pharmaceutics Article Antibacterial Activity of T22, a Specific Peptidic Ligand of the Tumoral Marker CXCR4 Naroa Serna 1,2,3,†,‡ , José Vicente Carratalá 1,2,3,‡ , Oscar Conchillo-Solé 1,2 , Carlos Martínez-Torró 1 , Ugutz Unzueta 2,3,4,5 , Ramón Mangues 3,4,5 , Neus Ferrer-Miralles 1,2,3 , Xavier Daura 1,3,6, *, Esther Vázquez 1,2,3 and Antonio Villaverde 1,2,3, * 1 Institut de Biotecnologia i de Biomedicina, Universitat Autònoma de Barcelona, 08193 Cerdanyola del Vallès, Spain; srnaroa@gmail.com (N.S.); josevicente.carratala@uab.cat (J.V.C.); ocs@bioinf.uab.es (O.C.-S.); carlosmartineztorro@gmail.com (C.M.-T.); neus.ferrer@uab.cat (N.F.-M.); Esther.Vazquez@uab.es (E.V.) 2 Departament de Genètica i de Microbiologia, Universitat Autònoma de Barcelona, 08193 Cerdanyola del Vallès, Spain; uunzueta@santpau.cat 3 CIBER de Bioingeniería, Biomateriales y Nanomedicina (CIBER-BBN), 28029 Madrid, Spain; rmangues@santpau.cat 4 Biomedical Research Institute Sant Pau (IIB-Sant Pau), Hospital de la Santa Creu i Sant Pau, 08025 Barcelona, Spain  5 Josep Carreras Research Institute, 08916 Barcelona, Spain  6 Catalan Institution for Research and Advanced Studies (ICREA), 08010 Barcelona, Spain Citation: Serna, N.; Carratalá, J.V.; * Correspondence: Xavier.Daura@uab.cat (X.D.); Antoni.Villaverde@uab.cat (A.V.) Conchillo-Solé, O.; Martínez-Torró, † Present address: Nanoligent SL. Edifici Eureka, Universitat Autònoma de Barcelona, C.; Unzueta, U.; Mangues, R.; 08193 Cerdanyola del Vallès, Spain. ‡ These authors contribute equally to this paper. Ferrer-Miralles, N.; Daura, X.; Vázquez, E.; Villaverde, A. Antibacterial Activity of T22, a Abstract: CXCR4 is a cytokine receptor used by HIV during cell attachment and infection. Overex- Specific Peptidic Ligand of the pressed in the cancer stem cells of more than 20 human neoplasias, CXCR4 is a convenient antitumoral Tumoral Marker CXCR4. drug target. T22 is a polyphemusin-derived peptide and an effective CXCR4 ligand. Its highly selec- Pharmaceutics 2021, 13, 1922. tive CXCR4 binding can be exploited as an agent for the cell-targeted delivery and internalization https://doi.org/10.3390/ of associated antitumor drugs. Sharing chemical and structural traits with antimicrobial peptides pharmaceutics13111922 (AMPs), the capability of T22 as an antibacterial agent remains unexplored. Here, we have detected T22-associated antimicrobial activity and biofilm formation inhibition over Escherichia coli, Staphylo- Academic Editors: Claudia Monteiro, coccus aureus and Pseudomonas aeruginosa, in a spectrum broader than the reference AMP GWH1. In Viorica Patrulea, Martijn Riool and contrast to GWH1, T22 shows neither cytotoxicity over mammalian cells nor hemolytic activity and Maria Nowakowska is active when displayed on protein-only nanoparticles through genetic fusion. Under the pushing Received: 14 September 2021 need for novel antimicrobial agents, the discovery of T22 as an AMP is particularly appealing, not Accepted: 5 November 2021 only as its mere addition to the expanding catalogue of antibacterial drugs. The recognized clinical Published: 13 November 2021 uses of T22 might allow its combined and multivalent application in complex clinical conditions, such as colorectal cancer, that might benefit from the synchronous destruction of cancer stem cells Publisher’s Note: MDPI stays neutral and local bacterial biofilms. with regard to jurisdictional claims in published maps and institutional affil- Keywords: antimicrobial peptides; nanoparticles; fusion proteins; inhibition of biofilm formation; iations. multivalent drugs Copyright: © 2021 by the authors. 1. Introduction Licensee MDPI, Basel, Switzerland. The peptide T22, also known as [Tyr5,12,Lys7]-polyphemusin II, is a 18-mer derivative This article is an open access article of the horseshoe crab cationic antimicrobial peptide (AMP) polyphemusin I, in which three distributed under the terms and amino acid replacements enable it for a precise binding to the cell surface chemokine recep- conditions of the Creative Commons tor CXCR4 [1,2]. Since CXCR4 is an HIV co-receptor [3], T22 was developed as an anti-HIV Attribution (CC BY) license (https:// peptide potentially effective in antiretroviral therapies, blocking the fusion between the vi- creativecommons.org/licenses/by/ ral envelope and the cell membrane and thus preventing viral infection [1,4]. From another 4.0/). Pharmaceutics 2021, 13, 1922. https://doi.org/10.3390/pharmaceutics13111922 https://www.mdpi.com/journal/pharmaceutics Pharmaceutics 2021, 13, 1922 2 of 16 point of view and taking advantage of its selective CXCR4 binding, T22 has been largely exploited as a targeting agent, for precision therapies against diverse CXCR4+ human cancers. Among others, these include leukemia, lymphoma, head and neck and colorectal cancer [5–14], in which CXCR4 is overexpressed in metastatic cancer stem cells. In this con- text, when T22 is engineered as an N-terminal peptide in H6-tagged proteins it promotes, due to its cationic character [15], protein self-assembly into homomeric nanoparticles [16], which include around 10 monomers positioned in a regular, toroidal architecture [17]. The multivalent display of T22 on the particle surface and the nanometric size of these constructs (usually ranging from 12 to 30 nm, depending of the domain composition of the fusion protein) enhances CXCR4 binding and the consequent penetrability into CXCR4+ cells, while preventing the renal filtration of chemically coupled drugs [18]. By a combina- tion of all these properties, T22 ensures the architectonic stability of the protein material in the bloodstream and allows a selective intracellular accumulation of T22-empowered protein-only nanoparticles and associated drugs into CXCR4-overexpressing cancer stem cells [17–19]. Then, upon the systemic administration of the reporter protein T22-GFP-H6 and derived cytotoxic constructs, a precise in vivo biodistribution is observed, with the destruction of CXCR4-overexpressing cancer tissues and metastatic foci in the absence of side toxicities [18]. Despite polyphemusins generically displaying potent antimicrobial activities [4,20–23], these functionalities and their interactivity with bacterial cell membranes largely depend on the precise amino acid sequence, which is highly sensitive to even a few amino acid substitutions [24,25]. This is because even point mutations can alter its amphipathic- ity and hydrophobicity, features that have been postulated to be pivotal in maintaining the right balance between toxicity and antimicrobial activity [26–28]. Although several structural variants of polyphemusin peptides have been tested for antimicrobial proper- ties [24,25], T22 has been never explored in this regard. Considering the growing need for new antimicrobial agents and the proved clinical potential of T22 in cell-targeted drug delivery [11,18,29], the detection of any new antimicrobial activities in this peptide would be of broad interest and deserves a thorough investigation. More so, these functionali- ties might be conserved in T22 peptides displayed on multimeric protein nanoparticles, since antimicrobial activities largely benefit from nanostructured and multivalent presen- tations [30,31]. Additionally, the combination of anticancer and antimicrobial properties could be of special relevance in many cancers and cancer-linked conditions in which bacteria have a predominant or even triggering role [32–36]. 2. Material and Methods 2.1. Peptides, Proteins and Protein Nanoparticles T22 (RRWCYRKCYKGYCYRKCRK(5,6-FAM)) was synthetized by Caslo Aps (Caslo Aps Kongens, Lyngby, Denmark) and GWH1 (GYNYAKKLANLAKKFANALWC) by NZYTech (NZYTech, Lisboa, Portugal). An additional C-terminal lysine was included in the synthetic T22 for potential functionalization. On the other hand, production and purification of the fusion proteins T22-GFP-H6 [37], T22-PE24-H6 [38], GWH1-GFP-H6 [39] and GFP-H6 [40] have been precisely described elsewhere. Briefly, these are fusion polypep- tides consisting of three functional modules (with the exception of the bimodular construct GFP-H6). In them, the hexahistidine H6 is placed at the carboxy terminus, and the cationic peptides T22 or GWH1 at the amino terminus. The core module is either the enhanced GFP or the de-immunized catalytic domain of Pseudomonas aeruginosa exotoxin A (PE24), which displays a potent cytotoxic activity. All proteins were produced by the expression of codon-adapted synthetic genes carried by the expression vector pET22b, in Escherichia coli Origami B (BL21, OmpT−, Lon−, TrxB−, Gor−, Novagen, Merck, Darmstadt, Germany), at yields ranging between 3 and 71 mg/L. Proteins were then purified through the H6 domain by Immobilized Metal Affinity Chromatography (IMAC) using a HiTrap Chelat- ing HP 1 mL column (GE Healthcare, Piscataway, NJ, USA), in an AKTA purifier FPLC Pharmaceutics 2021, 13, 1922 3 of 16 (GE Healthcare, Piscataway, NJ, USA). The eluted proteins showed a purity level usually over 95%. 2.2. Bacterial Growth and Determination of the Minimum Inhibitory Concentration The effects of the different antimicrobial agents were evaluated against E. coli ATCC 25922, S. aureus ATCC 29737 and P. aeruginosa ATCC-27853. The assay was performed using a broth microdilution method. In 96-well plates, after a two-fold dilution process, each well contained a specific amount of the corresponding peptide, ranging from 2 to 32 µmol/L for GWH1 and 2 to 64 µmol/L for T22 in Mueller Hinton Broth Cation-adjusted medium (MHB- 2, Sigma-Aldrich, Saint Louis, MO, USA). Then, 50 µL of MHB-2 containing 106 colony forming units per mL (CFU/mL) was inoculated in each well. After inoculation, the plates were incubated without agitation at 37 ◦ C for 18 h. Bacterial growth was measured by OD620 . The effect of protein nanoparticles in bacterial growth was analyzed following the same protocol with concentrations ranging from 2 to 16 µmol/L, in S. aureus ATCC 29737. Maximal growth was achieved in control wells with no protein and each concentration was evaluated in duplicate. To determine the minimum inhibitory concentration (MIC) of the agents, the lowest concentration showing no bacterial growth, evaluated by visual inspection, was taken. The raw numerical data for all the experimental can be found in the dataset in the Supplementary Materials. 2.3. Time-Killing Kinetic Assay Different concentrations of GWH1 and T22 were distributed in 96-well plates and incubated with a suspension (in Mueller Hinton Broth Cation-adjusted medium, MHB-2, Sigma-Aldrich, Saint Louis, MO, USA) containing 106 CFU/mL of E. coli ATCC 25922 or S. aureus ATCC 29737. Plates were incubated without agitation at 37 ◦ C. At the indicated times (0, 0.5, 1, 2, 3, 4, 5 and 24 h), an aliquot of 10 µL (out of a total of 200 µL per well) was serially diluted (10-fold) in a different 96-well plate and subsequently seeded in LB plates to evaluate the bacterial viability by CFU counting. Each concentration was evaluated in triplicate and each dilution was seeded in duplicate; therefore, a maximum of six individual counts were used to determine the final CFU for each concentration. A control was included to evaluate bacterial growth in absence of the peptides. 2.4. Evaluation of Biofilm Formation Biofilms were formed by addition of 106 CFU mL−1 of the bacterial suspension (E. coli ATCC 25922 or S. aureus ATCC 29737) in sterile, flat-bottomed, 96-well polystyrene micro- well plates (100 µL per well) and incubated in a static condition for 18 h at 37 ◦ C. To determine the antibiofilm activity, different concentrations of the peptides GWH1 and T22 were added to the wells to prevent cell adherence. After incubation, the total biomass of the biofilm was analyzed using the crystal violet (CV) staining method [41]. The contents of the wells were discarded and washed three times with distilled water to remove the planktonic bacteria. Then, biofilms formed by adherent sessile bacteria in the plate wall were fixed by air-drying at 60 ◦ C for 60 min and stained for 15 min with 150 µL of (CV) solution at 0.1%. The stained biofilms were again washed with distilled water and dried for 30 min at 37 ◦ C. Finally, the adhered biofilms were extracted with 200 µL of 30% acetic acid. The biofilm quantification was determined by the photometric measurement of the CV intensity at 550 nm using the multilabel plater Reader VICTOR3 (PerkinElmer, Inc., Waltham, MA, USA). Each concentration was evaluated in duplicate. 2.5. Mammalian Cell Viability Assay The potential cytotoxicity of peptides was tested in murine embryo (NIH3T3 cells) and human lung (MRC-5 cells) fibroblasts and in cervical cell line (HeLa). NIH 3T3 ATCC CRL-1658 cells were maintained in Dulbecco’s Modified Eagle’s Medium (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) and MRC-5 ATCC CCL-171 cells and HeLa cells were maintained in Eagle’s Minimum Essential Medium (Gibco, Thermo Fisher Scientific, Pharmaceutics 2021, 13, 1922 4 of 16 Waltham, MA, USA). All cell lines were supplemented with 10% fetal bovine serum (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) and incubated in a humidified atmosphere at 37 ◦ C and 5% of CO2 . A total of 5000 cells/well for fibroblasts and 3000 cells/well for cervical cells were cultured in opaque-walled 96-well plates for 24 h at 37 ◦ C until reaching 70% confluence and were then exposed to peptides at 8, 16, 32 and 64 µmol/L for 48 h. After incubation, CellTiter-Glo® Luminescent Cell Viability Assay (Promega, Madison, WI, USA) was used to determine the potential peptide cytotoxicity. The luminescent signal, proportional to the amount of ATP present in the sample, was measured in a conventional microplate reader VICTOR3 (PerkinElmer, Inc., Waltham, MA, USA). The cell viability experiments were performed in triplicate. 2.6. Hemolysis Assay Freshly drawn human erythrocytes were harvested by centrifugation for 5 min at 1500 g and washed three times with PBS (137 mM NaCl, 2.7 mM KCI, 10 mM Na2 HPO4 , 1.8 mM de KH2 PO4 ). Subsequently, a work solution was prepared by diluting the washed erythrocytes with PBS (1%, v/v). In a 96-well conical bottom plate, the 1% (v/v) erythro- cyte suspension was incubated for 1 h at 37 ◦ C with different concentrations (16, 32 and 64 µmol/L of the GWH1 and T22 peptides. After incubation, the plates were centrifuged for 5 min at 1500× g, and the supernatant was transferred to a new 96-well plate to mea- sure the absorbance at 405 nm in a multilabel plater Reader VICTOR3 (PerkinElmer, Inc., Waltham, MA, USA). Two controls were included, PBS as a non-hemolysis control and Triton X-100 as a 100% hemolysis control. Experiments were performed in triplicate. 2.7. Measurement of the Nanoparticle Size Size distribution of protein samples was determined by dynamic light scattering (DLS). Average values were obtained after the independent measurement of protein samples in triplicate, at 633 nm, in a Zetasizer Nano ZS (Malvern Instruments Ltd., Malvern, UK). 2.8. Structure-Based Calculations and Molecular Modeling Polyphemusin-I (PM1; Protein Data Bank –PDB– code: 1RKK) [24], Tachyplesin-1 (TL1; PDB code: 1WO0); http://dx.doi.org/10.2210/pdb1wo0/pdb, accessed on 23 July 2021), Arenicin-3 (AR3; PDB code: 5V0Y) [42]; Protegrin-1 (PG1, PDB code: 1PG1) [43], Gomesin (GM, PDB code: 1KFP) [44], Thanatin (TT, PDB code: 8TFV) [45], PV7 (PM1 synthetic structural variant) [25] and T22 (a polyphemusin II analog) [4] were used as model β-hairpin antimicrobial peptides. For each multimodel NMR structure down- loaded from the PDB, the model closest to the average was selected as representative (https://swift.cmbi.umcn.nl/servers/html/bestml.html, accessed on 23 July 2021). Foldx v5 [46] RepairPDB function was applied to all selected structures. PV7 was modeled from PM1 using the BuildModel function from FoldX v5 (Fundació Centre de Regulació Genòmica, Barcelona, Catalonia, Spain), T22 was modeled from PM1 too, PyMOL builder tool (The PyMOL Molecular Graphics System, Version 2.0 Schrödinger, LLC. New York, NY, USA. https://pymol.org, accessed on 4 January 2021) was used to add the extra lysine at the C-terminus and then FoldX v5 RepairPDB and BuildModel functions were applied. The 3D- HM web application (Karlsruhe Institute of Technology, Karlsruhe, Baden-Württemberg, Germany) [47] was used to generate the GWH1 structure as well as to calculate the net charge, average absolute electrostatic potential at the peptide’s surface and Hydrophobic Moment (HM) vector (a representation of the distribution of polar and nonpolar parts in a molecular surface) for all analyzed structures. VMD version 1.9.4a51 [48] and PyMOL ver- sion 2.4.2 were used to generate figures with these results. PDB2PQR version 3.1.0 [49] was used to add hydrogen atoms to all structures before submitting them to 3D-HM. PDBparam (https://www.iitm.ac.in/bioinfo/pdbparam/index.html, accessed on 20 July 2021) [50] was used to calculate the surface hydrophobicity and what the authors call hydrophobic free energy (a solvent-accessible-area-based estimate of the non-polar component of the change in solvation free energy upon folding). Superposition of structures for analysis or represen- July 2021) [50] was used to calculate the surface hydrophobicity and what the authors call hydrophobic free energy (a solvent-accessible-area-based estimate of the non-polar com- Pharmaceutics 2021, 13, 1922 ponent of the change in solvation free energy upon folding). Superposition of structures5 of 16 for analysis or representation in figures was based on the main chain N, CA, C and O atoms with the McLachlan algorithm [51] as implemented in ProFit v3.3 (http://www.bi- oinf.org.uk/software/profit/, accessed on 22 March 2021) and using residue equivalences tation in figures was based on the main chain N, CA, C and O atoms with the McLachlan as obtained algorithmwith jCE, [51] as the java implementation implemented of the CE method [52]. in ProFit v3.3 (http://www.bioinf.org.uk/software/profit/, accessed on 22 March 2021) and using residue equivalences as obtained with jCE, the java 3. Results implementation of the CE method [52]. The generated T22 model reproduces an early NMR structure of T22 [4], which re- 3. Results vealed the conservation of the hairpin structure that promotes the interaction with the The generated T22 model reproduces an early NMR structure of T22 [4], which membrane and antimicrobial activity of PM1 and other AMPs [28]. A closer analysis of revealed the conservation of the hairpin structure that promotes the interaction with the the membrane T22 structure (Figure 1) demonstrates a similarity of traits with other AMPs, including and antimicrobial activity of PM1 and other AMPs [28]. A closer analysis of length, hydrophobicity, the T22 structure (Figure net 1) charge and amphipathicity, demonstrates as with a similarity of traits reflected otherby the Hydrophobic AMPs, including Moment length,(HM) vector. Although hydrophobicity, theand net charge majority of its properties amphipathicity, are naturally as reflected closer to those by the Hydrophobic of other Momentβ-hairpin-forming AMPsthe (HM) vector. Although (Figure 1A,of majority allitsbut GWH1),are properties innaturally particular and to closer asthose expected the of close relative other PM1 and itsAMPs β-hairpin-forming derivative (FigurePV7 (Figure 1A, all 1A,B),init particular but GWH1), stands out andthat T22 presents as expected the close an HM relative (a magnitude PM1 and its derivative quantity PV7 (Figure that increases with the 1A,B), it standsofout unbalance that the T22 presents distribution of po- an HM magnitude (a quantity that increases with the unbalance of the distribution lar and nonpolar surface areas in the molecule, i.e., amphipathicity) closest to GWH1. of polar and nonpolar surface areas in the molecule, i.e., amphipathicity) closest to GWH1. When When the corresponding HM vectors are aligned with the membrane normal it is shown the corresponding HM vectors are aligned with the membrane normal it is shown that T22 that T22 and GWH1 present very similar tilts (Figure 1C, the alignment of the HM vector and GWH1 present very similar tilts (Figure 1C, the alignment of the HM vector and the andmembrane the membranenormalnormal provides provides an indication an indication of the orientation of the orientation that the that the peptide maypeptide adopt may adopt when when inserted inserted in theinmembrane). the membrane). FigureFigure 1. Structure 1. Structure of GWH1, of GWH1, T22T22andand related related AMPs.(A) AMPs. (A)Columns Columns contain contain the thefollowing following data: data:peptide name peptide (β-hairpin- name (β-hairpin- formingforming peptides peptides withwith background background coloredasasin colored in(B); (B); length length ininnumber number of of amino acids; amino peptide acids; sequence peptide alignedaligned sequence with T22with (all T22 (all butbut GWH1),residues GWH1), residues used usedfor forsuperposition superposition in bold; hydrophobic-moment-vector in bold; hydrophobic-moment-vector magnitude; average absolute magnitude; averageelectrostatic absolute elec- potential trostatic on the potential peptide’s on the surface; peptide’s net charge; surface; RMSDRMSD net charge; from T22 superposition; from surface hydrophobicity; T22 superposition; hydrophobic surface hydrophobicity; free hydropho- bic free energy (see methods for definition); FoldX total energy after repair. (B) Superposition of all β-hairpin AMPs energy (see methods for definition); FoldX total energy after repair. (B) Superposition of all β-hairpin AMPs used in thisused in thisstudy study and and their their calculated calculated HM vectors HM vectors against against model T22model (RMSDT22 and(RMSD colors forand eachcolors peptideforgiven eachin peptide given in A). (C) (A). (C) Superposition Superposition of the calculated of the calculated HM vectorHM vector for T22 andfor T22 and GHW1, GHW1, oriented oriented parallel to theparallel membraneto the membrane normal. normal. (D) of (D) Representation Represen- the tation surface of the surface electrostatic electrostatic potential potential for T22 and for T22 and GWH1 GWH1 (scale: top red = −10, (scale: top red top = -10, blue top blue = 10). = 10). Pharmaceutics 2021, 13, 1922 6 of 16 In view of these AMP-like physicochemical properties, T22 was tested for its antimi- crobial activity over liquid cultures of three bacterial pathogens, namely Escherichia coli, Staphylococcus aureus and Pseudomonas aeruginosa. GWH1, of similar length (Figure 1A), was used as control. GWH1 is a non-natural peptide, developed as AMP [53,54], that adopts an amphipathic helical structure when bound to a membrane and its GFP fusion construct (GWH1-GFP-H6) self-assembles similarly to T22-GFP-H6 [55]. Upon exposure to bacterial cultures, both peptides promoted a clear drop of optical density in a dose-dependent manner (Figure 2A), with E. coli and S. aureus being the most sensitive species. GWH1 was superior to T22 in terms of its antibacterial activity, with lower MIC values in all cases (Figure 2B). While the biological impact of GWH1 was immediate, T22 required longer times to reach comparable disruptive effects over bacterial cells (Figure 3A,B). In addition, both peptides inhibited biofilms formed by E. coli and S. aureus (Figure 3B), which we value as a promising feature regarding the potential applicability of T22 as an AMP. Due to their interaction with biological membranes, many AMPs show hemolytic or cytotoxic activities over mammalian cells, bottlenecking a widespread use of these materials as safe drugs [56,57]. In this context and as expected, GWH1 showed a mild cytotoxicity (Figure 4A) and a dose-dependent hemolysis (Figure 4B) that compromise its clinical use. In contrast, T22 shows only a moderate or absent cytotoxicity in several cell rmaceutics 2021, 13, x FOR PEER REVIEW 7 of lines (Figure 4A) and a complete absence of hemolytic activity up to the very high doses tested here (Figure 4B), which represents a clear competitive advantage over the control peptide GWH1. Figure 2.of Figure 2. Impact Impact of peptides peptides T22 andT22 GWH1 and GWH1 onon bacterialgrowth bacterial growth ininliquid culture. liquid (A) Bacterial culture. of E. coli,of growthgrowth (A) Bacterial S. aureus E. coli, S. aureus P. aeruginosa and P. aeruginosa cultures, measured by their optical density at 620 nm, treated with T22 and control GWH1inpeptides in and cultures, measured by their optical density at 620 nm, treated with T22 and control GWH1 peptides serial 2-fold dilutions at 37 ◦ C for 18 h. Each point represents an average of at least two different values and error bar serial 2-fold dilutions at 37 °C for 18 h. Each point represents an average of at least two different values and error bar indicates standard deviation. (B) Minimum inhibitory concentration (MIC) of the peptides for E. coli and S. aureus. The indicates standard deviation. (B) Minimum inhibitory concentration (MIC) of the peptides for E. coli and S. aureus. The lowest concentration showing no bacterial growth (evaluated by visual inspection) in the broth microdilution method was lowest concentration showing no bacterial growth (evaluated by visual inspection) in the broth microdilution method was taken as the MIC. Significant differences between groups are indicated as * p < 0.01, one-way analysis of variance (ANOVA) taken as the MIC. Significant differences between groups are indicated as * p < 0.01, one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparisons test. followed by Tukey´s multiple comparisons test. Pharmaceutics 2021, 13, x FOR PEER REVIEW 8 of 17 Pharmaceutics 2021, 13, 1922 7 of 16 Figure 3. Time- and concentration-dependent impact of the peptides on bacterial growth and biofilm Figure 3. Time- formation, and concentration-dependent respectively. impact (A) Time-kill kinetics of E. ofand coli (up) theS.peptides on bacterial aureus (down) exposedgrowth to T22 andand bio- film formation, GWH1 peptidesrespectively. (A) Time-killfor at different concentrations kinetics of2,E.3, coli 0, 0.5, 1, 4, 5 (up) and 24and S. aureus h. Each point (down) represents exposed an to T22average and GWH1 peptides at different concentrations for 0, 0.5, 1, 2, 3, 4, 5 and 24 h. of at least two different values and error bars indicate standard deviation. Control representsEach point rep- resents an average the bacterial growth of without at least peptide two different values exposure. andindicate Figures error bars indicate in concentration standard µmol/L.deviation. (B) Effect Con- trolofrepresents the bacterial growth without peptide exposure. Figures indicate peptide concentration on biofilm formation on the surface of the microtiter wells. Significant concentration in µmol/L. (B) Effect of peptide concentration on biofilm formation on the surface differences over bacterial control are indicated as * p < 0.05, one-way analysis of variance (ANOVA) of the microtiter wells. Significant differences over bacterial control are indicated as * p < 0.05, one-way analysis of followed by Tukey’s multiple comparisons test. variance (ANOVA) followed by Tukey´s multiple comparisons test. Pharmaceutics Pharmaceutics 13, x13, 2021,2021, FOR1922PEER REVIEW 8 of 9 16 of 17 Figure 4. Cytotoxicity of peptides T22 and GWH1 on mammalian cells. (A) Cell viability in the presence of peptides over cultured Figure mammalian 4. Cytotoxicity cells, recorded of peptides 48 hGWH1 T22 and after exposure. (B) Hemolytic on mammalian activity cells. (A) associated Cell viability in with peptides of the presence over humanover peptides erythrocytes. cultured Significant mammalian differences48 cells, recorded over cell control h after are indicated exposure. as * p < activity (B) Hemolytic 0.01, one-way analysis associated of variance with peptides(ANOVA) over human followed by erythrocytes. Tukey’s multiple Significant comparisons differences over celltest. Figures control areindicate concentration indicated in µmol/L. as * p < 0.01, one-way analysis of variance (ANOVA) followed by Tukey´s multiple comparisons test. Figures indicate concentration in µmol/L. If T22 would keep its antimicrobial activities when presented in assembled, protein- only nanoparticles, it would have a potential dual application as a CXCR4-targeting agent and AMP. It is widely recognized that bacterial infections not only represent further complications in solid tumors [58–60] but that they also participate in tumor formation or as triggering agents in several human neoplasias [32,33]. Such triggering effects are specially suspected in organs such as colon that are continuously exposed to microbiome components that might largely contribute to, or modulate, the initiation and progression of colorectal cancer [61–65]. Particularly, the formation of E. coli biofilms has been recently pointed out as an oncogenic driver in colorectal cancer development [66], and as shown above (Figure 3B), T22 is a good inhibitor of E. coli biofilm formation. On the other hand, T22, in the form of fusion proteins assembled as nanoparticles, has proved to be highly effective in the targeted delivery of antitumoral drugs. In this context, the P. aeruginosa exotoxin A has been genetically inserted in a T22-based protein construct, thus generating the build-in cytotoxic, CXCR4-targeted nanoparticle T22-PE24-H6. In animal models of metastatic human cancers, T22 confers selectivity for CXCR4-overexpressing cancer stem cells while the bacterial toxin PE24 causes cell death and cancer remission [6,8]. The combination of the toxin and T22 is then a clinically promising concept [18,67]. Envisaging a dual role of T22 in protein constructs, the T22-empowered fusions T22- GFP-H6 and T22-PE24-H6, assembled as regular nanoparticles (Figure 5A), were tested for their potential activity as antibacterial agents against S. aureus (Figure 5B). Interestingly, the display of T22 in oligomers (Figure 5C) is not only maintained but it also tends to enhance the antibacterial capacity of the free peptide, particularly for T22-PE24-H6 (compare data from Figures 2A and 5B). In the oligomers, 10 copies of the T22 loop are predicted to be exposed to the solvent with their HM pointing in perfect sense and direction to allow interaction with the membrane (Figure 5C). In this orientation, T22 would keep the β- hairpin structure associated with AMP activity (Figure 5D). The observation of AMP Figure 5. Antibacterial activityactivity linkedstructure and possible to T22 inofthe form of multimeric a recombinant nanoparticles, T22 displayed on proteininnanoparticles. which the peptide (A) Vol-is genetically fused to the building block, opens interesting routes for ume size distribution of T22-PE24-H6, T22-GFP-H6 and GWH1-GFP-H6 nanoparticles and unassembled GFP-H6, as de- the engineering of antimicrobial termined by dynamic light scattering. peptides Data are in more represented as effective formulations mean ± standard easily error on the reachable mean (SEM).through simple (B) Bacterial integrity of S. aureus measuredgenetic engineering. by the optical density at 620 nm, after incubation of GWH1-GFP-H6, T22-GFP-H6 and T22- PE24-H6 nanoparticles and GFP-H6 in serial 2-fold dilutions at 37 °C for 18 h. Protein concentration at the X axis refers to monomers. Significant differences over GFP control protein are indicated as * p < 0.01, one-way analysis of variance (ANOVA) followed by Tukey´s multiple comparisons test. (C) A proposed model for the T22-GFP-H6 nanoparticle ac- cording to a previous approach [17]. Each T22 peptide has been colored differently and its calculated HM vector has been drawn in black. (D) Left: detail of (C) for a single nanoparticle monomer. Right: Closeup of the T22 region with PM1- selected model superposed (RMSD 1.34 calculated with PyMOL’s “super” function). Figure 4. Cytotoxicity of peptides T22 and GWH1 on mammalian cells. (A) Cell viability in the presence of peptides over cultured Pharmaceutics mammalian 2021, 13, 1922 cells, recorded 48 h after exposure. (B) Hemolytic activity associated with peptides over human9 of 16 erythrocytes. Significant differences over cell control are indicated as * p < 0.01, one-way analysis of variance (ANOVA) followed by Tukey´s multiple comparisons test. Figures indicate concentration in µmol/L. Figure Figure 5.5. Antibacterial Antibacterial activity activityand andpossible possiblestructure structureofofa arecombinant recombinant T22 T22 displayed displayed onon protein protein nanoparticles. nanoparticles. (A)(A) Vol- Volume ume size distribution of T22-PE24-H6, T22-GFP-H6 and GWH1-GFP-H6 nanoparticles and unassembled size distribution of T22-PE24-H6, T22-GFP-H6 and GWH1-GFP-H6 nanoparticles and unassembled GFP-H6, as determined GFP-H6, as de- termined by dynamic light scattering. Data are represented as mean ± standard error on the mean (SEM). (B) Bacterial by dynamic light scattering. Data are represented as mean ± standard error on the mean (SEM). (B) Bacterial integrity of integrity of S. aureus measured by the optical density at 620 nm, after incubation of GWH1-GFP-H6, T22-GFP-H6 and T22- S. aureus measured by the optical density at 620 nm, after incubation of GWH1-GFP-H6, T22-GFP-H6 and T22-PE24-H6 PE24-H6 nanoparticles and GFP-H6 in serial 2-fold dilutions at 37 °C for 18 h. Protein concentration at the X axis refers to nanoparticlesSignificant and GFP-H6 in serial 2-fold dilutions ◦ C for 18 h. Protein concentration at the X axis refers to monomers. at 37protein monomers. differences over GFP control are indicated as * p < 0.01, one-way analysis of variance Significant differences (ANOVA) followed byover GFP control Tukey´s multipleprotein are indicated comparisons test. (C) p