted genes from the C. majus genome, so the CM-AMP1 sequence was manually added to a protein fasta file for analysis through a SignalP / Cysmotif search workflow in an attempt to categorize CM-AMP1 into a known AMP family. After processing the appended C. majus genome, 139 AMPs were predicted (Table 1). CM-AMP1 was categorized into the “Cys-rich” motif, meaning that its sequence does not conform to the patterns of the other defined categories. To ensure predicted AMPs were not the source of bioactivity, fractions from the C. majus extract were reduced, alkylated, and digested. Following digestion, fractions were analyzed via MS and searched against the predicted AMP database. Only four peptides were identified in this process (Table 1, S2), one of them being CM-AMP1. The other three identified Cys-rich peptides were identified from inactive fractions, and unlikely to contribute to the observed bioactivity.

Table 1.

Output summary of SignalP and Cysmotif searcher bioinformatic analysis.

Class	# motifs predicted	# identified
Cys-rich	72	4
α- hairpin	25	-
Lipid transfer proteins	19	-
Defensins	15	-
Snakins	8	-
Hevein-like	1	-
Total	140	4

Disulfide Bonds in CM-AMP1 Confer Proteolytic Stability.

Disulfide bonds provide important biological and structural functionalities for antimicrobial peptides. Specifically, they aid in the stability of the AMP in complex biological environments and protect against proteolytic degradation. To determine if the disulfide bonds prevent proteolytic degradation, CM-AMP1 with reduced and alkylated cysteines was incubated in the presence of Proteinase K, a non-specific protease, and compared to native CM-AMP1 with intact disulfide bonds. Analysis via LC-MS revealed nearly complete recovery of the native CM-AMP1 after proteinase K exposure, while the reduced and alkylated CM-AMP1 sample resulted in nearly complete degradation of the peptide (Figure 4). This analysis confirms that disulfide bonds in CM-AMP1 are important for the proteolytic stability of the AMP.

Figure 4.

Proteolytic stability of CM-AMP1 against Proteinase K. Extracted ion chromatograms (XICs) of the +5 charge state of CM-AMP1 before and after 10 min of exposure to Proteinase K. Disulfide-bound CM-AMP1 (top) is compared to CM-AMP1 with reduced and alkylated Cys residues.

Because the CM-AMP1 sequence shows a weak alignment to a family of serine protease inhibitors, data from the Proteinase K digestion was further examined to determine if CM-AMP1 was inhibiting protease activity, or if the disulfide bonds solely increase its resistance to proteolytic degradation. While CM-AMP1 is the primary species in fraction 24, there are other low-abundance peptides in the native sample that can be leveraged to assess the ability of native CM-AMP1 to inhibit protease activity. A low-abundant peptide from the native sample (m/z = 471.51) was selected as a proxy for determining the ability for CM-AMP1 to inhibit protease activity. The sequence of this peptide was determined through PEAKS de novo sequencing (Figure S4A). It was selected as a suitable candidate due to the high sequence coverage, and lack of Cys residues available for alkylation. The lack of cysteine residues results in this peptide being detected in both native and reduced and alkylated samples as the same m/z (Figure S4B). It stands that if CM-AMP1 serves as a serine protease inhibitor, this peptide would still be detected in the native + Proteinase K sample, but not detected in the reduced and alkylated + Proteinase K sample. In both samples containing Proteinase K this peptide was depleted (Figure S4B), suggesting CM-AMP1 does not inhibit protease function.

Bioactivity of Synthetic Variants of CM-AMP1.

The core motif of CM-AMP1 (KSVDKKRRFFIPT) consists of a chemically diverse set of amino acids. Due to the charge density of basic residues in concert with the hydrophobic residues present in the core motif, it was postulated that this motif was a likely driver of antimicrobial activity. To confirm bioactivity of the predicted CM-AMP1 sequence and the importance of the core motif, two analogs of CM-AMP1 were synthesized: sCM-AMP1 and sCM-MAP1 core (Figure 5A). Since precise DSB pattern is convoluted in native CM-AMP1, we attempted a native chemical ligation to cyclize the peptide backbone prior to oxidation in an attempt to position the residues for disulfide bonding. However, the resulting cyclized peptide was insoluble at aqueous concentrations required for activity. To prevent aberrant side chain reactions of the C-terminal Asn during ligation,³⁶ we moved the N-terminal “CG” residues to the C-terminus and included an C-terminal acyl hydrazide to be used in the native chemical ligation step. This yielded a peptide with the sequence: H₂N-CYCKSVDKKRRFFIPTCLRSCCNCG-NHNH. In addition to the full-length sequence, the truncated core motif was synthesized to test for activity. For activity assays, both the full-length CM-AMP1 and truncated analog were tested in oxidized, reduced, and alkylated forms (Figure 5B). There was little difference in activity between the reduced and oxidized forms, likely due to the reversibility of cleavage and formation of the disulfide bonds. However, a significant loss of activity was observed upon alkylation of the cysteine residues in both peptide analogs. This suggests that the availability of the cysteine residues is highly important for bioactivity of the peptide.

Figure 5.

Bioactivity of synthetic CM-AMP1 and CM-AMP1 core. A) Sequence of sCM-AMP1 and sCM-AMP1 core analogs. Cysteine residues are highlighted in red. B) Antibiotic activity of sCM-AMP1 and the sCM-AMP1 variant at different concentrations (n = 3) assayed against E. coli ATCC 25922. Activity was calculated relative to ampicillin controls (100 μg / mL). Error bars represent +/− 1 standard deviation away from the mean values.

Characterizing the Biological Effect of CM-AMP1 in E. coli.

Understanding how bacteria respond to antibiotics can inform about the mechanism of action, and is key to understanding resistance mechanisms within bacteria. E. coli cultures were treated with a sub-lethal dose of the full-length reduced and oxidized synthetic CM-AMP1 and grown to log phase. Label-free quantitative proteomics was used to establish protein abundance changes between the treated cultures, and untreated controls. 858 proteins were quantified in this experiment, with 43 proteins displaying a significant difference between conditions (Figure 6A, Table S3). The most dramatic increase in abundance in the CM-AMP1 treated sample was the protein, BasS (Figure 6A, P30844, FC = 83.8). BasS is part of a two-component system responsible for regulating cell wall charge.³⁷ Basic residues of membrane-permeabilizing peptides will interact electrostatically with the negatively charged bacterial membrane.³⁸ To combat this, bacteria often modify their outer membrane through the addition of phosphoethanolamine (PEtN) and/or 4-amino-4-deoxy-β-L-arabinose(L-Ara4N) moieties to the 1′ and 4′ headgroups on lipid A.^{38, 39} CM-AMP1 has a highly basic core motif, and this proteomic data suggests that this peptide likely acts through membrane permeabilization, similar to other antimicrobial peptides.

Figure 6.

Biological characterization of CM-AMP1 against E. coli. A) Volcano plot visualizing global proteomic changes in the E. coli proteome following treatment with 5 μg / mL linear reduced CM-AMP1. Vertical lines represent the cutoff used to determine significant proteomic changes (|log₂(FC)| > 1). Horizontal lines represent the threshold used to determine significant false discovery rates determined by an FDR-corrected two-sided t-test (-log₁₀(q-value) > 1.3). B) Outer membrane permeabilization assay with NPN after treatment with synthetic peptide. % NPN uptake is calculated relative to the positive (Polymyxin B, PmB) and negative (H₂O + 0.1% DMSO) controls.

This potential mechanism of action was confirmed through an N-phenylnaphthalen-1-amine (NPN) uptake assay, which measures the amount of NPN that is able to transverse through the outer membrane of bacteria. Similar to the antimicrobial assays, oxidized, reduced, and alkylated peptides were tested for outer membrane permeability. Interestingly, all three of these peptides showed significant amounts of permeability compared to the negative control (H₂O + 0.1% DMSO) (Figure 6B). These results suggest that cysteines have a significant effect on antimicrobial activity of the peptide. While some known antimicrobial peptides have intracellular targets, the cysteines may just be important for conferring stability after transversing the outer membrane. Disulfides in native CM-AMP1 were found to help confer proteolytic stability (Figure 4), and alkylated CM-AMP1 would likely not encounter any proteases until transversing the outer membrane into the periplasm.

CONCLUSIONS

C. majus has long been of medicinal interest, but characterization of antimicrobial peptides within this species had not yet been explored. From an aerial tissue extract, a highly abundant peptide, deemed CM-AMP1 was detected. Multiple MS techniques were used to establish the full sequence of this new antimicrobial peptide. The sequence of CM-AMP1 is highly unique, with very little sequence similarity across the NCBI non-redundant protein database. At the core, a highly basic motif plays a major role in antimicrobial activity and the ability of the peptide to traverse the membrane. Quantitative proteomics was used to characterize the bacterial response to sub-inhibitory concentrations of CM-AMP1. This study highlights the discovery of a potent antimicrobial peptide and provides in-depth molecular and biological characterization.

EXPERIMENTAL

Plant Material.

Chelidonium majus seeds were purchased through Strictly Medicinal Seeds and planted in nutrient rich soil. Plants were grown in a greenhouse under standard temperatures (25 °C) and light cycle (14 h). The plants were grown for approximately 12 weeks before the aerial tissue was harvested and flash frozen.

Peptide Extraction and Creation of Peptide Library.

Across two extractions, a total of 378.7 g of plant aerial tissue (180.1 g and 198.6 g, respectively) was extracted in acetic acid as described previously.²⁸ Size exclusion steps were included to remove large proteins (>30,000 kDa) and small molecules (<500 Da). Neutral and negatively charged molecules were removed through a strong cation exchange fractionation, in which the flow-through was discarded. The SCX retained peak(s) were desalted using C₁₈ Sep-Pak cartridges (Waters, 500 mg), and subjected to an additional round of fractionation via RPLC. Fractions across the gradient were collected every minute, dried via vacuum centrifugation, and resuspended in 60–120 μL of LCMS grade H₂O.

Antimicrobial Assays.

E. coli ATCC 25922 and S. aureus LAC were streaked onto Mueller-Hinton agar plates and incubated at 37 °C for 16 h. Starter cultures were inoculated into 5 mL of MHB media and grown for 16 h with shaking at 37 °C. Assays were performed in technical triplicate by adding 10 μL of the resuspended RPLC fractions with 40 μL of bacterial culture at a starting OD₆₀₀ = 0.1 in a 96-well plate. Ampicillin and erythromycin were used as positive controls for E. coli and S. aureus, respectively, at 100 μg / mL. Plates were incubated at 37 °C with shaking for 4 h, or until growth was observed in the negative control (H₂O). 1 μL of 50 mM resazurin was added to each well, and the plates were incubated for an additional 1 h at 37 °C with shaking. A fluorescence measurement with 544 nm excitation and 590 nm emission was collected for each well and used to calculate % activity relative to the positive controls.

LC-MS/MS Analysis for CM-AMP1 Sequencing.

All experiments described were analyzed on a Sciex ZenoTOF 7600 mass spectrometer operating in positive mode, however specific MS methods varied depending on the experimental requirements.

LC gradient.

Peptides were injected from a Waters M-class UHPLC onto a Phenomenex Kinetex XB-C₁₈ column (2.6 μm, 150×0.3 mm). A flow rate of 5 μL / min was used for the entire method. Mobile phase A consisted of H₂O with 0.1% formic acid (FA) and mobile phase B consisted of MeCN with 0.1% FA. After injection, the gradient was held at 3% B for 5 m. Peptides were separated using an increasing linear gradient of 3–50% B over 13 min, then increased to 85% B for 5 min before returning to 3% B and re-equilibrating for 8 min.

CID acquisition parameters.

Samples were acquired with an information-dependent acquisition (IDA) method consisting of a 100 ms survey scan from 350 to 2000 m/z. The top 8 features above 100 cps from charge states 2 – 7 were fragmented with 100 ms accumulation time. Additionally, a 5 s dynamic exclusion window was applied for precursor selection.

EAD acquisition parameters.

Samples were acquired with an IDA method consisting of a 100 ms survey scan from 350 to 2000 m/z. The electron energy for the alternative fragmentation in the EAD cell was set to a value of 7 eV, with an electron beam current of 5000 nA. Fragment spectra were acquired with an accumulation time of 75 ms and a reaction time of 35 ms.

De novo Sequencing, and Ile/Leu Determination.

1 μL of fraction 24 was diluted in 20 μL of 50mM Tris-HCl, reduced with 3 mM dithiothreitol for 10 min at 25 °C and subsequently alkylated with 5 mM iodoacetamide for 15 min at 25 °C. To aid in de novo sequencing efforts, 0.5 μg of LysC was added to one aliquot of reduced and alkylated sample and incubated at 25 °C for 3 h. All samples were desalted with C₁₈ Zip-Tips prior to MS analysis.

Proteolytic Stability Assay.

Two aliquots of 1 μL of fraction 24 were diluted in 20 μL of 50 mM Tris-HCl. One aliquot was reduced and alkylated, as described above. To both aliquots, 0.5 μg of Proteinase K was added and incubated for 10 min at 25 °C. Both samples were desalted with C₁₈ Zip-Tips prior to MS analysis.

SignalP and Cysmotif Searcher.

SignalP-5.0 with default options was used to identify proteins containing signal peptides in the C. majus protein database generated using C.majus genome assembled in Department of Molecular Virology, Faculty of Biology, AMU, Poznań, Poland. Proteins that contained signal peptides, predicted cleavage sites, and export a FASTA file with these proteins after signal peptide cleavage. This FASTA file was used with Cysmotif Searcher to predict antimicrobial peptides in the C. majus genome.

Peptide Synthesis.

Peptide synthesis was performed as previously reported using a semi-automated flow chemistry instrument built in-house.⁴⁰ For CM-AMP1 full sequence synthesis, 2-chlorotrityl hydrazine resin (0.80 mmol/g, 200–400 mesh) was used to afford C-terminal acyl hydrazide product. For CM-AMP1 core synthesis, Rink amide resin (1.2 mmol/g, 100–200 mesh) was used. Coupling of each Fmoc amino acid (1 mmol) as a 0.38 M HBTU in DMF (2.5 mL) solution with 450 mL of DIPEA (or 250 mL with His, Cys, or Trp) was performed, followed by Fmoc-deprotection with 20% piperidine in DMF. DMF was used for resin washing between deprotection and coupling steps.

Finalized synthetic peptide was cleaved using a trifluoroacetic acid (TFA) / EDT / TIPS / H₂O (94:2.5:2.5:1.0) cleavage mixture (10 mL) with a 10-min incubation at 60 °C. The solution was filtered and reduced under a stream of N₂ gas, followed by precipitation with diethyl ether at 4 °C. The crude peptide pellet was collected by centrifugation (3000 rpm, 4 °C, 10 min) and further purified by semi preparative HPLC (10–60% B over 20 min, 5 mL/min). Oxidative folding was performed by dissolving reduced peptide (0.1 mg/mL) in 0.1% acetic acid with 5% DMSO (v/v) and adjusting to pH 8.0 for gentle mixing with analytical HPLC monitoring.³¹ Upon completion, the peptide solution was changed to pH 4.0 with 1.0% aqueous TFA and purified by reversed-phase preparative HPLC.

Proteomic Response of E. coli.

Bacterial growth.

Colonies of E. coli ATCC 25922 were inoculated into 5 mL of MHB media and grown for 16 h with shaking at 37 °C. Cultures were back-diluted to and OD₆₀₀ = 0.1, and CM-AMP1 was added to a concentration of 5 μg / mL. Cultures were incubated for 6 h, and the pellets were harvested via centrifugation.

Protein extraction.

Cell pellets were resuspended in 200 μL of lysis buffer (100 mM Tris-HCl, 0.1% (v/v)TritonX-100, pH 8.0) and shaken for 20 min at 37 °C. Following lysis, protein was precipitated in 5x cold 100 mM ammonium acetate in MeOH. Samples were incubated at −20 °C for 30 min, and proteins were pelleted by centrifugation for 10 min at 20 000 RCF. Precipitated protein was resuspended in 4 M urea and 100 mM Tris-HCl at pH 8 and quantified with a CB-X assay (Biosciences) against BSA standards. Proteins (100 μg) were reduced (10 mM dithiothreitol, 30 min, 25 °C) and alkylated (30 mM iodoacetamide, 45 min, 25 °C, dark). Reduced and alkylated proteins were precipitated with 10x cold acetone and collected via centrifugation (20 000 RCF, 10 min). Precipitated protein was resuspended in 2 M urea and 100 mM Tris-HCl at pH 8. Reduced and alkylated proteins were enzymatically digested with Trypsin gold (1:25 enzyme/protein; Promega) for 16 h, shaking at 850 rpm at 25 °C. Samples were acidified using TFA to a pH < 3, desalted using 50 mg C18 Sep-Paks (Waters), and dried in a CentriVap (Labconco).

LC-MS/MS analysis.

500 ng of peptides were injected from a Waters M-class UHPLC onto a Phenomenex Kinetex XB-C₁₈ column (2.6 μm, 150×0.3 mm). A flow rate of 5 μL / min was used for the entire method. Mobile phase A consisted of H₂O with 0.1% formic acid (FA) and mobile phase B consisted of MeCN with 0.1% FA. After injection, the gradient was held at 3% B for 5 m. Peptides were separated using an increasing linear gradient of 3–33% B over 37 min, then increased to 85% B for 5 min before returning to 3% B and re-equilibrating for 8 min. Samples were acquired with an information-dependent acquisition (IDA) method consisting of a 100 ms survey scan from 350 to 1500 m/z. The top 25 features above 100 cps from charge states +2 – +7 were fragmented with 30 ms accumulation time. Additionally, a 5 sec dynamic exclusion window was applied for precursor selection after two occurrences.

Bioinformatic analysis.

Raw data files were imported into Progenesis QI (Nonlinear Dynamics, v.2.0) for peak alignment and quantification. Spectra were searched in Mascot (Matrix Science, v.2.5.0) against the E. coli K12 reference proteome (UniProt, accessed 9/16/2023) using a precursor / fragment tolerance of 25 ppm / 0.1 Da, trypsin specificity, two possible missed cleavages, fixed modification cysteine carbamidomethylation, and variable modifications of methionine oxidation and protein N-term acetylation. Significant peptide identifications above the homology threshold were adjusted using the embedded Mascot percolator algorithm and uploaded to Progenesis for peak matching before exporting protein measurements. A students t-test with a Benjamini-Hochberg FDR correction was used to determine proteins that were significantly changing between conditions (q-value < 0.05, | log₂(fold change) | > 1).

Outer Membrane Permeabilization Assay.

E. coli ATCC 25922 was streaked on an LB plate and incubated for 16 h at 37 °C. A bacterial colony was incubated overnight in MHB and then diluted to OD₆₀₀ = 0.25 to be incubated for an additional hour. Cells were washed with 5 mM HEPES buffer containing 5 mM glucose at pH 7.4, and resuspended in a 5 mL aliquot of the same buffer. NPN was added to a final concentration of 10 μM and the culture was incubated for 30 min in the dark, at room temperature. The assay was performed in technical triplicate in a 96-well plate, using 40 μL of culture and 10 μL peptide. Polymyxin B (PmB) was used as the positive control, with a final concentration of 10 μg / mL, and H₂O was used as the negative control. After additions of each solution to the wells, fluorescence was measured at an excitation wavelength of 350 nm and an emission wavelength of 420 nm.

SAFETY STATEMENT

No unexpected or unusually high safety hazards were encountered in the experiments described.