Next Article in Journal
Precise CRISPR-Mediated Editing of the TGFBI R555W Mutation in Patient-Derived Peripheral Blood Mononuclear Cells
Previous Article in Journal
Extraction, Characterization, and Biological Evaluation of Atranorin Against Diabetes-Induced Reproductive Dysfunction Through Modulation of Oxidative Stress, Inflammatory Pathways and Key Reproductive Enzymes
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
IJMS
Volume 27
Issue 5
10.3390/ijms27052417
ijms-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Combining Yeast Display and Bacterial Genomic Library for the Unbiased Isolation of Novel Polysaccharide-Binding Peptides

by
Angela Stabile
1,†,
Gaia Scaramella
1,†,
Simone Puccio
2,3,
John Brady
4,
Lise Goltermann
5,
Tim Tolker-Nielsen
5,
Barbara Bellich
6,
Simone De Zotti
1,
Cristina Lagatolla
1,
Fortunato Ferrara
7,
Roberto Rizzo
1,
Paola Cescutti
1 and
Daniele Sblattero
1,*
1
Department of Life Sciences, University of Trieste, Via L. Giorgieri 5, 34127 Trieste, Italy
2
Laboratory of Translational Immunology, Istituto di Ricovero e Cura a Carattere Scientifico Humanitas Research Hospital, Via Manzoni 56, 20089 Rozzano, Italy
3
Institute of Genetic and Biomedical Research, National Research Council, Via Manzoni 113, 20089 Rozzano, Italy
4
Food Science Department, Cornell University, 101A Stocking Hall, Ithaca, NY 14853, USA
5
Costerton Biofilm Center, Department of Immunology and Microbiology, University of Copenhagen, Blegdamsvej 3B, DK-2200 Copenhagen, Denmark
6
Institute for Maternal and Child Health, Istituto di Ricovero e Cura a Carattere Scientifico “Burlo Garofolo”, Via dell’Istria 65/1, 34137 Trieste, Italy
7
Specifica Inc., an IQVIA Laboratories Company, Santa Fe, NM 87501, USA
*
Author to whom correspondence should be addressed.
†
These authors contributed equally to this work.
Int. J. Mol. Sci. 2026, 27(5), 2417; https://doi.org/10.3390/ijms27052417
Submission received: 27 January 2026 / Revised: 1 March 2026 / Accepted: 3 March 2026 / Published: 6 March 2026
(This article belongs to the Special Issue Interactions and Functions Between Peptides and Polysaccharides)
Browse Figures
Versions Notes

Abstract

Here, we present a novel yeast surface display-based platform for the discovery of biofilm-associated exopolysaccharide-binding peptides. Unlike conventional synthetic libraries, our approach utilizes the genomic diversity of Burkholderia multivorans strain C1576 through open-reading frame-filtered genomic fragment libraries, thereby enriching for naturally encoded carbohydrate-binding domains. By iterative fluorescence-activated cell sorting, we identified 21 peptides with confirmed binding to two structurally distinct rhamnose-rich polysaccharides: the exopolysaccharide Epol C1576 and the capsular polysaccharide CPS KpB-1. Biophysical characterization revealed that these peptides adopt predominantly α-helical or disordered conformations and undergo structural rearrangements upon polysaccharide binding. Functional assays demonstrated that selected peptides modulate biofilm architecture and bacterial viability in a species-specific manner, although they do not have a direct bactericidal effect against planktonic cells. This proof-of-concept study establishes yeast surface display as a powerful tool for the discovery of biofilm-targeting peptides and provides a basis for development of novel diagnostics and therapeutics to combat biofilm-associated infections.

1. Introduction

The ability to select proteins and peptides that specifically interact with a target of interest has been revolutionized over the past two decades by in vitro display technologies, such as phage display [1] and yeast display [2]. These platforms rely on the construction of libraries of potential binders displayed on the surface of biological entities, followed by controlled selection against a defined target. A key advantage of these systems is the physical linkage between phenotype (the displayed protein or peptide) and genotype (the encoding DNA), which enables straightforward identification, cloning, and downstream characterization of positive binders.
Recent advances have further expanded the capabilities of display technologies. Integration with next-generation sequencing (NGS) [3,4] allows in-depth profiling of library diversity and selection dynamics, while in silico design approaches—including computational modeling and machine learning—facilitate rational optimization of library composition and binder properties. Among the most notable successes of these technologies is the discovery of the therapeutic antibody [5,6,7]. Beyond antibodies, display platforms have enable the selection of diverse proteins and peptides classes [8]. For example, macrocyclic peptides with high affinity and specificity toward a wide range of targets have recently been isolated using these methods [9], underscoring the versatility of in vitro display technologies in ligand discovery.
A major biomedical challenge where such technologies may have significant impact is the targeting of bacterial biofilms. Many microorganisms form biofilms—structured communities of cells embedded in a self-produced matrix known as extracellular polymeric substances (EPSs) [10]. This matrix constitutes the majority of the biofilm volume [11], and typically contains polysaccharides, proteins, nucleic acids, and lipids [12]. Among these components, exopolysaccharides (Epols) are predominant structural elements. Often negatively charged and composed of repeating monosaccharide units [11], Epols promote cell adhesion, structural integrity, and protection against environmental stressors, including dehydration and antibiotics [13]. They may also directly interact with antibiotics reducing their bioavailability [14,15] and contributing to treatment failure. In this context, antimicrobial peptides (AMPs) have emerged as promising alternatives to conventional antibiotics [16,17], sparking interest in the development of dedicated AMP-discovery pipelines. AMPs exhibit broad-spectrum antimicrobial activity and act through multiple mechanisms including inhibition of biofilm formation and disruption of established biofilms [18,19]. Compared to classical antibiotics, AMPs often mediate rapid pathogen clearance and display a lower propensity for resistance development. Additionally, some AMPs act synergistically with antibiotics, enhancing therapeutic efficacy while reducing required dosages and associated toxicity [20]. Particularly promising—yet still underexplored—are AMPs that target carbohydrates, especially biofilm-associated exopolysaccharides [21].
The isolation of non-natural peptides commonly relies on synthetic random libraries; however, such libraries present inherent limitations. The stochastic nature of sequence generation frequently results in a high proportion of non-functional or poorly folded peptides, reducing selection efficiency. The absence of structural bias can hinder the identification of candidate with favorable biophysical properties. Furthermore, the limited chemical diversity of purely random peptides restricts exploration of complex binding interfaces, especially when targeting structurally challenging epitopes. As a result, substantial downstream optimization is often required, increasing time and resources investment. To address these limitations, we explored bacterial genomes as a natural reservoir of genes encoding carbohydrate-binding proteins, including lectins and carbohydrate-binding modules (CBMs) found in carbohydrate-active enzymes (CAZymes) [22]. Many such domains remain poorly annotated, and metagenomic display approaches have demonstrated that previously uncharacterized carbohydrate-binding domains can be successfully identified from genomic sources [23]. We hypothesized that bacterial genomes encode functional carbohydrate-binding domains capable of interacting with biofilm exopolysaccharides.
Based on this rationale we constructed a domain library from the genome of Burkholderia multivorans, a bacterium known to produce clinically relevant biofilms [24]. To enable systematic identification of biologically relevant protein–polysaccharide interactions, we employed the “interactome-seq” pipeline [25,26]—a two-step strategy previously validated in our laboratory. First, we generated a “domainome”—a library enriched in functional protein domains—by fragmenting bacterial genomic DNA and applying selection pressure to enrich for open reading frames encoding properly folded domains. This approach takes advantage of the high coding density of prokaryotic genomes and the modular nature of protein domains. Second, the domainome was screened using yeast surface display to identify functional binders against selected polysaccharide targets.
Protein–carbohydrate interactions are typically characterized by low intrinsic affinity, and multivalency is often required to achieve sufficient binding strength. Yeast surface display enables the presentation of thousands of copies of a protein on the surface of Saccharomyces cerevisiae, thereby enhancing apparent affinity through avidity effects [27]. When combined with fluorescence-activated cell sorting (FACS), this system provides a powerful high-throughput platform for isolating binders against complex carbohydrate structures such as biofilm-associated polysaccharides. We focused on two clinically relevant biofilm-associated polysaccharides. The first, Epol C1576, is produced by Burkholderia multivorans strain C1576—a clinical isolate from a cystic fibrosis patient [28]. The second, KpB-1 capsular polysaccharide (CPS), is produced by Klebsiella pneumoniae strain KpB-1, isolated from a patient with a carbapenem-resistant infection [29].
In this study, we aimed to evaluate the potential of combining our prokaryotic open reading frame (ORF) filtering with yeast display for the unbiased isolation of peptides that specifically recognize biofilm-associated polysaccharides from clinically relevant pathogens. Selections against the two target polysaccharides yielded 21 novel peptides with specific binding activity. These candidates were characterized through three-dimensional structural modeling, circular dichroism spectroscopy, and their binding kinetics. Finally, their antimicrobial and antibiofilm properties were assessed using in vitro assays. Collectively our findings demonstrate that genome-derived domain libraries integrated with yeast display provide an effective strategy for isolating carbohydrate-binding proteins with potential utility as functional agents to inhibit microbial growth and biofilm formation.

2. Results

2.1. Strategy of the Study

Our initial objective was to identify peptides and/or proteins capable of specifically bind the biofilm-associated exopolysaccharides Epol C1576, from B. multivorans C1576 and CPS KpB-1 from K. pneumoniae (Supplementary Figure S1) using an unbiased selection approach. These two polysaccharides were chosen due to their distinctive primary structure, which is rich in rhamnose residues—sugars that are less polar than conventional monosaccharides such as glucose. Both polysaccharides exhibit aggregation in aqueous solution, suggesting a propensity for self-interaction rather than extensive solvation. This behavior is likely due to hydrophobic interactions among rhamnose residues, contributing to structural integrity of the biofilm architecture [30]. As a negative control, we chose dextran, a commercially available polysaccharide, that consists exclusively of glucose units and, unlike Epol C1576 and CPS KpB-1 [30], does not interact with quorum-sensing molecules.
To identify binding peptides, we applied a sequential strategy combining open reading frame filtering and yeast surface display. Given that carbohydrate-binding modules (CBMs) are typically between 30 and 200 amino acids [22,31], we constructed an ORF library with fragments encoding between 50 and 150 aa, derived from B. multivorans genomic DNA. These fragments were subcloned into a yeast display system to enable surface expression and facilitate the selection of peptides that specifically bind the target polysaccharides. Yeast was chosen both for its low sequence bias and capacity for multivalent expression, which helps to overcome the generally weak affinity of individual protein–carbohydrate interactions. Finally, positively selected peptides were synthesized and fully characterized both biochemically and for their in vitro activity (Figure 1).

2.2. Construction of Yeast Display Burkholderia multivorans ORFs Library

To initiate our experimental strategy, we employed a previously established pipeline for the generation and selection of ORF libraries derived from whole bacterial genomes [32]. This approach integrates an ORF filtering procedure design to enrich for folded protein domains [25], followed by an in vitro display system for the selection of interacting domains [4], and concludes with deep sequencing to identify enriched clones [26]. Genomic DNA from B. multivorans C1576 was extracted and fragmented into 200–800 bp segments. The fragments were blunt-ended and cloned into our pFILTER vector, electroporated into E. coli cells and plated on chloramphenicol, resulting in a library with a diversity exceeding 2 × 106 clones.
ORF filtering was performed by plating on ampicillin-containing plates, resulting in approximately 1 × 105 colonies. To assess library diversity and the efficacy of the filtering process, both the original and the filtered libraries were analyzed by deep sequencing [26], yielding 648,000 and 526,000 sequences, respectively. The insert size in both libraries was in the lower range of the starting fragments with an average of approximately 175 bp. Based on the 6259 annotated coding sequences (CDSs) in B. multivorans C1576, we found that ~64% of CDSs were represented in the library at a conservative threshold of ≥10 read, increasing to ~ 87% at ≥1 read. Notably, post-filtering analyses revealed that the library was not limited to known CDSs, but it also contained ORF fragments mapped to genomic regions that were not yet annotated as coding, suggesting the potential for discovery of novel domains.

2.3. Construction and Selection of Yeast Display Peptide Library Against Epol C1576 and CPS KpB-1

Following validation of the diversity of the filtered ORF library, we constructed a yeast surface display library. DNA inserts were excised from the filtered ORF pool, recloned into a yeast display plasmid, and transformed into yeast, resulting in a library of approximately 1 × 105 clones. PCR analysis of randomly selected clones confirmed an average insert size of 150–200 bp. To validate the ORF filtering procedure of the cloned fragments, we measured the surface expression of the V5 tag on the yeast surface, which is possible only in the presence of a cloned fragment coding for an ORF. As shown in Figure 2C, in the unselected yeast library, approximately 50% of the yeast cells express a stable peptide on the surface. This percentage is lower than expected but is consistent with the fact that the use of a yeast display system introduces much stronger quality control, as the cloned fragment must not only be an ORF but also be efficiently secreted to be displayed on the yeast surface.
The library was then subjected to iterative selection cycles to enrich for peptides capable of binding biofilm-associated polysaccharides. Selection was performed via fluorescence-activated cell sorting (FACS), which require biotinylated target molecules. To this end, purified EpoI C1576 and CPS KpB-1 were biotinylated, and successful modification was confirmed by 1H-NMR spectroscopy and dot-blot analysis with streptavidin detection (Figure 2A).
In each selection round, yeast cells were incubated with biotinylated EpoI C1576 or CPS KpB-1, and binding was assessed by flow cytometry as shown in Figure 2B,D. Sorting gates were defined (red boxes in Figure 2B,D), and yeast cells within these gates were sorted, collected, and expanded for subsequent rounds. The selection process was repeated three times and data are summarized in Figure 2C. Over the course of the selection rounds, the proportion of V5-positive cells—indicating successful surface expression—increased significantly from 45% to 83–86% for both targets. The MFI signal, indicating successful target binding, also increased during the rounds. The percentage of double-positive clones increased from 1.3% to 46.2% for Epol C1576 and from 1% to 55.3% for CPS KpB-1, suggesting successful enrichment of clones with both high expression and specific binding to the target polysaccharides.

2.4. Identification of Polysaccharide Binding Peptides

At the end of the selection process, individual yeast colonies were screened to identify clones expressing peptides with specific binding to biofilm-associated polysaccharides. Twenty clones were randomly selected from the EpoI C1576 selection (referred to A clones), and forty from the CPS KpB-1 selection (referred to B clones). Each clone was analyzed by flow cytometry to assess binding specificity towards the two target polysaccharides (EpoI C1576 and CPS KpB-1) and the control polysaccharide dextran. Representative binding profiles for one clone from each selection are shown in Figure 3A,C. All twenty A clones exhibited specific binding to Epol C1576, while 34 out of 40 of the B clones showed specific binding to CPS KpB-1. Importantly, none of the clones demonstrated significant binding to dextran, confirming target specificity.
From the initial pool of 54 positive clones, 21 were selected based on their highest binding signal to the target polysaccharide and lowest signal to dextran: 11 from the Epol C1576 selection (Supplementary Figure S2) and 10 from the CPS KpB-1 selection (Supplementary Figure S3). The Epol C1576-derived clones (Figure 3B) consistently showed strong binding to the target polysaccharide, with negligible interaction with dextran. In contrast, CPS KpB-1-derived clones (Figure 3D) showed greater variability in binding strength but maintained high specificity with minimal cross-reactivity to Epol C1576 or dextran.
All 21 selected clones were sequenced, confirming that all DNA inserts indeed corresponded to valid ORFs. The resulting peptide sequences are listed in Table 1.
Sequence homology analysis revealed that only three peptides matched annotated proteins (Supplementary Table S1), while the remaining 18 showed no recognizable homology to known CDSs. Domain analysis using the InterPro server [33] identified conserved domains in two of the three annotated peptides, but none of these were associated with known carbohydrate-binding proteins. The remaining peptides were classified as “consensus disorder predictions” (Supplementary Table S2). To further investigate structural properties, all 21 peptide sequences were submitted to AlphaFold3 [34] for structure prediction and to AIUPred for the prediction of protein disorder [35]. The resulting 3D models (Supplementary Figure S4) predominantly exhibited alpha-helical or disordered conformations, with low PlDDT scores indicating limited structural confidence. This lack of a clearly defined structure was confirmed by the AIU prediction (Supplementary Figure S5) with 18 out of 21 clones having an average of prediction profiles values largely higher that the cutoff (>0.5) that indicate disordered regions. This lack of defined structural motifs is consistent with the absence of homology to known carbohydrate-binding domains. Overall, these results confirm the successful identification of novel peptides capable of specifically recognizing and binding biofilm-associated exopolysaccharides, despite the lack of canonical carbohydrate-binding features.

2.5. Peptide Sequence Analysis

Thousands of peptides, both naturally occurring and synthetically engineered, have been reported to exhibit antimicrobial activity or specific binding to bacterial biofilms. Given that our peptide sequences were identified through an unbiased in vitro selection process, we sought to compare their composition and chemical properties with those of known antimicrobial peptides (AMPs) described in the literature.
To establish a reference dataset, we extracted AMP sequences from three publicly available databases: DBAASP [36], APD3 [37], and DRAMP4.0 [38]. The collected sequences were filtered based using the following criteria: (a) natural origin, (b) minimum length of four amino acids, and (c) inclusion of only standard amino acids. After filtering and removing duplicates, we identified 7042 unique AMP sequences shared across the databases (Figure 4A).
First, we compared the length distribution of the selected peptides with that in the reference databases. As shown in Figure 4B, the average length of our peptides (46 amino acids) was consistent with the average length distribution of the ORFs in the library but was significantly longer than that of the AMPs (35 amino acid). An even more pronounced difference was observed in the distribution of isoelectric point (pI) (Figure 4C), with all but two (B2 and B4) of our peptides having a pI above 11, while only 24% of the reference AMPs exceeded a pI of 10.
Next, we analyzed the relative amino acid composition of the two sets of peptides. Figure 4D,E show the individual amino acid distributions for our peptides and the reference AMPs, respectively, while Figure 4F summarizes the comparative differences. Among the charged residues, amino acids were similarly represented in both groups; however, notable differences were observed among the basic residues. Histidine levels were comparable, but arginine was significantly overrepresented, and lysine underrepresented in our peptide set. Non-polar amino acids were generally less abundant in our group, with particularly marked reduction in leucine and glycine. Among the polar residues, serine was the most abundant in our dataset.
Taken together, these analyses reveal that the peptides identified in our study possess distinct composition and physicochemical properties compared to previously characterized AMPs. These differences suggest that our peptides may represent a new class of biofilm-targeting molecules with unique structural and functional properties.

2.6. Peptide Structural Analysis by Circular Dichroism Spectroscopy

To evaluate the behavior of the selected peptides independently of their display in the yeast system, we synthesized them for in vitro testing. Fourteen peptides were chosen from the original pool of 21, based on their binding reactivity, and used for all subsequent analysis. We first employed circular dichroism spectroscopy (CD) to assess the secondary structure of the peptides and to detect conformational changes upon interaction with polysaccharides. CD spectra were recorded for each peptide in 10 mM phosphate buffer, both alone and in the presence of one of the three polysaccharides: C1576, KpB-1, and dextran. The spectra of the individual peptides are shown in Figure 5.
All peptides showed a characteristic negative peak between 200 and 210 nm, indicating a random coil structure, consistent with the computational predictions. Upon addition of dextran, most peptides showed minimal spectral changes, suggesting limited or no interaction in contrast, the presence of the target polysaccharides led to a remarkable change in CD spectra for nearly all peptides—except A1 and A4—which was related to a decrease in the intensity of the 200 nm CD band.
Peptides A10 and B8 displayed particularly pronounced spectral shifts. For A10, the interaction with CPS KpB-1 resulted in the disappearance of the 200 nm signal and the appearance of a new band near 220 nm, suggesting a transition to a more ordered secondary structure. Peptide B8 exhibited a shift of the 200 nm band to higher wavelengths upon binding to both Epol C1576 and CPS KpB-1, indicative of a conformational change, possibly towards a helical structure. CD experiments showed that in the absence of polysaccharides, the peptides presented a not-ordered probably quite elongated structure as suggested by AIUPred analysis (Supplementary Figure S5), likely due to the high number of positive charges which produced intra-chain repulsion. The peptide–polysaccharide interactions may occur with the formation of hydrogen bonding, hydrophobic interaction among rhamnose residues, and non-polar amino acids and, in the case of KpB-1, also through ionic linkages. These interactions may overcome the intrachain electrostatic repulsion in peptides thus driving the formation of an ordered secondary structure. The formation of peptides’ secondary structure in the presence of polysaccharides has already been reported for LL-37 antimicrobial peptide interacting with alginate [39]. Considering the type of bonds involved in the peptide–polysaccharide interactions, it is not surprising that the yeast display platform selected peptides rich in Arg and Ser. In fact, Arg residues contain the guanidinium group which can interact through its positive charge as well as hydrogen bonds; the hydroxyl function of Ser residues is a hydrogen bond forming group
In summary, the observed changes in CD spectra upon polysaccharide binding provide evidence of peptide–polysaccharide interactions. Peptides A10 and B8 demonstrated the most significant structural changes. Notably, CPS KpB-1 contains negatively charged moieties, which could facilitate electrostatic interactions with the highly basic peptides identified in our selection.

2.7. Evaluation of Peptide Binding Specificity via Biolayer Interferometry

Biolayer interferometry (BLI) analysis was performed to further validate the peptide–polysaccharide interaction and to assess binding specificity. Biotinylated peptides were immobilized on a streptavidin-coated biosensor, and the binding signals were measured in the presence of dextran (negative control) and the two target polysaccharides. As shown in Figure 6, all peptides exhibited low or negligible binding to dextran, comparable to the buffer-only control. The only exception was peptide B5 which showed a modest signal with dextran, at the same level as the negative control but still significantly lower than its binding to the target polysaccharides C1576 and KpB1. These results are consistent with the CD data and confirm the specificity of the selected peptides.
All class A peptides bound to Epol C1576, the polysaccharide used for their selection. Although binding kinetics varied slightly among clones, the maximum signal was comparable across the group, except for the A10 clone, which showed a slightly reduced response. When tested against CPS KpB-1, two A-class peptides (A5 and A6) exhibited similar binding kinetics and signal intensity, while the remaining clones showed weaker but detectable binding.
The B-class peptides displayed more diverse binding reactivities. While peptides B5 and B6 showed strong binding to CPS KpB-1, consistent with their selection origin, peptide B8 demonstrated preferential binding to Epol C1576. Interestingly peptide B4, the only peptide with a relatively low pI, did not show specific binding to any polysaccharides tested. Notably, regardless of the signal reached many of the peptides exhibit a two-phase dissociation, with a first, faster phase followed by a much slower second phase, which may suggest some degree of stickiness of the polysaccharides to the peptide-coated surface.
In summary, BLI confirmed that all tested peptides selectively bind their respective target polysaccharide, with minimal interaction with the control polysaccharide dextran, further confirming the success of the selection strategy.

2.8. Evaluation of Antimicrobial and Antibiofilm Activity

To evaluate the functional relevance of the selected peptides, we investigated their antimicrobial and antibiofilm activities against the bacterial strains producing the target polysaccharides. The antimicrobial activity was tested against Burkholderia multivorans strain C1576 and Klebsiella pneumoniae strain KpB-1. As shown in Supplementary Table S3, none of the peptides exhibited antimicrobial activity at the highest tested concentration of 32 µM.
Antibiofilm activity was subsequently assessed using B. multivorans C1576 and K. pneumoniae strain KMn-7, which was used in place of KpB-1. Although both strains produce the same capsular polysaccharide, KMn-7 was selected because it yields more reproducible results due to its consistent biofilm formation. Cultures were grown in microtiter trays with or without peptides, and biofilm biomass was quantified via crystal violet staining. As shown in Figure 7A,B, the peptides showed different effects on biofilm formation. In B. multivorans C1576, peptides A3 and B6 enhanced biofilm formation by up to 150% compared to the untreated control, while peptides A2 and B5 reduced it to 52% and 38%, respectively. In K. pneumoniae KMn-7, most peptides reduced biofilm biomass, with A9 and A2 being the most effective, reducing biofilm formation to 55% and 32%, respectively. Since the crystal violet assay does not provide information about the viability of the biofilm, to further investigate these effects, confocal laser scanning microscopy was used to visualize biofilm formed in the presence of peptides. Peptide A2, which reduced biofilm formation in the crystal violet assay, and peptide A3, which appeared to increase it, were chosen for detailed analysis.
All biofilms were stained with a live/dead stain to assess cell viability as shown in Figure 8. Consistent with the crystal violet assay results, peptide A2 reduced the amount of biofilm attached to the well surface for both species. For B. multivorans, the A2 increased the proportion of dead cells in the biofilm, while in K. pneumoniae, it increased the proportion of filamentous bacteria in the biofilm, Bacterial filamentation is a known response to stressors such as antibiotic treatment at subinhibitory concentrations [40]. Incubation with peptide A3 resulted in a high proportion of biofilm cells being stained as dead or dying for both bacterial species. The dead cells appeared to remain attached to the surface, suggesting that the increased crystal violet signal does not reflect an actual increase in viable biomass, but rather the absence of detachment even after the biofilm cells have been killed.
To exclude general growth inhibition as a confounding factor, planktonic growth rates were measured in the presence of peptides A2 or A3 at the same concentrations used in the biofilm assays (Supplementary Figure S6). Although all cultures grew to saturation, some differences were observed in the maximum growth rate measured during the early exponential phase. Peptide A3 had some effect (however not statistically significant) on the growth rate of B. multivorans C1576 possibly caused by the fact that the peptide formed small precipitates when added to the growth medium. The growth rate of either strain was not affected by peptide A2.

3. Discussion

In this study, we present a novel strategy to identify peptides that selectively bind to extracellular polysaccharides within the biofilm matrix. Our approach uniquely integrates genomic open reading frame (ORF) filtering with yeast surface display technology, enabling the targeted exploration of naturally encoded carbohydrate-binding domains. In contrast to conventional peptide libraries, which are often based on synthetic or combinatorial sequences, our method utilizes the genomic diversity of Burkholderia multivorans strain C1576—a clinically relevant biofilm-forming bacterium—by constructing a library of randomly fragmented genomic DNA, which is subsequently filtered to obtain only ORF-containing clones.
The rationale for using genomic material from B. multivorans lies in the hypothesis that bacterial genomes harbor proteins with intrinsic carbohydrate-binding capabilities, that may have evolved for biofilm formation and maintenance [41]. By utilizing this natural repertoire, our approach increases the likelihood of identifying biologically relevant interactions with biofilm polysaccharides.
The resulting ORF-derived peptide library was screened against two distinct polysaccharide targets through three iterative rounds of fluorescence-activated cell sorting (FACS). This enrichment process yielded a focused set of 21 novel peptides with confirmed polysaccharide-binding activity. The success of this strategy underscores its potential as a powerful platform for discovering biofilm-targeting ligands, with implications for diagnostics, therapeutics, and biofilm biology research.
Biofilms have garnered significant attention in the context of antibiotic resistance due to their exceptional resilience to environmental stressors and antimicrobial agents, as well as their capacity to colonize a wide range of surfaces. Within this framework, antimicrobial peptides (AMPs) have emerged as promising alternative therapeutics. Despite extensive research on biofilm biology, the molecular interactions between proteins and the polysaccharide components of the biofilm matrix remain poorly characterized. This gap highlights the need for novel approaches to identify specific binders capable of targeting these carbohydrate-rich structures.
In this proof-of-concept study, we validate our peptide discovery pipeline using two structurally distinct polysaccharides: Epol C1576 and CPS KpB-1. Epol C1576, produced by Burkholderia multivorans, consists of linear tetrasaccharide repeats composed of equimolar amounts of D-mannose and D-rhamnose, with a methyl substituent on carbon 3 of approximately half of the two-linked rhamnose residues [42]. The low polarity segments formed by rhamnose residues are hypothesized to contribute to intramolecular tethering within the biofilm matrix, enhancing its structural stability.
The second target, CPS KpB-1, is a capsular polysaccharide produced by Klebsiella pneumoniae strain KpB-1, a clinical isolate that causes carbapenem-resistant infections [29]. CPS KpB-1 is composed of linear pentasaccharide units containing L-rhamnose, D-galacturonic acid, and D-galactose, with a single L-rhamnose branching residue [43]. We have selected these two Epols as they represent an interesting target, since they are rich in rhamnose residues. The methyl group on carbon 6 of rhamnose units lowers the polymer hydrophilicity and can form low polarity regions on the polysaccharide chains. These regions promote chain–chain interactions that contribute to the biofilm matrix stability but may also act as binding sites for low-polarity molecules, aiding their mobility through the hydrated matrix. In fact, both Epol C1576 and CPS KpB-1 have been shown to interact with Quorum Sensing components [30], suggesting a potential role in the sequestration or delivery of signaling molecules within the biofilm matrix. To target biofilm-associated polysaccharides, we employed a hybrid selection strategy combining solution-phase binding with multiple rounds of fluorescence-activated cell sorting (FACS), facilitated by biotinylated antigens. This approach enabled the enrichment of specific binders from a yeast display peptide library. Yeast display was found to be particularly advantageous for polysaccharide targets due to the high density of peptides displayed, which enhances binding through avidity effects, an aspect that can be particularly critical when interactions between individual peptides and polysaccharides are inherently weak. This property is particularly important in early rounds of selection, where low abundance clones might otherwise be overlooked.
The selection protocol offers substantial flexibility for optimization, particularly in tuning specificity and affinity. Beyond adjusting target concentration to influence binding kinetics, environmental parameters such as temperature, pH, ionic strength, and buffer composition can significantly affect peptide–polysaccharide interactions and should be carefully considered in future iterations.
To characterize the selected peptides, we compared them to a curated reference dataset compiled from three major antimicrobial peptide databases: DRAMP, DBAASP, and ADP3. The selected peptides exhibited distinct physicochemical profiles, including reduced hydrophobicity and lysine content, and increased alanine and arginine representation. These compositional differences suggest a unique binding repertoire tailored to polysaccharide recognition.
Structural modeling using AlphaFold3 predicted that most peptides adopt either alpha-helical or disordered conformations. Fourteen of the twenty-one peptides were synthesized for further characterization. Biolayer interferometry confirmed binding to the target polysaccharides for all but one peptide. Circular dichroism spectroscopy validated their structural features and revealed conformational changes upon polysaccharide interaction.
We next assessed the antibacterial properties of the peptides. Minimum inhibitory concentration (MIC) assays showed no activity against planktonic bacteria at 32 µM, suggesting that their primary mode of action may involve biofilm modulation rather than direct antimicrobial effects. Crystal violet staining revealed diverse impacts on biofilm formation: while some peptides had no effect, others—such as peptide A2—significantly reduced biofilm formation in both Klebsiella and Burkholderia. Conversely, certain peptides appeared to enhance biofilm formation, potentially due to stress-induced biofilm responses.
To better understand these effects, we employed confocal laser scanning microscopy (CLSM) with live/dead staining to assess structure and cell viability within biofilms. Peptides A2 and A3 were selected for this analysis. Peptide A2 reduced biofilm biomass, particularly in B. multivorans, and induced filamentation in K. pneumoniae, suggesting biological activity. Peptide A3 increased the proportion of dead cells in both species’ biofilms, despite an apparent increase in biomass by crystal violet staining, likely due to retention of stain by dead cells and cellular debris. Neither peptide inhibited planktonic growth at biofilm assay concentrations, although A3 modestly affected B. multivorans growth rate, possibly due to peptide aggregation in culture medium. A2 showed no detectable impact on planktonic growth. These findings underscore the importance of microscopy-based analysis to accurately interpret peptide effects on biofilm structure and cell viability.
Biofilm matrix components can in some cases sequester molecules such as antibiotics and thereby reduce the concentration of the antibiotics around the bacteria leading to antibiotic tolerance [44]. However, recent studies suggest that in some cases biofilm matrix components may attract antibiotics into the biofilm interior, effectively increasing their local concentration near the bacterial cells [45]. A similar mechanism might explain why the peptides (at a concentration of 32 μM) did not kill planktonic bacteria whereas they did kill a fraction of the bacteria in the biofilms. Unlike the planktonic bacteria, the bacteria in biofilms synthesize exopolysaccharides that may attract peptides so that the local peptide concentration around the biofilm bacteria is higher and more lethal than that experienced by the planktonic bacteria.
Yeast surface display has proven to be a highly effective platform for the discovery and characterization of peptides that bind biofilm-associated polysaccharides. Beyond initial discovery, yeast display offers a versatile platform for downstream optimization as it can be integrated with other in vitro-directed evolution technologies. Peptides once isolated could undergo affinity maturation, increased stability, and improved epitope specificity by computationally designing a derivative library in silico and applying advanced screening strategies. Our yeast display approach was able to select functional peptides influencing biofilm formation and bacterial viability in complex ways. Confocal microscopy with live/dead staining revealed nuanced effects not captured by bulk assays like crystal violet staining, underscoring the importance of high-resolution analysis in biofilm research. Importantly, peptides that bind biofilm exopolysaccharides may disrupt polysaccharide–antibiotic interactions, potentially lowering the antibiotic concentration required for biofilm eradication. This synergistic potential warrants further investigation, particularly in the context of combination therapies. The human microbiome metagenome constitutes an immense and still largely unexplored reservoir of genetic diversity, encoding novel sequence classes with biochemical and functional properties. Among these, enzymes and carbohydrate-binding proteins represent particularly promising targets due to their central roles in microbial metabolism, host–microbe interactions, and biotechnological applications. Although functional metagenomic approaches have successfully identified a wide range of enzymatic activities, including oxidoreductases and flavonoid-modifying enzymes [46,47], as well as polysaccharide-degrading systems and carbohydrate-binding modules (CBMs) [48,49], a substantial portion of the functional landscape encoded within microbial communities remains uncharacterized. In this context, our discovery pipeline—combining ORF filtering strategies with phage display-based functional selection—proved effective in accessing previously hidden protein diversity. The identification of a novel family of β-helix proteins with cellulose- and/or acid-sugar-binding properties highlights the power of integrating sequence-based preselection with functional screening [32,50]. Carbohydrate-binding modules are of particular interest due to their specificity and modularity. Their capacity to discriminate among structurally diverse glycans positions them as versatile molecular tools. Fluorophore-labeled CBMs, for instance, could selectively stain biofilm matrices, enabling high-resolution confocal imaging and providing insights into the spatial organization of microbial communities and host–microorganism interactions. More importantly, the selective isolation of peptides capable of recognizing only a specific polysaccharide opens new perspectives for diagnostics. Such molecules could allow not only species-level pathogen identification but potentially strain-level discrimination, thereby enhancing the precision of microbial detection strategies. In this framework, yeast display represents a highly adaptable platform for the functional characterization and affinity maturation of carbohydrate-binding ligands. Its quantitative screening capabilities and compatibility with complex libraries make it particularly suitable for exploring metagenome-derived diversity. By enabling the systematic refinement of binding specificity and affinity, yeast display can accelerate the translation of naturally occurring ligands into diagnostic or therapeutic tools.

4. Materials and Methods

4.1. Bacterial Strains

Burkholderia multivorans C1576 (LMG 16660) is a reference strain from the panel of Burkholderia cepacia Complex strains [28]. and it was purchased from BCCM™ bacteria collection (Dept. of Biochemistry and Microbiology, Faculty of Sciences of Ghent University, Belgium). B. multivorans C1576 has been used for ORF library construction, to produce the polysaccharide Epol C1576 and for in vitro tests. Klebsiella pneumoniae strain KpB-1, isolated from a patient with a carbapenem-resistant infection [29] was used to produce the capsular polysaccharide (CPS) KpB-1. Klebsiella pneumoniae strain KMn-7, a clinical strain isolated from an inpatient in an Italian hospital [43], was used for in vitro tests. Escherichia coli DH5αF’ (Gibco BRL), F’/endA1 hsd17 (rK− mK+) supE44 thi-1recA1 gyrA (Nalr) relA1 (lacZYA-argF) U169 deoR [F80dlacD-(lacZ)M15], was used for library preparation.

4.2. Purification, Depolymerization, and Biotinylation of Polysaccharides

The Epol C1576 was produced and purified as previously described [30]. The purified Epol C1576 was depolymerized upon treatment with 0.5 M TFA for 2 h at 80 °C; after cooling, the solution was taken to dryness under reduced pressure by rotoevaporation and washed four times with methanol. The dried sample was dissolved in water, taken to pH 6.8 and dialysed (cut-off 1 kDa). The degraded Epol C1576 was separated by medium pressure size exclusion chromatography using a Sephacryl S-300 HR column (GE Healthcare, Uppsala, Sweden). Elution was carried out with 0.15 M NaCl using a flow rate of 1 mL/min and monitored with a refractive index detector (Smartline 2300, Knauer, Lab-Service Analitica, Berlin, Germany) which was interfaced with a computer via Clarity software v.10.1.1. Fractions were collected at 30 s intervals. The Sephacryl S-300 HR column was calibrated with dextran standards and the fractions corresponding to about 10 kDa were pooled together, dialyzed, and freeze-dried. The CPS KpB-1 was obtained and purified as previously described [30]. The purified CPS KpB-1 (1 g/L solution) was depolymerized by sonication using a Branson sonifier equipped with a microtip at 2.8 Å. The sample was cooled in an ice bath, sonicated using 10 bursts of 1 min each separated by 1 min intervals, and recovered by lyophilization. HPSEC analysis of the degraded polysaccharides was performed on an Agilent Technologies 1200 series HPLC (Agilent Technologies, Inc., Santa Clara, CA, USA) equipped with three TSK gel columns (Tosoh Bioscience, Tokyo, Japan) in series: G3000PW (<50,000 Da), G5000PW (<1 × 106 Da) and G6000PW (<8 × 106 Da), with internal diameter of 7.5 mm, length of 30 cm and kept at 40 °C. Elution was performed with 0.15 M NaCl, using a flow rate of 0.5 mL/min and monitored using a refractive index detector. HPSEC analysis established the CPS KpB-1 to be about 2.6 × 105 Da. Dextran 12 kDa (Fluka, version 4-5.1), degraded Epol C1576 and CPS KpB-1 were used for the biotinylation reaction, after verifying their purity by 1H NMR spectroscopy. Before biotinylation, the polysaccharides were partially oxidized with NaIO4 to introduce aldehyde functions necessary for the subsequent reaction. After dissolving 8 mg of Dextran (12 kDa, Fluka) 10.4 mg of CPS KpB-1, and 7 mg of Epol C1576 in 0.1 M pH = 5 sodium acetate buffer, a volume of 2.45 × 10−2 M NaIO4 was added in order to have a final concentration of 7.5 × 10−3 M for dextran, 7.7 × 10−4 M for CPS KpB-1 and 4.3 × 10−3 M for Epol C1576. The final polysaccharides concentration was 10 mg/mL. After stirring the solutions for 2 h at 25 °C in the dark, excess of ethylen glycol was added to stop the reactions, followed by dialysis. The oxidized polysaccharides were biotinylated with EZ-Link™ Biotin-LC-Hydrazide (MW = 371.50 Da, Thermo Fisher Scientific, Waltham, MA, USA) and NaBH3CN, following the protocol by Grün et al. [51]. To verify the success of the reaction, the biotinylated polysaccharides were analyzed by 1H NMR spectroscopy.

4.3. ORFeome Peptide Library Construction and Sequencing

B. multivorans ORFeome-library was prepared as previously described [25,32]. Briefly, after extraction of the genomic DNA, it was fragmented by ultra-sonication to obtain fragments between 200 and 800 bp in length. The fragments were then gel purified, blunted, and phosphorylated. Finally, the genomic fragments were ligated into the previously blunt-cut PFILTER3 vector. The ligation was transformed into DH5aF’ cells by electroporation. After 1 h at 37 °C, the bacteria were plated on 2 × TY agar plates containing 34 μg/mL chloramphenicol. The next day, colonies were collected and replated on 2xTY agar plates supplemented with 25 μg/mL ampicillin (selective marker for ORFs) and grown at 37 °C for 16 h. The size of the cloned fragments was determined by PCR on random colonies, and finally both libraries were analyzed by deep sequencing [26].

4.4. Yeast Display Library Construction and Selection

The DNA fragments were excised from the pFILTER vector using the restriction enzymes BssHII and NheI and cloned into the PDNL6 vector using the same restriction enzymes. The ligation was transferred into the yeast cell by electroporation. The positive clones present in the library were selected according to the standard procedure. The yeast cells were cultured in a selective medium SDCAA for 24 h at 30 °C. Subsequently, the cells were diluted in SGRCAA medium with an optical density of 0.5, which induced the expression of the peptide on the yeast surface for another 24 h at a temperature of 20 °C. The following day, 2 × 106 cells were centrifuged and resuspended in 200 µL PBS containing 0.5 µg/mL conjugated anti-V5-PE, the target polysaccharides at a concentration of 1 µM for Epol C1576 or 200 nM for kpB1, and 2.5 µg/mL conjugated streptavidin-AlexaFluor488, and incubated on ice in the dark for 1 h, washed with yeast wash buffer (0.5% (w/v) BSA and 2mM EDTA in PBS) and resuspended in PBS to be analyzed in flow cytofluorometry.
Yeast cells showing a positive signal for V5 and binding with the target were sorted by FACS. The sorted cells were cultured at 30 °C for 48 h and then induced as described above for analysis and a new round of sorting. The procedure was repeated three times to obtain the final population.

4.5. Single Clone Characterization

After the third selection cycle, yeast cells were plated on solid SDCAA and incubated at 30 °C to obtain single colonies. A total of 20 random colonies from the Epol C1576 selection and 40 from the CPS KpB-1 selection were picked and grown individually. After induction, each clone was tested in flow cytofluorimetry, for expression of the peptide on the surface with PE-conjugated anti-V5 and for binding to the biotinylated polysaccharides. Epol C1576-derived clones were tested with 1 μM Epol C1576 and 200nM CPS KpB-1 and dextran. Each clone resulting from selection with CPS KpB-1 was tested for binding of all polysaccharides at 200 nM. Binding was detected with streptavidin-AlexaFluor488 conjugated. The MFI values for each clone were measured and plotted and used to compare the binding capacity for the target polysaccharide and for control polysaccharides (dextran). The clones with the highest specific signal were identified. The selected clones were then sequenced and analyzed.

4.6. Comparison of the Yeast Sequences with Online Peptide Databases

Peptide sequences from the APD3 [37], DBAASP [36], and DRAMP4.0 [38] databases were downloaded according to the following filters: (i) sequences of natural origin; (ii) peptide length > 4 aa; (iii) containing only conventional amino acids. As the resulting sequence sets largely overlapped, a unified database was created from which all duplicate sequences were removed. The final sequences, 7042 in total, were then analyzed to determine (i) the length distribution; (ii) the isoelectric point using the Expasy tool Compute pI/Mw; (iii) the net charge at pH 7.4 by the ProtPi Protein Tool; (iv) the relative frequency of each amino acid per peptide and the mean relative frequency of each amino acid across the dataset. The same analysis was performed for the dataset consisting of the newly isolated peptides. The resulting data were used to calculate the differences between the two peptide databases.

4.7. Peptide Synthesis and Circular Dichroism Spectroscopy

Out of the initial 21 clones 14 sequences were selected and ordered at NovoPro Bioscience Inc. (Bellevue, WA, USA) as synthetic peptides. All peptides are >95% pure by HPLC both in the biotinylated (N-terminal) and un biotinylated form. The secondary structure of the peptides was analyzed using circular dichroism (CD) spectroscopy on a Jasco J-810 spectropolarimeter (JASCO Corporation, Tokyo, Japan). Peptide samples at concentrations of 12.5 µM and 25 µM in phosphate buffer (PB) pH 7.4 10 mM were loaded into a quartz cuvette with a 1 cm path length. Spectra were recorded at 25 °C over the 190–260 nm range, with a scanning speed of 20 nm/min, and were obtained by averaging 10 scans. Each peptide was analyzed in separate assays, both alone and in combination with a single polysaccharide (C1576, KpB-1, or dextran) used at a concentration of 300 µM, calculated with respect to the repeating unit. Control spectra of the polysaccharides alone were also recorded, and the corresponding blanks were subtracted. The observed ellipticity values reported on the y-axis were normalized based on the length and the concentration of each tested peptide, and the optical path length.

4.8. Biolayer Interferometry (BLI)

BLI was performed using the Octet® N1 System, Sartorius, Varedo Italy and the biotinylated peptides were immobilized on streptavidin tips. For each peptide, the tip was first rehydrated in sample diluent for 10 min, and then the peptide was immobilized in the following steps: baseline for 30s in sample diluent, association for 120 s with peptides at a concentration of 10 ng/mL, and dissociation for 120 s in sample diluent. The coated tip was then incubated with the polysaccharide at 2000 nM to check the binding capacity. Binding was evaluated according to the following scheme: baseline for 30 s, incubation with the polysaccharide for 120 s, dissociation phase for 60 s in sample diluent and then the bound polysaccharide was removed with 60 s in a solution of 2.5 M NaCl. Finally, the baseline was recovered for 60 s in sample diluent. For each polysaccharide–peptide interaction, the maximum signal was calculated.

4.9. Antimicrobial and Antibiofilm Activity

Each peptide was tested for antimicrobial activity with B. multivorans strain C1576 and K. pneumoniae strain KpB-1. A total of 5 × 105 bacterial cells were incubated with the peptide at a concentration of 32 µM in MHB media in a volume of 100 μL overnight at 37 °C and 90 rpm on a 96-well round-bottom plate in technical triplicate.
Biofilm activity was tested with B. multivorans strain C1576 and K. pneumoniae strain KMn-7. Based on our previous results, in order to have consistent biofilm formation, bacteria were cultured in a 96-well flat bottom plate in static conditions for 24 h for K. pneumoniae and for 48 h for B. multivorans. Bacteria were the removed and the biofilm was washed three times in PBS. The biofilm was then fixed at 60 °C for one hour and stained with crystal violet 2% Hucker (2 mg of Crystal violet, 20 mL Ethyl alcohol, 0.8 mg Ammonium oxalate in water). The biofilm was then washed three times with water. The following day, the crystal violet was dissolved in 100 μL of acetic acid at a concentration of 33% and the optical density was read at 570 nm. The data obtained were normalized in percentage to the control growth and reproduced in biological triplicate.

4.10. Confocal Laser Scanning Microscopy

Cultures of B. multivorans strain C1576 and K. pneumoniae strain KMn-7 were grown overnight in LB medium at 37 °C with shaking. The inoculum was prepared by adjusting the OD600 to 0.1 in MHB-II medium resulting in 108 cfu/mL. The suspension was further diluted in MHB-II to 5 × 105 cfu/mL. A total of 100 μl bacterial suspension was aliquoted into each well of a clear-bottom 96-well plate along with 1.6 μl of 2 mM peptide stock resulting in a final concentration of 32 μM. The plate was incubated at 37 °C, 90 rpm, for 22 h (Klebsiella pneumoniae KMn-7) or 48h (Burkholderia multivorans C1576). Planktonic cultures were removed carefully, and the remaining biofilm in the wells was washed carefully 1–3 times with 100 mL saline. An amount of 100 μL live/dead stain (5 μm SYTOX green + 20 μM PI) was added to each well and after 20–30 min staining, and the biofilms were imaged with a 63× #1.4 oil objective using a Zeiss LSM 880 inverted confocal laser scanning microscope (Zeiss, Oberkochen, Germany). Images were processed with Imaris 10.1 (Bitplane, Zürich, Switzerland) for generating 3D projections.

4.11. Peptide Susceptibility and Growth Rate Testing

MIC values were determined by broth microdilution according to standard protocols. Briefly, 190 μL of bacterial cell culture containing approx 105 CFU/mL was dispensed into a 96-well plate along with 10 μL peptide stock. Incubation and measurements were carried out in a TECAN Sunrise plate reader (Tecan Austria GmbH, Grödig, Austria) at 37 °C every 20 min for a period of 18–24 h with intermediate shake cycles in linear mode.

4.12. Statistic

Flow cytometric analyses were conducted using FlowJo software 1.8.1, BLI data were analyzed with Octet N1 software 1.4.0.13. Statistical analyses were carried out with GraphPad Prism v10.1.0 and Microsoft Excel Office 365, using predefined evaluation protocols. Data are presented as mean ± standard error of the mean (SEM), unless otherwise indicated. Information on the number of biological replicates (n) and the statistical tests employed is provided in the corresponding figure legends. Each experiment was independently repeated at least three times with consistent results.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ijms27052417/s1.

Author Contributions

Conceptualization, D.S. and P.C.; methodology, A.S., G.S. and S.P.; software, S.P.; investigation, A.S., G.S., L.G., B.B., S.D.Z. and C.L.; resources, D.S., P.C. and J.B.; writing—original draft preparation, A.S. and G.S.; writing—review and editing, D.S., L.G., T.T.-N., F.F., J.B., R.R. and P.C.; supervision, D.S. and P.C.; project administration, D.S. and P.C.; funding acquisition, D.S., P.C. and J.B. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by the Department of Food Science at Cornell University and a grant from the National Institutes of Health (USA), (Grant Number GM123283) and a grant from the Novo Nordisk Foundation (NNF22OC0080132).

Data Availability Statement

Data supporting reported results are available from the coordinator of the study, Daniele Sblattero, corresponding author.

Conflicts of Interest

The authors declare no conflicts of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results. Fortunato Ferrara is employed at Specifica an IQVIA Laboratories Company. The authors declare that this research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Correction Statement

This article has been republished with a minor correction to the reference 23. This change does not affect the scientific content of the article.

Abbreviations

The following abbreviations are used in this manuscript:
EPSextracellular polymeric substances
Epolsexopolysaccharides
AMPsantimicrobial peptides
CPScapsular polysaccharide
ORFopen reading frame
FACSfluorescence-activated cell sorting

References

  1. Smith, G.P.; Petrenko, V.A. Phage Display. Chem. Rev. 1997, 97, 391–410. [Google Scholar] [CrossRef]
  2. Kang, B.H.; Lax, B.M.; Wittrup, K.D. Yeast Surface Display for Protein Engineering: Library Generation, Screening, and Affinity Maturation. Methods Mol. Biol. 2022, 2491, 29–62, Correction in Methods Mol. Biol. 2022, 2491, C1. [Google Scholar] [CrossRef]
  3. Mejias-Gomez, O.; Braghetto, M.; Sørensen, M.K.D.; Madsen, A.V.; Guiu, L.S.; Kristensen, P.; Pedersen, L.E.; Goletz, S. Deep Mining of Antibody Phage-Display Selections Using Oxford Nanopore Technologies and Dual Unique Molecular Identifiers. New Biotechnol. 2024, 80, 56–68. [Google Scholar] [CrossRef]
  4. Di Niro, R.; Sulic, A.-M.; Mignone, F.; D’Angelo, S.; Bordoni, R.; Iacono, M.; Marzari, R.; Gaiotto, T.; Lavric, M.; Bradbury, A.R.M.; et al. Rapid Interactome Profiling by Massive Sequencing. Nucleic Acids Res. 2010, 38, e110. [Google Scholar] [CrossRef] [PubMed]
  5. Jahandar-Lashaki, S.; Farajnia, S.; Faraji-Barhagh, A.; Hosseini, Z.; Bakhtiyari, N.; Rahbarnia, L. Phage Display as a Medium for Target Therapy Based Drug Discovery, Review and Update. Mol. Biotechnol. 2025, 67, 2161–2184. [Google Scholar] [CrossRef] [PubMed]
  6. França, R.K.A.; Studart, I.C.; Bezerra, M.R.L.; Pontes, L.Q.; Barbosa, A.M.A.; Brigido, M.M.; Furtado, G.P.; Maranhão, A.Q. Progress on Phage Display Technology: Tailoring Antibodies for Cancer Immunotherapy. Viruses 2023, 15, 1903. [Google Scholar] [CrossRef] [PubMed]
  7. Hutchings, C.J.; Sato, A.K. Phage Display Technology and Its Impact in the Discovery of Novel Protein-Based Drugs. Expert. Opin. Drug Discov. 2024, 19, 887–915. [Google Scholar] [CrossRef]
  8. Jaroszewicz, W.; Morcinek-Orłowska, J.; Pierzynowska, K.; Gaffke, L.; Węgrzyn, G. Phage Display and Other Peptide Display Technologies. FEMS Microbiol. Rev. 2022, 46, fuab052. [Google Scholar] [CrossRef]
  9. Linciano, S.; Mazzocato, Y.; Romanyuk, Z.; Vascon, F.; Farrera-Soler, L.; Will, E.; Xing, Y.; Chen, S.; Kumada, Y.; Simeoni, M.; et al. Screening Macrocyclic Peptide Libraries by Yeast Display Allows Control of Selection Process and Affinity Ranking. Nat. Commun. 2025, 16, 5367. [Google Scholar] [CrossRef]
  10. Branda, S.S.; Vik, S.; Friedman, L.; Kolter, R. Biofilms: The Matrix Revisited. Trends Microbiol. 2005, 13, 20–26. [Google Scholar] [CrossRef]
  11. Allison, D.G. The Biofilm Matrix. Biofouling 2003, 19, 139–150. [Google Scholar] [CrossRef] [PubMed]
  12. Fong, J.N.C.; Yildiz, F.H. Biofilm Matrix Proteins. Microbiol. Spectr. 2015, 3, 201–222. [Google Scholar] [CrossRef]
  13. Limoli, D.H.; Jones, C.J.; Wozniak, D.J. Bacterial Extracellular Polysaccharides in Biofilm Formation and Function. Microbiol. Spectr. 2015, 3, 223–247. [Google Scholar] [CrossRef]
  14. Sutherland, I. Biofilm Exopolysaccharides: A Strong and Sticky Framework. Microbiology 2001, 147, 3–9. [Google Scholar] [CrossRef] [PubMed]
  15. Nichols, W.W.; Dorrington, S.M.; Slack, M.P.; Walmsley, H.L. Inhibition of Tobramycin Diffusion by Binding to Alginate. Antimicrob. Agents Chemother. 1988, 32, 518–523. [Google Scholar] [CrossRef]
  16. Neubauer, D.; Jaśkiewicz, M.; Migoń, D.; Bauer, M.; Sikora, K.; Sikorska, E.; Kamysz, E.; Kamysz, W. Retro Analog Concept: Comparative Study on Physico-Chemical and Biological Properties of Selected Antimicrobial Peptides. Amino Acids 2017, 49, 1755–1771. [Google Scholar] [CrossRef] [PubMed]
  17. Erdem Büyükkiraz, M.; Kesmen, Z. Antimicrobial Peptides (AMPs): A Promising Class of Antimicrobial Compounds. J. Appl. Microbiol. 2022, 132, 1573–1596. [Google Scholar] [CrossRef]
  18. Ulm, H.; Wilmes, M.; Shai, Y.; Sahl, H.-G. Antimicrobial Host Defensins—Specific Antibiotic Activities and Innate Defense Modulation. Front. Immunol. 2012, 3, 249. [Google Scholar] [CrossRef]
  19. Park, C.B.; Kim, H.S.; Kim, S.C. Mechanism of Action of the Antimicrobial Peptide Buforin II: Buforin II Kills Microorganisms by Penetrating the Cell Membrane and Inhibiting Cellular Functions. Biochem. Biophys. Res. Commun. 1998, 244, 253–257. [Google Scholar] [CrossRef]
  20. Pletzer, D.; Coleman, S.R.; Hancock, R.E. Anti-Biofilm Peptides as a New Weapon in Antimicrobial Warfare. Curr. Opin. Microbiol. 2016, 33, 35–40. [Google Scholar] [CrossRef]
  21. De Los Santos, L.; Beckman, R.L.; DeBarro, C.; Keener, J.E.; Torres, M.D.T.; de la Fuente-Nunez, C.; Brodbelt, J.S.; Fleeman, R.M. Polyproline Peptide Targets Klebsiella pneumoniae Polysaccharides to Collapse Biofilms. Cell Rep. Phys. Sci. 2024, 5, 101869. [Google Scholar] [CrossRef]
  22. Lombard, V.; Golaconda Ramulu, H.; Drula, E.; Coutinho, P.M.; Henrissat, B. The Carbohydrate-Active Enzymes Database (CAZy) in 2013. Nucleic Acids Res. 2014, 42, D490–D495. [Google Scholar] [CrossRef]
  23. Akhtar, A.; Lata, M.; Sunsunwal, S.; Yadav, A.; Lnu, K.; Subramanian, S.; Ramya, T.N.C. New Carbohydrate Binding Domains Identified by Phage Display Based Functional Metagenomic Screens of Human Gut Microbiota. Commun. Biol. 2023, 6, 371, Erratum in Commun. Biol. 2024, 7, 1190. [Google Scholar] [CrossRef] [PubMed]
  24. Leitão, J.H.; Feliciano, J.R.; Sousa, S.A.; Pita, T.; Guerreiro, S.I. Burkholderia cepacia Complex Infections Among Cystic Fibrosis Patients: Perspectives and Challenges. In Progress in Understanding Cystic Fibrosis; IntechOpen: London, UK, 2017; ISBN 978-953-51-3292-9. [Google Scholar]
  25. Soluri, M.F.; Puccio, S.; Caredda, G.; Grillo, G.; Licciulli, V.F.; Consiglio, A.; Edomi, P.; Santoro, C.; Sblattero, D.; Peano, C. Interactome-Seq: A Protocol for Domainome Library Construction, Validation and Selection by Phage Display and Next Generation Sequencing. J. Vis. Exp. 2018, 56981. [Google Scholar] [CrossRef]
  26. Puccio, S.; Grillo, G.; Consiglio, A.; Soluri, M.F.; Sblattero, D.; Cotella, D.; Santoro, C.; Liuni, S.; Bellis, G.D.; Lugli, E.; et al. InteractomeSeq: A Web Server for the Identification and Profiling of Domains and Epitopes from Phage Display and next Generation Sequencing Data. Nucleic Acids Res. 2020, 48, W200–W207. [Google Scholar] [CrossRef]
  27. Feldhaus, M.J.; Siegel, R.W. Yeast Display of Antibody Fragments: A Discovery and Characterization Platform. J. Immunol. Methods 2004, 290, 69–80. [Google Scholar] [CrossRef]
  28. Mahenthiralingam, E.; Coenye, T.; Chung, J.W.; Speert, D.P.; Govan, J.R.; Taylor, P.; Vandamme, P. Diagnostically and Experimentally Useful Panel of Strains from the Burkholderia cepacia Complex. J. Clin. Microbiol. 2000, 38, 910–913. [Google Scholar] [CrossRef]
  29. Cannatelli, A.; Di Pilato, V.; Giani, T.; Arena, F.; Ambretti, S.; Gaibani, P.; D’Andrea, M.M.; Rossolini, G.M. In Vivo Evolution to Colistin Resistance by PmrB Sensor Kinase Mutation in KPC-Producing Klebsiella pneumoniae Is Associated with Low-Dosage Colistin Treatment. Antimicrob. Agents Chemother. 2014, 58, 4399–4403. [Google Scholar] [CrossRef]
  30. Bellich, B.; Cacioppo, M.; De Zorzi, R.; Rizzo, R.; Brady, J.W.; Cescutti, P. Interactions of Biofilm Polysaccharides Produced by Human Infective Bacteria with Molecules of the Quorum Sensing System. A Microscopy and NMR Study. Int. J. Biol. Macromol. 2024, 281, 136222. [Google Scholar] [CrossRef] [PubMed]
  31. Shoseyov, O.; Shani, Z.; Levy, I. Carbohydrate Binding Modules: Biochemical Properties and Novel Applications. Microbiol. Mol. Biol. Rev. 2006, 70, 283–295. [Google Scholar] [CrossRef] [PubMed]
  32. D’Angelo, S.; Velappan, N.; Mignone, F.; Santoro, C.; Sblattero, D.; Kiss, C.; Bradbury, A.R. Filtering “Genic” Open Reading Frames from Genomic DNA Samples for Advanced Annotation. BMC Genom. 2011, 12, S5. [Google Scholar] [CrossRef]
  33. Blum, M.; Andreeva, A.; Florentino, L.C.; Chuguransky, S.R.; Grego, T.; Hobbs, E.; Pinto, B.L.; Orr, A.; Paysan-Lafosse, T.; Ponamareva, I.; et al. InterPro: The Protein Sequence Classification Resource in 2025. Nucleic Acids Res. 2025, 53, D444–D456. [Google Scholar] [CrossRef]
  34. Abramson, J.; Adler, J.; Dunger, J.; Evans, R.; Green, T.; Pritzel, A.; Ronneberger, O.; Willmore, L.; Ballard, A.J.; Bambrick, J.; et al. Accurate Structure Prediction of Biomolecular Interactions with AlphaFold 3. Nature 2024, 630, 493–500. [Google Scholar] [CrossRef]
  35. Erdős, G.; Dosztányi, Z. AIUPred: Combining Energy Estimation with Deep Learning for the Enhanced Prediction of Protein Disorder. Nucleic Acids Res. 2024, 52, W176–W181. [Google Scholar] [CrossRef] [PubMed]
  36. Pirtskhalava, M.; Amstrong, A.A.; Grigolava, M.; Chubinidze, M.; Alimbarashvili, E.; Vishnepolsky, B.; Gabrielian, A.; Rosenthal, A.; Hurt, D.E.; Tartakovsky, M. DBAASP v3: Database of Antimicrobial/Cytotoxic Activity and Structure of Peptides as a Resource for Development of New Therapeutics. Nucleic Acids Res. 2021, 49, D288–D297. [Google Scholar] [CrossRef] [PubMed]
  37. Wang, G.; Li, X.; Wang, Z. APD3: The Antimicrobial Peptide Database as a Tool for Research and Education. Nucleic Acids Res. 2016, 44, D1087–D1093. [Google Scholar] [CrossRef]
  38. Ma, T.; Liu, Y.; Yu, B.; Sun, X.; Yao, H.; Hao, C.; Li, J.; Nawaz, M.; Jiang, X.; Lao, X.; et al. DRAMP 4.0: An Open-Access Data Repository Dedicated to the Clinical Translation of Antimicrobial Peptides. Nucleic Acids Res. 2025, 53, D403–D410. [Google Scholar] [CrossRef]
  39. Benincasa, M.; Mattiuzzo, M.; Herasimenka, Y.; Cescutti, P.; Rizzo, R.; Gennaro, R. Activity of Antimicrobial Peptides in the Presence of Polysaccharides Produced by Pulmonary Pathogens. J. Pept. Sci. 2009, 15, 595–600. [Google Scholar] [CrossRef]
  40. Ultee, E.; Ramijan, K.; Dame, R.T.; Briegel, A.; Claessen, D. Chapter Two—Stress-Induced Adaptive Morphogenesis in Bacteria. In Advances in Microbial Physiology; Poole, R.K., Ed.; Academic Press: Cambridge, MA, USA, 2019; Volume 74, pp. 97–141. [Google Scholar]
  41. Passos da Silva, D.; Matwichuk, M.L.; Townsend, D.O.; Reichhardt, C.; Lamba, D.; Wozniak, D.J.; Parsek, M.R. The Pseudomonas Aeruginosa Lectin LecB Binds to the Exopolysaccharide Psl and Stabilizes the Biofilm Matrix. Nat. Commun. 2019, 10, 2183. [Google Scholar] [CrossRef]
  42. Dolfi, S.; Sveronis, A.; Silipo, A.; Rizzo, R.; Cescutti, P. A Novel Rhamno-Mannan Exopolysaccharide Isolated from Biofilms of Burkholderia multivorans C1576. Carbohydr. Res. 2015, 411, 42–48. [Google Scholar] [CrossRef] [PubMed]
  43. Bellich, B.; Lagatolla, C.; Rizzo, R.; D’Andrea, M.M.; Rossolini, G.M.; Cescutti, P. Determination of the Capsular Polysaccharide Structure of the Klebsiella pneumoniae ST512 Representative Strain KPB-1 and Assignments of the Glycosyltransferases Functions. Int. J. Biol. Macromol. 2020, 155, 315–323. [Google Scholar] [CrossRef]
  44. Ciofu, O.; Tolker-Nielsen, T. Tolerance and Resistance of Pseudomonas Aeruginosa Biofilms to Antimicrobial Agents-How P. Aeruginosa Can Escape Antibiotics. Front. Microbiol. 2019, 10, 913. [Google Scholar] [CrossRef]
  45. Osondu-Chuka, G.O.; Schandl, S.; Scheberl, A.; Guillaume, O.; Ovsianikov, A.; Reimhult, E. Beyond the Matrix: Rethinking Antibiotic Tolerance in CF Biofilms Using 3D Models. NPJ Biofilms Microbiomes 2025, 12, 3. [Google Scholar] [CrossRef] [PubMed]
  46. Knietsch, A.; Waschkowitz, T.; Bowien, S.; Henne, A.; Daniel, R. Construction and Screening of Metagenomic Libraries Derived from Enrichment Cultures: Generation of a Gene Bank for Genes Conferring Alcohol Oxidoreductase Activity on Escherichia coli. Appl. Environ. Microbiol. 2003, 69, 1408–1416. [Google Scholar] [CrossRef] [PubMed]
  47. Rabausch, U.; Juergensen, J.; Ilmberger, N.; Böhnke, S.; Fischer, S.; Schubach, B.; Schulte, M.; Streit, W.R. Functional Screening of Metagenome and Genome Libraries for Detection of Novel Flavonoid-Modifying Enzymes. Appl. Env. Microbiol. 2013, 79, 4551–4563. [Google Scholar] [CrossRef] [PubMed]
  48. Tasse, L.; Bercovici, J.; Pizzut-Serin, S.; Robe, P.; Tap, J.; Klopp, C.; Cantarel, B.L.; Coutinho, P.M.; Henrissat, B.; Leclerc, M.; et al. Functional Metagenomics to Mine the Human Gut Microbiome for Dietary Fiber Catabolic Enzymes. Genome Res. 2010, 20, 1605–1612. [Google Scholar] [CrossRef]
  49. Liu, N.; Li, H.; Chevrette, M.G.; Zhang, L.; Cao, L.; Zhou, H.; Zhou, X.; Zhou, Z.; Pope, P.B.; Currie, C.R.; et al. Functional Metagenomics Reveals Abundant Polysaccharide-Degrading Gene Clusters and Cellobiose Utilization Pathways within Gut Microbiota of a Wood-Feeding Higher Termite. ISME J. 2019, 13, 104–117. [Google Scholar] [CrossRef]
  50. Close, D.W.; D’Angelo, S.; Bradbury, A.R.M. A New Family of β-Helix Proteins with Similarities to the Polysaccharide Lyases. Acta Cryst. D 2014, 70, 2583–2592. [Google Scholar] [CrossRef]
  51. Grün, C.H.; van Vliet, S.J.; Schiphorst, W.E.C.M.; Bank, C.M.C.; Meyer, S.; van Die, I.; van Kooyk, Y. One-Step Biotinylation Procedure for Carbohydrates to Study Carbohydrate-Protein Interactions. Anal. Biochem. 2006, 354, 54–63. [Google Scholar] [CrossRef]
Ijms 27 02417 g001
Figure 1. Schematic overview of the selection and validation procedure. STEP 1: An ORF-filtered yeast peptide library was created from B. multivorans genomic DNA. STEP 2: The yeast library was incubated with the biotinylated target polysaccharide, allowing visualization of binding via flow cytometry using streptavidin-conjugated fluorophores. After incubation, cells were analyzed by flow cytometry, and those displaying peptides capable of binding the target were isolated by fluorescence-activated cell sorting (FACS). STEP 3: Individual clones from the selection were produced as peptides and validated by BLI, CD, and in vitro biofilm tests. Created in BioRender Stabile, A. (2026) https://BioRender.com/bcv0w7e (accessed on 2 March 2026).
Figure 1. Schematic overview of the selection and validation procedure. STEP 1: An ORF-filtered yeast peptide library was created from B. multivorans genomic DNA. STEP 2: The yeast library was incubated with the biotinylated target polysaccharide, allowing visualization of binding via flow cytometry using streptavidin-conjugated fluorophores. After incubation, cells were analyzed by flow cytometry, and those displaying peptides capable of binding the target were isolated by fluorescence-activated cell sorting (FACS). STEP 3: Individual clones from the selection were produced as peptides and validated by BLI, CD, and in vitro biofilm tests. Created in BioRender Stabile, A. (2026) https://BioRender.com/bcv0w7e (accessed on 2 March 2026).
Ijms 27 02417 g001
Ijms 27 02417 g002
Figure 2. Selection performed with yeast display technology. (A) Dot-blot assay of the three polysaccharides: 500 ng of dextran, KpB-1, and C1576 were applied to nitrocellulose as an unmodified or biotinylated molecule. Biotinylation was detected by incubation with streptavidin-AP conjugate. (B–D) Selection of yeast expressing peptides that bind to polysaccharides. The yeast library was incubated with an anti-V5 tag-PE conjugate to observe the expression of the peptides and with the Biotinylated target molecules: (B) with Epol C1576 and (D) with CPS KpB1, and binding to the target was detected with Alexa-conjugated streptavidin. The sorting gates are represented in red and were optimized for each of the three selection cycles. (C) Diagram showing the relationship between the round of selection and the percentage of cells displaying a peptide, the V5-positive fraction, or the degree of target binding (Alexa Fluor signal (MFI).
Figure 2. Selection performed with yeast display technology. (A) Dot-blot assay of the three polysaccharides: 500 ng of dextran, KpB-1, and C1576 were applied to nitrocellulose as an unmodified or biotinylated molecule. Biotinylation was detected by incubation with streptavidin-AP conjugate. (B–D) Selection of yeast expressing peptides that bind to polysaccharides. The yeast library was incubated with an anti-V5 tag-PE conjugate to observe the expression of the peptides and with the Biotinylated target molecules: (B) with Epol C1576 and (D) with CPS KpB1, and binding to the target was detected with Alexa-conjugated streptavidin. The sorting gates are represented in red and were optimized for each of the three selection cycles. (C) Diagram showing the relationship between the round of selection and the percentage of cells displaying a peptide, the V5-positive fraction, or the degree of target binding (Alexa Fluor signal (MFI).
Ijms 27 02417 g002
Ijms 27 02417 g003
Figure 3. Cytofluorimetric analysis of single clones. The cytofluorimetric profiles of representative clones are shown for illustration: (A) Clone A3 from the Epol C1576 selection and (C) Clone B7 from the CPS KpB-1 selection. (B) Single clone test results from selection with Epol C1576; (D) Single clone test results from selection with CPS KpB-1. The boxplot shows the median fluorescence intensity (MFI) of each selected clone tested in cytofluorimetric analysis against three polysaccharides: Epol C1576, CPS KpB-1, and dextran; buffer only represents the background signal with no polysaccharide present. t-test was performed, * p < 0.05 vs. Dextran; **** p < 0.0001 vs. Dextran, ns indicate not significant.
Figure 3. Cytofluorimetric analysis of single clones. The cytofluorimetric profiles of representative clones are shown for illustration: (A) Clone A3 from the Epol C1576 selection and (C) Clone B7 from the CPS KpB-1 selection. (B) Single clone test results from selection with Epol C1576; (D) Single clone test results from selection with CPS KpB-1. The boxplot shows the median fluorescence intensity (MFI) of each selected clone tested in cytofluorimetric analysis against three polysaccharides: Epol C1576, CPS KpB-1, and dextran; buffer only represents the background signal with no polysaccharide present. t-test was performed, * p < 0.05 vs. Dextran; **** p < 0.0001 vs. Dextran, ns indicate not significant.
Ijms 27 02417 g003
Ijms 27 02417 g004
Figure 4. Selected peptide comparison with online databases. (A) Scheme of the process applied to recover sequences from the three databases ADP3, DBAASP, and DRAMP. The datasets were filtered to include only peptides longer than four amino acids, natural origin and composed by conventional amino acids. The three databases largely overlap, the replicates were removed obtaining a final dataset containing a total of 7042 unique peptides; (B) length distribution of the database and the selected peptides; (C) distribution of the isoelectric point of the database and the selected peptides; (D) relative abundance of each amino acid in the selected peptides set; (E) relative abundance of each amino acid in the database; (F) difference in relative abundance of each amino acid between the selected peptides and the databases. The positive bars (green) indicate a higher presence in the selected peptides, while the negative bars (red) indicate an underrepresentation in the set compared to the database.
Figure 4. Selected peptide comparison with online databases. (A) Scheme of the process applied to recover sequences from the three databases ADP3, DBAASP, and DRAMP. The datasets were filtered to include only peptides longer than four amino acids, natural origin and composed by conventional amino acids. The three databases largely overlap, the replicates were removed obtaining a final dataset containing a total of 7042 unique peptides; (B) length distribution of the database and the selected peptides; (C) distribution of the isoelectric point of the database and the selected peptides; (D) relative abundance of each amino acid in the selected peptides set; (E) relative abundance of each amino acid in the database; (F) difference in relative abundance of each amino acid between the selected peptides and the databases. The positive bars (green) indicate a higher presence in the selected peptides, while the negative bars (red) indicate an underrepresentation in the set compared to the database.
Ijms 27 02417 g004
Ijms 27 02417 g005
Figure 5. CD spectra of the selected peptides. Spectra were recorded on peptide alone (black line) and in the presence of Epol C1576 (red line), CPS KpB-1 (green line), or dextran (blue line). Each peptide was incubated with 300 µM of the respective polysaccharide, referring to the concentration of the repeating unit. The signal obtained by the polysaccharides alone were subtracted, and the spectra were smoothed by average.
Figure 5. CD spectra of the selected peptides. Spectra were recorded on peptide alone (black line) and in the presence of Epol C1576 (red line), CPS KpB-1 (green line), or dextran (blue line). Each peptide was incubated with 300 µM of the respective polysaccharide, referring to the concentration of the repeating unit. The signal obtained by the polysaccharides alone were subtracted, and the spectra were smoothed by average.
Ijms 27 02417 g005
Ijms 27 02417 g006
Figure 6. Biolayer interferometry of peptides. Each peptide was fixed on a streptavidin tip and then tested on binding buffer (gray), Dextran 2000 nM (blue), C1576 2000nM (red), and KpB-1 2000nM (green). The Y-axes represent the binding signal between the peptides and the polysaccharides, while the X-axes indicate the time. First phase 0–30 s baseline in the buffer, second phase 30–150 s association with the polysaccharide, third phase 150–270 s dissociation in the buffer.
Figure 6. Biolayer interferometry of peptides. Each peptide was fixed on a streptavidin tip and then tested on binding buffer (gray), Dextran 2000 nM (blue), C1576 2000nM (red), and KpB-1 2000nM (green). The Y-axes represent the binding signal between the peptides and the polysaccharides, while the X-axes indicate the time. First phase 0–30 s baseline in the buffer, second phase 30–150 s association with the polysaccharide, third phase 150–270 s dissociation in the buffer.
Ijms 27 02417 g006
Ijms 27 02417 g007
Figure 7. Quantification of biofilm produced from (A) B. multivorans C1576 and (B) K. pneumoniae KpB-1, when incubated with each of the selected peptides, compared to the growth in media. The biofilm was stained with crystal violet and then the optical density at 570 nm was measured to obtain a quantification. The OD570 was normalized to the control growth in the media without peptide, expressed in percentage. One Way ANOVA was performed, * p < 0.05 vs. control; ** p < 0.005 vs. control; *** p < 0.001 vs. control; **** p < 0.0001 vs. control.
Figure 7. Quantification of biofilm produced from (A) B. multivorans C1576 and (B) K. pneumoniae KpB-1, when incubated with each of the selected peptides, compared to the growth in media. The biofilm was stained with crystal violet and then the optical density at 570 nm was measured to obtain a quantification. The OD570 was normalized to the control growth in the media without peptide, expressed in percentage. One Way ANOVA was performed, * p < 0.05 vs. control; ** p < 0.005 vs. control; *** p < 0.001 vs. control; **** p < 0.0001 vs. control.
Ijms 27 02417 g007
Ijms 27 02417 g008
Figure 8. Confocal laser scanning micrographs of live/dead stained biofilms. Living cells stain green while red indicates dying bacteria with compromised membranes or extracellular DNA. Images are presented as 3D projections and single slices at the surface of the substratum. The size bars correspond to 20 µm.
Figure 8. Confocal laser scanning micrographs of live/dead stained biofilms. Living cells stain green while red indicates dying bacteria with compromised membranes or extracellular DNA. Images are presented as 3D projections and single slices at the surface of the substratum. The size bars correspond to 20 µm.
Ijms 27 02417 g008
Table 1. Sequence of the peptides obtained from the selection.
Table 1. Sequence of the peptides obtained from the selection.
CloneSequenceLength (aa)Isoelectric PointNet Charge (pH 7.4)
A1 *PSAPSSRAARSDCDAASASCTSRRRARRPR3011.85.7
A2 *ARDNGAPVDTPRVDTPGVGTPRIATGRLARCVRRTFRSRLSRGIADRGRAADATCRTSD5911.76.4
A3 *GCVAAARAVRQDDRREELPERARHAAPDRVLDVRHAAARRFRSGRRARRERLCARHQ5711.78.5
A4 *APRRERRSSATARSKCCRTASASSARRKCRTSRAPTPVDSVA4211.89.2
A5 *ALPAETTTARDSRACAASRTAHGRSSVRSPCRRRRESARAVGDPRAAAAATARARS5611.98.4
A6 *SAARPSAGTGRDTDRRDRACVSARRESSARACRRPRNRAGRAAPSSG4711.98.0
A7ARPSPIPIEPWPPASAASTSSACAATTRARARSCTSTRSCRPRRTR4611.97.3
A8 *GGPVAQGAEERTAGAVPPRRPNTARRSRGGARAGRGRV3812.36.6
A9 *AADRKCAVRTGQRARRPYERKCGKYRVGDVSG 3210.56.4
A10 *RRRGRRRHRAAHHAGREHARIAPTRNLTCSTSFT3412.48.3
A11 *GPDRPSACRSAARCPSGSAGNARAARKACRSRSACRSAPTAAAAGPRAPA5011.78.1
B1RGARRDAGRMPSVPAERSRSAVLLRPAAARTRDRRIARDASAA4312.27.2
B2RDSIYRWIDRKGLPAHRVGRLWKFQVSEVDEWVRAGGTDEN418.50.2
B3AGLRKVCCLRCGHGVALRRIAACNRLRIVERPK3311.37.2
B4 *ALASANANSPKGAIEFGPQHYQLYTNDQASQASQYRDLVVAYRNGAGVR498.40.6
B5 *SISQPCSVNRLTITRPCARASSRGRRSRRDRNRGLRPRRRSAGRDL4612.312
B6 *GAASCIPSARPTKTARATSRCTAFRPGSRCRSSTRTPRPRSSIT4412.19.2
B7VGRTRGPAADSRDRNAARRRRRPRSAARRAHATSGIPVNMAALDELKAASERKVFLDAQSADL6311.87.6
B8 *TRSAGANRVHPYAAAAEFVRGHTRHHDGRGLGGRIGRIRGGLRALARNRRDVDDHAAARL6011.97.5
B9PPAGSTASSSSCRSSAAPTGATPHRGSSTSCRSARGPASSRGRGCARRR4912.18.6
B10CSTTSKARWTRSSATRSAPATCSCCAISARKAGPACRKCSRRRRRSSARG5011.813.0
* indicates peptides that were synthetically produced for further investigation.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Stabile, A.; Scaramella, G.; Puccio, S.; Brady, J.; Goltermann, L.; Tolker-Nielsen, T.; Bellich, B.; De Zotti, S.; Lagatolla, C.; Ferrara, F.; et al. Combining Yeast Display and Bacterial Genomic Library for the Unbiased Isolation of Novel Polysaccharide-Binding Peptides. Int. J. Mol. Sci. 2026, 27, 2417. https://doi.org/10.3390/ijms27052417

AMA Style

Stabile A, Scaramella G, Puccio S, Brady J, Goltermann L, Tolker-Nielsen T, Bellich B, De Zotti S, Lagatolla C, Ferrara F, et al. Combining Yeast Display and Bacterial Genomic Library for the Unbiased Isolation of Novel Polysaccharide-Binding Peptides. International Journal of Molecular Sciences. 2026; 27(5):2417. https://doi.org/10.3390/ijms27052417

Chicago/Turabian Style

Stabile, Angela, Gaia Scaramella, Simone Puccio, John Brady, Lise Goltermann, Tim Tolker-Nielsen, Barbara Bellich, Simone De Zotti, Cristina Lagatolla, Fortunato Ferrara, and et al. 2026. "Combining Yeast Display and Bacterial Genomic Library for the Unbiased Isolation of Novel Polysaccharide-Binding Peptides" International Journal of Molecular Sciences 27, no. 5: 2417. https://doi.org/10.3390/ijms27052417

APA Style

Stabile, A., Scaramella, G., Puccio, S., Brady, J., Goltermann, L., Tolker-Nielsen, T., Bellich, B., De Zotti, S., Lagatolla, C., Ferrara, F., Rizzo, R., Cescutti, P., & Sblattero, D. (2026). Combining Yeast Display and Bacterial Genomic Library for the Unbiased Isolation of Novel Polysaccharide-Binding Peptides. International Journal of Molecular Sciences, 27(5), 2417. https://doi.org/10.3390/ijms27052417

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop