pharmaceutics Communication The Diphtheria Toxin Translocation Domain Impairs Receptor Selectivity in Cancer Cell-Targeted Protein Nanoparticles Eric Voltà-Durán 1,2,3 , Julieta M. Sánchez 1,2,3,4 , Eloi Parladé 1,2,3 , Naroa Serna 1,2,3,† , Esther Vazquez 1,2,3 , Ugutz Unzueta 2,3,5,6, * and Antonio Villaverde 1,2,3, * 1 Institut de Biotecnologia i de Biomedicina, Universitat Autònoma de Barcelona, Bellaterra, 08193 Barcelona, Spain 2 CIBER de Bioingeniería, Biomateriales y Nanomedicina (CIBER-BBN), Instituto de Salud Carlos III, Bellaterra, 08193 Barcelona, Spain 3 Departament de Genètica i de Microbiologia, Universitat Autònoma de Barcelona, Bellaterra, 08193 Barcelona, Spain 4 Instituto de Investigaciones Biológicas y Tecnológicas (IIBYT), CONICET-Universidad Nacional de Córdoba, Av. Velez Sarsfield 1611, Córdoba X5016GCA, Argentina 5 Institut d’Investigació Biomèdica Sant Pau (IIB SANT PAU), Sant Quintí 77-79, 08041 Barcelona, Spain 6 Josep Carreras Leukaemia Research Institute, 08025 Barcelona, Spain * Correspondence: uunzueta@santpau.cat (U.U.); antoni.villaverde@uab.cat (A.V.) † Present address: Nanoligent SL, Edifici Eureka, Campus Universitari, Universitat Autònoma de Barcelona, Cerdanyola del Vallés, 08193 Barcelona, Spain. Abstract: Protein-based materials intended as nanostructured drugs or drug carriers are progressively gaining interest in nanomedicine, since their structure, assembly and cellular interactivity can be tailored by recruiting functional domains. The main bottleneck in the development of deliverable protein materials is the lysosomal degradation that follows endosome maturation. This is especially Citation: Voltà-Durán, E.; Sánchez, disappointing in the case of receptor-targeted protein constructs, which, while being highly promising J.M.; Parladé, E.; Serna, N.; Vazquez, and in demand in precision medicines, enter cells via endosomal/lysosomal routes. In the search for E.; Unzueta, U.; Villaverde, A. The Diphtheria Toxin Translocation suitable protein agents that might promote endosome escape, we have explored the translocation do- Domain Impairs Receptor Selectivity main (TD) of the diphtheria toxin as a functional domain in CXCR4-targeted oligomeric nanoparticles in Cancer Cell-Targeted Protein designed for cancer therapies. The pharmacological interest of such protein materials could be largely Nanoparticles. Pharmaceutics 2022, 14, enhanced by improving their proteolytic stability. The incorporation of TD into the building blocks 2644. https://doi.org/10.3390/ enhances the amount of the material detected inside of exposed CXCR4+ cells up to around 25-fold, pharmaceutics14122644 in absence of cytotoxicity. This rise cannot be accounted for by endosomal escape, since the lysosomal Academic Editors: Eleftherios degradation of the new construct decreases only moderately. On the other hand, a significant loss in Halevas and Chrysoula Kokotidou the specificity of the CXCR4-dependent cellular penetration indicates the unexpected role of the toxin segment as a cell-penetrating peptide in a dose-dependent and receptor-independent fashion. These Received: 3 October 2022 data reveal that the diphtheria toxin TD displayed on receptor-targeted oligomeric nanoparticles Accepted: 23 November 2022 partially abolishes the exquisite receptor specificity of the parental material and it induces nonspecific Published: 29 November 2022 internalization in mammalian cells. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in Keywords: recombinant proteins; nanoparticles; diphtheria toxin; self-assembling; cell-penetrating peptide published maps and institutional affil- iations. 1. Introduction Copyright: © 2022 by the authors. Protein drugs benefit from the structural versatility and broad and multiple func- Licensee MDPI, Basel, Switzerland. tionality of natural polypeptides, whose activities cover most biological events in living This article is an open access article systems [1–3]. In addition, protein functions can be tuned by genetic engineering and distributed under the terms and combined in single-chain modular proteins [4]. Therefore, apart from plain protein drugs, conditions of the Creative Commons such as insulin versions, interferons, hormones, antibodies and enzymes, the possibility to Attribution (CC BY) license (https:// combine, for instance, a cell-targeting domain and a cytotoxic protein in a single molecular creativecommons.org/licenses/by/ product is the basis for ongoing research in many clinical areas, especially in oncology [5]. 4.0/). In this context, fusion proteins can be designed to recruit complementing capabilities in Pharmaceutics 2022, 14, 2644. https://doi.org/10.3390/pharmaceutics14122644 https://www.mdpi.com/journal/pharmaceutics Pharmaceutics 2022, 14, 2644 2 of 13 multifunctional constructs [4,6], whose active segments are selected from the vast catalogue of functional protein domains so far identified. Since proteins are produced at an industrial scale as recombinant versions in a diversity of cell factories, protein drugs are progressively increasing their presence in the list of approved pharmaceuticals [2,7–9]. Among the main obstacles for the use of drug proteins are their potential immunogenic- ity (when from non-human sources) and the proteolytic degradation in the lysosomal com- partment. De-immunization by site-directed mutagenesis and other engineering strategies have allowed for a large reduction in immunotoxicities, such as the ones reported in several interesting products [10–14]. However, lysosomal degradation remains the main bottleneck in the design of effective protein drugs intended for intracellular activities [15–18]. Regard- ing the emerging interest in receptor-mediated, cell-targeted drug delivery in oncology and other fields (aimed to enhance specificity and to reduce non-desired side-effects) [19–21], endosomal uptake of protein drugs results in the degradation of most of the internal- ized product and in the consequent reduction of effectivity [15,18,22,23]. In this context, endosomal-escape protein segments are desirable components of protein-based drugs. Among these peptides, short amino acid sequences (for instance, histidine-rich peptides) or more complex protein domains, some of them from pathogenic microorganisms, have been explored and/or adapted with different levels of success to protein constructs aimed at cell-targeted delivery [16,18,22,24,25]. In particular, the diphtheria toxin (DT) produced by Corynebacterium diphtheriae kills mammalian cells by inactivating the elongation factor EF-2 [26]. This protein exhibits a modular architecture (Figure 1A) in which the transloca- tion domain (TD, also called T domain) plays a critical role in allowing the catalytic domain (CD) to reach the cytosol from endosomal compartments [27,28]. At a late endosome acidic pH, the DT TD undergoes conformational changes that drive its insertion in the endosomal membrane, resulting in the formation of a small pore by which the CD is transported to the cytosol. The DT TD has been used for the active translocation of protein modules to the cytosol, proving its ability to assist in the membrane crossing of polypeptides different from CD [29,30]. Although the precise mechanism by which proteins are translocated to the cytosol by the DT TD is a controversial subject [31,32], the DT TD, free from the active catalytic domain, it is an agent of potential interest in drug delivery [33,34]. In addition, as far as we know, there are no existing reports about the potential role of TD in promoting cell surface internalization previous to receptor-mediated endosomal uptake. We have explored here the applicability of the DT TD as a functional component in self-assembling protein-only nanoparticles targeted to the cell surface cytokine receptor CXCR4, which is a relevant molecular marker for anti-cancer drug targeting [35–38]. The highly precise and selective targeting is provided by the peptide T22, an exclusive ligand of CXCR4 [39–41]. This has been done through the incorporation of the TD into the well-characterized protein T22-GFP-H6, which self-assembles as regular homo-oligomeric nanoparticles of 12 nm and is used as a nanoscale drug vehicle in cancer therapies [42,43]. T22-GFP-H6 enters cells via the surface-exposed cytokine receptor CXCR4, but it undergoes important lysosomal degradation estimated to affect between 85–90% of the engulfed material [25]. While the extent of protein destruction may only moderately affect the activities of attached conven- tional chemical drugs, such as floxuridine, cytarabine or auristatin E [43–46], it might be a limitation for the therapeutic outcome of protein-only pharmaceuticals [5], including H6-tagged versions with self-assembling properties [47]. Pharmaceutics 2022, 14, x FOR PEER REVIEW 3 of 13 Pharmaceutics 2022, 14, 2644 3 of 13 Figure Figure 1. 1. Modular Modular organization organization of of T22-empowered T22‐empoweredproteins. proteins.(A) (A)Modular Modulardisposition dispositionand and3D3Dmodels mod‐ els of T22‐GFP‐TD‐H6 and T22‐DITOX‐H6. T22‐GFP‐TD‐H6 was engineered of T22-GFP-TD-H6 and T22-DITOX-H6. T22-GFP-TD-H6 was engineered to mimic the CXCR4- to mimic the CXCR4‐ targeted toxin T22‐DITOX‐H6 through the replacement of the catalytic domain of the toxin (CD) by targeted toxin T22-DITOX-H6 through the replacement of the catalytic domain of the toxin (CD) by the green fluorescent protein (GFP). The translocation domain (TD) was maintained. Both proteins the green fluorescent protein (GFP). The translocation domain (TD) was maintained. Both proteins display, from the N to the C terminus, the CXCR4 ligand T22, a flexible linker (L; GGSSRSS), two display, furin from the cleavable N to sites (F;the C terminus,recruited GNRVRRSV) the CXCR4 ligand from T22, a flexible the original DT, andlinker (L; GGSSRSS), two the hexa‐histidine furin tag (H6), cleavable all sites (F; represented in GNRVRRSV) color code. Furin recruited fromsites cleavable the original DT, and have proved thefunctional to be hexa-histidine tagfamily in this (H6), all of represented modular in color code. constructions Furin [48]. cleavable Models in 3D sites werehave provedintoAlphaFold2 predicted be functional in represented and this family ofwith modular Chi‐ constructions meraX. (B) The[48]. Models modular in 3D were disposition predicted of the DT and in AlphaFold2 T22‐GFP‐H6 areand alsorepresented with ChimeraX. shown as references. In the center (B) Theismodular the T22‐GFP‐H6 nanoparticle disposition of the DTasandpreviously modelled T22-GFP-H6 [49],shown are also with each monomer In as references. represented the center in a different is the color and T22-GFP-H6 the stabilizing nanoparticle divalent cations as previously modelled as red [49],spheres with each(reprinted monomer from [49] with per‐ represented in a mission different color and the stabilizing divalent cations as red spheres (reprinted from [49] withused from Elsevier). (C) Molecular weight and isoelectric point of the three proteins in this permission study, predicted by ProtParam (Expasy, http://web.expasy.org/protparam; accessed on 20 Novem‐ from Elsevier). (C) Molecular weight and isoelectric point of the three proteins used in this study, ber 2022). predicted by ProtParam (Expasy, http://web.expasy.org/protparam; accessed on 20 November 2022). 2. 2. Materials Materials and and Methods Methods 2.1. Genetic 2.1. Genetic Design Design and and Protein Protein Production Production and Purification Escherichia coli Origami B (BL21, OmpT−, Escherichia OmpT−,Lon−,Lon−TrxB−, , TrxB−Gor−) , Gor− ) (Novagen) (Novagen) were were usedused to to produce produce T22-GFP-H6, T22‐GFP‐H6, T22-GFP-TD-H6 T22‐GFP‐TD‐H6 andand T22-DITOX-H6 T22‐DITOX‐H6 (Figure (Figure 1A).1A). T22-GFP-TD- T22‐GFP‐TD‐H6 H6 gene gene was designed was designed in-house in‐house as codon-optimized as codon‐optimized genes forgenes E. coli, E. coli, synthetized forsynthetized and insertedand inserted in pET22binplasmids pET22b plasmids by Geneart by(Thermo GeneartFisher (Thermo Fisher Waltham, Scientific, Scientific, MA, Waltham, USA).MA, USA). T22‐GFP‐ T22-GFP-TD-H6 TD‐H6 was conceived was conceived to include,tofrom include, from the N-terminus the N‐terminus to the C-terminus, to the C‐terminus, the following the following modules: T22, a peptide that targets the cytokine receptor modules: T22, a peptide that targets the cytokine receptor CXCR4 [38]; a flexible linker CXCR4 [38]; a flexible linker (GGSSRSS), (GGSSRSS), to provide to provide structural structural flexibility flexibility to T22; to T22; the the DTDT furin furin cleavablesite cleavable site(F, (F, GNRVRRSV); the green fluorescent protein (GFP); a second furin GNRVRRSV); the green fluorescent protein (GFP); a second furin cleavable site (F); the cleavable site (F); the DT translocation DT translocationdomain domain(TD); (TD);andandthe thehexa‐histidine hexa-histidine (H6)(H6) tag.tag. TheTheH6 H6 tag tag waswas usedusedfor for both single-step 2+ -based affinity chromatography purification and for ion-assisted both single‐step Ni2+Ni ‐based affinity chromatography purification and for ion‐assisted as‐ assembly sembly of the of the monomers monomers intointo oligomeric oligomeric nanoparticles, nanoparticles, becausebecause of theof the cross-linking cross‐linking prop‐ erties of H6 [47]. Origami B cells were transformed by heat shock, grown ingrown properties of H6 [47]. Origami B cells were transformed by heat shock, in LB LB medium medium (supplemented with 1 × 10 −1 g·L−1 ampicillin, 1.25 × 10−2 g·L−1 tetracycline (supplemented with 1 × 10 g∙L ampicillin, 1.25 × 10 g∙L tetracycline and 1.5 × 10−2 −1 −1 −2 −1 and−11.5 g∙L × 10−2 g·LT22‐GFP‐TD‐H6 kanamycin). −1 kanamycin). T22-GFP-TD-H6 was produced at 20 ◦ C overnight upon was produced at 20 °C overnight upon induction with −3 M isopropyl β-D-1-thiogalactopyranoside (IPTG). Two rounds 1 × 10β‐D‐1‐thiogalactopyranoside 1induction × 10−3 M with isopropyl (IPTG). Two rounds of French Press Pharmaceutics 2022, 14, x FOR PEER REVIEW 4 of 13 Pharmaceutics 2022, 14, 2644 4 of 13 (Thermo Fisher Scientific FA‐078A, Waltham, MA, USA) at 18,000 psi were applied to ob‐ tain bacterial lysates, in the presence of a wash buffer (20 mmol L−1 Tris‐HCl, 500 mmol of French Press (Thermo Fisher Scientific FA-078A, Waltham, MA, USA) at 18,000 psi L−1 NaCl, 10 mmol L−1 imidazole, pH 8.0) and a protease inhibitor cocktail (cOmplete were applied to obtain bacterial lysates, in the presence of a wash buffer (20 mmol L−1 EDTA‐free, Roche Diagnostics, Basel, Switzerland). Soluble fractions were collected by Tris-HCl, 500 mmol L−1 NaCl, 10 mmol L−1 imidazole, pH 8.0) and a protease inhibitor centrifugation (45 min, cocktail (cOmplete 15,000×Roche EDTA-free, g). T22‐GFP‐TD‐H6 Diagnostics, Basel,was purified by Switzerland). immobilized Soluble fractions metal affinity chromatography were collected (IMAC)(45 by centrifugation using min,HisTrap 15,000×HP g). 1T22-GFP-TD-H6 mL columns (Cytiva, Marlborough, was purified by MA, USA). Elution immobilized metal was achieved affinity by a linear chromatography gradient (IMAC) of HisTrap using elution buffer HP 1 mL(2 ×columns 10−2 M Tris‐ (Cytiva, HCl, 5 × 10Marlborough, −1 M NaCl, 1 M MA, USA). Elution imidazole, was pH 8.0) achieved and by a linear the purified gradient protein of elution was finally dialyzed (2 × 10−bicarbonate buffersodium against 2 M Tris-HCl, 5 × 10−1−3M NaCl, 1 M imidazole, pH 8.0) and the purified (166 × 10 M NaHCO3, pH 8.0) with or without sodium chlo‐ protein ride (333 ×was 10−3finally dialyzed M NaCl). against sodium T22‐GFP‐H6 × 10−produced bicarbonate (166 were and−T22‐DITOX‐H6 3 M NaHCO , pH 8.0) and 3 purified by 3 with or without sodium chloride (333 × 10 M NaCl). T22-GFP-H6 and T22-DITOX-H6 slight variations of this protocol. Precise details can be found elsewhere [48]. The oligo‐ were produced and purified by slight variations of this protocol. Precise details can be meric organization of T22‐GFP‐H6 is shown in Figure 2B and the main proteomic traits of found elsewhere [48]. The oligomeric organization of T22-GFP-H6 is shown in Figure 2B alland these theconstructs main proteomicare depicted in these traits of all Figure 1C. constructs are depicted in Figure 1C. Figure Figure CharacterizationofofT22‐GFP‐TD‐H6 2. 2.Characterization T22-GFP-TD-H6 nanoparticles nanoparticles and and their their cellular cellularaccumulation. accumulation. (A) Protein gel electrophoresis (A) Protein gel electrophoresis of the three of the proteins three proteinsused usedininthe thestudy study(T22‐GFP‐H6, green;T22- (T22-GFP-H6, in green; T22‐GFP‐ TD‐H6, in dark GFP-TD-H6, in pink; T22‐DITOX‐H6, dark pink; T22-DITOX-H6, inindark darkyellow). (B)Protein yellow). (B) Proteingelgel electrophoresis electrophoresis of T22‐GFP‐ of T22-GFP- TD‐H6, TD-H6, before before(pre) (pre)and and after (post)exposure after (post) exposure to to recombinant recombinant furin. furin. (C) Intensity (C) Intensity size distribution size distribution of of T22‐GFP‐H6 T22-GFP-H6and and T22‐GFP‐TD‐H6, T22-GFP-TD-H6, measured measuredbyby Dynamic Dynamic Light Light Scattering Scattering (DLS). (DLS). Inset values Inset values indicate indicate mean mean size sizemeasured measuredby byintensity intensity (nm) (nm) and polydispersityindex and polydispersity index(PDI) (PDI)±±standard standarderror. error. (D) Relative intracellular (D) Relative accumulation intracellular accumulation of of T22‐GFP‐H6 T22-GFP-H6and and T22‐GFP‐TD‐H6 T22-GFP-TD-H6 inin HeLa HeLa cells cells after after 24 h24 h of exposure, of exposure,at four different at four differentconcentrations concentrations (100 (100nM, nM,10001000nM,nM, 2000 2000 nM,nM, 30003000 nM).nM). Data Data are shown are shown as as mean ± standard error. Fluorescence values are referred to those of background mean ± standard error. Fluorescence values are referred to those of background cells. Significant sta- cells. Significant statistical differences tistical differences between between datadata pairs pairs (T22‐GFP‐H6 (T22-GFP-H6 vs. T22‐GFP‐TD‐H6 vs. T22-GFP-TD-H6 at a concentration) at a respective respective concen‐ tration) are represented are represented by * at p