Shelf-life extension of lettuce using cell-penetrating peptide-peptide nucleic acid conjugates targeting Sphingomonaspaucimobilis Authors: Su-jin Yum, Seon Yeong Yu, Seung Min Kim, Seockmo Ku, Hee Gon Jeong Published: 2026 ABSTRACT Sphingomonas is a dominant bacterial genus within the indigenous microbiota of plant-derived foods. S. paucimobilis, a food spoilage species and potential human pathogen, is frequently found on plant surfaces, in soils, and in freshwater environments, and raises food safety concerns. Conventional preservation strategies are often insufficient for the selective control of specific spoilage-associated microorganisms without disrupting the surrounding microbiota, underscoring the need for precision-targeted spoilage control approaches. Here, we present a peptide nucleic acid (PNA)-based antimicrobial strategy that employs antisense inhibition for selective suppression of S. paucimobilis. To enhance their cellular uptake, PNAs were conjugated with cell-penetrating peptides (CPPs), and several candidates were compared to identify the most effective delivery system. Among the tested CPPs, (KFF)3K-PNA conjugates exhibited the strongest antimicrobial activity. PNAs targeting the essential genes gyrA and rpoA showed potent sequence-specific activity, with (KFF)3K-gyrA PNAs demonstrating greater efficacy (≥0.5 μM) than (KFF)3K-rpoA PNAs (≥2 μM), whereas mismatch conjugates showed no detectable effect. Quantitative transcriptional analysis confirmed significant downregulation of target genes, particularly gyrA, after treatment. Application of (KFF)3K-gyrA PNA (2 μM) to lettuce significantly reduced discoloration and suppressed S. paucimobilis growth. These results highlight the high target specificity and antimicrobial efficacy of (KFF)3K-PNA conjugates and suggest their potential use as precision biocontrol tools for enhancing the microbial stability and shelf life of plant-derived foods. Introduction Minimally processed foods, such as fresh vegetables, fruits, meats, seafood, and dairy products, are particularly vulnerable to microbial spoilage because their high water and nutrient contents promote microbial growth. Among these foods, fresh vegetables are widely consumed raw and undergo only minimal processing to preserve flavor and nutrition (Sirsat and Neal, 2013; Sun et al., 2022). These features make fresh produce especially susceptible to microbial contamination, which degrades quality (taste, odor, texture) and increases the risk of foodborne illness, leading to economic losses across the supply chain and contributing to environmental pollution through food waste (Snyder et al., 2024). Extensive research has examined microbial spoilage in fresh vegetables and strategies for its prevention, with the spoilage mechanisms associated with Pseudomonas spp. receiving particular attention mechanisms (Federico et al., 2015; Lagaert et al., 2009; Lee et al., 2013). Microbial community analyses conducted during lettuce processing and storage have identified further diverse taxa implicated in spoilage (Sun et al., 2022), while parallel efforts have advanced physical (e.g., pasteurization, freezing, refrigeration) and chemical (e.g., organic acids, salt, preservatives) preservation methods for inhibiting microbial growth and extending produce shelf life (Doyle, 2009; Juneja et al., 2018; Mustapha and Lee, 2017). Recent advances have included the use of synthetic biology and other molecular approaches, which have opened new avenues for precision microbial control. For example, CRISPR-Cas9 technology has been applied to eliminate virulence genes in foodborne pathogens, and silver nanoparticles with potent antimicrobial activity have shown efficacy in eliminating microbes from food-contact surfaces (Sridhar et al., 2021). The use of metagenomics has also improved our understanding of spoilage community diversity and function and now enables rapid detection of spoilage microbes and targeted interventions (Jeon et al., 2019; Yu et al., 2018). In parallel, nonpathogenic microbes have been engineered to biosynthesize antimicrobials that support food preservation (Amit et al., 2017). However, these preservation interventions are limited by their lack of taxonomic specificity and raise potential risks of disrupting beneficial microbiota, accelerating resistance development, and/or altering sensory quality. For these reasons, new techniques are needed that can specifically target spoilage microbes. Among newly emerging tools for precision microbial control, the use of peptide nucleic acids (PNAs) has attracted considerable attention because they provide a sequence-specific modality for suppressing selected microbial genes (MacLelland et al., 2024). PNAs are synthetic analogs of DNA that bind to complementary DNA or RNA sequences in an antisense manner, thereby blocking transcription or translation and silencing target gene expression (MacLelland et al., 2024). The potential applications of PNAs have been actively explored in medical contexts, including disease treatment, and in diagnostic settings for detecting harmful microorganisms, including foodborne pathogens (Kim et al., 2023＀＀S. K. Kim et al., 2023; MacLelland et al., 2024). Recent studies have increasingly focused on PNAs as novel antimicrobials that suppress essential or virulence-associated genes in target microbes (Patenge et al., 2013; Barkowsky et al., 2022; Kim et al., 2024). However, research applying PNAs to extend food shelf life or improve food safety remains extremely limited. Compared with conventional antimicrobials, PNAs provide high specificity, thereby minimizing off-target effects and environmental disruption. Unfortunately, their practical application has been constrained by poor cellular uptake due to their neutral backbone. However, this limitation can be overcome by conjugating PNAs with cell-penetrating peptides (CPPs), which facilitate transport across bacterial membranes. Although CPPs may exhibit lower transport effectiveness in eukaryotic systems, they are highly effective in bacteria because of their structural compatibility with bacterial membranes. Furthermore, the transient intracellular stability of CPP-PNA conjugates-often a drawback in therapeutic applications-can be advantages in food systems as it enables effective bacterial suppression while reducing residual accumulation or raising environmental safety concerns. In this study, metagenomic sequencing data on plant-derived foods were used to identify Sphingomonas paucimobilis as a target organism. S. paucimobilis has been detected in raw or plant-derived food environments and may be associated with microbial quality changes in fresh produce (Kim et al., 2024). It has also been reported as an opportunistic pathogen in clinical and environmental contexts (Alkhatib et al., 2022; Ionescu et al., 2022). Here, a specific PNA targeting S. paucimobilis was designed and conjugated with CPP to improve delivery. The antimicrobial activity was evaluated both in vitro and on lettuce to assess its spoilage-delaying effect. This work provides a foundation for developing precision-targeting technologies against spoilage-associated or potentially harmful microorganisms in raw food materials. Material and methods Reagents and bacterial culture S. paucimobilis ATCC 29837 was cultivated in nutrient broth (NB, Difco, Detroit, MI, USA) at 37°C with shaking at 200 rpm to determine its antimicrobial activities in a planktonic culture. Nutrient agar (NA, Difco, Detroit, MI, USA) was used for the colony forming unit (CFU) determination and standard spot assay. Polystyrene 96-well plates and petri dishes used to confirm antimicrobial activity were purchased from SPL Life Sciences (Pocheon, Korea). Analysis of the microbial community and S. paucimobilis quantification In this study, the raw sequences obtained from previous studies were reanalyzed (Jeon et al., 2019; Seo et al., 2022; Yu et al., 2018; S.-J. Yum et al., 2023; S. J. Yum et al., 2025). As described in previous studies, the resident microbiota were analyzed in five types of plant-based foods (260 samples; 80 chicory, 60 Chinese chives, 60 kale, 30 lettuce, and 30 perilla leaves) using 16S rRNA gene sequencing based on the Illumina MiSeq system (Illumina, San Diego, CA, USA) with paired-end sequencing (2×300 bp). The raw sequences were processed to remove adapter/primer sequences and chimeric sequences using MOTHUR software (ver. 1.38.1). Subsequently, operational taxonomic units (OTUs) were classified using the SILVA database (ver. 123, clustered at 99% sequence similarity) within the CLC Genomics Workbench software (ver. 9.5.3, CLC bio-Qiagen, Aarhus, Denmark). Our analysis of the resident microbiota structure excluded plant archaea, chloroplasts (Streptophyta), and mitochondria (Raphanus) sequences present in the plant-based foods. Quantitative analysis of S. paucimobilis was conducted as described in previous studies (kh Mustafa et al., 2018). Briefly, quantitative polymerase chain reaction (qPCR) was performed using the TDP-glucose pyrophosphorylase gene-based specific primer set (F: TATCCCCTGTCCGTGCTGAT, R: GGATCGTCGACATGGTAGGC) with the CFX Connect Optics Module (Bio-Rad, Hercules, CA, USA) and KOD SYBR qPCR mix (TOYOBO, Osaka, Japan). The bacterial genomic DNA extraction was performed on 1 mL of S. paucimobilis suspension using the NucleoSpin microbial DNA mini kit (Macherey-Nagel, Düren, Germany) according to the manufacturer's instructions. PCR reaction mixtures were assembled from 10 μL KOD SYBR qPCR Mix (TOYOBO, Osaka, Japan), 10 μM each primer, and 1 μL metagenome DNA template or distilled water (DW; negative control group) in a final volume of 20 μL. The PCR conditions were as follows: 98°C for3 min initially; 40 cycles of 98°C for10 s, 55°C for 30 s, and 72°C for 30 s; and 72°C for 3 min. The standard curves were generated using 10-fold dilutions of this bacterial gDNA (101 to 104). The CFX Manager software (BioRad, Hercules, CA, USA) was used to determine the cycle threshold (Ct) values after adjusting the baseline to 500. The regression coefficient (r2) value of the standard curve was above 0.98, and the Ct value was compared with the standard curve to quantify the bacterial loads. All samples were analyzed in triplicate. Antisense CPP-PNA synthesis and determination of CPP-PNA antimicrobial activities All CPP-PNA conjugates were synthesized and purified by HLB Panagene (Daejeon, Korea). The PNA oligomers were designed as complementary to mRNA around the start codon and its neighboring region (10 mer). The flexible linker (L) for the CPPs was coupled directly to the PNA part on the sequence. The antimicrobial PNA targeted the mRNA transcripts of the gyrA and rpoA genes in S. paucimobilis and other potential pathogenic bacteria, which encode DNA gyrase and RNA polymerase, respectively. Among various essential genes, gyrA and rpoA were selected as targets because they play key roles in DNA replication and transcription, are highly conserved, and have been widely validated as effective PNA targets in previous studies (Kim et al., 2023, Kim et al., 2023; Pal and Seleem, 2023). To examine sequence similarity of these target genes across a broad range of microbial taxa detected in plant-derived foods, representative gene sequences were retrieved from the NCBI database using BLASTn. The retrieved sequences were aligned using the ClustalW, and multiple sequence alignments were subsequently used for comparative analysis. Given the high binding specificity of PNA oligomers, where even a single mismatch can markedly destabilize the complex, we designed a mutant PNA (3 mer) to verify the specificity of our gene targeting (Otsuka et al., 2016) (Table 1). The CPP-PNA conjugates were dissolved in DW to produce a stock solution (1 mM) and stored at −20°C.Table 1CPP-PNA conjugates for antimicrobial activity studies in S. paucimobilis.Table 1CPPCPP-PNA designationCPP-PNA sequenceReferences(RXR)4XB(RXR)4XB-anti-gyrA-PNARXRRXRRXRRXRXB-L-CCAA AGTCTG-NH2Barkowsky et al. (2022)(KFF)3K(KFF)3K-anti-gyrA-PNAKFFKFFKFFK-L-CCAA AGTCTG-NH2(KFF)3K-anti-rpoA-PNAKFFKFFKFFK-L-ACACGTGGTT-NH2Popella et al. (2022)Mismatch-(KFF)3K-anti-gyrA-PNAKFFKFFKFFK-L-CGTA AGGCTG-NH2OligolysinOligolysin-anti-gyrA-PNAKKKKKKKK-L-CCAA AGTCTG-NH2Barkowsky et al. (2022) Table 1CPP-PNA conjugates for antimicrobial activity studies in S. paucimobilis.Table 1CPPCPP-PNA designationCPP-PNA sequenceReferences(RXR)4XB(RXR)4XB-anti-gyrA-PNARXRRXRRXRRXRXB-L-CCAA AGTCTG-NH2Barkowsky et al. (2022)(KFF)3K(KFF)3K-anti-gyrA-PNAKFFKFFKFFK-L-CCAA AGTCTG-NH2(KFF)3K-anti-rpoA-PNAKFFKFFKFFK-L-ACACGTGGTT-NH2Popella et al. (2022)Mismatch-(KFF)3K-anti-gyrA-PNAKFFKFFKFFK-L-CGTA AGGCTG-NH2OligolysinOligolysin-anti-gyrA-PNAKKKKKKKK-L-CCAA AGTCTG-NH2Barkowsky et al. (2022) CPP-PNA conjugates for antimicrobial activity studies in S. paucimobilis. S. paucimobilis were precultured in NB at 37°C with shaking at 200 rpm for overnight and diluted 1:1000 with buffer peptone water to obtain a bacterial concentration of 1.86 × 105 colony forming units (CFU)/mL. Diluted S. paucimobilis cells were used to investigate the antimicrobial activity of various CPP-PNA conjugates. For CPP screening and MIC testing, the S. paucimobilis diluents were transferred to a polystyrene 96-well plate (SPL Life Sciences, Korea), and the different CPP-PNA conjugates were then added to the plate. After CPP screening and minimal inhibitory concentration (MIC) testing, various concentrations of CPP-PNA conjugates ((RXR)4XB, (KFF)3K, and oligolysin (0-5 μM) were added to the 96-well plate. The CPPs were chosen as representative candidates based on their previously demonstrated ability to facilitate PNA delivery in gram-negative bacteria and their structural diversity (Wojciechowska et al., 2020). The cultures were incubated at 37°C for 12 h, followed by optical density measurements at 600 nm using a UV/Vis spectrophotometer (Thermo Fisher Scientific, USA). For concentration-dependent spotting assays, 10 μL of culture was spotted onto agar plates, incubated for 10 h, and photographed. MICs were determined following the protocol of the Clinical and Laboratory Standard Institute (CLSI). The MIC values were recorded as the lowest concentration that showed no turbidity in the wells. The antimicrobial effect was determined using standard spot assays and bacterial cell counting, by adding 30 μL CPP-coupled antisense PNAs ((RXR)4XB-anti-gyrA PNA, (KFF)3K-anti-gyrA PNA, (KFF)3K-anti-rpoA PNA, or mutant-(KFF)3K-anti-gyrA PNA) to 270 μL of bacterial suspension for a final PNA concentration of 0-2 μM and incubating at 37°C for 6 h in a 96-well plate. Samples were collected at 1, 3, and 6 h and serially diluted for CFU measurements and spot analyses. For the standard spot assays, 5 μL of each bacterial diluent was dropped onto NA plates and incubated at 37°C for 16 h. A 100 μL of diluent was spread on an NA plate incubated at 37°C overnight, followed by CFU counting. Extraction of RNA and real-time quantitative reverse transcription PCR (qRT-PCR) The antimicrobial activities of the CPP-PNA conjugates at the transcriptional level were assessed using RNAprotect bacteria reagent, RNeasy mini kits, and SsoAdvanced Universal SYBR Green Supermix (purchased from Qiagen, Valencia, CA, USA and Bio-Rad, Hercules, CA, USA). A 1470 μL volume of bacterial suspension (6.52 × 104 CFU/mL) was prepared for each experimental group and tested using 30 μL CPP-PNA by incubating in a glass test tube at 37°C for 6 h with shaking at 7 rpm. Total bacterial RNA was isolated and purified according to the manufacturer's instructions, and the RNA concentration and purity were assessed using the Nabi UV/Vis Nano spectrophotometer. A 12 μL sample of undiluted RNA in a reaction volume of 20 μL was used to synthesize cDNA using the iSCript cDNA synthesis kit (Bio-Rad, USA) according to the manufacturer's instructions. The final cDNA was diluted if larger volumes were required for analysis of target genes. Thermal cycling conditions were as follows (per the kit instructions): 5 min at 25°C, 20 min at 46°C, 1 min at 95°C, hold at 4°C. The qRT-PCR was performed using an CFX Connect Optics Module (Bio-Rad, USA) with SsoAdvanced Universal SYBR Green Supermix, primer sets. The primer sequences used were as follows: gyrA_F (5′-CTGTCCTCCGCCTATGAATATC-3′), gyrA_R (5′- GATCGGTGACGAGCATCAAT-3′), rpoA_F (5′-AACTGCCTCAAGAACGACAA-3′), rpoA_R (5′-GACGACAGCACTTCCTTGAT-3′), and 16S rRNA_F (CACGTGCTACAATGGCAACT), 16S rRNA_R (GTACAAGGCCTGGGAACGTA). Reaction mixtures were assembled from 5 μL SsoAdvanced Universal SYBR Green Supermix reagent (Bio-Rad, USA), 0.5 μM of each primer, 2 μL nuclease-free water, and 1 μL cDNA template or DW (negative control group) in a final volume of 10 μL. All reactions were carried out in triplicate. The qRT-PCR conditions were 30 s at 98°C, 15 s at 98°C, 30 s at 55-60°C (rpoA, gyrA, and 16S rRNA), and 30 s at 72°C for 40 cycles, followed by an additional heating step from 65 to 95°C to generate a melting curve. The gyrA and rpoA genes were used as internal controls. The gene expression levels were analyzed using the comparative delta-delta Ct (2−ΔΔCt) method. Antimicrobial effect of CPP-PNAs on lettuce infected with S. paucimobilis Lettuce (Lactuca sativa L.) samples were purchased from a retail market (Homplus, Daejeon, Korea) and exposed to UV irradiation for 15 min on each side on a clean bench. The lettuce leaves were then inoculated with S. paucimobilis (log 7.74 CFU) using a spotting method and then air dried for 90 min to allow bacterial colonization. The different CPP-conjugated PNAs - (KFF)3K-anti-gyrA-PNA, and (KFF)3K-anti rpoA-PNA - were spotted at 2 μM onto the inoculated lettuce surfaces, using DW as the experimental control. After incubation at 25°C for 7 days, the antimicrobial activities of CPP-PNA on the treated samples were determined using the total surface color difference (ΔE) over time. Images were converted from RGB to CIE Lab∗ color space using Photoshop, and ΔE values were computed from L∗, a∗, and b∗ coordinates relative to day 0. Here, L∗ represents lightness (+lighter, − darker), a∗ indicates the red-green axis (+redder, − greener), and b∗ represents the yellow-blue axis (+yellower, − bluer). ΔE was then calculated using the following formula:ΔE=(L2∗−L1∗)2+(a2∗−a1∗)2+(b2∗−b1∗)2 Statistical analysis Statistical analyses were performed using GraphPad Prism (v5.02; GraphPad Software, Boston, MA, USA) and SAS software (v9.4; SAS Institute, Cary, NC, USA). All experiments were conducted in triplicate with independent biological and technical replicates, and data are presented as mean ± standard deviation (S.D.). The MIC and ΔE values were analyzed using appropriate statistical tests, with significance determined at p < 0.05. Heat map visualization of microbial abundance at the genus level was generated using the “heatmap.plus” package in RStudio (v1.1.463; R Foundation for Statistical Computing, Vienna, Austria). DNA sequence alignment was performed with ClustalW using Multalin (v5.4.1; http://multalin.toulouse.inra.fr/multalin/). Results and discussion Microbiota composition of plant-derived foods and S. paucimobilis quantification Ecologically relevant target organisms were identified and the potential off-target impacts of CPP-PNA treatment were evaluated by first characterizing the resident microbiota of five different plant-derived foods (chicory, Chinese chive, kale, lettuce, and perilla). This baseline analysis provided critical context for selecting S. paucimobilis as the target organism and for assessing the specificity of the designed PNAs within complex microbial communities. Specifically, we re-analyzed 16S rRNA gene sequences obtained in previous work on the resident microbiota of these species (Jeon et al., 2019; Seo et al., 2022; Yu et al., 2018; S.-J. Yum et al., 2023; S. J. Yum et al., 2025). The top 10 genera, while showing low proportions in individual produce types, accounted for an average relative abundance above 2% across all samples (Fig. 1A). Dominant genera included Pantoea, Enterobacter, Pseudomonas, Methylobacterium, Acinetobacter, Sphingomonas, and Escherichia from the phylum Proteobacteria, and Exiguobacterium, Bacillus, and Weissella from Firmicutes. Notably, Exiguobacterium (Lettuce, Perilla, Kale; 22.84 ± 15.73%), Pantoea (lettuce, perilla, Chinese chive, kale, chicory; 13.74 ± 4.31%), Pseudomonas (lettuce, Chinese chive, chicory; 15.18 ± 11.37%), and Bacillus (lettuce, perilla, kale; 10.78 ± 3.96%) were widely distributed across various plant tissues and environments, consistent with previous reports (S.-J. Yum et al., 2023).Fig. 1Taxonomic analysis and quantification of S. paucimobilis in plant-derived foods. (A) Heatmap showing the common top 10 genera across plant-derived food samples. Each column represents genus-level relative abundance. (B) Quantification of S. paucimobilis by qRT-PCR, expressed as target gene copies relative to total 16S rRNA gene copies. Data represent mean ± S.D. from three replicates. Taxonomic analysis and quantification of S. paucimobilis in plant-derived foods. (A) Heatmap showing the common top 10 genera across plant-derived food samples. Each column represents genus-level relative abundance. (B) Quantification of S. paucimobilis by qRT-PCR, expressed as target gene copies relative to total 16S rRNA gene copies. Data represent mean ± S.D. from three replicates. Spoilage-related genera, such as Pseudomonas and Bacillus, are relatively abundant in spoiled produce, including lettuce, spinach, potato, and onion, indicating their strong association with plant-derived food spoilage (Karanth et al., 2023). Pseudomonas has been shown to accelerate spoilage by secreting protein- and lipid-degrading enzymes, forming biofilms, and growing at refrigeration temperatures, while producing green pigments that deteriorate seafood and meat quality (Chmielewski and Frank, 2003; Gill and Newton, 1977; Gram and Huss, 1996). Bacillus also survives in heat-processed foods through spore formation and toxin production, further contributing to spoilage and foodborne disease outbreaks (Ehling-Schulz et al., 2004; Logan and Vos, 2015; Vilain and Brözel, 2006). However, the contribution of Sphingomonas to plant-derived food spoilage is not well understood. To close this knowledge gap, we used qRT-PCR to identify and quantify S. paucimobilis as a potential pathogenic bacterium on our produce samples. The Sphingomonas genus is known to produce pectinolytic enzymes and contribute to the spoilage of plant-derived foods (Fig. 1). Consistent with its relative abundance, S. paucimobilis was detected more frequently on perilla (33.33%, 1.24 × 107 CFU/g; relative abundance and detected CFU) and chicory (27.50%, 1.04 × 107 CFU/g) than on lettuce (6.67%, 3.18 × 107 CFU/g), Chinese chive (15.00%, 5.72 × 105 CFU/g), or kale (5.00%, 2.92 × 107 CFU/g), although the overall cell counts were comparable among the samples. The microbial production of pectinolytic enzymes plays a critical role in the decay of fruits and vegetables by breaking down plant cell walls (Rehman, 2023). Therefore, S. paucimobilis likely contributes substantially to spoilage in plant-derived foods, highlighting the need for targeted inhibition strategies. Appropriate approaches could include biological control using beneficial microbes, precision targeting via PNAs, optimized packaging, or controlled storage to suppress bacterial growth and enzyme activity. Thus, S. paucimobilis was selected for further study on lettuce to evaluate the feasibility of using CPP-PNA technology for precision microbial control while minimizing unintended impacts on the resident lettuce microbiota. Comparative screening of CPPs for efficient intracellular delivery of antisense PNAs in S. paucimobilis Given the significant role of S. paucimobilis in plant-based food spoilage, this study on the antimicrobial efficacy of CPP-conjugated antisense PNAs targeted the gyrA and rpoA genes, which are both indispensable for DNA replication and transcription and are therefore critical for bacterial growth and survival. The gyrA and rpoA genes encode the A subunit of DNA gyrase and the α subunit of RNA polymerase, respectively (Bai et al., 2012a, Bai et al., 2012b), and previous studies have identified these enzymes as effective targets for antisense PNAs and other antimicrobial compounds (El-Fateh et al., 2024). The sequence similarities of gyrA and rpoA were examined across a broad range of microbial taxa detected in plant-derived foods (Fig. 2). No identical sequences were identified in non-target microorganisms, suggesting that the designed PNAs were expected to preferentially recognize S. paucimobilis at the sequence level. This high degree of sequence specificity ensured minimal off-target effects, regardless of whether the surrounding microbiota include spoilage-associated, pathogenic, or harmless species. This type of precision is particularly important in food systems, where maintaining the native microbial balance may help preserve product quality and reduce unintended ecological shifts. Accordingly, our PNAs were designed to bind to approximately 10 bp regions flanking the start codons of gyrA and rpoA to provide specific inhibition of S. paucimobilis.Fig. 2Comparison of gyrA and rpoA sequences from diverse bacteria potentially present in plant-derived foods. (A) Alignment of gyrA and (B) rpoA sequences, including regions containing the start codons, was performed across multiple bacterial species. The analysis confirmed that the PNA target region (box) of S. paucimobilis does not share identical sequences with those of other bacteria. Comparison of gyrA and rpoA sequences from diverse bacteria potentially present in plant-derived foods. (A) Alignment of gyrA and (B) rpoA sequences, including regions containing the start codons, was performed across multiple bacterial species. The analysis confirmed that the PNA target region (box) of S. paucimobilis does not share identical sequences with those of other bacteria. Although other nucleic acid analogs, such as Locked Nucleic Acid (LNA), Bridged Nucleic Acid (BNA), and Zip Nucleic Acid (ZNA), have been developed for gene regulation, their microbial control applications remain limited due to their low delivery efficiency or complex synthesis requirements (Kuijper et al., 2021; Roberts et al., 2020). In contrast, CPP-PNA conjugates show effective intracellular delivery, specificity, and stability, as validated in multiple studies (Nielsen and Shiraishi, 2011; Tsylents et al., 2023). Based on these attributes, a CPP-PNA strategy was selected as the most suitable approach for targeted control of food-spoilage bacteria. Three CPPs - (RXR)4XB, (KFF)3K, and oligolysin - were compared for their ability to deliver antisense PNAs into S. paucimobilis (Fig. 3). In this experiment, PNAs targeting the gyrA gene, which had been reported in previous studies to exhibit the consistent antimicrobial activity, were conjugated to the CPPs. The sequences of the CPP-PNA conjugates are summarized in Table 1. Optical density measurements and MIC assays confirmed significant growth inhibition by all CPP-anti-gyrA PNA groups relative to the DW control. Notably, (RXR)4XB and (KFF)3K showed inhibition at concentrations as low as 1 μM (p < 0.05), with stronger effects at 2.5 and 5 μM (p < 0.001). The corresponding optical density reductions were approximately 0.07-0.17 for (RXR)4XB and 0.08-0.17 for (KFF)3K, while oligolysin caused significant inhibition only at concentrations above 2.5 μM (reductions 0.14-0.18; p < 0.001).Fig. 3Screening of CPPs for growth inhibition of S. paucimobilis by antisense PNA targeting gyrA. Bacterial growth was measured by O.D.600 nm after treatment with (RXR)4XB-, (KFF)3K-, and Oligolysin-conjugated antisense PNAs targeting gyrA at different concentrations. Data represent mean ± S.D. from three replicates. (∗, p < 0.05; ∗∗∗, p < 0.001). Screening of CPPs for growth inhibition of S. paucimobilis by antisense PNA targeting gyrA. Bacterial growth was measured by O.D.600 nm after treatment with (RXR)4XB-, (KFF)3K-, and Oligolysin-conjugated antisense PNAs targeting gyrA at different concentrations. Data represent mean ± S.D. from three replicates. (∗, p < 0.05; ∗∗∗, p < 0.001). The MIC determinations via broth microdilution revealed effective concentrations of <0.05 μM for (KFF)3K, <0.1 μM for (RXR)4XB, and <0.5 μM for oligolysin, confirming (KFF)3K as the most potent inhibitor (Table 2). These results indicate that the ability of CPPs to permeabilize microbial membranes enhanced intracellular PNA delivery. However, the penetration efficiency may vary among gram-negative bacteria other than S. paucimobilis, as the CPPs tested here have shown broad activity in prior studies. For example, (KFF)3K-PNA conjugates inhibited E. coli at ≥ 3 μM, whereas (RXR)4XB-PNA conjugates were active against P. aeruginosa at ≥ 1 μM (Howard et al., 2017; Wojciechowska et al., 2020).Table 2Minimum inhibitory concentrations (MICs) of CPP-PNAs against S. paucimobilis.Table 2CPP-anti-gyrA-PNA conjugatesConcentration of PNA (μM)00.010.050.10.51(RXR)4XB-gyrA0.22 ± 0.02 aa0.19 ± 0.01 a0.17 ± 0.02 a0.16 ± 0.01 b0.10 ± 0.00 c0.09 ± 0.01 c(KFF)3K-gyrA0.19 ± 0.01 a0.16 ± 0.01 b0.13 ± 0.01 c0.13 ± 0.02 c0.07 ± 0.01 dOligolysin-gyrA0.18 ± 0.02 a0.18 ± 0.01 a0.16 ± 0.04 a0.14 ± 0.01 b0.11 ± 0.01 baDifferent superscripts indicate statistically significant difference (p < 0.05). All values are mean ± standard deviation of three repeated processes. Table 2Minimum inhibitory concentrations (MICs) of CPP-PNAs against S. paucimobilis.Table 2CPP-anti-gyrA-PNA conjugatesConcentration of PNA (μM)00.010.050.10.51(RXR)4XB-gyrA0.22 ± 0.02 aa0.19 ± 0.01 a0.17 ± 0.02 a0.16 ± 0.01 b0.10 ± 0.00 c0.09 ± 0.01 c(KFF)3K-gyrA0.19 ± 0.01 a0.16 ± 0.01 b0.13 ± 0.01 c0.13 ± 0.02 c0.07 ± 0.01 dOligolysin-gyrA0.18 ± 0.02 a0.18 ± 0.01 a0.16 ± 0.04 a0.14 ± 0.01 b0.11 ± 0.01 baDifferent superscripts indicate statistically significant difference (p < 0.05). All values are mean ± standard deviation of three repeated processes. Minimum inhibitory concentrations (MICs) of CPP-PNAs against S. paucimobilis. Different superscripts indicate statistically significant difference (p < 0.05). All values are mean ± standard deviation of three repeated processes. Consistent with previous studies, the antimicrobial potency observed in this study depended strongly on the CPP composition. (RXR)4XB, composed of repetitive arginine sequences, interacts with negatively charged cell membranes to facilitate PNA delivery (Barkowsky et al., 2022; Greenberg et al., 2010), whereas oligolysin and (RXR)4XB can bypass transporter requirements and traverse the inner membrane through membrane potential-dependent mechanisms (Mulkern et al., 2024). Conversely, (KFF)3K, which is rich in lysine and phenylalanine, efficiently penetrates both gram-positive and gram-negative membranes and exhibits low cytotoxicity, making it a promising vehicle for antimicrobial PNA delivery (da Silva et al., 2021; Macyszyn et al., 2023; Wojciechowska et al., 2020). (KFF)3K-PNA conjugates are believed to exploit ATP-binding cassette (ABC) transporters, thereby explaining their high intracellular uptake and efficacy (El-Fateh et al., 2024b). These mechanistic differences in membrane transport likely underlie the observed variations in inhibitory activity in different gram-negative species (Hashimoto et al., 2005; Popella et al., 2022). Although CPPs may have mild intrinsic antimicrobial effects, the concentrations used here were below their inhibitory thresholds, confirming that the observed effects stemmed from antisense PNA activity rather than CPP toxicity. Taken together, these results indicate that (KFF)3K is the most effective carrier for delivering antisense PNAs against S. paucimobilis. Importantly, this suggests the potential for its use in developing precision biocontrol strategies for plant-derived foods, either alone or in combination with optimized storage conditions and packaging technologies that mitigate spoilage risks. Target-specific antimicrobial effects of CPP-PNA conjugates against S. paucimobilis In this study, (RXR)4XB and (KFF)3K, which were identified as the most potent CPPs in preliminary screening, were conjugated to antisense PNAs targeting the essential genes gyrA and rpoA to evaluate their antimicrobial potential against S. paucimobilis. Quantification of the inhibitory effects of these CPP-PNA conjugates using either colony counting or spot assays confirmed consistent trends across all treatments (Fig. 4). Relative to the untreated control, both conjugates produced significant reductions in bacterial counts, confirming their effectiveness at inhibiting bacterial growth.Fig. 4Antimicrobial activity of CPP-PNA conjugates targeting gyrA and rpoA in S. paucimobilis. Bacterial growth inhibition was assessed by colony counting and standard spot assays after treatment with (RXR)4XB-anti-gyrA PNA (A), (KFF)3K-anti-gyrA PNA (B), and (KFF)3K-anti-rpoA PNA (C) at different concentrations and time points. Data represent mean ± SD from three replicates. (∗, p < 0.05; ∗∗, p < 0.01; ∗∗∗, p < 0.001). Antimicrobial activity of CPP-PNA conjugates targeting gyrA and rpoA in S. paucimobilis. Bacterial growth inhibition was assessed by colony counting and standard spot assays after treatment with (RXR)4XB-anti-gyrA PNA (A), (KFF)3K-anti-gyrA PNA (B), and (KFF)3K-anti-rpoA PNA (C) at different concentrations and time points. Data represent mean ± SD from three replicates. (∗, p < 0.05; ∗∗, p < 0.01; ∗∗∗, p < 0.001). The (RXR)4XB-anti-gyrA PNA exhibited inhibition at concentrations ≥1 μM after 6 h of treatment, with viable cell count reduction of 0.33-0.79 log CFU/mL (p < 0.05; Fig. 4A). In contrast, the (KFF)3K-anti-gyrA PNA at 0.5 μM achieved detectable inhibition within 1 h, resulting in cell count reductions of 0.25-1.37 log CFU/mL after 6 h (p < 0.001; Fig. 4B). This superior efficacy of (KFF)3K suggests a more efficient intracellular delivery of the antisense PNA and stronger suppression of gyrA, a key gene in bacterial DNA replication. The antimicrobial activity of (KFF)3K-anti-rpoA PNA, targeting rpoA (encoding the α-subunit of RNA polymerase), was also evaluated. Treatment with 2 μM of this conjugate reduced viable counts by 0.53, 0.38, and 0.43 log CFU/mL at 1, 3, and 6 h, respectively (Fig. 4C). However, the (KFF)3K-anti-gyrA PNA displayed significantly stronger inhibition, identifying gyrA as a more effective target under the tested conditions. DNA gyrase (encoded by gyrA and gyrB) is indispensable for replication, whereas RNA polymerase (including the α-subunit encoded by rpoA) governs transcription (Balas et al., 1998; Mamudu et al., 2024). Interference with rpoA hampers transcription initiation and may destabilize the RNA polymerase complex, but direct disruption of DNA replication via gyrA inhibition likely accounts for the greater antibacterial efficacy observed. Our findings are supported by previous reports showing that CPP-PNA conjugates targeting gyrA and rpoA have broad-spectrum activity, as (RXR)4XB-anti-gyrA inhibited Streptococcus pneumoniae at ≥ 5 μM, whereas (RFR)4XB-anti-rpoA suppressed Clostridioides difficile at ≥ 8 μM (Barkowsky et al., 2022; Pal and Seleem, 2023). The concentration ranges and inhibitory patterns observed here are therefore consistent with prior studies. Collectively, our results indicate that although both gyrA and rpoA encode essential bacterial enzymes, direct inhibition of DNA replication through gyrA yields stronger antimicrobial effects under the tested conditions. Future studies should validate these observations across diverse bacterial taxa and environmental contexts to generalize target specificity and efficacy. Mismatch-based evaluation of (KFF)3K-anti-gyrA PNA specificity and uptake We also assessed the sequence specificity of (KFF)3K-anti-gyrA PNA using colony counting and spot assays for a mutant version containing three base mismatches (Fig. 5). This experiment was designed to distinguish whether the previously observed antimicrobial activity originated from CPP-mediated membrane penetration or from the antisense sequence itself. As shown in Fig. 5B, treatment with the mismatch (KFF)3K-anti-gyrA PNA (0.2-2 μM) produced no significant reduction in S. paucimobilis cell counts compared to the untreated DW control. These findings confirm that the antimicrobial effect of (KFF)3K-anti-gyrA PNA is highly sequence-specific rather than due to nonspecific CPP activity.Fig. 5Sequence specificity of (KFF)3K-anti-gyrA PNA against S. paucimobilis. (A) Alignment of the wild-type gyrA target sequence and the mutant (3 bp mismatch) sequence used to assess specificity. Mismatched bases are highlighted red. (B) Antimicrobial activity was evaluated using colony counting and standard spot assays. (n.s. indicates not significant). Sequence specificity of (KFF)3K-anti-gyrA PNA against S. paucimobilis. (A) Alignment of the wild-type gyrA target sequence and the mutant (3 bp mismatch) sequence used to assess specificity. Mismatched bases are highlighted red. (B) Antimicrobial activity was evaluated using colony counting and standard spot assays. (n.s. indicates not significant). Comparable results have been reported in previous studies using other bacterial models. For example, (KFF)3K-anti-acpP PNA exhibited a minimum inhibitory concentration (MIC) of 1.25 μM against E. coli, whereas the scrambled control required 10 μM to achieve inhibition (Popella et al., 2022). A subsequent spot assay confirmed that (KFF)3K-anti-acpP PNA retained antimicrobial activity for 24 h, while the scrambled conjugate showed no detectable effect. Similarly, (RXR)4XB-anti-rpoD PNA (5 μM) completely eliminated viable E. coli cells, whereas a mismatch variant failed to reduce bacterial density even at 30 μM (Bai et al., 2012). Collectively, these findings underscore that the antimicrobial activity of CPP-PNA conjugates depends strictly on antisense sequence complementarity. The results obtained in this study are consistent with these earlier reports, reinforcing that (KFF)3K-anti-gyrA PNA exerts its antibacterial effect through precise, sequence-specific gene inhibition in S. paucimobilis. Transcriptional repression of gyrA and rpoA by (KFF)3K-conjugated PNAs in S. paucimobilis We verified that (KFF)3K-conjugated PNAs exerted transcriptional repression on their target genes by quantifying the expression levels of gyrA and rpoA in S. paucimobilis using qRT-PCR (Fig. 6). Expression levels were normalized to the transcription of the 16S rRNA gene. S. paucimobilis cultures were treated with 2 μM of CPP-conjugated PNAs, a concentration previously shown to exert antimicrobial effects.Fig. 6Relative expression of gyrA and rpoA in S. paucimobilis following CPP-PNA conjugate treatment. Gene expression was quantified by qRT-PCR and normalized to 16S rRNA levels. Treatment with CPP-PNA conjugates reduced expression of both gyrA and rpoA, with more efficient inhibition observed for gyrA. (∗, p < 0.05; ∗∗∗, p < 0.001). Relative expression of gyrA and rpoA in S. paucimobilis following CPP-PNA conjugate treatment. Gene expression was quantified by qRT-PCR and normalized to 16S rRNA levels. Treatment with CPP-PNA conjugates reduced expression of both gyrA and rpoA, with more efficient inhibition observed for gyrA. (∗, p < 0.05; ∗∗∗, p < 0.001). Following treatment with (KFF)3K-anti-gyrA PNA, the gyrA transcript abundance decreased to 64.31% of the control level (0 μM treatment; p < 0.001). In contrast, treatment with (KFF)3K-anti-rpoA PNA reduced rpoA expression to 18.29% (p < 0.05). Although both targets were significantly downregulated, the stronger repression of gyrA suggests that inhibition of DNA replication has a more direct and pronounced antibacterial effect than is achieved following transcriptional interference via rpoA. Comparable transcript-level reductions have been reported in other bacterial systems. For instance, treatment of Streptococcus pneumoniae with 7.5 μM TAT-anti-gyrA PNA and 10 μM (RXR)4XB-anti-gyrA PNA resulted in target gene expression levels of 29% and 45% of the control, respectively (Barkowsky et al., 2022). Consistent with these reports, the present results demonstrate that antisense PNAs that target essential genes such as gyrA and rpoA achieve precise and effective transcriptional silencing in S. paucimobilis, reinforcing their potential as sequence-specific antimicrobial tools. Inhibition of visual spoilage on fresh lettuce by (KFF)3K-conjugated PNAs targeting S. paucimobilis We examined whether transcriptional repression of gyrA and rpoA results in measurable preservation effects by applying the CPP-PNA conjugates to a lettuce model susceptible to S. paucimobilis surface spoilage and evaluating produce quality by monitoring visual appearance and spoilage progression in the inoculated lettuce leaves (Fig. 7). No discoloration was observed in the non-inoculated group throughout the 5-day period. By contrast, both the S. paucimobilis-inoculated group and the inoculated DW control group exhibited partial discoloration beginning on day 2, with clear signs of spoilage (most notably localized yellowing or browning) appearing from day 3 onward. In groups treated with (KFF)3K-anti-gyrA PNA and (KFF)3K-anti-rpoA PNA, only slight surface yellowing and discoloration were first observed on day 5 and day 3, respectively, and the extent of these changes remained minimal.Fig. 7Effect of PNA treatment on S. paucimobilis infection in lettuce. Phenotypic response of lettuce leaves following inoculation with S. paucimobilis. Each leaf was spot-inoculated with a bacterial suspension and subsequently treated with D.W., (KFF)3K-anti-gyrA PNA, or (KFF)3K-anti-rpoA PNA, while non-inoculated leaves served as the control. Samples were stored at 25°C for 7 days under controlled conditions, and photographs were taken daily to monitor changes. Effect of PNA treatment on S. paucimobilis infection in lettuce. Phenotypic response of lettuce leaves following inoculation with S. paucimobilis. Each leaf was spot-inoculated with a bacterial suspension and subsequently treated with D.W., (KFF)3K-anti-gyrA PNA, or (KFF)3K-anti-rpoA PNA, while non-inoculated leaves served as the control. Samples were stored at 25°C for 7 days under controlled conditions, and photographs were taken daily to monitor changes. These trends were corroborated by quantitative colorimetry (Table 3). Total color difference (ΔE) values, calculated as the change from day 0, differed significantly among treatments during storage, reflecting differences in microbial growth and surface quality. The S. paucimobilis inoculation group and the DW control inoculation group showed higher ΔE values than the non-inoculation group.Table 3Changes in CIELAB color parameters (L, a, b) and total color difference (ΔE) of lettuce leaves inoculated with S. paucimobilis and treated with CPP-PNA conjugates.Table 3GroupColor parameters (mean)Time (day)01234567Non-inoculationL49.61949.94150.03743.48841.12241.14941.31851.295a−18.067−18.099−16.653−14.698−15.902−12.795−15.682−13.956b27.50431.80629.38823.82523.69721.96529.28524.326ΔE-4.31 ± 6.522.39 ± 5.78∗7.90 ± 2.99 b9.56 ± 2.51∗11.41 ± 3.54∗∗8.82 ± 3.91∗∗5.46 ± 6.14∗∗S. paucimobilis inoculationL51.86651.5753.4254.26457.80557.38262.36860.925a−19.09−16.042−17.488−14.136−13.274−9.499−10.011−9.498b33.41731.1840.44636.0938.22136.10342.40934.796ΔE-3.79 ± 3.827.37 ± 1.04∗∗6.12 ± 3.09 b9.60 ± 3.69∗11.39 ± 4.94∗∗16.54 ± 4.86∗∗13.27 ± 1.06∗∗S. paucimobilis inoculation/D.W.L51.31455.0655.60463.68260.34863.0963.20756.527a−19.897−15.57−16.141−11.679−8.607−5.97−4.782−5.444b31.26834.08936.02238.86835.72935.23235.71828.153ΔE-6.38 ± 5.147.42 ± 2.0416.68 ± 2.3015.13 ± 3.7918.67 ± 3.1519.74 ± 3.3915.68 ± 2.59S. paucimobilis inoculation/gyrA PNAL44.64144.64142.2447.5752.7649.78150.2554.983a−15.912−15.912−16.23−13.188−17.198−12.566−16.202−16.471b26.34126.34124.53625.03731.44126.98438.46734.38ΔE-0.00 ± 4.343.02 ± 3.78∗4.21 ± 4.259.67 ± 11.13∗∗6.17 ± 6.23∗∗13.36 ± 5.75∗13.11 ± 8.71∗S. paucimobilis inoculation/rpoA PNAL49.45146.43348.08852.49853.12253.07358.19661.232a−18.55−17.717−19.019−17.689−17.82−13.78−14.855−15.301b29.84330.27131.54836.12637.70236.24445.41541.62ΔE-3.16 ± 4.772.23 ± 2.67 ∗7.04 ± 2.848.70 ± 3.54∗8.77 ± 3.28∗∗18.24 ± 4.9216.97 ± 2.68a, The ΔE values of each treatment group were compared with those of the S. paucimobilis inoculation/DW-treated vehicle control group within the same storage day using Student's t-test for planned pairwise comparisons. Asterisks indicate significant differences compared with the vehicle control group within the same day (∗, p < 0.05; ∗∗, p < 0.01). All values are expressed as mean ± standard deviation. Table 3Changes in CIELAB color parameters (L, a, b) and total color difference (ΔE) of lettuce leaves inoculated with S. paucimobilis and treated with CPP-PNA conjugates.Table 3GroupColor parameters (mean)Time (day)01234567Non-inoculationL49.61949.94150.03743.48841.12241.14941.31851.295a−18.067−18.099−16.653−14.698−15.902−12.795−15.682−13.956b27.50431.80629.38823.82523.69721.96529.28524.326ΔE-4.31 ± 6.522.39 ± 5.78∗7.90 ± 2.99 b9.56 ± 2.51∗11.41 ± 3.54∗∗8.82 ± 3.91∗∗5.46 ± 6.14∗∗S. paucimobilis inoculationL51.86651.5753.4254.26457.80557.38262.36860.925a−19.09−16.042−17.488−14.136−13.274−9.499−10.011−9.498b33.41731.1840.44636.0938.22136.10342.40934.796ΔE-3.79 ± 3.827.37 ± 1.04∗∗6.12 ± 3.09 b9.60 ± 3.69∗11.39 ± 4.94∗∗16.54 ± 4.86∗∗13.27 ± 1.06∗∗S. paucimobilis inoculation/D.W.L51.31455.0655.60463.68260.34863.0963.20756.527a−19.897−15.57−16.141−11.679−8.607−5.97−4.782−5.444b31.26834.08936.02238.86835.72935.23235.71828.153ΔE-6.38 ± 5.147.42 ± 2.0416.68 ± 2.3015.13 ± 3.7918.67 ± 3.1519.74 ± 3.3915.68 ± 2.59S. paucimobilis inoculation/gyrA PNAL44.64144.64142.2447.5752.7649.78150.2554.983a−15.912−15.912−16.23−13.188−17.198−12.566−16.202−16.471b26.34126.34124.53625.03731.44126.98438.46734.38ΔE-0.00 ± 4.343.02 ± 3.78∗4.21 ± 4.259.67 ± 11.13∗∗6.17 ± 6.23∗∗13.36 ± 5.75∗13.11 ± 8.71∗S. paucimobilis inoculation/rpoA PNAL49.45146.43348.08852.49853.12253.07358.19661.232a−18.55−17.717−19.019−17.689−17.82−13.78−14.855−15.301b29.84330.27131.54836.12637.70236.24445.41541.62ΔE-3.16 ± 4.772.23 ± 2.67 ∗7.04 ± 2.848.70 ± 3.54∗8.77 ± 3.28∗∗18.24 ± 4.9216.97 ± 2.68a, The ΔE values of each treatment group were compared with those of the S. paucimobilis inoculation/DW-treated vehicle control group within the same storage day using Student's t-test for planned pairwise comparisons. Asterisks indicate significant differences compared with the vehicle control group within the same day (∗, p < 0.05; ∗∗, p < 0.01). All values are expressed as mean ± standard deviation. Changes in CIELAB color parameters (L, a, b) and total color difference (ΔE) of lettuce leaves inoculated with S. paucimobilis and treated with CPP-PNA conjugates. a, The ΔE values of each treatment grou p were compared with those of the S. paucimobilis inoculation/DW-treated vehicle control group within the same storage day using Student's t-test for planned pairwise comparisons. Asterisks indicate significant differences compared with the vehicle control group within the same day (∗, p < 0.05; ∗∗, p < 0.01). All values are expressed as mean ± standard deviation. Here, DW was used as the solvent for PNA treatment, and the same volume was added to this group as a vehicle control. In particular, the DW-treated group exhibited the greatest increase, from 6.38 ± 5.14 on day 1 to 19.74 ± 3.39 on day 6, consistent with the highest bacterial counts (p < 0.05). This indicated that moisture promoted bacterial proliferation through enhanced surface water film formation and nutrient diffusion. In contrast, the (KFF)3K-anti-gyrA PNA and (KFF)3K-anti-rpoA PNA treatment groups exhibited significantly lower ΔE values (6.17 ± 6.23 and 8.77 ± 3.28, respectively) compared with the S. paucimobilis inoculation of the DW-treated group (vehicle control, 18.67 ± 3.15) (p < 0.05 and p < 0.01). Overall, ΔE remained lower in the(KFF)3K-anti-gyrA PNA group from days 2 to 7 and in the (KFF)3K-anti-rpoA PNA group from days 3 to 5, indicating sustained suppression of discoloration during storage. This reduction reflects effective inhibition of gyrA and rpoA expression to suppress bacterial growth and delay surface discoloration. These results demonstrate that antisense PNA can effectively limit S. paucimobilis proliferation and slow quality deterioration under high-moisture conditions, thereby supporting its potential as a targeted control strategy. Given the lack of identical target sequences among non-target taxa detected in our datasets, these effects are expected to be sequence-dependent; however, broader selectivity across co-resident microbiota during storage should be validated in future studies. Regarding the uptake mechanism, the (KFF)3K efficiently delivered anti-gyrA PNA to S. paucimobilis. This PCC consists of repeating lysine-derived cationic and phenylalanine-driven hydrophobic motifs, and internalization via peptide transporters such as SbmA has been reported in some bacteria (Ghosal et al., 2013). However, plant tissues possess a cuticle and a pectin-rich, negatively charged cell wall that forms an extracellular diffusion barrier, and wall properties such as surface charge, swelling, and pore size can promote nonspecific adsorption of cationic CPPs (Shin et al., 2021). These surface-barrier features suggest that CPP-PNA conjugates may act more selectively on target bacteria attached to the surface of plant-derived foods during production and distribution, rather than penetrating deeply into plant cells. Consistent with this, (KFF)3K-anti-gyrA PNA exhibited potent, sequence-specific antimicrobial activity against S. paucimobilis, indicating potential applicability for extending the shelf life of other plant-derived foods when coupled with appropriate formulation and process parameters (e.g., ionic strength, pH, Ca2+). Overall, CPP-PNAs provide a potential precision strategy by silencing essential genes of target microorganisms via antisense mechanisms, although their broader selectivity within indigenous microbiota requires further validation. Remaining limitations include the synthesis costs, scalability, and the need for robust delivery in complex food environments. Continued advances in oligonucleotide synthesis, CPP engineering, and integration with standard preservation strategies are likely to overcome these barriers, positioning CPP-PNA based interventions as a promising tool for food storage and safety. Conclusion The genus Spingomonas has been identified as one of the dominant genera in the indigenous microbiota of plant-derived foods, and the opportunistic pathogen S. paucimobilis has been detected as a common spoilage microorganism in various plant-derived foods. To control this microorganism, we used peptide nucleic acids (PNAs) conjugated with the peptides (RXR)4XB and (KFF)3K targeting gyrA and rpoA, which are thought to be essential for bacterial growth. The antimicrobial activity of these peptide-conjugated PNAs was evaluated both in vitro and in an infected food model. We provide the first evidence that gyrA and rpoA are potential targets for antisense PNAs and that PNA conjugates with cell-penetrating peptides are highly effective against S. paucimobilis. Additionally, in vitro experiments demonstrated precise control over S. paucimobilis, and favorable spoilage-inhibitory effects were observed in a lettuce model. (KFF)3K-anti-gyrA PNA may provide a strategy to enhance food safety and shelf life by targeting S. paucimobilis in plant-derived foods, although its broader selectivity within indigenous microbiota should be further validated. Declaration of generative AI and AI-assisted technologies in the writing process During the preparation of this work the author(s) used ChatGPT in order to improve the readability and language of the manuscript. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the published article. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.