This is the accepted version of the journal article: López-Laguna, Hèctor; Parladé, Eloi; Álamo, Patricia; [et al.]. «In vitro fabrica- tion of microscale secretory granules». Advanced Functional Materials, Vol. 31, Issue 21 (May 2021), art. 2100914. DOI 10.1002/adfm.202100914 This version is available at https://ddd.uab.cat/record/264010 under the terms of the license In vitro fabrication of microscale secretory granules Hèctor López-Laguna, Dr. Eloi Parladé, Dr. Patricia Álamo, Dr. Julieta M. Sánchez, Eric Voltà-Durán, Dr. Naroa Serna, Dr. Laura Sánchez-García, Dr.Olivia Cano-Garrido, Dr. Alejandro Sánchez-Chardi, Prof. Antonio Villaverde*, Prof. Ramón Mangues*, Dr. Ugutz Unzueta* and Dr. Esther Vázquez Hèctor López-Laguna, Dr. Eloi Parladé, Dr. Julieta M. Sánchez, Eric Voltà-Durán, Dr. Naroa Serna, Dr. Laura Sánchez-García, Dr. Esther Vázquez, Prof. Antonio Villaverde Institut de Biotecnologia i de Biomedicina, Universitat Autònoma de Barcelona, Bellaterra, 08193 Barcelona, Spain Hèctor López-Laguna, Dr. Eloi Parladé, Dr. Julieta M. Sánchez, Eric Voltà-Durán, Dr. Naroa Serna, Dr. Laura Sánchez-García, Dr. Ugutz Unzueta, Dr. Esther Vázquez, Prof. Antonio Villaverde Departament de Genètica i de Microbiologia, Universitat Autònoma de Barcelona, Bellaterra, 08193 Barcelona, Spain Hèctor López-Laguna, Dr. Eloi Parladé, Dr. Julieta M. Sánchez, Eric Voltà-Durán, Dr. Naroa Serna, Dr. Laura Sánchez-García, Prof. Ramón Mangues, Dr. Ugutz Unzueta, Dr. Esther Vázquez, Prof. Antonio Villaverde CIBER de Bioingeniería, Biomateriales y Nanomedicina (CIBER-BBN), C/ Monforte de Lemos 3-5, 28029 Madrid, Spain Dr. Julieta M. Sánchez Instituto de Investigaciones Biológicas y Tecnológicas (IIBYT) (CONICET‐Universidad Nacional de Córdoba), ICTA & Cátedra de Química Biológica, Departamento de Química, FCEFyN, UNC. Av. Velez Sarsfield 1611, X 5016GCA Córdoba, Argentina Dr. Olivia Cano-Garrido, Nanoligent SL, Edifici Eureka, Universitat Autònoma de Barcelona, Bellaterra, 08193 Barcelona, Spain Dr. Alejandro Sánchez-Chardi Servei de Microscòpia, Universitat Autònoma de Barcelona, Bellaterra, 08193 Barcelona, Spain Dr. Alejandro Sánchez-Chardi Departament de Biologia Evolutiva, Ecologia i Ciències Ambientals, Facultat de Biologia, Universitat de Barcelona, Av. Diagonal 643, 08028 Barcelona, Spain Dr. Patricia Álamo, Prof. Ramón Mangues, Dr. Ugutz Unzueta Institut d'Investigacions Biomèdiques Sant Pau and Josep Carreras Research Institute, Hospital de la Santa Creu i SantPau, 08041 Barcelona, Spain Keywords: protein materials, microparticles, divalent cations, secretory granules; drug release * Corresponding authors: AV antoni.villaverde@uab.cat RM rmangues@santpau.cat UU uunzueta@santpau.cat 1 In diverse clinical settings, the application of advanced medical treatments and drug delivery approaches require the development of fully biocompatible materials with the ability to release functional drugs in a time-prolonged way. Ideally, the delivered molecules should be self-contained in the form of a chemically homogenous entity, to prevent the use of potentially toxic scaffolds or holding matrices. In nature, peptidic hormones are self-stored in protein-only secretory granules formed by the reversible coordination of Zn2+ and histidine residues. Inspired by this concept, we have developed, analyzed and comparatively applied an in vitro transversal procedure for the fabrication of protein-only secretory granules at the microscale, aimed to sustained protein drug delivery. These materials can be produced out from any polyhistidine-tagged protein by using physiological concentrations of Zn2+ as a potent and versatile glue-like agent. By the screening of granules formed by 12 engineered and non-engineered proteins at different Zn2+ concentrations, we have defined optimal fabrication conditions and determined the consequent further release kinetics. Moreover, we have assessed the functional and structural properties of the delivered protein using a drug- targeting protein platform in a mouse model of human colorectal cancer. Finally, the incorporation of short histidine tags allows the packaging of structurally and functionally dissimilar polypeptides, what validates the proposed fabrication method as a standard, generic and powerful protocol extensible to diverse scenarios in which slow protein drug delivery is required. The capability of a material to release drugs in a sustained way is appealing in the clinical setting, as its use might provide steady drug concentrations at local or systemic levels without repeated dose administration or under a lower dosing frequency. [1] Because of saved costs and improved patient comfortability new drug delivery platforms based on this concept are 2 under development. For that, matrices constructed with selected materials are being generated so that when arranged in a sponge-like form, they are ready to entrap different types of drug molecules for their time-prolonged delivery, once implanted in the body. [2] Therefore, the final form of such usable structures is a hybrid composite formed by the scaffold and the drug. As in the case of nanocarriers for intracellular drug delivery, [3] the material used as container might present toxicity issues thus posing severe regulatory concerns [4] and then, its use should be desirably avoided. In the case of nanoconjugates, ‘taking the vehicle out’ [3] of the delivery system would involve the design of self-assembled, self-delivered drugs, a goal that has been reached at different extents and by taking different strategies, [5, 6] and that is particularly suited for protein-based therapies of cancer. [7] Regarding sustained delivery systems, avoiding a scaffold in the drug reservoir would imply developing platforms in which the drug is self-clustered in absence of any supportive matrix, into mechanically stable but disintegrable materials from which the building blocks should be progressively released (Figure 1). In nature, the appropriate levels of peptide and protein hormones are reached by their delivery into blood from secretory granules, [8] namely non-toxic amyloidal clusters formed by the hormone itself. [9, 10] The Zn ion (Zn2+) is the molecular glue involved in hormone packaging through its divalent coordination with histidine residues from different polypeptides. [11] Being Zn2+ a nontoxic metal largely occurring in the human body for multiple functions, [12] its in vitro use as a glue-like agent for the generation of protein-only clusters, mimicking secretory granules, should be perfectly feasible. In fact, the exploration of Zn2+ for the generation of artificial protein complexes at both nano- or microscales has already started from different approaches and aimed at different biomedical uses, [13, 14, 15] resulting in progressively developed procedures for the fabrication of promising protein materials. [14] Recently, Zn2+ and other divalent cations have been adapted to generate protein-only microparticles through the coordination with overhanging histidine rich tails or even with 3 intrinsic histidine-rich regions.[16, 17] In vitro, these protein clusters release the forming protein in a time sustained way during a slow disintegration process, [16, 17] that is mirrored in vivo upon subcutaneous implantation of the material. [17] The released protein is functional and capable of clinically-oriented activities such as precise tumor cell targeting and selective destruction of cancer tissues through highly cytotoxic microbial toxins. [17] In fact, these protein clusters, sizing a few micrometers, have been named artificial inclusion bodies (ArtIBs),[17] as they are structurally and functionally similar to a category of non-toxic functional amyloids (namely natural IBs) found in recombinant bacterial cells [18, 19]. Upon injection in animal models, such natural IBs act as dynamic depots of the forming protein, [20] but their heterogenous composition (that includes undetermined amounts of bacterial macromolecules) prevent their systemic use in humans. Regarding the ability of artificial protein granules to disintegrate in functional building blocks, it would be interesting to explore to which extent, the capability to release functional protein can be tailored, envisaging particular clinical purposes for which different protein drug releasing rates would be convenient. For that, we have constructed Zn2+-induced clusters of T22-GFP-H6 to investigate them as depots of releasable protein and to test the regulatable nature of the disintegration process. T22-GFP-H6 (Figure 2 A) is a modular protein previously developed as a targeting agent in cancer therapies. [21, 22] Zn2+ was mixed with protein solutions at increasing concentrations of the metal by a simple procedure (Figure 2 A), rendering the fast formation of increasing amounts of insoluble aggregates (Figure 2 B). A complete precipitation of the soluble protein was already achieved at around 10 mM of the metal, conforming pellets that were fully fluorescent in all cases (Figure 2 B). This fact indicated a proper folding of GFP even in such clustered form. In this regard, it has been recently shown that the GFP oligomerization does not disturb the emission capacity of the protein, whose stability might be even enhanced. [23] The aggregated material appeared as discrete fluorescent particles with slightly irregular surfaces (Figure 2 C), with round or 4 oblong shapes and of around 4-5 µm in diameter (Figure 2 D). In these particles Zn2+ was detected homogeneously only in the bulk structure and clearly dependent on the Zn2+ concentration used for fabrication (Figure 2 D). The surface of the microparticles slightly changed from rough to smooth, at increasing concentrations of the crosslinking metal (Figure 2 D), indicative of a more compact architecture and protein packaging as the availability of metal increases. During incubation with conventional physiological phosphate-buffered saline (PBS) and under gentle agitation, microparticles formed by different Zn2+ amounts progressively released the forming protein until the loss of 100 % of the material and the full disintegration of the cluster (Figure 3 A). The lowest concentration of Zn2+, however, generates granules with a fast protein leakage, while the highest metal doses result in more stable materials with a smoother protein release profile. This finding indicates how the compactness and stability of the original material can be regulated by adjusting the amount of the linking divalent cation during the fabrication process. Concentrations of Zn2+ around 10 mM showed an intermediate behavior, resulting in materials with tightly attachment but still highly releasable protein. In all cases, the solubilized T22-GFP-H6 occurs as small oligomers of around 13 nm in size (Figure 3 B), similar to the form adopted by the soluble T22-GFP-H6. This protein, when purified from producing bacterial cells, forms regular toroid nanoparticles of this size. [24] It must be noted that the protein packaged as granules showed slightly lower fluorescent emission than the soluble version (Figure 3 C). This fact could be due to a self-quenching event occurring in protein clusters, although minor conformational changes linked to dissociation from the bulk material cannot be discarded, as the protein is structurally flexible. [25] Granule disintegration was, as expected, stimulated by EDTA but progressively inhibited by Zn2+ (Figure 3 D) at increasing concentrations, an observation that fully confirmed the structural role of the metal in the cluster architecture. 5 All the above data clearly demonstrated the feasibility to control both the architecture at the microscale and the protein release profile of Zn2+-driven microparticles, which behave as depot materials with stable secretion properties. However, to evaluate at which extent the proposed technological platform is extensible to proteins other than the used model, a set of protein species related and unrelated to T22-GFP-H6 was selected for further testing (Figure 4). Most of them carry a conventional hexahistidine (H6) tag either placed at the N- or C- terminus or accommodated in the protein core, whereas other ones carry the unconventional H3A, H5T and H5E tags, with a lower number of histidine residues. Moreover, the histidine- containing, tag-less human serum albumin (HSA) and bovine serum albumin (BSA) were also included. When tested for their capability to precipitate as Zn2+-driven microparticles, all these proteins responded to the presence of the divalent cation by aggregating in concentration-dependent profiles comparable to that of the control T22-GFP-H6 (Figure 5). However, the responsiveness to the crosslinking properties of Zn2+ was clearly distinguishable among the protein set, ranging from a very sensitive, aggregation-prone protein (namely T22- GFP-H6LOOP, with a non-terminal solvent-exposed H6 tag) to the hardly clustering tag-less HSA and BSA (Figure 5). Proteins carrying the unconventional histidine tags were equally responsive than T22-GFP-H6. Protein sensitivity to Zn2+ as a crosslinker was numerically determined by calculating a max P% (the maximum percentage of protein precipitated by a defined Zn2+ concentration), that rendered a scale of tendencies by which proteins could be ordered. A numerical analysis of these data confirmed that the presence of a histidine tag (versus the naturally occurring histidines in the primary protein sequence) dramatically favored aggregation (Figure 6A). On the other hand, neither the scaffold (Figure 6B), the presence of a highly cationic peptide (Figure 6C) or the particular number of histidine residues in the protein sequence (Figure 6D) presented a statistically significant influence on microparticle formation. However, clear tendencies were indeed observed, where the capacity to reach maximum precipitation decreased compared to the model T22-GFP-H6 (Figure 6). 6 To fully assess the proposed methodology as a feasible fabrication process for usable pharmacological biomaterials, we selected T22-GFP-H6 as a fluorescent model protein for a final analysis. This polypeptide self-assemble as CXCR4-targeted, 12 nm-nanoparticles, that have been recently used to selectively deliver conventional antitumoral drugs [21] and cytotoxic fusion proteins [26] to CXCR4+ human cancers. A 9.8 mM cationic Zn concentration (corresponding to a metal: protein ratio of 50:1), was selected for the fabrication test. This cation dose promotes an efficient aggregation process (Figure 1B) that renders stable, well formed microparticles (Figure 1D), with a robust, time-prolonged protein release profile (Figure 3A). In vitro, protein liberation initiated at 5 h, and all the forming protein was released within a 6-7 day period. Then, the granules constructed under these conditions appear as an optimal intermediate architecture between that promoting an immediate disintegration of the material (< 8 mM) and that showing too tight protein retention (> 19 mM) for a drug depot (Figure 3A). The fabrication process of T22-GFP-H6 granules rendered a very repetitive efficiency in the packaging of the protein as regular microparticles (Figure 7A), with a packaging efficiency higher than 75 % and with less than 10 % variability between six independent fabrication batches. The granule size was also reproducible comparing those batches (around 1 micron, Figure 7B), with only a moderate reduction upon sonication (Figure 7B). This fact reflects the mechanical stability inherent to the protein material, that was confirmed by the stability of its fluorescence emission (Figure 7A, inset). The protein released in vitro by these granules is organized in 11-12 nm oligomers, which resulted also stable in front of the mechanical stress posed by sonication (Figure 7C). The released nanoparticles kept the CXCR4-targeting conferred by the N-terminal peptide T22, [27] as demonstrated by the inhibition of protein internalization into CXCR4+ cells by the CXCR4 antagonist AMD3100 [28] (Figure 7D). Then, since the T22 targeting was fully functional in the released material, we decided testing the tumor targeting properties of T22-GFP-H6 administered to animals in form of secretory 7 granules. For that, 1 mg of secretory granules were subcutaneously injected in a mouse model of human CXCR4+ colorectal cancer, in an injection point remote to the tumor. At different times, GFP fluorescence emission from tumor, liver and kidney was recorded ex vivo. As observed (Figure 7E), the fluorescence levels at the injection point decreased progressively at least during 10 days, while a fluorescent signal concomitantly increased in tumor but not in liver or kidney (Figure 7E). The weight of the administered animals increased normally without differences comparing to a control group injected only with buffer (Figure 7F), indicative of absence of acute toxicity associated to the subcutaneous administration of secretory granules. Divalent cation coordination mainly based on Zn2+ but also on other cations (such as Cu2+ and Ca2+) is a natural mechanism to assemble proteins, and in fact, it is the molecular basis of amyloid formation and stable architecture. [29] Among non-functional amyloids, whose occurrence and relevance as natural regulators is progressively revealed in a diversity of organisms and systems, [10, 30, 31] an important fraction of them acts as physiological protein depots, whose components are progressively released to the medium upon appropriate stimuli. [8, 11, 32] Mainly because of their structural properties, amyloids are explored as components of novel biomaterials with supportive functions, [31, 33] but their potential as disintegrable depots has been much more neglected. Here we demonstrate the feasibility to generate, in a controlled fabrication process, microscale protein-only particles by using physiological concentrations of Zn ion and by means of simple procedures. These particles mimic in morphology and molecular architecture the category of bacterial amyloids named IBs [18] but also the secretory granules of the human endocrine system whose activities are modulated by the Zn2+-histidine coordination. By the appropriate setting of the in vitro fabrication conditions, both the protein clustering (Figure 2) and further release (Figure 3, 6) kinetics can be easily controlled. In addition, the comparative examination of 12 related and unrelated 8 proteins (Figure 4, 5) has revealed important key factors in the performance of such artificial secretory granules, mainly related to the presence, nature and positioning of a histidine tag (Figure 6). By the appropriate selection of the metal concentration, secretory granules with a release profile convenient for in vivo applications have been constructed, through a simple, robust and reproducible fabrication process. In this context, the feasibility of the applicability of the proposed platform has been fully validated from the biofabrication side and also regarding the in vivo performance and biosafety of the resulting protein depots (Figure 7). Of course, the platform proposed here is restricted to protein drugs and it excludes chemicals (Figure 1), but the increasing prevalence of proteins approved by clinical use [34] shows a wide spectrum of potential clinical landscapes and markets. These insights should make possible a semi-rational design of protein-only artificial secretory granules (formed by the drug as the unique component) for sustained protein drug release, aimed to overcome the obstacles potentially posed by external drug holders such as hydrogels, porous materials and polymers regarding biodegradability and biosafety issues. Conclusion The feasibility to produce artificial secretory granules as functional depots for protein drugs has been experimentally supported by the controlled coordination of ionic Zn2+ and histidine rich regions, naturally present or engineered in such proteins. By simple physicochemical procedures combined with plain protein engineering, the aggregation properties of a target protein and its ability to be released from the aggregates can be controlled. The data set presented here totally supports the development of disintegrable microscale protein materials, fully usable in vivo, in which the unique building block is the protein drug itself. The convenience of these secretory granules for clinical application relies in the absence of xenobiotic or potentially toxic holding matrices combined with the enormous versatility of the 9 Zn 2+-His platform, which is suited to fine tailoring to specific purposes by simple protein engineering approaches. Experimental Section Protein genetic design, production and purification Tested proteins were grouped in three different major categories defined as Green fluorescent protein-, Toxin- and Serum Albumin-based constructs. Firstly, HSA (hemoderivative) and BSA (Bovine Serum Albumin Fraction V) proteins without any histidine tag were commercially provided by Grifols and Merck respectively. The other proteins (except HSA- H6) were designed as Escherichia coli codon optimized genes, introduced into a pET22b using NdeI and HindIII restriction enzymes and provided by GeneArt (ThermoFisher). Recombinant plasmids were then transformed in E. coli Origami B cells (BL21, OmpT−, Lon−, TrxB, Gor−; Novagen) by thermal shock for 45 sec at 42 ºC, bacterial cells grown in Lysogenic Broth (LB) medium supplemented with 100 µg mL-1 ampicillin, 12.5 µg mL-1 tetracycline and 15 µg mL-1 kanamycin, and protein produced at 20 ºC overnight (T22-GFP- H6, GFP-H6, T22-GFP-H6LOOP, T22-GFP-H3A, T22-GFP-H5T, T22-GFP-H5E, T22-PE24- H6 and T22-DITOX-H6) or 37 ºC during 3 h (H6-GFP-T22) upon induction with 0.1 mM of isopropyl-β-D-thiogalactopyranoside (IPTG). Cells were then harvested by centrifugation for 15 min at 5,000 g, resuspended in Wash buffer (20 mM Tris-HCl, 500 mM NaCl, 10 mM imidazole, pH 8) in presence of protease inhibitors (cOmplete EDTA-free, Roche Diagnostics) and disrupted by two rounds of 1200 psi in a French Press (Thermo FA-078A). Soluble fraction from bacterial cell extract was finally collected by centrifugation (45 min at 15,000 g) for subsequent protein purification. HSA-H6 protein was also designed as a human codon optimized gene and provided by GeneArt (ThermoFisher) into a pTriex6 plasmid. Transient transfection of human HEK-293F cells was achieved upon addition of PEI:plasmid 10 complexes (3:1 ratio) in Freestyle 293 medium when cell concentration reached 106 cells mL- 1 . Cells were subsequently grown and secretion signal peptide containing protein produced in presence of 4 mM Valproic acid for 6 days (37 ºC, 70 % humidity, 8 % CO2 and 120 rpm). Finally, extracellular medium was collected by centrifugation (15 min at 300 g) for subsequent purification of the secreted protein. All proteins were purified by Immobilized Metal Affinity Chromatography (IMAC) using HisTrap HP 1 or 5 mL columns (GE Healthcare) in an Äkta Pure System (GE Healthcare). Protein elution was achieved by a linear gradient of Elution buffer (20 mM Tris-HCl, 500 mM NaCl, 500 mM imidazole, pH 8) and recovered proteins dialyzed against sodium carbonate (166 mmol L−1 NaHCO3, pH 8) with or without salt (333 mM NaCl) buffers. Protein concentration, purity and integrity Protein purity was determined after purification by SDS-PAGE using TGX Stain-FreeTM FastCastTM Acrylamide Kit, 12 % (BioRad), followed by a subsequent transference into PVDF membranes by Trans-Blot® TurboTM Transfer System (BioRad). Protein bands were finally immunodetected by Western Blot (WB) using the corresponding monoclonal antibodies (1/5,000 dilution for both an Anti-GFP Ab for GFP scaffold-based; Clontech, and an anti-His Ab for toxin or serum albumin-based proteins; Santa Cruz Biotechnology). Protein concentration was determined by Bradford assay and integrity assessed by Matrix-assisted Laser Desorption Ionization Time-of-Flight (MALDI-TOF) mass spectrometry. Fabrication of protein particles Pure soluble protein was initially adjusted at 2 mg mL−1 and aliquoted in fixed final volumes of 250 µL. Then, a 0.22 µm filtered version of ZnCl2 solution was added in each Eppendorf tube at increasing divalent cation:protein proportions in which 1:1 corresponds to 0.196 mM. The overall selected precipitation conditions were (5:1, 20:1, 40:1, 50:1, 70:1, 100:1, 150:1) 11 besides for some constructions was extended to (5:1, 10:1, 20:1, 30:1, 40:1, 50:1, 70:1, 100:1, 150:1) or (5:1, 10:1, 20:1, 30:1, 40:1, 42:1, 44:1, 46:1, 48:1, 50:1, 70:1, 100:1, 150:1) depending on experimental needs. Mixtures were then gently vortexed for 5 s and centrifuged for 15 min at 15,000 g and soluble and insoluble fractions isolated. Remaining protein in the soluble fraction was quantified by Bradford assay and percentage of precipitated protein thus estimated. Pellets were then stored at -80 ºC for further use. Physicochemical characterization Fluorescence of different concentrations (from 0.2 to 1 mg mL-1) of T22-GFP-H6 particle samples was determined at 510 nm upon excitation at 488 nm with an 5 mm emission slit in a Cary Eclipse Fluorescence Spectrophotometer (Agilent Technologies) in comparison to its respective T22-GFP-H6 soluble protein version by using quartz cells with a 10 mm path of light. The same samples were afterwards imaged and the fluorescence detected at 0.5 mg mL-1 final concentration and 10x magnifications in a Nikon-Eclipse Ts2R-FL fluorescence microscope by using the same previous wavelength conditions. For stability experiments, protein particles were resuspended in 250 µL of respective storage buffer and sonicated (0.5 on and 0.5 off) for 40 sec at 10 % of amplitude in a Bransons Digital Sonifier. Morphometric and analytical imaging High-resolution imaging of granule morphology (size and shape) and elemental analyses at nearly native state were performed with a rapid method for field emission scanning electron microscopy (FESEM). Pellets of the material were resuspended in deionized water, and 10 µL of each sample were placed in silicon wafers (Ted Pella) for 2 min, air-dried and observed in a FESEM Merlin (Zeiss). The microscope was equipped with a standard secondary electron 12 (SE) detector operating at 1 kV for morphological assessment and an energy dispersive X- rays (EDX) detector (Oxford Instruments) working at 18 kV for analytical evaluation. For the same particles, images of morphological and both 2D (line scan, 300 s) and 3D (mapping, 500 s) analyses of Zn localization were obtained to link morphology with elemental distribution. Release of soluble protein Among all manufacturing conditions, 40:1, 50:1, 100:1 and 150:1 ratios were selected to test the capacity of the resulting particles to release soluble protein. Stored pellets were thawed at room temperature for several min and softly resuspended in 1 mL of PBS (Phosphate- Buffered Saline). Samples were placed at 37ºC without agitation (mimicking subdermal conditions in living systems) and aliquots of 75 µL collected at 0, 0.5, 1, 3, 5, 24, 72, 120, 168 and 240 h. Soluble and insoluble fractions were isolated after centrifugation for 15 min at 15,000 g and protein mobility and integrity assessed by SDS-PAGE using TGX Stain-Free TM FastCastTM Acrylamide Kit, 12 % (BioRad). Protein bands were immunodetected by Western Blot (WB). Band intensity was then quantified by Image Lab software, what allowed to calculate the percentage of released protein. Rational control of soluble protein release Among all manufacturing conditions, 100:1 was selected to test the capacity to release soluble protein in presence of different chemical compounds, namely EDTA and ZnCl2, acting as stimulators or inhibitors of protein delivery respectively. Stored pellets were thawed at room temperature for several minutes and softly resuspended in 1 mL of PBS in presence of 19.6 mM of EDTA or increasing ZnCl2 concentrations (3.92, 11.76 and 15.68 mM). Samples were placed at 37ºC without agitation as previously mentioned and aliquots of 75 µL collected at times 0, 24, 72, 120 and 168 h. Then, soluble and insoluble fractions were isolated thanks to a centrifugation process for 15 min at 15,000 g and protein mobility and integrity equally 13 assessed by SDS-PAGE using TGX Stain-Free TM FastCastTM Acrylamide Kit, 12% (BioRad) and by WB. Image Lab software was used to quantify band intensity and the difference (∆) of released protein (in percentage) between control and treated samples were calculated. Size characterization of microparticles and released protein The hydrodynamic diameter (in nm) of soluble protein released in previous experiments was determined by Dynamic Light Scattering (DLS) at 25 ºC and 633 nm in a Zetasizer NanoZS (Malvern Instruments Limited) using ZEN2112 3 mm quartz batch cuvettes. The run duration value was 10 s for released protein and 4 s for microparticles. Size was determined by volume for released protein nanoparticles and by intensity for microparticles. The polydispersion index (PDI) was also obtained. Internalization of released protein Protein internalization was monitored in SW1417 colorectal cancer cell lines both patient- derived CXCR4+ and ATCC CCL-238 CXCR4. Cells were cultured in 24-well plates (120,000 cells/well for CXCR4+ and 180,000 cells/well for CXCR4-) for 24 h in DMEM (1X) GlutaMAXTM-I (4.5 g/L D-Glucose and Pyruvate; Gibco) medium supplemented with fetal bovine (FBS) at 37 ºC and 5% CO2 humidified atmosphere until a confluence of 70 % was reached. Protein was incubated at 0.1 µM for 1h in both cell lines. Cells were then detached, and external attached protein removed by adding Trypsin-EDTA (Gibco) at 1 mg.mL-1 for a harsh digestion for 15 min at 37 ºC. Such protocol has been designed to remove all potential externally attached protein. Protein internalization, detected as intracellular fluorescence, was then determined by flow cytometry using a FACS-Canto system (Becton Dickinson) at 15 mW with an air-cooled argon laser exciting at 488 nm. Receptor mediated internalization was assessed by adding AMD3100 at 10 µM, that blocks CXCR4 and prevents its interaction with the CXCR4 ligand T22, 1 h prior protein incubation. 14 Standardization of the precipitation curve and statistical analyses A protein precipitation curve was represented for each modular protein type as percentage of material formation in respect of Zn2+ concentration (in mM), and statistically analyzed by an asymmetric sigmoidal 5 parameter logistic (PL) model. From curve fitting, R square (R2) and maximum percentage of precipitated protein (max P%) were numerically obtained and graphically represented as displayed in Figure 1B. Max P% protein values were compared with the presence or absence of T22 ligand, histidine tag, protein scaffold or number of histidines in the primary sequence, respectively. An initial variety of normality and lognormality tests (Anderson-Darling, D’Agostino & Pearson, Shapiro-Wilk and Kolmogorov-Smirnov) were used to determine the well-modeled normal distribution of all data sets. For parametric data, a t-test (single comparisons) or one/two-way ANOVA (multiple comparisons) tests were used, meanwhile for non-parametric data, a Mann-Whitney test was used. All statistical analyses were performed in comparison with the control or initial specified time = 0 respectively (*p ≤ 0.05), expressed as mean ± standard error of the mean (SEM) and at least in triplicate (n=3). In vivo kinetic biodistribution from subcutaneously implanted protein particles In vivo experiments were approved by the Animal Ethics Committee at the Hospital de la Santa Creu i Sant Pau and performed according to European Council directives. Four-week- old female mice of the Swiss nude strain, in the 18-20 g body weight range (Charles River, L- Abreslle, France), maintained in pathogen-free conditions, were used for the in vivo experiments. To generate the CXCR4+ SW1417 CRC cancer model, 5 × 106 CXCR4+ SW1417 human CRC cells, resuspended in 150 μL of media, were injected in the mouse flank (n=12). When tumors reached approximately a 120-200 mm3 volume, animals were randomly allocated to the different groups and administered in the subcutis of 15 the mouse lumbar region, at the site contralateral to the tumor site, with a pellet of 1mg of T22-GFP-H6 Zn2+ granules suspended in a 150 μL PBS buffer. Buffer injection was used as a negative control. Ten minutes, and 5 or 10 days after the administration, we drew approximately 1 ml of blood in EDTA anticoagulated collection tubes. In these samples we measured the exact volume of obtained plasma and their fluorescent emission at each studied time, and calculated the concentration of nanoparticle as referred to the total fluorescence emitted by the administered dose. We euthanized each mouse, resected the kidney, liver and tumor and the tissue surrounding the injection point. Following, we ex vivo registered the intensity of the fluorescence emitted by the nanoparticles released by the protein particles and biodistributed to tumor and non-tumor organs and also the fluorescence remaining at the IP, using the IVIS® 200‐Spectrum (PerkinElmer). Fluorescent signal was digitalized, displayed as pseudocolor overlay, and expressed as radiant efficiency. Statistical analysis was performed in GraphPad Prism (v8.0.2). Unpaired t-test was used to evaluate differences between means of the average radiant efficiencies extracted from each IVIS images, and at least three replicates were analyzed for each time-point. Acknowledgements We are indebted to Agencia Estatal de Investigación (AEI) and to Fondo Europeo de Desarrollo Regional (FEDER) (grant BIO2016-76063-R, AEI/FEDER, UE) to AV, AEI (PID2019-105416RB-I00/AEI/10.13039/501100011033) to EV and PI18/00650 AEDI/FEDER to RM), AGAUR (2017SGR-229 to AV and 1027SGR-865 to RM) and CIBER-BBN (project NANOPROTHER granted to AV, project 4NanoMets to RM and project NANOREMOTE granted to EV). We are also indebted to the Networking Research Center on Bioengineering, Biomaterials and Nanomedicine (CIBER-BBN) that is an initiative funded by the VI National R&D&I Plan 2008–2011, Iniciativa Ingenio 2010, Consolider Program, CIBER Actions and financed by the Instituto de Salud Carlos III, with assistance from the European Regional Development Fund. Protein production has been partially performed by the ICTS “NANBIOSIS”, more specifically by the Protein Production Platform of CIBER in Bioengineering, Biomaterials & Nanomedicine (CIBER-BBN)/ IBB, at the UAB sePBioEs scientific-technical service (http://www.nanbiosis.es/portfolio/u1-protein- production-platform-ppp/), the in vivo experimental work by the Nanotoxicology Unit (http://www.nanbiosis.es/portfolio/u18-nanotoxicology-unit/) and the nanoparticle size analysis by the Biomaterial Processing and Nanostructuring Unit. Electron microscopy studies were performed by the Servei de Microscòpia in the UAB. AV received an ICREA ACADEMIA award. UU is supported by Miguel Servet fellowship (CP19/00028) from ISCIII 16 (cofounding “Fondo Social Europeo”). LSG and HLL were supported by a predoctoral fellowship from AGAUR (2018FI_B2_00051 and 2019FI_B00352, respectively) and EVD by a predoctoral fellowship from Ministerio de Ciencia, Innovación y Universidades (FPU18/04615). Received: ((will be filled in by the editorial staff)) Revised: ((will be filled in by the editorial staff)) Published online: ((will be filled in by the editorial staff)) References [1] C. Helary, M. F. Desimone, Current pharmaceutical biotechnology 2015, 16, 635; G. Acharya, K. Park, Advanced drug delivery reviews 2006, 58, 387; A. S. Hoffman, Journal of controlled release : official journal of the Controlled Release Society 2008, 132, 153. [2] J. Li, D. J. Mooney, Nature reviews. Materials 2016, 1; T. A. Esquivel-Castro, M. C. Ibarra-Alonso, J. Oliva, A. Martinez-Luevanos, Materials science & engineering. C, Materials for biological applications 2019, 96, 915; R. Narayan, U. Y. Nayak, A. M. Raichur, S. Garg, Pharmaceutics 2018, 10; A. Ali, S. Ahmed, International journal of biological macromolecules 2018, 109, 273; N. T. T. Uyen, Z. A. A. Hamid, N. X. T. Tram, N. Ahmad, International journal of biological macromolecules 2020, 153, 1035. [3] J. Shen, J. Wolfram, M. Ferrari, H. Shen, Materials today 2017, 20, 95. [4] H. Park, K. Park, Pharmaceutical research 1996, 13, 1770; S. Al-Jawadi, P. Capasso, M. Sharma, Pharmaceutical development and technology 2018, 23, 953; D. da Silva, M. Kaduri, M. Poley, O. Adir, N. Krinsky, J. Shainsky-Roitman, A. Schroeder, Chemical engineering journal 2018, 340, 9. [5] L. Sanchez-Garcia, N. Serna, P. Alamo, R. Sala, M. V. Cespedes, M. Roldan, A. Sanchez-Chardi, U. Unzueta, I. Casanova, R. Mangues, E. Vazquez, A. Villaverde, Journal of controlled release : official journal of the Controlled Release Society 2018, 274, 81. [6] P. Huang, D. Wang, Y. Su, W. Huang, Y. Zhou, D. Cui, X. Zhu, D. Yan, Journal of the American Chemical Society 2014, 136, 11748; Y. Zhao, W. Wang, S. Guo, Y. Wang, L. Miao, Y. Xiong, L. Huang, Nat Commun 2016, 7, 11822; J. E. Chung, S. Tan, S. J. Gao, N. Yongvongsoontorn, S. H. Kim, J. H. Lee, H. S. Choi, H. Yano, L. Zhuo, M. Kurisawa, J. Y. Ying, Nature nanotechnology 2014, 9, 907; N. Serna, P. Alamo, P. Ramesh, D. Vinokurova, L. Sanchez-Garcia, U. Unzueta, A. Gallardo, M. V. Cespedes, E. Vazquez, A. Villaverde, R. Mangues, J. P. Medema, Journal of controlled release : official journal of the Controlled Release Society 2020, 320, 96. [7] N. Serna, L. Sanchez-Garcia, U. Unzueta, R. Diaz, E. Vazquez, R. Mangues, A. Villaverde, Trends in biotechnology 2018, 36, 318. [8] S. K. Maji, M. H. Perrin, M. R. Sawaya, S. Jessberger, K. Vadodaria, R. A. Rissman, P. S. Singru, K. P. Nilsson, R. Simon, D. Schubert, D. Eisenberg, J. Rivier, P. Sawchenko, W. Vale, R. Riek, Science 2009, 325, 328. [9] M. P. Badtke, N. D. Hammer, M. R. Chapman, Science signaling 2009, 2, pe43. [10] D. Otzen, R. Riek, Cold Spring Harbor perspectives in biology 2019, 11, pii: a033860. [11] R. S. Jacob, S. Das, S. Ghosh, A. Anoop, N. N. Jha, T. Khan, P. Singru, A. Kumar, S. K. Maji, Scientific reports 2016, 6, 23370; R. Jacob, A. Anoop, S. Maji, Protein Nanofibrils as Storage Forms of Peptide Drugs and Hormones, Vol. 1174 Springer, Singapore, Singapore 2019. 17 [12] G. A. Rutter, P. Chabosseau, E. A. Bellomo, W. Maret, R. K. Mitchell, D. J. Hodson, A. Solomou, M. Hu, The Proceedings of the Nutrition Society 2016, 75, 61; M. A. Zoroddu, J. Aaseth, G. Crisponi, S. Medici, M. Peana, V. M. Nurchi, Journal of inorganic biochemistry 2019, 195, 120; Y. Cherasse, Y. Urade, International journal of molecular sciences 2017, 18, pii: E2334; Y. Ogawa, M. Kinoshita, S. Shimada, T. Kawamura, Nutrients 2018, 10, pii: E199; K. Kerns, M. Zigo, P. Sutovsky, International journal of molecular sciences 2018, 19, pii: E4097; H. Wu, Y. Shen, D. Wang, H. Herrmann, R. D. Goldman, D. A. Weitz, bioRxiv (CSH Laboratory) 2019, https://doi.org/10.1101/844167. [13] H. K. Munch, J. Nygaard, N. J. Christensen, C. Engelbrekt, M. Ostergaard, T. Porsgaard, T. Hoeg-Jensen, J. Zhang, L. Arleth, P. W. Thulstrup, K. J. Jensen, Angewandte Chemie 2016, 55, 2378; I. W. Hamley, Biomacromolecules 2019, 20, 1829; B. Maniaci, C. H. Lipper, D. L. Anipindi, H. Erlandsen, J. L. Cole, B. Stec, T. Huxford, J. J. Love, Biochemistry 2019, 58, 2199; I. Tunn, M. J. Harrington, K. G. Blank, Biomimetics 2019, 4, 25; M. Yang, W. J. Song, Nat Commun 2019, 10, 5545; A. S. Knight, J. Larsson, J. M. Ren, R. Bou Zerdan, S. Seguin, R. Vrahas, J. Liu, G. Ren, C. J. Hawker, Journal of the American Chemical Society 2018, 140, 1409; S. Zechel, M. D. Hager, T. Priemel, M. J. Harrington, Biomimetics 2019, 4. [14] H. López-Laguna, J. Sánchez, U. Unzueta, R. Mangues, E. Vázquez, A. Villaverde, Trends in biochemical sciences 2020, 45, 992. [15] H. Lopez-Laguna, U. Unzueta, O. Conchillo-Sole, A. Sanchez-Chardi, M. Pesarrodona, O. Cano-Garrido, E. Volta, L. Sanchez-Garcia, N. Serna, P. Saccardo, R. Mangues, A. Villaverde, E. Vazquez, Acta biomaterialia 2019, 83, 257. [16] T. Y. Chen, W. J. Cheng, J. C. Horng, H. Y. Hsu, Colloids and surfaces. B, Biointerfaces 2019, 110644. [17] J. M. Sanchez, H. Lopez-Laguna, P. Alamo, N. Serna, A. Sanchez-Chardi, V. Nolan, O. Cano-Garrido, I. Casanova, U. Unzueta, E. Vazquez, R. Mangues, A. Villaverde, Advanced science 2020, 7, 1902420. [18] A. de Marco, N. Ferrer-Miralles, E. Garcia-Fruitos, A. Mitraki, S. Peternel, U. Rinas, M. A. Trujillo-Roldan, N. A. Valdez-Cruz, E. Vazquez, A. Villaverde, FEMS microbiology reviews 2019, 43, 53. [19] U. Rinas, E. Garcia-Fruitos, J. L. Corchero, E. Vazquez, J. Seras-Franzoso, A. Villaverde, Trends in biochemical sciences 2017, 42, 726. [20] M. V. Cespedes, O. Cano-Garrido, P. Alamo, R. Sala, A. Gallardo, N. Serna, A. Falgas, E. Volta-Duran, I. Casanova, A. Sanchez-Chardi, H. Lopez-Laguna, L. Sanchez- Garcia, J. M. Sanchez, U. Unzueta, E. Vazquez, R. Mangues, A. Villaverde, Advanced materials 2020, 32, e1907348; M. Pesarrodona, T. Jauset, Z. V. Díaz‐Riascos, A. Sánchez‐ Chardi, M. E. Beaulieu, J. Seras‐Franzoso, L. Sánchez‐García, R. Baltà‐Foix, S. Mancilla, Y. Fernández, U. Rinas, S. Schwartz, L. Soucek, A. Villaverde, I. Abasolo, E. Vázquez, Advanced science 2019, https://doi.org/10.1002/advs.201900849; U. Unzueta, M. V. Cespedes, R. Sala, P. Alamo, A. Sanchez-Chardi, M. Pesarrodona, L. Sanchez-Garcia, O. Cano-Garrido, A. Villaverde, E. Vazquez, R. Mangues, J. Seras-Franzoso, Journal of controlled release : official journal of the Controlled Release Society 2018, 279, 29; M. V. Cespedes, Y. Fernandez, U. Unzueta, R. Mendoza, J. Seras-Franzoso, A. Sanchez-Chardi, P. Alamo, V. Toledo-Rubio, N. Ferrer-Miralles, E. Vazquez, S. Schwartz, I. Abasolo, J. L. Corchero, R. Mangues, A. Villaverde, Scientific reports 2016, 6, 35765. [21] M. V. Cespedes, U. Unzueta, A. Avino, A. Gallardo, P. Alamo, R. Sala, A. Sanchez- Chardi, I. Casanova, M. A. Mangues, A. Lopez-Pousa, R. Eritja, A. Villaverde, E. Vazquez, R. Mangues, EMBO molecular medicine 2018, 10, pii: e8772. [22] A. Falgas, V. Pallares, U. Unzueta, M. V. Cespedes, I. Arroyo-Solera, M. J. Moreno, J. Sierra, A. Gallardo, M. A. Mangues, E. Vazquez, A. Villaverde, R. Mangues, I. Casanova, Haematologica 2020, 105, 741; A. Falgas, V. Pallares, N. Serna, L. Sanchez-Garcia, J. Sierra, 18 A. Gallardo, L. Alba-Castellon, P. Alamo, U. Unzueta, A. Villaverde, E. Vazquez, R. Mangues, I. Casanova, Theranostics 2020, 10, 5169. [23] U. Unzueta, M. Roldan, M. Pesarrodona, R. Benitez, A. Sanchez-Chardi, O. Conchillo-Sole, R. Mangues, A. Villaverde, E. Vazquez, Acta biomaterialia 2020, 103, 272. [24] F. Rueda, M. V. Cespedes, O. Conchillo-Sole, A. Sanchez-Chardi, J. Seras-Franzoso, R. Cubarsi, A. Gallardo, M. Pesarrodona, N. Ferrer-Miralles, X. Daura, E. Vazquez, E. Garcia-Fruitos, R. Mangues, U. Unzueta, A. Villaverde, Advanced materials 2015, 27, 7816. [25] J. M. Sanchez, L. Sanchez-Garcia, M. Pesarrodona, N. Serna, A. Sanchez-Chardi, U. Unzueta, R. Mangues, E. Vazquez, A. Villaverde, Biomacromolecules 2018, 19, 3788. [26] N. C. Serna, M; Sánchez-García, L; Unzueta, U; Sala, R; Sánchez-Chardi, A; Cortés, F; Ferrer-Miralles, N; Mangues, R; Vázquez, E; Villaverde, A, Advanced Functional Materials 2017, 27, 1700919; N. Serna, O. Cano-Garrido, L. Sánchez-García, M. Pesarrodona, U. Unzueta, A. Sánchez-Chardi, R. Mangues, E. Vázquez, A. Villaverde, Particle & Particle Systems Characterization 2020, 37, 2000040. [27] H. Tamamura, R. Arakaki, H. Funakoshi, M. Imai, A. Otaka, T. Ibuka, H. Nakashima, T. Murakami, M. Waki, A. Matsumoto, N. Yamamoto, N. Fujii, Bioorganic & medicinal chemistry 1998, 6, 231; H. Tamamura, M. Imai, T. Ishihara, M. Masuda, H. Funakoshi, H. Oyake, T. Murakami, R. Arakaki, H. Nakashima, A. Otaka, T. Ibuka, M. Waki, A. Matsumoto, N. Yamamoto, N. Fujii, Bioorganic & medicinal chemistry 1998, 6, 1033. [28] S. Poty, P. Desogere, C. Goze, F. Boschetti, T. D'Huys, D. Schols, C. Cawthorne, S. J. Archibald, H. R. Maecke, F. Denat, Dalton transactions 2015, 44, 5004; Y. H. Jung, D. Y. Lee, W. Cha, B. H. Kim, M. W. Sung, K. H. Kim, S. H. Ahn, Head & neck 2016, 38, 1479. [29] P. Faller, C. Hureau, O. Berthoumieu, Inorganic chemistry 2013, 52, 12193; E. Jozsa, K. Osz, C. Kallay, P. de Bona, C. A. Damante, G. Pappalardo, E. Rizzarelli, I. Sovago, Dalton transactions 2010, 39, 7046; M. Rana, A. K. Sharma, Metallomics : integrated biometal science 2019, 11, 64. [30] L. Molina-Garcia, F. Gasset-Rosa, M. Moreno-Del Alamo, M. E. Fernandez- Tresguerres, S. Moreno-Diaz de la Espina, R. Lurz, R. Giraldo, Scientific reports 2016, 6, 25425; S. Shin, S. Cherry, Immunity 2017, 47, 604; M. P. Jackson, E. W. Hewitt, Biomolecules 2017, 7; A. Piscitelli, P. Cicatiello, A. M. Gravagnuolo, I. Sorrentino, C. Pezzella, P. Giardina, Biomolecules 2017, 7; A. Hewetson, H. Q. Do, C. Myers, A. Muthusubramanian, R. B. Sutton, B. J. Wylie, G. A. Cornwall, Biomolecules 2017, 7; N. Van Gerven, S. E. Van der Verren, D. M. Reiter, H. Remaut, Journal of molecular biology 2018; K. Schwartz, A. K. Syed, R. E. Stephenson, A. H. Rickard, B. R. Boles, PLoS pathogens 2012, 8, e1002744; P. Falabella, L. Riviello, M. Pascale, I. D. Lelio, G. Tettamanti, A. Grimaldi, C. Iannone, M. Monti, P. Pucci, A. M. Tamburro, M. Deeguileor, S. Gigliotti, F. Pennacchio, Insect biochemistry and molecular biology 2012, 42, 203; M. S. Dueholm, D. Otzen, P. H. Nielsen, PloS one 2013, 8, e76630; B. Guyonnet, N. Egge, G. A. Cornwall, Molecular and cellular biology 2014, 34, 2624; J. D. Taylor, S. J. Matthews, Frontiers in cellular and infection microbiology 2015, 5, 33; T. Seviour, S. H. Hansen, L. Yang, Y. H. Yau, V. B. Wang, M. R. Stenvang, G. Christiansen, E. Marsili, M. Givskov, Y. Chen, D. E. Otzen, P. H. Nielsen, S. Geifman-Shochat, S. Kjelleberg, M. S. Dueholm, The Journal of biological chemistry 2015, 290, 6457; S. M. Lyons, P. Anderson, Developmental cell 2016, 39, 131; R. D. Rameau, D. N. Jackson, A. Beaussart, Y. F. Dufrene, P. N. Lipke, mBio 2016, 7, e01815. [31] M. F. Gebbink, D. Claessen, B. Bouma, L. Dijkhuizen, H. A. Wosten, Nature reviews. Microbiology 2005, 3, 333. [32] R. A. Akhtar, M. C. Perry, Biochimica et biophysica acta 1979, 585, 117. [33] S. Mankar, A. Anoop, S. Sen, S. K. Maji, Nano reviews 2011, 2; T. P. J. Knowles, R. Mezzenga, Advanced materials 2016, 28, 6546; P. K. R. Tay, P. Q. Nguyen, N. S. Joshi, ACS 19 synthetic biology 2017, 6, 1841; G. Wei, Z. Su, N. P. Reynolds, P. Arosio, I. W. Hamley, E. Gazit, R. Mezzenga, Chemical Society reviews 2017, 46, 4661. [34] L. Sanchez-Garcia, L. Martin, R. Mangues, N. Ferrer-Miralles, E. Vazquez, A. Villaverde, Microbial cell factories 2016, 15, 33; H. A. Lagasse, A. Alexaki, V. L. Simhadri, N. H. Katagiri, W. Jankowski, Z. E. Sauna, C. Kimchi-Sarfaty, F1000Research 2017, 6, 113. [35] E. Voltà-Durán, C.-G. O, N. Serna, H. López-Laguna, L. Sánchez-García, M. Pesarrodona, A. Sánchez-Chardi, R. Mangues, A. Villaverde, E. Vázquez, U. Unzueta, Sicence China Materials 2019. [36] H. López-Laguna, R. Sala, J. M. Sánchez, P. Álamo, U. Unzueta, A. Sánchez-Chardi, N. Serna, L. Sánchez-García, E. Voltà-Durán, R. Mangues, A. Villaverde, E. Vázquez, Particle and Particle Characterization System 2019; H. López-Laguna, R. Cubarsi, U. Unzueta, R. Mangues, E. Vázquez, A. Villaverde, Science China Materials 2020, 63, 644. 20 Figure 1. The concept of artificial secretory granules for time-prolonged drug release. Conventional drug depots (top) are based on a scaffold structure (grey material) constructed by polymers, hydrogels, matrices or in general, porous composites that entrap the drug (yellow balls) in a reversible way. The drugs are then released along time in a passive fashion, while the scaffold itself remains as a stable entity. On the other hand, the scaffold imposes fabrication costs and additional complexity in the fabrication process, apart from biological risks associated to limited biocompatibility and toxicity of the material. On the contrary, the spectrum of payload drugs is potentially broad. In contrast, artificial secretory granules (bottom) are formed in absence of heterologous scaffolds, and polypeptides of a single species act as both scaffolds and releasable protein drugs. The material, that is mechanically stable, is constructed by the simple coordination of Zn2+ cations (small green dots) with histidine-rich stretches of the protein, in absence of additional agents or complex fabrication settings. The protein is slowly released under physiological conditions by still unknown mechanisms, which mimic the secretion of peptide hormones in the mammalian endocrine systems. Therefore, the whole granule disintegrates along with protein release, without leaving any remaining structure. The system is then limited to protein drugs, but it offers high versatility regarding the spectrum of suitable protein species as it is fully regulatable through the addition and engineering of Histidine tags to the desired protein. 21 Figure 2. Fabrication of secretory granules using the T22-GFP-H6 modular protein as a model. A. Schematic representation of the modular T22-GFP-H6 protein and the manufacturing process (30 min of duration as standard). B. Protein precipitation curve represented as the % of the formed material relative to Zn2+ concentration (in mM, black), with the respective R square value. The corresponding non-linear fitting curve (Asymmetric Sigmoidal 5 Parameters Logistic; PL) is represented in blue. The maximum % of protein in the aggregate is expressed as (max P%) and indicated in y axes with a red cross. Additionally, the picture displays the fluorescent pellet that is usually obtained by this process at different concentrations or the ionic Zn (mM). C. Optical imaging and fluorescence detection of granules. White bar corresponds to 100 µm. D. Representative high-resolution images linking the granule ultrastructure with 2D (line scan) and 3D (mapping) Zn2+ localization at different constructive doses of the metal. 22 Figure 3. In vitro physicochemical analysis of protein released from T22-GFP-H6 granules. A. Percentage of released protein (in regard to pellet signal after First Resuspension) at different times in granules formed at different Zn2+ doses (in mM, each colored line). B. Hydrodynamic diameter (in nm) of released protein from panel A. Polydispersion Indexes (PDI) and mean ± Standard Error of the mean (SEM) are also shown. C. Fluorescence of granules and soluble protein given in arbitrary units (au). D. Percentage of released protein from granules formed by 19.6 mM Zn2+ after different times of incubation in alternative buffers (+EDTA or +Zn2+). The increase or decrease % (Δ%) of soluble protein is also shown. Data are expressed as mean ± SEM, n = 3 and statistical comparisons in relation to the control group for panel A, C and D (*p < 0.05). 23 Figure 4. Domain organization of the fusion proteins used in the study and amino acid sequences of relevant modules. A. Modular organization of fusion proteins. The conventional His tag (hexahistidine or H6) is displayed in blue, the modified His tag versions in green and T22 peptide in black. Box sizes are only indicative. PE24 is a fragment of the Pseudomonas aeruginosa exotoxin and DITOX a fragment of the diphtheria toxin. [5] HSA is the human serum albumin and BSA the bovine serum albumin. Additional information about GFP-based constructs can be found elsewhere. [15, 24, 35] B. Sequences of protein scaffolds and peptidic tags. Note that a spacer sequence (GGSSRSS) was accommodated between T22 and each scaffold protein in all cases (not displayed in the modular scheme). More details about the modified His tags used here can be found elsewhere. [36] 24 Figure 5. Analytical modelling of granule formation. Protein precipitation curves represented as the % of material formation versus Zn2+ concentration in mM (black). Respective statistical fittings by non-linear curve model (Asymmetric Sigmoidal of 5 parameters Logistics; PL) are represented in continuous blue lines. For each plot, the R square and the maximum percentage of protein that can precipitate (max P%) are displayed. Data are expressed as mean ± SEM (n=3). 25 Figure 6. Statistical analysis comparing parameters involved in protein aggregation with intrinsic characteristics of each protein. A. Analysis of max P% depending on the presence, absence or modification of histidine tag. B. Analysis of max P% depending on the type of protein scaffold (GFP-based versus other scaffolds). C. Analysis of max P% depending on the presence or absence of T22. D. Analysis of max P% depending on the number of histidine residues in the protein sequence (more or less than 15). Data are expressed as scattered dot plots with mean ± SEM, n=3 and significance (*p ≤ 0.05). 26 Figure 7. Global validation of the depot/secretion platform, selecting 9.8 mM cationic Zn (metal: protein ration of 50:1) as a convenient model. A. Reproducibility of the fabrication process tested through the percentage of aggregated material. The insert displays the emitted fluorescence from replicates at 510 nm of both control (black) and sonicated (brown) versions of secretory granules. Data derive from fully independent processes in a sixtiplicate experiment. B. Hydrodynamic diameter, expressed in nm, of sonicated and non sonicated secretory granules. C. Hydrodynamic diameter, expressed in nm, of released protein nanoparticles. D. Protein (released nanoparticles) internalization into cultured SW1416 cells, overexpressing and non-expressing CXCR4, in presence and in the absence of the CXCR4 antagonist AMD3100. Intracellular material was determined by GFP fluorescence. E. Kinetic analysis of GFP fluorescence in the injection point, tumor and relevant major organs, at different times post-administration. At the right, an example of IVIS images at the injection point and the tumor, in a representative animal. F. Comparative weigth of treated and untreated animals before and after injection. Data expressed as mean ± SEM, n=6, and statistical analysis performed in relation to the control group (*p ≤ 0.05). 27 A novel type of microscale protein materials has been proposed to act as mechanically stable secretory granules. The use of ionic Zn to cluster histidine rich polypeptides keeps their functionality and allows the further slow release under physiological conditions. The versatility to control protein clustering and its further release profile supports the development of such materials as new biocompatible and disintegrable delivery systems for time-sustained release of protein drugs. Keyword secretory protein materials Hèctor López-Laguna, Dr. Eloi Parladé, Dr. Patricia Álamo, Dr. Julieta M. Sánchez, Eric Voltà-Durán, Dr. Naroa Serna, Dr. Laura Sánchez-García, Dr.Olivia Cano-Garrido, Dr. Alejandro Sánchez-Chardi, Prof. Antonio Villaverde*, Prof. Ramón Mangues*, Dr. Ugutz Unzueta* and Dr. Esther Vázquez In vitro fabrication of microscale secretory granules 28