A marine-derived fatty acid targets the cell membrane of Gram-positive bacteria

ABSTRACT The rapid evolution of antibiotic resistance is shrinking our stockpile of commercially available antibiotics. Therefore, new antimicrobials with novel mechanisms of action (MoAs) are desperately needed. The free fatty acid, (Z)-13-methyltetra-4-decenoic acid ((Z)-4C-14:1), isolated from a marine sediment bacterium Olleya marilimosa, displays strong inhibition of Gram-positive pathogens with minimal cytotoxic effects to mammalian cells. Here, we applied a combination of experimental approaches to identify the mechanism by which (Z)-4C-14:1 kills the Gram-positive bacterium Bacillus subtilis. Using quick and cost-effective bacterial cytological profiling (BCP), we established the cytological signatures of 17 antibiotics representing 6 general classes of antibiotic MoAs. We used BCP to demonstrate that while (Z)-4C-14:1-treated B. subtilis cells display unique morphological features compared to other antibiotic classes, (Z)-4C-14:1 mode of action shares substantial overlap with the fast-acting antibiotic colistin specifically in Gram-positive cells and daptomycin, both of which target bacterial permeability by destroying the cell membrane and causing extensive cell surface alterations. To further determine if the cell membrane of B. subtilis was the target of (Z)-4C-14:1, we used cell membrane- and peptidoglycan-specific diagnostic stains to investigate bacterial permeability. Our results indicate that (Z)-4C-14:1 destabilizes the cell membrane by pore formation in Gram-positive bacteria and emphasize the importance of mining the marine environment for naturally occurring antibiotics. IMPORTANCE With the lack of new antibiotics in the drug discovery pipeline, coupled with accelerated evolution of antibiotic resistance, new sources of antibiotics that target pathogens of clinical importance are paramount. Here, we use bacterial cytological profiling to identify the mechanism of action of the monounsaturated fatty acid (Z)-13-methyltetra-4-decenoic acid isolated from the marine bacterium Olleya marilimosa with antibacterial effects against Gram-positive bacteria. The fatty acid antibiotic was found to rapidly destabilize the cell membrane by pore formation and membrane aggregation in Bacillus subtilis, suggesting that this fatty acid may be a promising adjuvant used in combination to enhance antibiotic sensitivity.

4).Compounding the threat of antibiotic resistance is the drought in the antibiotic discovery pipeline (5,6) due to both the lack of biologically relevant chemical diversity in combination with synthetic approaches that have yet to supplant structural novelty of natural products (7).Therefore, searching for new antibacterial candidates with unique mechanisms of action (MoAs) is crucial to combat bacterial resistance mechanisms.
Recently, the discovery of new natural product-based antibiotics, teixobactin from soil (8) and darobactin from invertebrate symbionts (9), has reinvigorated antibiotic research in mining distinct environment niches for new chemical sources.In particular, the ocean serves as a diverse repository of untapped biological inventory whereby organisms have evolved the ability to synthesize a variety of potent chemical moieties resulting from competitive interactions between microbial life (10).Oceanic sediments represent a major global biome, harboring complex microbial communities likely structured by biologically active natural products as revealed by environmental metabolomic profiling (11).Even within an apparent homogeneous environment of sediment exists complex chemical landscapes associated with marine microbial communities, and paired metagenomic sequencing efforts hint at the vast, yet undescribed, natural product diversity awaiting discovery (12).
Here, we describe the isolation, synthesis, and characterization of (Z)-13-methylte tra-4-decenoic acid ((Z)-4C-14:1), a cis-monounsaturated novel free fatty acid (FFA) derived from the marine, Gram-negative bacterium Olleya marilimosa (A414) isolated from a marine sediment.Culture-based assays demonstrated that (Z)-4C-14:1 is a rapid-acting compound most effective against Gram-positive bacteria Bacillus subtilis, S. aureus, and Enterococcus faecalis.Moreover, significant cytotoxicity was observed in human liver (HepG2) cells only at concentrations above 800 µM.To elucidate the MoA of (Z)-4C-14:1, we employed fluorescence-based microscopy to discriminate between different MoAs of known antibiotics and our marine compound by analyzing the cell morphological profiles upon antibiotic treatment using bacterial cytological profiling (BCP) (13).BCP has proven effective in MoA studies of antibiotics (14), complex mixtures (15), and synergistically acting natural products (16).We show that (Z)-4C-14:1 has a unique cytological profile that clusters with those antibiotics targeting cell permeability, with specific overlap with antibiotics that inhibit cell membrane functioning.Diagnos tic fluorescence staining of B. subtilis cells exposed to (Z)-4C-14:1 confirmed the cell membrane to be significantly compromised while the peptidoglycan layer was still contiguous, suggesting that (Z)-4C-14:1 may have detergent-like capabilities.Together, these findings support further development of (Z)-4C-14:1 as a promising candidate to combat Gram-positive bacterial infections when used in combination with other classes of antibiotics to enhance susceptibility.

Identification and activity testing of (Z)-4C-14:1
Using the bacterial susceptibility assay, which identifies isolates capable of reducing an antibiotic's minimum inhibitory concentration (MIC) greater than or equal to fourfold, extract from the isolate A414 was found to potentiate the activity of erythromycin when tested against Escherichia coli multidrug resistant (MDR) strain MG1655 ΔBC/pXYM (17), overexpressing the efflux transporter MexXY-OprM, in the p-iodonitrotetrazolium chloride (INT) assay as described in Whalen et al. (18).Strain A414 was subsequently identified as O. marilimosa based on 16S rRNA gene sequencing (99.64%identity O. marilimosa strain Y23, NCBI accession number MN746117.1).A total of 30 L of A414 bacterial culture was grown yielding 4.5 g of crude organic extract, which was subjected to bioassay-guided fractionation through three rounds of chemical fractionation [i.e., silica column, SPE/ENVI-18 Sep-Pak, reverse-phase high-performance liquid chromatog raphy (HPLC)] as described in Fig. S1, with each semi-purified fraction being tested for activity in the bacterial susceptibility assay until a single pure compound was identified.The resulting active compound was isolated as a clear oil.The HRESIMS analysis established the molecular formula as C 15 H 28 O 2 (m/z 239.2011, [M-H] − calcd.239.2017).The 13 C nuclear magnetic resonance (NMR) spectrum, combined with the HSQC experiment (Fig. S2) and the degree of unsaturation of 2, determined by the molecular formula, indicated that the compound is a fatty acid with one double bond..00),along with the key HMBC correlations from H 2 -3 to C-1 (δ C 173.9) and from H-4 to C-2 (δ C 33.8), allowed for the location of the double bond at C-4/C-5 (Fig. S2).The double bond geometry at C-4/C-5 was assigned as Z by observing that the pattern of the olefinic proton signals in the 1 H NMR spectrum of the isolate was similar to that of methyl oleate (Z) rather than methyl elaidate (E) (19).Furthermore, the methyl signals at δ H 0.84 (6H, d, J = 6.5 Hz) as well as the molecular formula suggested the compound to be (Z)-4C-14:1.Only 8.9 mg of this compound was isolated from the initial 4.5 g of crude extract; therefore, (Z)-4C-14:1 was synthesized (Fig. S3) yielding 544 mg of a pure compound and was determined to be identical to the naturally occurring compound by NMR spectroscopy (Fig. S4 and S5).
Bacterial susceptibility screening with the isolated natural compound indicated that the MICs against E. coli MDR strain MG1655 ΔBC/pXYM overexpressing the efflux transporter MexXY-OprM (17) and the methicillin-sensitive clinical isolate S. aureus (DMS 1104) were 75 and 30 µg/mL, respectively.Due to the limited amount of isolated natural compound, all subsequent experimentation was performed with the synthetic version of (Z)-4C-14:1.Next, we examined the biological activity of (Z)-4C-14:1 against a panel of different pathogens.(Z)-4C-14:1 demonstrated activity exclusively against Gram-positive bacteria (Table 1).Additionally, (Z)-4C-14:1 was evaluated for its cytotoxicity against human liver HepG2 cells using a luminescence-based cell viability assay that detects the presence of ATP in viable cell populations.(Z)-4C-14:1 only exhibited substantial cytotoxic effects at concentrations above 800-µM (192.3 µg/mL) final concentration.The average of five independent assays was used to calculate an IC 50 value of 508.0 ± 29.5 µM (Fig. 1).

Evaluation of (Z)-4C-14:1 by BCP
B. subtilis cells were used in BCP experiments due to the dynamic change in cell morphol ogy upon exposures to various classes of antibiotics (20).B. subtilis was treated with (Z)-4C-14:1 and antibiotics across six diverse inhibition categories (Fig. 2): cell wall inhibition (ampicillin and linoleic acid), ionophores [calcimycin, carbonyl cyanide 3chlorophenylhydrazone (CCCP), monactin, and valinomycin], protein synthesis inhibition (chloramphenicol, erythromycin, gentamicin, and tetracycline), DNA/RNA inhibition (ciprofloxacin and rifampicin), fatty acid synthesis inhibition (cerulenin, platensimycin, and triclosan), and disruption of cytoplasmic membrane (daptomycin and colistin in Gram-positive cells specifically).A total of 39 cytological parameters were measured for each cell within each antibiotic treatment from four independent experiments.Variable reduction was performed using a principal component analysis (PCA) on a per-cell basis from (Z)-4C-14:1 treatments and from each other antibiotic with known MoAs in pairwise comparisons (Fig. 3A).Concordance of 95% confidence ellipses indicated that (Z)-4C-14:1-treated B. subtilis was morphologically most similar to antibiotics colistin, Full-Length Text Journal of Bacteriology linoleic acid, and daptomycin, all of which target bacterial permeability, albeit through different mechanisms (Fig. 3B).

Loss of membrane integrity
To examine the rapidity of morphological changes in cells exposed to (Z)-4C-14:1, B. subtilis cells were treated with (Z)-4C-14:1 at 5× MIC and visualized in 30-min intervals for 2 h using diagnostic fluorescence staining.Membrane disruption/aggregation and SYTOX staining were detected in (Z)-4C-14:1-treated cells as early as 30 min (Fig. 4), which mirrors the onset of morphological changes seen with colistin at 30-min postexposure and linoleic acid at 2-h post-exposure, respectively.Reports indicate that polymyxins, like colistin, can get through the peptidoglycan layer of Gram-positive cells and destroy the integrity of the cytoplasmic membrane (21), resulting in leakage of cellular contents (22).To confirm if (Z)-4C-14:1 also targets the cell membrane, we performed diagnostic fluorescence microscopy to assess the structural integrity of both the peptidoglycan layer stained with wheat germ agglutinin (WGA)-Alexa Fluor 488 and the cell membrane stained with FM4-64 in untreated and (Z)-4C-14:1-treated B. subtilis.Cells exposed to (Z)-4C-14:1 at 2× MIC for 2 h were observed to show membrane collapse via the formation of membrane pores and membrane aggregates, which were absent in untreated cells (Fig. 5).Additionally, even though the loss of membrane integrity was observed in (Z)-4C-14:1-treated cells, the peptidoglycan layer remained intact, indicating that (Z)-4C-14:1 initially targets the cell membrane of Gram-positive cells.

DISCUSSION
In particular, FFAs offer considerable antibiotic potential due to their broad activity spectrums effective against various Gram-positive and Gram-negative pathogens (23,24) and difficulties in bacterial resistance development (25).Due to their amphipathic properties, antibacterial FFAs are capable of destabilizing cell membranes, disrupting both the electron transport chain, and uncoupling oxidative phosphorylation (23,25).
(Z)-4C-14:1 is a cis-monounsaturated long-chain fatty acid compound isolated from marine O. marilimosa within the family Flavobacteriaceae.The first report of this branched-chain monoenoic fatty acid 13-methyl-4-14:1 was described from a Black Sea sponge, Dysidea fragilis (27), with a subsequent description of its isolation from an additional sponge (28); however, the biological activity of the compound was never explored.Marine sponges contain diverse microbial communities that can account for up to 40% of the host volume (29), and it is now recognized that associated bacteria are typically the source of these C 15 -C 19 fatty acids within sponge tissues (30).Consistent with a bacterial origin of sponge-associated, unsaturated fatty acids was the isolation of additional monounsaturated iso-branched fatty acids identified from members of the Cytophaga-Flavobacterium-Bacteroides group of bacteria (31) and the report of (Z)-4C-14:1 isolation from Flavobacterium within the family Flavobacteriaceae (32).
Concordance comparison of PCA results allowed us to narrow down the possible MoA of (Z)-4C-14:1 to bacterial permeability, with the highest similarities to antibiotics colistin (polymyxin E), linoleic acid, and daptomycin, albeit affecting bacterial permeability in different ways.Colistin and (Z)-4C-14:1 are both fast-acting antibiotics that showed similar rapid membrane aggregation/blebbing morphology changes on B. subtilis-trea ted cells within 30 min of application.Colistin is most notable for its ability to target the lipopolysaccharide component of the outer membrane of Gram-negative bacte ria by competing with divalent cations to destabilize the outer membrane, resulting in permeability (21).In Gram-positive bacteria that lack an outer membrane, once polymyxins are able to penetrate through the peptidoglycan layer, these compounds destroy the integrity of the membrane resulting in cell leakage (21,22).Initial investiga tions into linoleic acid's MoA describe the compound's action as a surfactant-like increase in membrane permeability (33); however, additional work in S. aureus suggests that linoleic acid can alter the expression of genes for peptidoglycan synthesis (34) and inhibit fatty acid synthesis (35), both of which can impact bacterial permeability.For daptomy cin, there are two general hypotheses as to this antibiotic's MoA.The first suggests that aggregates of daptomycin form oligomeric pore structures in the membrane, resulting in ion leakage and the loss of membrane potential (36).A second hypothesis suggests that daptomycin asserts a lipid-extracting effect whereby compound insertion into the membrane causes rapid aggregation of the lipid on the membrane surface, the Full-Length Text Journal of Bacteriology eventual release of lipid clusters from the membrane, and the formation of transient water pores that could explain the ion leakage observation seen previously (36).Because all three antibiotics with high concordance to (Z)-4C-14:1 have been shown to permeabilize bacterial cells, we specifically interrogated the structural integrity of both the peptidoglycan layer and membrane by fluorescence microscopy of B. subtilis cells exposed to (Z)-4C-14:1.We observed significant pore formation after 2-h exposure at 2× MIC, accompanied with membrane aggregation and no apparent loss of pepti doglycan integrity, indicating that (Z)-4C-14:1 initially targets phospholipid membrane stabilization.Our result is consistent with recent findings investigating various natural long-chain unsaturated fatty acids found to induce concentration-dependent membrane remodeling behavior in vitro (37).Finally, we did not see appreciable antibiotic activity of (Z)-4C-14:1 against the Gram-positive Streptococcus pneumoniae, which we postulate is likely as a result of the physical thickness of the polysaccharide capsule that may prevent physical contact of (Z)-4C-14:1 with the membrane.
Cis-arrangement (38,39), iso-branching (40), and long-chained unsaturated fatty acids (41) have a long history of antimicrobial activity against Gram-positive bacte ria (23).Specifically, unsaturated fatty acids with long chains consisting of 12 or more carbons demonstrate exceptional activity against Gram-positive bacteria (42,43).Moreover, unsaturated fatty acids have been shown to inhibit the growth of both Gram-negative and Gram-positive bacteria (reviewed in (24)).Interestingly, (Z)-4C-14:1 was originally isolated in a screen for antibiotic adjuvants targeting Gram-negative bacteria overexpressing resistance nodulation cell division (RND) transporters.This assay functions by screening complex chemical mixtures of bacterial exudates in E. coli overexpressing RND pumps grown in a background of antibiotic at one-fourth of their MIC concentration.This finding suggests that (Z)-4C-14:1 could exhibit synergistic effects with other antibiotics by permeabilizing the outer membrane of Gram-negative bacteria.This strategy of combining fatty acid adjuvants with antibiotics has been shown to be effective in overcoming bacterial drug resistance and inhibiting biofilm formation (9,26,(44)(45)(46)(47)(48)(49)(50)(51).For example, when the synergistic activities of 30 FFAs and 11 antibiotics were investigated against MRSA, the monounsaturated myristoleic acid (C14:1) most enhanced the bactericidal activities of four aminoglycoside antibiotics synergistically, including significantly decreasing biofilm formation by S. aureus (46).
In summary, the broad spectrum of activity of unsaturated FFAs, their low eukaryotic cytotoxicity, and lack of FFA resistance mutants make them attractive agents for further exploration for treating drug-resistant pathogens.Marine organisms are a rich source of novel FFAs with proven biomedical potential (52).Furthermore, the essential function of the cell membrane makes it a robust target for developing new antibacterial therapies.The use of membrane-disrupting unsaturated FFAs in synergy with approved therapeu tics is a promising strategy to combat bacterial infections.

Microbial isolation and identification
Marine bacterial isolate A414 was part of the Mincer Culture Collection housed at the Woods Hole Oceanographic Institution and originally isolated for a sediment sample from Clayoquot Sound, British Columbia, Canada onto a tryptone seawater (TSW) agar (TSW: 1-g/L tryptone; 1 L of 75:25 natural seawater: Milli-Q water; 15-g agar) plate and cultivated at 23℃ for 3 days.A single yellow colony was transferred to a fresh TSW plate to obtain a pure isolate of strain A414.
Genomic DNA was isolated from a bacterial pellet of A414 with the Qiagen DNeasy Blood and Tissue Kit (Qiagen, Valencia, CA, USA) following the manufacturer proto col with an enzymatic lysis buffer containing 50-mM Tris-Cl, 10-mM sodium-EDTA, 1.2% Triton X-100, and lysozyme at a final concentration of 20 mg/mL.The bacterial small subunit rRNA gene was amplified using PCR with oligonucleotide primers 27F (5′-AGAGTTTGATCMTGGCTCAG-3′) and 1492R (5′-TACGGYTACCTTGTTACGACTT-3′).A 50-µL PCR reaction contained 1.25 U of GoTaq(R) G2 Flexi DNA polymerase (Promega, Fitchburg, WI, USA), 1× GoTaq Flexi Buffer, 1.25-mM MgCl 2 , 200 µM of each dNTP, 200 nM of each primer, and 250 ng of the genomic template and was performed in a MyCycler thermal cycler (Bio-Rad Laboratories, Hercules, CA, USA).Amplification of PCR products was carried out according to the GoTaq(R) Flexi Kit, and cycling parameters were as follows: 95℃ for 2 min; 40 cycles of 95℃ for 20 s, 54℃ for 30 s, and 72℃ for 1.5 min; and 1 cycle of 72℃ for 5 min.The amplification product was subjected to gel electrophoresis in 1% agarose gels and purified using the QIAquick PCR Purification Kit (Promega, Madison, WI, USA).The product was sequenced in a single direction using the Sanger method by Eurofin MWG Operon Biotech.Sequences were manually inspected using Sequencher 4.8 (Gene Codes, Ann Arbor, MI, USA), and ambiguous nucleotides on the ends of the sequence were excluded from further analysis.Phylogenetic analysis indicated that the isolated strain A414 was O. marilimosa on the basis of 99.85% identity in the 16S rRNA gene sequence (GenBank accession number OP950677).The pure culture was cryopreserved in 10% sterile DMSO and stored at −80℃ until use.

Bacterial culture and extract production
A starter culture of A414 was prepared by inoculating 10 mL of TSW media with 100 µL of cryopreserved stock and grown for 3 days at 23℃ at 100 rpm.To scale up bacterial growth conditions, large batch cultures consisted of inoculating a 1.5-L TSY (1-g/L tryptone, 1-g/L yeast extract, and 1.5-L of 75:25 natural seawater: Milli-Q water) culture in a 2-L Fernbach flask with 2 mL of starter culture.This large batch culture was grown at 18℃ at 100 rpm for 8 days.Twenty-four hours prior to culture filtration (Day 7), 20 mL of a 1:1 mixture of sterile, washed Amberlite XAD-7 and XAD-16 resin was added to the cultures.On Day 8, the resin was filtered under vacuum, desalted by rinsing in 300-mL Milli-Q water, and dried overnight at room temperature before storing at −20℃ until use.The secreted metabolites were eluted from the resin first in 100 mL of (1:1) methanol/dichloromethane, followed by 100 mL of methanol/20 mL of XAD resin.Both organic fractions were combined and dried under vacuum centrifugation (Thermo Savant) and stored at −20℃ until bacterial susceptibility assays.

Bacterial susceptibility determinations with isolated (Z)-4C-14:1
Initially, the crude extract from A414 was screened in a whole-cell assay described in Whalen et al. (18) against E. coli strains engineered to overexpress RND transporters with the aim of identifying either an efflux pump inhibitor or antimicrobial compound(s).This bacterial susceptibility assay was used to identify marine microbial isolates capable of reducing antibiotic MICs greater than or equal to fourfold in three strains overexpress ing three archetype RND transporters (AcrAB-TolC, MexAB-OprM, and MexXY-OprM) common in Gram-negative pathogens.The crude extract from A414 was tested in duplicate at 0.25 mg/mL and visualized by rapid p-iodonitrotetrazolium (INT) chloride colorimetric assay in 96-well microtiter plates in a final volume of 200 µL.Test E. coli strains were grown in the presence of the extract and either chloramphenicol or erythromycin at one-fourth of their MICs.An extract was considered active if it was able to reduce the MIC of the co-administered antibiotic at least fourfold.The extract from A414 was chemically fractionated as described above, and at each step, the semi-purified fractions and final pure compound were screened in the whole-cell assay against E. coli for activity.
Initially, the MIC of the isolated (Z)-4C-14:1 was determined for S. aureus (DSM 1104), a methicillin-sensitive clinical isolate.The MIC was determined by broth microdilution in Luria Broth (LB) medium as previously described (53).S. aureus (DSM 1104) was obtained from the American Type Culture Collection (ATCC 25923).
In step 2, a solution of triphenylphosphine in anhydrous CH 2 Cl 2 was added dropwise over 30 min to an ice-H 2 O bath-cooled solution of the alcohol and CBr 4 in anhydrous CH 2 Cl 2 .The reaction was allowed to warm to room temperature as the cooling bath melted.Stirring was continued for 3 h, and the reaction was analyzed by TLC (10% EtOAc in hexane, KMnO 4 stain) indicating complete consumption of the starting material.The clear, pale solution was concentrated to dryness under reduced pressure.The residue was triturated with hexane, then supernatants were combined with silica gel, and the mixture was concentrated to dryness under reduced pressure.Flash column chromatog raphy (RediSepR f , SiO 2 , 100% hexane, fractions visualized with PMA stain) gave the product as a clear, colorless oil. 1 H NMR (400 MHz, CDCl 3 ): δ 3.93 (t, J = 7.6 Hz, 2H), 1.84 (m, J = 7.2 Hz, 2H), 1.48 (m, J = 7.2 Hz, 1H), 1.38 (m, 2H), 1.26 (m, 6H), and 1.14 (m, 2H).This material was used in the next step with no further characterization.
In step 5, solid pyridinium dichromate was added in one portion to an ice-H 2 O bath-cooled solution of the alcohol in anhydrous DMF.The reaction was allowed to warm to room temperature as the cooling bath melted with stirring for 16 h.Analysis of the reaction by TLC (10% EtOAc in hexane, KMnO 4 stain) after that period of time indicated complete consumption of the starting material.The reaction was diluted into a solution of 10% citric acid.The mixture was extracted with MTBE.The combined organic extracts were washed with saturated aqueous NaCl and then added to silica gel.The resulting mixture was concentrated to dryness under reduced pressure.Flash column chromatog raphy (RediSepR f , SiO 2 , 100% hexane → 10% EtOAc in hexanes, fractions visualized with KMnO 4 stain) gave the product as a clear colorless oil (0.54 g, 30% yield). 1 H NMR (500 MHz, CDCl 3 ): δ 5.43 (m, 1H), 5.34 (m, 1H), 2.39 (m, 3H), 2.04 (q, J = 7.2 Hz, 2H), 1.50 (m, J = 6.8 Hz, 1H), 1.37-1.24(m, 10H), 1.15 (m, 2H), and 0.86 (d, J = 6.8 Hz, 6H). 13C NMR (125 MHz, CDCl 3 ): δ 178.6, 131.9, 126.9, 39.0, 29.9, 29.6 (two signals: 29.62 and 26.59), 29.3, 28.0, 27.4,27.2, 22.7, and 22.5 (Fig. S4 and S5).The MIC of (Z)-4C-14:1 against five strains of bacterial pathogens (B.subtilis ATCC 23857, E. faecalis ATCC 29212, E. coli ATCC 25922, Pseudomonas aeruginosa ATCC 27853, S. aureus ATCC 29213, and S. pneumoniae ATCC 49619) was determined in triplicate by standard microdilution method in microtiter plates in cation-adjusted Muller-Hinton broth (CAMHB) as defined by the National Committee for Clinical Laboratory Standards (54)(55)(56).All antibiotics tested were prepared fresh on the day of testing.All strains were allowed to grow at 37°C on freshly streaked tryptic soy broth (BD Bacto) agar plates except E. faecalis and S. pneumoniae, which were streaked onto tryptic soy broth plates supplemented with 5% defibrinated sheep blood.S. pneumoniae was always grown at 37°C supplemented with 5% CO 2 .The following day, single colonies for each strain were inoculated into 10 mL of CAMHB and grown at 37°C, 120 rpm until the desired OD 600 was reached equivalent to 1.5 × 10 8 CFU/mL corresponding to 0.5 McFarland standard.Test compounds were prepared in serial twofold dilutions across the 96-well plate with final concentrations ranging from 0.06 to 128 µg/mL.The culture was diluted 100-fold in CAMHB to reach a concentration of 1.5 × 10 6 CFU/mL, and 10 µL of this inoculum was added to 90 µL of test compounds in CAMHB, yielding a final volume of 100 µL.Plates were incubated at 37°C for 18 h (or 22 h with 5% CO 2 for S. pneumoniae), and the MIC value was defined as the lowest concentration producing no visible growth.

Pathogen strains and MIC testing
B. subtilis ATCC 23857 was used in the BCP assay due to the bacterium's large size and susceptibility to (Z)-4C-14:1.The growth dynamics of B. subtilis (Fig. S6) were determined as follows.A single colony of B. subtilis was used to inoculate 10-mL CAMHB and grown at 37°C, 120 rpm for approximately 4 h until an exponential phase was reached (OD 600 > 0.1).The culture was diluted in CAMHB to an OD 600 = ~0.02 in a final volume of 60 mL.The OD 600 of this culture was monitored for over 8 h using a spectrophotometer.At each timepoint, the concentration of bacteria in CFU/mL was determined by plating 100 µL of diluted bacterial culture on LB agar in triplicate.CFU/mL was then calculated using the equation: The amount of culture plated mL The MICs for 17 antibiotics were determined for B. subtilis ATCC 23857 (Table S1) by standard microdilution method in microtiter plates in CAMHB as described above.All antibiotics tested were prepared fresh on the day of testing.Briefly, B. subtilis was streaked on LB agar plate, and a single colony was inoculated in 10-mL CAMHB and allowed to grow at 37°C, 120 rpm for 3.5 h until the OD 600 reached 0.497, equivalent to 1.5 × 10 8 cells/mL.Antibiotic compounds were prepared in serial twofold dilutions across the 96-well plate with final concentrations ranging from 0.006 to 256 µg/mL.The 96-well plate was incubated for 18 h at 37°C, and the MIC value was defined as the lowest concentration producing no visible growth.

Fluorescence microscopy
B. subtilis was streaked on an LB agar plate and incubated overnight at 32°C for 18 h.Single colonies were used to inoculate 10-mL CAMHB and incubated at 37°C, 120 rpm overnight.Overnight, cultures were then diluted in CAMHB to an OD 600 0.100 and incubated at 37 °C, 120 rpm for ~1.5 h until OD 600 0.195, equivalent to 3 × 10 7 cells/mL (exponential growth).Exponential phase cells in 1-mL aliquots were then treated at 2× or 5× MIC concentrations (Table S2) at 37°C on a rotator until distinct morphology changes in the cells were observed ranging between 30 min and 6 h (13).For the BCP assay, an incubation time of only 30 min for colistin at 5× MIC was deemed the maximum time of exposure of B. subtilis to this antibiotic due to significant cell loss observed.Moreover, incubation times longer than 2-3 h in the BCP assay were needed for selecting antibiotics in order to obtain a distinct cell morphology compared to untreated control cells.
Bacterial samples in 1-mL aliquots were centrifuged for a minute at 6,000 rpm, 900 µL of supernatant was removed, and the cell pellet was resuspended in 100 µL of remaining CAMHB.A 12-µL aliquot of resuspended cells was incubated with 3 µL of tri-dye mix for 20-25 min at room temperature in the dark.The final concentrations of the dyes within the sample mixture were 6-µg/mL FM4-64, 0.5-µM SYTOX Green, and 10-µg/mL DAPI.The samples were then pipetted onto a glass coverslip affixed to a 35-mM diameter dish and covered with 1.2% ultrapure agarose (Invitrogen) pads to immobilize cells (20) and immediately imaged.
Cells were imaged using a Zeiss Axio Observer seven equipped with Zeiss Axio cam 702 monochrome CMOS and Zeiss Axiocam 503 color CCD cameras using the oil immersion 100× objective lens at the Cell and Developmental Biology Microscopy Core within the University of Pennsylvania.Images were taken in z-stacks with 13 slices measuring 0.24-µm thick.Raw images were deconvolved using the imaging program Zeiss ZEN Blue constrained iterative deconvolution software (version 2.5.75.6).GPU processing accelerations were used during deconvolution, and fluorescence decay corrections were applied to all deconvolved images.The median focal plane is shown.Imaging parameters and exposure times for each dye were kept constant (DAPI was imaged using a 385-nm LED line with an exposure time of 100 ms, SYTOX Green was imaged using a 475-nm LED line with an exposure time of 250 nm, and FM4-64 was imaged using a 555-nm LED line with an exposure time of 300 nm).For each antibiotic or untreated sample, four replicates were prepared and imaged.Each sample replicate contained a minimum of 100 cells that were processed for quantitative measurements.All deconvolved and non-deconvolved images were saved in the CZI format.

Generation of cell morphology measurements
Non-deconvolved images were imported into ImageJ Fiji 2 (2.3.0 version), and the FM4-64 channel was used to determine the medial focal plane using the Focus LP plug-in (1.0.x version).TIFFs were created in Fiji of the focused slice for all three channels for both non-deconvolved and deconvolved images.Images were then analyzed using CellProfiler (version 4.2.5)(57).Cells were initially identified using the deconvolved FM4-64 images, and the objects were expanded using the phase image as a guide to obtain final cell objects.Nucleoids were identified separately and associated with corresponding cells.Cells were then filtered based on the presence of nucleoids in the deconvolved DAPI channel, and only cells with DAPI fluorescence were included in the analysis.Images were run through a pipeline (File S1) that quantified 39 separate cell morphological and fluorescence-based parameters per cell.The measured parameters used for analysis and their definitions are listed in Table S3.Each replicate yielded at minimum 100 cells from which quantitative parameters were generated.Fluorescence intensity measurements were measured only on non-deconvolved images.All CellProfiler data were exported and saved as a CSV file.

Principal component analysis
Individual cells with missing data across the 39 morphological parameters were removed, and cells with values over five standard deviations from the within-treatment mean for any of the 39 parameters were removed.To weigh each treatment identically, 217 cells were randomly selected from each treatment, as this was the lowest number of cells available in a single treatment (File S2).PCA was calculated for pairwise sets of (Z)-4C-14:1-treated cells and each antibiotic-treated cells using R's prcomp function, with morphological parameters scaled and centered.The first two principal components and 95% confidence ellipse for each treatment were graphed using ggplot2.We calculated pairwise concordance as the number of cells found in both 95% confidence ellipses divided by the total number of cells in either of the 95% confidence ellipses.This analysis was completed using R v4.

Cytotoxicity assay
The cytotoxicity assay was performed using a human liver cell line (HepG2, ATCC HB-8065) cultured according to methods described in Donato et al. (58).Briefly, cell cultures were grown at 37℃/5% CO 2 in Dulbecco's Modification of Eagle's Medium (DMEM) supplemented with 10% fetal bovine serum (FBS), with final concentrations of 4-mM penicillin/streptomycin and 50-U/mL L-glutamine, and passaged every 6 days at a ratio of 1:6.Cell viability was determined via the trypan blue exclusion method (59) and averaged 96% viability before all experiments were performed.
The cytotoxicity of (Z)-4C-14:1 was evaluated in HepG2 cells with the vehicle DMSO and Triton X (0.1%) as the negative control and positive control, respectively.Briefly, after log-phase growth, HepG2 cells were seeded at 2 × 10 4 cells/well in a 96-well microtiter plate and left overnight to establish attachment.Following attachment, the media was removed, and cells were dosed in triplicate wells for 48 h with a range of (Z)-4C-14:1 concentrations (1-2,218 µM of final concentrations), DMSO (0.9%) or Triton X. Viability of cells was evaluated using a commercial ATP assay (CellTiterGlo 2.0, Promega, Madison, WI, USA) where 100 µL of CellTiterGlo 2.0 reagent was added to each well containing 100 µL of media and cells and then incubated for 10 min, after which 100 µL of the well contents was transferred to a microfuge tube and luminescence was measured on the Promega Glomax 20/20 using 1-s integration.Each assay was repeated five times and viability measurements were normalized to paired vehicle controls and used to calculate the IC 50 value by nonlinear regression using a variable slope with four parameters in Graphpad Prism (version 9.5.1).

Diagnostic fluorescence staining of cells treated with (Z)-4C-14:1
FM4-64 (6 µg/mL) and WGA-Alexa Fluor 488 (Molecular Probes) (100 µg/mL), which stain the cell membrane and peptidoglycan layer, respectively, were used to stain B. subilis cells exposed to (Z)-4C-14:1 at 2× MIC for 2 h or untreated cells.WGA is a lectin that binds to oligomers of N-acetylglucosamine and N-acetylmuramic acid (60).B. subtilis cells were grown, stained, and mounted as described above.Cells were imaged on a Leica Stellaris 5 DMi8 series laser scanning confocal microscope equipped with Power HyD S detectors and 63× oil objective.

FIG 1 Full 4 FIG 2
FIG1 Evaluation of (Z)-4C-14:1 cytotoxicity in the HepG2 cell line.HepG2 cells were treated with (Z)-4C-14:1 for 48 h, and the relative cytotoxicity was determined by luminescence using the CellTiter-Glo 2.0 (ATP) assay.Percent viability was calculated as a percentage of control cells (0.9% DMSO only).The mean ± SEM of five replicates are shown.

FIG 3
FIG 3 Pairwise PCA of B. subtilis cytological profiles.(A) Pairwise PCA showing the first two principal components of the cytological profiles for all cells from (Z)-4C-14:1-treated B. subtilis to antibiotic-treated B. subtilis.95% confidence ellipses for (Z)-4C-14:1-treated B. subtilis are in blue and for the other treatment in orange.(B) Proportion of concordance for (Z)-4C-14:1-treated B. subtilis with antibiotic-treated B. subtilis.Using the pairwise PCAs, we calculated concordance as the proportion of cells within either 95% confidence ellipse that were within both individual 95% confidence ellipses for the two treatments.

FIG 5
FIG 5 Specific labeling of the cell membrane and peptidoglycan layer of untreated and (Z)-4C-14:1-treated B. subtilis.B. subtilis cells were exposed to (Z)-4C-14:1 at 2× MIC or a solvent control for 2 h and then visualized live.Cell membranes were stained with 6 µg/mL of FM4-64 (magenta), and the peptidoglycan layer was stained with 100 µg/mL of N-acetylglucosamine (GlcNAc)-specific WGA lectin conjugated to Alexa Fluor 488 (cerulean).White arrow heads indicate loss of cell membrane integrity without a corresponding loss of peptidoglycan layer integrity in cells treated with (Z)-4C-14:1.Scale bar represents 5 µm.
NMR spectroscopies including 1D ( 1 H and13 C) and 2D (COSY, HSQC, and HMBC) were performed on a 500-MHz Varian Inova NMR spectrometer, and the chemical shifts were recorded as d values (ppm) referenced to solvent residual signals [DMSO-d 6 (δ H 2.50 ppm, δ C 39.520 ppm)].Active HPLC fractions underwent untargeted chemical profiling on an Agilent G6125BW liquid chromatography-mass spectrometer (LC-MS) equipped with an Agilent 1260 Infinity II series HPLC with a Phenomenex Kinetex 2.6 µm, C18, 100 Å, LC column (150 × 2.1 mM) as the stationary phase.Spectra were collected in both positive ((+)-HRESI) and negative ((−)-HRESI) ionization modes.All HPLC-MS experiments used a flow rate of 0.2 mL/min.The instrument was equipped with OpenLAB CDS ChemStation Edition (Rev.C.01.08) software.Antibiotic isolation was accomplished using vacuum liquid chromatography with silica gel, pore size 60 Å, particle size 40-60 µm (Sigma-Aldrich).All solvents used throughout the project were OPTIMA grade (Fisher Scientific).Semipreparative HPLC was carried out on an Agilent 1260 Infinity II series equipped with an autosampler, a diode array detector, a quaternary pump, and a 96-well plate fraction collector with a Phenomenex Kinetex 5 µm C18, 100 Å column (150 × 10 mM) as the stationary phase.All semipreparative HPLC experiments used a flow rate of 4 mL/min.

TABLE 1
MIC of (Z)-4C-14:1 compared to tetracycline against select pathogens in CAMHB a a MIC determinations were performed in triplicate for each pathogen.