López‑Cano et al. Microbial Cell Factories https://doi.org/10.1186/s12934-022-01803-7 (2022) 21:77 Microbial Cell Factories RESEARCH Open Access Exploring the impact of the recombinant Escherichia coli strain on defensins antimicrobial activity: BL21 versus Origami strain Adrià López‑Cano1, Marc Martínez‑Miguel2,3, Judith Guasch2,3,4, Imma Ratera2,3, Anna Arís1* and Elena Garcia‑Fruitós1* Abstract The growing emergence of microorganisms resistant to antibiotics has prompted the development of alternative antimicrobial therapies. Among them, the antimicrobial peptides produced by innate immunity, which are also known as host defense peptides (HDPs), hold great potential. They have been shown to exert activity against both Gram-positive and Gram-negative bacteria, including those resistant to antibiotics. These HDPs are classified into three categories: defensins, cathelicidins, and histatins. Traditionally, HDPs have been chemically synthesized, but this strategy often limits their application due to the high associated production costs. Alternatively, some HDPs have been recombinantly produced, but little is known about the impact of the bacterial strain in the recombinant prod‑ uct. This work aimed to assess the influence of the Escherichia coli strain used as cell factory to determine the activity and stability of recombinant defensins, which have 3 disulfide bonds. For that, an α-defensin [human α-defensin 5 (HD5)] and a β-defensin [bovine lingual antimicrobial peptide (LAP)] were produced in two recombinant back‑ grounds. The first one was an E. coli BL21 strain, which has a reducing cytoplasm, whereas the second was an E. coli Origami B, that is a strain with a more oxidizing cytoplasm. The results showed that both HD5 and LAP, fused to Green Fluorescent Protein (GFP), were successfully produced in both BL21 and Origami B strains. However, differences were observed in the HDP production yield and bactericidal activity, especially for the HD5-based protein. The HD5 protein fused to GFP was not only produced at higher yields in the E. coli BL21 strain, but it also showed a higher quality and stability than that produced in the Origami B strain. Hence, this data showed that the strain had a clear impact on both HDPs quantity and quality. Keywords: Host defense peptides, Escherichia coli, Strain, Recombinant protein Background the top 10 global public health threats facing humanity Infections caused by antimicrobial resistant (AMR) bac- [3]. To tackle this global challenge affecting both human teria are continuously growing, whereas available drugs and animal health, research efforts are directed to the for their treatment are limited and, in some cases, non- generation of alternative antimicrobial therapies includ- existent [1, 2]. The current situation has led the World ing phage therapy [4], lysins [5], probiotics [6], antibod- Health Organization (WHO) to declare AMR as one of ies [7], and antimicrobial proteins [8]. Among them, host defense peptides (HDPs) outstand for their broad- spectrum bactericidal activity [9, 10]. HDPs are short, *Correspondence: anna.aris@irta.cat; elena.garcia@irta.cat cationic peptides, which are naturally produced by the 1 Department of Ruminant Production, Institut de Recerca i Tecnologia innate immunity of organisms of all life forms, being key Agroalimentàries (IRTA), 08140 Caldes de Montbui, Spain Full list of author information is available at the end of the article molecules for the prevention and overcoming infections © The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. The Creative Commons Public Domain Dedication waiver (http://​creat​iveco​ mmons.​org/​publi​cdoma​in/​zero/1.​0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. López‑Cano et al. Microbial Cell Factories (2022) 21:77 Page 2 of 10 [11–13]. Besides, their fast and multiple mechanisms of Other authors have proven that defensins produced in action hamper the development of resistances [14–17]. E. coli Origami are active against different pathogenic The different HDPs have been classified into three strains [44, 45]. However, the comparison of the quality groups: defensins, cathelicidins, and histatins [18, 19]. (activity) of defensins produced in these two strains has Defensins are one of the most remarkable groups, widely not yet been evaluated. Thus, in this study, we have deter- distributed in animals and plants. Whereas invertebrate mined the production yields and activity of an α-defensin and plant defensins contain a common structure com- and a β-defensin recombinantly produced in both oxidiz- prising an α-helix linked to a β-sheet by two disulfide ing and reducing E. coli cytoplasm. For that, we have used bridges (CSαβ-motif ) [20], mammalian defensins are both the soluble form and aggregated protein forming characterized by an antiparallel β-sheet structure, stabi- inclusion bodies (IBs) of the human α-defensin 5 (HD5) lized by three disulfide bonds [13]. In addition, mamma- and β-defensin lingual antimicrobial peptide (LAP). IBs lian defensins are divided into α- and β-defensins, which are mechanically stable protein-based nanoparticles mainly differ in length, location, and connectivity of their formed during recombinant protein production pro- three pairs of intramolecular disulfide bonds, as well as in cesses [46]. These aggregates have already been shown to their unique consensus sequences [21]. The α-defensins, be a low-cost drug delivery system for different applica- which are mainly produced by neutrophils and Paneth tions, including biocatalysis [47, 48] or biomedicine, such cells in the small intestine, are 29–35 residues long, con- as antimicrobial therapy [49]. taining six cysteines which are linked as follows: C1–C6, C2–C4, and C3–C5 [22–26]. In contrast, β-defensins Results produced by epithelial cells are 38–42 residues long with Two different defensins, HD5 and LAP were selected C1–C5, C2–C4, C3–C6 pairs forming disulfide bonds to perform this study (Table 1). Both HDPs, which are [24–27]. The conserved cysteines of defensins have led peptides with hydrophobic regions as well as positively to the conclusion that correct disulfide bond formation charged amino acids, were fused to GFP as protein could be critical for biological activity, structuration, and carrier. stability of these peptides [28]. Both HD5-GFP-H6 and LAP-GFP-H6 defensins were Most studies done with defensins have used synthetic successfully produced in E. coli BL21 and Origami B forms of these peptides. However, some of them have strains, although the production profile was different also been recombinantly produced [29–32]. Unlike depending on the HDP and the strain used (Fig. 1). In chemical synthesis, the recombinant production of pep- both cases, the proteins were produced in soluble (Fig. 1 tides is an efficient and fully scalable process with no top) and insoluble (Fig. 1 bottom) forms, but the aggrega- limits in peptide length [33–36]. Generally, when using tion ratio was higher for HD5-GFP-H6 than LAP-GFP- the recombinant production strategy, defensins (and H6, especially when using the Origami B strain (Fig. 1B). in general HDPs) are fused to carrier proteins to avoid Soluble LAP-GFP-H6 had similar levels of production proteolysis [37] and minimize the toxicity of these short in both BL21 and Origami B strains, being in both cases peptides [38–40]. Some examples of fusion carriers are time- dependent (p < 0.0001) (Fig. 1 top). In contrast, the thioredoxin, glutathione S-transferase (GST), small ubiq- production kinetics of HD5-GFP-H6 showed that the uitin-related modifier (SUMO) and Green Fluorescent soluble form is produced at higher levels in BL21 than in Protein (GFP). Although different production strategies the Origami B strain (Fig. 1 top; p = 0.040). However, the have been explored to optimize defensins production, lit- aggregated form of both LAP-GFP-H6 and HD5-GFP-H6 tle is known about the disulfide bond formation of HDPs showed no differences between strains at the any produc- under recombinant conditions. This is particularly rel- tion time (Fig. 1 bottom). evant in bacterial hosts and more specifically in Escheri- Taking 3 h as production time, the two defensins were chia coli (E. coli) because it has a reducing cytoplasmic produced and purified in their soluble form, and the environment maintained by the glutaredoxin and thiore- antimicrobial activity was tested against two bacterial doxin pathways, that hampers disulfide bond formation pathogens (Fig. 2). Both defensins at 5 μM significantly [41, 42]. Commercial E. coli strains such as Origami reduced methicillin-resistant Staphylococcus aureus (Novagen), in which the thioredoxin reductase (trxB) and -MRSA- and Pseudomonas aeruginosa survival (Fig. 2), glutathione reductase (gor) genes are deleted have been decreasing bacterial survival up to 99% in both organ- used to produce defensins in a more oxidizing environ- isms. By contrast, GFP alone did not show any antimi- ment. For example, Wang and coworkers have compared crobial activity (Additional file 1: Fig. S1). Comparing the production of human α-defensin 6 (HD6) in E. coli the activity of the proteins produced in a reducing envi- BL21 and Origami strains, determining that higher pro- ronment (BL21 strain) and under more oxidizing condi- duction yields are reached when using Origami [43]. tions (Origami B strain), no differences were observed for López‑Cano et al. Microbial Cell Factories (2022) 21:77 Page 3 of 10 Table 1 LAP (V25-K64) and HD5 (A63-R94) sequences with the disulfide cysteine pairing The proportion of hydrophobic residues, peptide M.W, net charge and pI are also shown HD5 human defensin 5, LAP lingual antimicrobial peptide, M.W molecular weight, pI Isoelectric point a The number of hydrophobic residues include amino acids with aliphatic side chains b pI was theoretically calculated according to Expasy ProtParam tool Fig. 1 A Production kinetics and distribution of soluble (top) and IBs (bottom) of LAP-GFP-H6 and HD5-GFP-H6 proteins in mg/L at 1, 3, and 5 h in E. coli BL21 (dark grey) and Origami B (light grey) strains. The ratio of aggregation (at 3 h) for each HDP and strain is indicated in Table (B) López‑Cano et al. Microbial Cell Factories (2022) 21:77 Page 4 of 10 Fig. 2 Bacterial survival of MRSA and P. aeruginosa in the presence of 5 μM of soluble LAP-GFP-H6 and HD5-GFP-H6 (top) and insoluble (IBs) LAP-GFP-H6 and HD5-GFP-H6 (bottom) produced in E. coli BL21 (dark grey) and Origami B (light grey). Different letters depict statistical differences between proteins and producer strain (MRSA soluble) p = 0.0024; (P. aeruginosa soluble) p < 0.0001; (MRSA IBs) p = 0.0108; (P. aeruginosa IBs) p = 0.094 LAP-GFP-H6 (Fig. 2). However, HD5-GFP-H6 produced BL21 strain (Fig. 3). In the case of LAP-GFP-H6, no dif- in BL21 showed a higher bactericidal effect against both ferences were observed between the protein produced in MRSA and P. aeruginosa (Fig. 2 top) than that produced both strains and forms. in an Origami B strain. In terms of protein stability, the analysis of the soluble To analyze the protein quality of the insoluble protein LAP-GFP-H6 and HD5-GFP-H6 at 37 °C showed that the fraction of LAP-GFP-H6 and HD5-GFP-H6, bacterial IBs producer strain had an impact on protein stability of the produced in both BL21 and Origami B strains, were puri- α-defensin, while LAP-GFP-H6 was not affected (Fig. 4). fied, and their activity was tested. The results shown in Harnessing the presence of GFP as a carrier fused to Fig. 2 (bottom) proved that defensin-based IBs showed the defensins, we evaluated the potential correlation values of antimicrobial activity to levels that are compa- between fluorescence (Fig. 5) and antimicrobial activity rable with the soluble fraction (Fig. 2 top). As observed of both soluble (Fig. 2 top) and insoluble (Fig. 2 bottom) with the soluble form, LAP-GFP-H6 had the same activ- of the HDPs used. However, the linear correlation values ity against MRSA regardless of the producer strain, ­(R2) were in all cases lower than 0.22, which indicates a whereas HD5-GFP-H6 IBs showed higher bactericidal lack of interrelation between both parameters. activity when it was produced in a reducing environment (BL21 strain) (Fig. 2 bottom). Discussion The analysis of free cysteines in LAP-GFP-H6 and The bactericidal capacity of defensins, and in general of HD5-GFP-H6 produced in E. coli BL21 and Origami B HDPs, has aroused the interest of the scientific commu- strains revealed some differences (Fig. 3). Surprisingly, nity for these short peptides [13, 50]. They are part of the both soluble and insoluble (IBs) HD5-GFP-H6 had more innate immunity and they have shown broad-spectrum free cysteines when using Origami as producer strain in activity against Gram-positive and Gram-negative bac- spite of their apparent oxidizing environment than with teria, including MDR microorganisms, making them López‑Cano et al. Microbial Cell Factories (2022) 21:77 Page 5 of 10 Fig. 3 Analysis of free-cysteines in soluble (A) and insoluble (IBs) (B) LAP-GFP-H6 and HD5-GFP-H6 produced in E. coli BL21 (dark grey) and Origami B (light grey). Different letters depict statistical differences between proteins and strains A p = 0.0008, B p = 0.0345 expressed in tongue, mammary gland, intestine and res- piratory tract [53]. Both α and β-defensins are cationic and amphiphilic peptides with a net positive charge, hydrophobicity and amphiphilic nature which allows them to interact with negatively charged bacterial cell surfaces [54]. After this interaction, HDPs have two mechanisms of action: physical disruption of bacterial cell surface and action on internal targets [54]. Different articles, in which chemically synthesized pep- tides have been used, reported contradictory information regarding the importance of disulfide bond formation in HDP bactericidal activity [55–60]. In terms of recombinant protein production, little is known about the impact of the producer strain in the HDPs antimicrobial activity. Classical E. coli strains, such as BL21 used as recombinant cell factory, have a reduc- ing cytoplasm, while the mutant strain E. coli Origami has an oxidizing intracellular environment which should favor disulfide bond formation [42]. To explore the Fig. 4 Antimicrobial activity of soluble LAP-GFP-H6 and HD5-GFP-H6 importance of cytoplasmic environment of E. coli strains against P. aeruginosa at 5 μM after a 0, 5, 24, 48, 72 h and 1-week in HDP recombinant production, we have here studied incubation at 37 °C. Dark grey bars represent the HDPs produced in E. coli BL21 and light grey bars represent proteins from E. coli Origami the production and activity of two HDPs (an α- and a B strain. Different letters indicate significant statistical differences β-defensins) in two different cytoplasmic environments. between proteins and producer strains. p < 0.0001. W week The results proved that both production yields and pro- tein activity are not only determined by the bacterial strain used, but also by the tested peptide (Figs. 1 and 2). a promising alternative to antibiotic therapy [51, 52]. Whereas the β-defensin LAP fused to GFP was efficiently Structurally, HDPs have 6 cysteines that form 3 con- produced (Fig. 1A) and showed comparable activities served disulfide bonds. In humans, among the group of when using both BL21 and Origami B strains (Fig. 2), α-defensins, there are two that are produced by epithe- HD5-GFP-H6 showed significant differences when pro- lial intestinal Paneth cells (HD5 and HD6), being HD5 duced in the two different bacterial backgrounds (Figs. 1 the most abundant enteric HDPs. These α-defensins and 2). The soluble form of the HD5-GFP-H6 showed are important as host defense against pathogens, but a decrease in the production yields (Fig. 1) and also a also maintaining intestinal homeostasis. In the group of lower bactericidal activity when using an E. coli strain β-defensins, LAP was one of the first described and it is with an oxidizing environment (Origami B) (Fig. 2 top). López‑Cano et al. Microbial Cell Factories (2022) 21:77 Page 6 of 10 Fig. 5 Specific GFP fluorescence (relative fluorescence units per ng of peptide) of soluble (A) and inclusion bodies (B) LAP-GFP-H6 and HD5-GFP-H6 produced in E. coli BL21 (dark grey) and Origami B (light grey). Different letters indicate statistical differences between proteins and strains A (p < 0.0001) B (p < 0.0001) The greater activity of the soluble α-defensin produced in produced in BL21 strain (Fig. 2 bottom). Moreover, both the BL21 strain against both Gram-positive and Gram- soluble (Fig. 2 top) and insoluble proteins (Fig. 2 bottom) negative microorganisms indicated that, contrary to have the same behavior in terms of protein activity. This expectations, this strain produced a protein with better is in line with a previous publication describing that pro- conformational quality than that produced by Origami B tein conformational quality of both soluble and insolu- strain (Fig. 2 top B). Interestingly, the soluble HD5-GFP- ble (IB) fractions takes place in parallel Thus, the factors H6 produced in BL21(DE3) is also more active than the affecting the conformational protein quality of the solu- synthetic HDP (Additional file 1: Fig. S2). ble form also affect the IBs [59]. The difference observed in activity for HD5-GFP-H6 Besides, this study has also proven that GFP is a good in the two cytoplasmic environments is in agreement carrier protein for the production of HDPs, as other with the free cysteine profile observed in Fig. 3A. When proteins such as thioredoxin, glutathione S-transferase comparing the HD5-GFP-H6 produced in E. coli BL21 (GST), small ubiquitin-related modifier (SUMO), or and Origami B strains, the number of free cysteines is PurF fragment [60]. Indeed, GFP did not just protect the higher in the second case (Fig. 3A). This correlates with resultant HDP-based proteins from proteolytic degrada- the resulting lower antimicrobial activity against the two tion, but also simplifies protein tracking during the whole pathogenic microorganisms tested (Fig. 2) and the dimin- production and purification process. However, the results ished stability (Fig. 4). Therefore, this data supports pre- shown in Fig. 5 indicated that this fluorescent protein vious works describing the importance of disulfide bonds cannot be used as antimicrobial activity reporter, since on α-defensins stability. Tanabe et al. and Maemoto et al. the differences observed in bactericidal activity (Fig. 2) reported that the disruption of disulfide bonds of HD5 did not correlate with differences in fluorescence emis- and mouse α-defensin cryptdin-4, increased peptide pro- sion (Fig. 5). pensity to be proteolyzed and, in consequence, the activ- ity of these peptide variants decreased [56–58]. Thus, Conclusions this shows that disulfide bonds have an important role This study proved that the strain used for the production in protein stabilization. The protein stability analysis also of HDP-based proteins had an impact on both the pro- showed that all the HDPs with low free-cysteines pro- duction yields and protein quality, being the E. coli BL21 duced are highly stable, keeping the bactericidal activity strain an optimal background for the recombinant pro- for at least 1 week at 37 °C (Fig. 4). This data is highly rel- duction of HDPs. evant in terms of applicability and storage of these bacte- ricidal peptides. Methods In the same line, when the protein aggregates (IBs) Bacterial strains and medium were analyzed, we could observe that even though Escherichia coli BL21 (DE3) and Origami B (DE3) (­TetR, in all the cases IBs were formed (Fig. 1), the activity of ­KanR) strains were used for heterologous protein expres- HD5-GFP-H6 IBs was again significantly higher when sion. For the antibacterial assay, the strains used were P. López‑Cano et al. Microbial Cell Factories (2022) 21:77 Page 7 of 10 aeruginosa (ATCC-10145) and methicillin-resistant S. by western blot using a monoclonal anti-His antibody aureus (MRSA. ATCC-33592). E. coli strains were grown (His-probe, Santa Cruz), and their purity was evaluated in Luria–Bertani (LB) medium, whereas P. aeruginosa by a Coomassie blue staining assay. Both outcomes were and S. aureus were grown in Brain–Heart Infusion (BHI) evaluated by ImageJ software to determine protein quan- broth (Scharlau, Barcelona, Spain). tity and purity. Genetic construct design Soluble antimicrobial protein purification Constructs consisting in the mature form of bovine lin- Cultures (1 L) of each fusion construct were grown and gual antimicrobial peptide (LAP; Uniprot entry Q28880, induced with IPTG, as described in the previous sec- V25-K64) or human defensin 5 (HD5, Uniprot entry tion. After 3 h of production, the whole culture was Q01523, A63-R94) were fused to green fluorescent pro- harvested (6000×g, 15 min, 4 °C). Pellets from 500 mL tein (GFP) [61] using a linker sequence (SGGGSGGS) culture of LAP-GFP produced in both BL21 and Origami and named LAP-GFP and HD5-GFP, respectively. Each strains and HD5-GFP produced in the BL21 strain were construct was C-terminally fused to a 6-histidine tag for resuspended in 30 mL of binding buffer (500 mM NaCl, and HD5-GFP-H6 were codon-optimized (­GeneArt®, purification and quantification purposes. LAP-GFP-H6 20 mM Tris, 20 mM imidazole) with EDTA-free pro- tease inhibitor (Roche). Bacteria were sonicated (4 cycles, Lifetechnologies, Regensburg, Germany) and cloned in 5 min, 0.5 s on, 0.5 s off at 10% amplitude, Branson pET22b ­(AmpR) (Novagene, Darmstadt, Germany) vec- SFX550 Sonifier) and centrifugated (15,000×g, 45 min, tor. pET22b vector has a T7 promoter. The plasmid with 4 °C), collecting the supernatant, which contains solu- each construct (LAP-GFP-H6 HD5-GFP-H6) was trans- ble protein. Culture samples (1 L) of HD5-GFP produced formed into competent E. coli BL21 (DE3) and Origami in Origami strain was harvested (6000×g, 15 min, 4 °C) B (DE3). DE3 indicates that the host is a λDE3 lyogen, at 3 h post-induction, and the pellet was resuspended which carries a copy of T7 RNA polymerase gene under in 60 mL of PBS, sonicated as previously described, and the control of lacUV promoter, which is induced by the centrifugated (15,000×g, 45 min, 4 °C). The supernatant presence of isopropyl-β-d-thiogalactoside (IPTG). The was discarded and the pellet, containing the IBs, was production of T7 RNA polymerase induce the synthe- washed with ­dH2O and centrifugated (15,000×g, 45 min, sis of the protein encoded in pET22b vector with a T7 4 °C). Then, the supernatant was discarded again, and promoter. the pellet was weighted, adding 40 mL of solubilization The GFP fused to a 6 His-tag (GFP-H6) previously buffer (0.2% N-lauroylsarcosine mild detergent, 40 mM developed [61] was used as a control. Tris) per gram of pellet. Next, the pellet was solubilized for 40 h at RT continuously stirred. Solubilized protein Kinetics of soluble protein and inclusion body production was recovered after centrifugation (15,000×g, 45 min, Escherichia coli BL21/pET22b cultures (0.5 L) with each 4 °C), and samples were equilibrated at 500 mM NaCl antimicrobial fusion (LAP-GFP-H6, and HD5-GFP-H6) and 20 mM imidazole for purification. were grown overnight (O/N) in shake flasks at 37 °C All soluble proteins (obtained from supernatant or and 250 rpm in LB broth with ampicillin at 100 μg/mL. solubilized IBs) were filtered using a pore diameter of Escherichia coli Origami B/pET22b with each antimicro- 0.2 μm and purified by Immobilized Metal Affinity Chro- bial fusion (LAP-GFP-H6 and HD5-GFP-H6) were grown matography (IMAC) in an ÄKTA Start (GE Healthcare) at the same conditions with ampicillin, kanamycin, and using 1 mL HiTrap chelating HP columns (GE Health- tetracycline at 100, 25, and 12.5 μg/mL, respectively. The care). Protein was loaded with binding buffer (20 mM O/N were used as inoculum in fresh LB medium, start- Tris, 500 mM NaCl, 20 mM Imidazole) and eluted using a ing at O ­ D600 = 0.05. Recombinant protein expression linear gradient with elution buffer (20 mM Tris, 500 mM was induced by 1 mM IPTG when cultures reached an NaCl, 500 mM Imidazole). Protein buffer exchange was ­OD600 = 0.4–0.6. Culture samples of 25 mL were with- done by dialysis in acetic 0.01% (v/v) O/N at 4 °C with was determined by NanoDrop™, and the integrity and drawn at 0, 1, 3, and 5 h post-induction, and they were gentle agitation. The yield of purified soluble protein collected by centrifugation at 6000×g for 15 min at 4 °C. Pellets were resuspended in 500 μL PBS with EDTA-free purity of the protein were analyzed by Western blot and protease inhibitor (Roche) and bacteria were disrupted Coomassie. by sonication (2 cycles of 3 min, 0.5 s on, 0.5 s off at 10% amplitude) (Branson SFX550 Sonifier). Soluble and insol- IB purification uble fractions were split by centrifugation (15,000×g, As described before, at 3 h post-induction, culture 15 min, 4 °C). Quantifications of LAP-GFP-H6 and HD5- was harvested (6000×g, 15 min, 4 °C). The superna- GFP-H6 in both BL21 and Origami strains were obtained tant was discarded and the pellet was stored at − 80 °C López‑Cano et al. Microbial Cell Factories (2022) 21:77 Page 8 of 10 (minimum 16 h). Then, cells were thawed at RT, soni- Supernatant was removed and the bacteria pellet was cated (2 cycles, 1.5 min, 0.5 s on, 0.5 s off at 10% ampli- resuspended in 150 μL of each treatment (acetic buffer— tude, Branson SFX550 Sonifier), and stored at − 80 °C negative control, soluble proteins (LAP-GFP-H6 and O/N. Next, samples were thawed, and 0.2% (v/v) Triton HD5-GFP-H6) at 5 μM and IBs (LAP-GFP-H6 and HD5- X-100 (Sigma Aldrich) was added, incubating for 1 h at GFP-H6) at 5 μM). Samples were incubated in a sterile RT and 250 rpm, sample then was frozen at − 80 °C. An polypropylene 96-well microtiter plate 5 h at 37 °C. Next, ume of BacTiter-Glo™ reagent on a sterile 96-well white extra frozen/thawed cycle was recommended. Next, a 100 µL of each sample were mixed with the same vol- contamination control was performed, 100 μL of sample was platted on LB-agar plate and incubated at 37 °C O/N. opaque microtiter plate. Plates were incubated for 5 min Freeze/thaw cycles were repeated until no viable bacteria and luminescence was measured in a microplate lumi- bated with 250 μL NP-40 (ThermoScientific™) for 1 h at were observed in control plates. Further, IBs were incu- nometer (LumiStar, Omega). The measured arbitrary luminescence values were normalized against the control 4 °C and 250 rpm. Afterward, 0.6 μg/mL DNase I (Roche) (KPi treatment). and 0.6 μg/mL M ­ gSO4 were added, and the sample was incubated for 1 h at 37 °C and 250 rpm. Then, the IBs Protein stability assay were collected by centrifugation (15,000×g, 15 min, 4 °C) To evaluate protein stability, soluble LAP-GFP-H6 and and the supernatant was discarded. After, IBs were resus- HD5-GFP-H6 were incubated at different timescales (0 h, pended in lysis buffer (100 mM NaCl, 50 mM Tris, 1 mM 5 h, 24 h, 48 h, 72 h and 1 W) at 37 °C and then, antimi- EDTA, 0.5% Triton X-100), followed by a contamination crobial activity was evaluated as previously described. control as previously described. Then, IBs were harvested (15,000×g, 15 min, 4 °C) and frozen − 80 °C after super- Fluorescence measurements natant was removed. Finally, IBs were washed in 10 mL Fluorescence of the GFP fused with the antimicrobial PBS, aliquoted, and centrifuged (15,000×g, 15 min, 4 °C). peptides was recorded in a fluorescence spectrophotom- The supernatant was removed and the pellets, which con- eter (LumiStar, Omega). LAP-GFP-H6 and HD5-GFP- tained purified IBs, were kept at 80 °C until use. Purity H6 in both soluble and IBs formats produced in both E. and quantity of purified IBs were assessed by Western coli BL21 and Origami B strains were analyzed, being Blot and Coomassie assay. Shortly, the samples of soluble the samples diluted when required. They were excited at and IBs proteins were boiled at 95 °C for 10 and 45 min, 480 nm and the emission was recorded at 510 nm. Spe- respectively, to ensure that the protein is completely cific fluorescence was calculated using the amount of denaturalized. Next, samples were resuspended with Lae- protein in each sample. mmli loading buffer (100 mM Tris, 8% glycerol, 55 mM SDS, 4% β-mercaptoethanol, and 1.6% urea) and subse- Sulfhydryl determination quent analyzed by electrophoresis (SDS-PAGE). Pro- Sulfhydryls (–SH) not-forming disulfide bonds (–S–S–) teins bands were electroblotted into PVDF membranes were determined according to a previously established at 1.3 A and 25 V for 10 min, followed by a blocked step protocol based on 4,4′-diothiodipyridine (DTDP) [62]. at 4ºC O/N with bovine serum albumin (BSA) at 5% in DTDP is a small, amphiphilic uncharged molecule, capa- TBST buffer (10 mM Tris, 150 mM NaCl, 0.05% Tween ble of quickly reacting with poorly accessible sulfhydryls. 20). Next, the membranes were incubated 2 h at RT in In this case, samples were diluted to a final sulfhydryl primary antibody (Anti-His, Sant Cruz Biotechnology) concentration ≤ 40 μM in 1 mL (calculated by multiply- at 1:1000 dilution, followed by 3 washes in TBST buffer. ing protein moles by the number of SH) and mixed with Finally, membranes were incubated with secondary anti- 200 μL of sodium buffer (100 mM ­NaH2PO4, 0.2 mM body (anti-mouse IgG-alkaline phosphatase (Sigma) at EDTA, pH 6.8 adjusted with NaOH). After the addition 1/20,000 dilution for 1 h at RT and three time washed. of 50 μL of 4 mM DTDP, samples were vortexed and Proteins bands were revealed after alkaline phosphatase incubated for 5 min at RT. Next, samples were read at an (NBT/BCIP, Thermo Scientific) substrate incubation and absorbance of 324 nm (A324) against a water blank. For protein was quantified with ImageJ software. the reagent blank (A324r), 1 mL of potassium phosphate buffer was mixed with 200 μL of sodium buffer and 50 μL Antibacterial activity assay of DTDP reagent. For the protein blank (A324p), 50 μL of ter-Glo ™ Microbial Cell Viability kit (Promega). Briefly, Antimicrobial activity was determined with the Bacti- water was added instead of DTDP reagent in the sample with a 200 μL sodium buffer. For proteins in insoluble an O/N culture of MRSA and P. aeruginosa was diluted form, the A­ 324 increase was monitored over time after the 100-fold in 10 mM KPi (10 mM), aliquoted in 150 μL addition of DTDP to the sample, to achieve an increment eppendorf, and centrifugated (6200×g, 15 min, 4 °C). of ­A324 over 5 min. López‑Cano et al. Microbial Cell Factories (2022) 21:77 Page 9 of 10 Statistical analysis Received: 15 February 2022 Accepted: 20 April 2022 All experiments were performed in triplicate and rep- resented as the mean of non-transformed data ± non- transformed standard error of the mean (SEM). Data were previously checked for normality (JMP, SAS Insti- References 1. Blair JM, Webber MA, Baylay AJ, Ogbolu DO, Piddock LJ. Molecular tute Inc.) and p-values and letters correspond to the mechanisms of antibiotic resistance. Nat Rev Microbiol. 2015;13(1):42–51. ANOVA analyses and Tukey test analyses respectively, https://​doi.​org/​10.​1038/​nrmic​ro3380. using transforming data when required. 2. Brown ED, Wright GD. Antibacterial drug discovery in the resistance era. Nature. 2016;529(7586):336–43. https://​doi.​org/​10.​1038/​natur​e17042. 3. World Health Organization. Antimicrobial resistance: global report on Supplementary Information surveillance. Geneva: World Health Organization; 2014. The online version contains supplementary material available at https://​doi.​ 4. Brives C, Pourraz J. Phage therapy as a potential solution in the fight org/​10.​1186/​s12934-​022-​01803-7. against AMR: obstacles and possible futures. Palgrave Commun. 2020;6(1):1–11. 5. Fischetti VA. Bacteriophage lysins as effective antibacterials. Curr Opin Additional file 1: Figure S1. Bacterial survival of P. aeruginosa in the pres‑ Microbiol. 2008;11(5):393–400. https://​doi.​org/​10.​1016/j.​mib.​2008.​09.​012. ence and absence of 5 μM of soluble GFP-H6. Figure S2. Antimicrobial 6. Silva DR, Sardi JdCO, de Souza Pitangui N, Roque SM, da Silva ACB, activity of soluble HD5-GFP-H6 (dark grey) and synthetic HD5 (light grey) Rosalen PL. Probiotics as an alternative antimicrobial therapy: current against MRSA and P. aeruginosa. Different letters indicate statistical differ‑ reality and future directions. J Func Foods. 2020;73:104080. ences between tested strains and proteins P = 0.0001. 7. Zurawski DV, McLendon MK. Monoclonal antibodies as an antibacterial approach against bacterial pathogens. Antibiotics. 2020. https://​doi.​org/​ 10.​3390/​antib​iotic​s9040​155. Acknowledgements 8. Levy O. Antimicrobial proteins and peptides of blood: templates for novel The authors are indebted to CERCA Programme (Generalitat de Catalunya) antimicrobial agents. Blood. 2000;96(8):2664–72. and the European Social Fund for supporting our research. 9. Gupta S, Bhatia G, Sharma A, Saxena S. Host defense peptides: an insight into the antimicrobial world. J Oral Maxillofac Pathol. 2018;22(2):239–44. Author contributions https://​doi.​org/​10.​4103/​jomfp.​JOMFP_​113_​16. AL performed all the experiments, analysis and contributed in writing—origi‑ 10. Mookherjee N, Anderson MA, Haagsman HP, Davidson DJ. Antimicrobial nal draft. MM, JG and IR contributed to the Sulfhydryl assay determination and host defence peptides: functions and clinical potential. Nat Rev Drug subsequent analysis. AR and EG performed the conceptualization, supervision Discov. 2020;19(5):311–32. https://​doi.​org/​10.​1038/​s41573-​019-​0058-8. and writing—review and editing. All authors read and approved the final 11. Haney EF, Mansour SC, Hancock RE. Antimicrobial peptides: an intro‑ manuscript. duction. Methods Mol Biol. 2017;1548:3–22. https://​doi.​org/​10.​1007/​ 978-1-​4939-​6737-7_1. Funding 12. Hancock RE, Lehrer R. Cationic peptides: a new source of antibiotics. This work was funded by Ministerio de Ciencia, Innovación y Universidades Trends Biotechnol. 1998;16(2):82–8. https://​doi.​org/​10.​1016/​s0167-​ Grants (PID2019-107298RB-C21) to AA and EG-F, PID2019-105622RB-I00 to IR 7799(97)​01156-6. and through the ‘‘Severo Ochoa’’ Programme for Centers of Excellence in R&D 13. Zasloff M. Antimicrobial peptides of multicellular organisms. Nature. (FUNFUTURE CEX2019-000917-S and SEV-2017-0706). The authors are also 2002;415(6870):389–95. https://​doi.​org/​10.​1038/​41538​9a. grateful to Marató de TV3 foundation for the Grant 201812-30-31-32-33 to 14. Yeung AT, Gellatly SL, Hancock RE. Multifunctional cationic host EG-F and IR, and the Networking Research Center on Bioengineering, Bioma‑ defence peptides and their clinical applications. Cell Mol Life Sci. terials and Nanomedicine (CIBER-BBN). AL-C received a pre-doctoral fellowship 2011;68(13):2161–76. https://​doi.​org/​10.​1007/​s00018-​011-​0710-x. from Generalitat de Catalunya (FI-AGAUR), EG-F a post-doctoral fellowship 15. Boparai JK, Sharma PK. Mini review on antimicrobial peptides, sources, from INIA and JG a Ramón y Cajal fellowship from MINN (RyC-2017-22614). mechanism and recent applications. Protein Pept Lett. 2020;27(1):4–16. https://​doi.​org/​10.​2174/​09298​66526​66619​08221​65812. Availability of data and materials 16. Brogden KA. Antimicrobial peptides: pore formers or metabolic inhibitors All data analysed during the current study are available from the correspond‑ in bacteria? Nat Rev Microbiol. 2005;3(3):238–50. https://​doi.​org/​10.​1038/​ ing author on reasonable request. nrmic​ro1098. 17. Hancock REW. Concerns regarding resistance to self-proteins. Microbiol‑ Declarations ogy. 2003;149(Pt 12):3343–4. https://​doi.​org/​10.​1099/​mic.0.​C0122-0. 18. De Smet K, Contreras R. Human antimicrobial peptides: defensins, catheli‑ Ethics approval and consent to participate cidins and histatins. Biotechnol Lett. 2005;27(18):1337–47. https://​doi.​org/​ Not applicable. 10.​1007/​s10529-​005-​0936-5. 19. Zhang LJ, Gallo RL. Antimicrobial peptides. Curr Biol. 2016;26(1):R14–9. Consent for publication https://​doi.​org/​10.​1016/j.​cub.​2015.​11.​017. Not applicable. 20. Vizioli J, Salzet M. Antimicrobial peptides from animals: focus on inverte‑ brates. Trends Pharmacol Sci. 2002;23(11):494–6. https://​doi.​org/​10.​1016/​ Competing interests s0165-​6147(02)​02105-3. The authors declare that they have no competing interests. 21. Papagianni M. Ribosomally synthesized peptides with antimicrobial properties: biosynthesis, structure, function, and applications. Biotechnol Author details Adv. 2003;21(6):465–99. https://​doi.​org/​10.​1016/​s0734-​9750(03)​00077-6. 1 Department of Ruminant Production, Institut de Recerca i Tecnologia Agroali‑ 22. Selsted ME, Harwig SS. Determination of the disulfide array in the mentàries (IRTA), 08140 Caldes de Montbui, Spain. 2 Department of Molecular human defensin HNP-2. A covalently cyclized peptide. J Biol Chem. Nanoscience and Organic Materials, Institut de Ciència de Materials de 1989;264(7):4003–7. Barcelona (ICMAB-CSIC), Campus UAB, 08193 Bellaterra, Spain. 3 Networking 23. Shi J, Ross CR, Chengappa MM, Sylte MJ, McVey DS, Blecha F. Antibacterial Research Center On Bioengineering, Biomaterials and Nanomedicine (CIBER- activity of a synthetic peptide (PR-26) derived from PR-39, a proline- BBN), Madrid, Spain. 4 Dynamic Biomimetics for Cancer Immunotherapy, Max arginine-rich neutrophil antimicrobial peptide. Antimicrob Agents Planck Partner Group, ICMAB-CSIC, Campus UAB, 08193 Bellaterra, Spain. Chemother. 1996;40(1):115–21. https://​doi.​org/​10.​1128/​AAC.​40.1.​115. López‑Cano et al. Microbial Cell Factories (2022) 21:77 Page 10 of 10 24. Lehrer RI, Lichtenstein AK, Ganz T. Defensins: antimicrobial and cytotoxic defensin 114. J Anim Sci Biotechnol. 2019;10:60. https://​doi.​org/​10.​1186/​ peptides of mammalian cells. Annu Rev Immunol. 1993;11:105–28. https://​ s40104-​019-​0367-0. doi.​org/​10.​1146/​annur​ev.​iy.​11.​040193.​000541. 45. Guillén-Chable F, Arenas-Sosa I, Islas-Flores I, Corzo G, Martinez-Liu C, Estrada 25. Selsted ME, Ouellette AJ. Mammalian defensins in the antimicrobial G. Antibacterial activity and phospholipid recognition of the recombinant immune response. Nat Immunol. 2005;6(6):551–7. https://​doi.​org/​10.​1038/​ defensin J1–1 from Capsicum genus. Protein Expr Purif. 2017;136:45–51. ni1206. https://​doi.​org/​10.​1016/j.​pep.​2017.​06.​007. 26. Yang D, Biragyn A, Kwak LW, Oppenheim JJ. Mammalian defensins in immu‑ 46. García-Fruitós E, González-Montalbán N, Morell M, Vera A, Ferraz RM, Arís A, nity: more than just microbicidal. Trends Immunol. 2002;23(6):291–6. https://​ et al. Aggregation as bacterial inclusion bodies does not imply inactivation doi.​org/​10.​1016/​s1471-​4906(02)​02246-9. of enzymes and fluorescent proteins. Microb Cell Fact. 2005;4:27. https://​doi.​ 27. Tang YQ, Selsted ME. Characterization of the disulfide motif in BNBD-12, an org/​10.​1186/​1475-​2859-4-​27. antimicrobial beta-defensin peptide from bovine neutrophils. J Biol Chem. 47. de Marco A, Ferrer-Miralles N, Garcia-Fruitós E, Mitraki A, Peternel S, Rinas U, 1993;268(9):6649–53. et al. Bacterial inclusion bodies are industrially exploitable amyloids. FEMS 28. Zhang L. Different dynamics and pathway of disulfide bonds reduction of Microbiol Rev. 2019;43(1):53–72. https://​doi.​org/​10.​1093/​femsre/​fuy038. two human defensins, a molecular dynamics simulation study. Proteins. 48. García-Fruitós E, Vázquez E, Díez-Gil C, Corchero JL, Seras-Franzoso J, Ratera I, 2017;85(4):665–81. https://​doi.​org/​10.​1002/​prot.​25247. et al. Bacterial inclusion bodies: making gold from waste. Trends Biotechnol. 29. Schroeder BO, Wu Z, Nuding S, Groscurth S, Marcinowski M, Beisner J, et al. 2012;30(2):65–70. https://​doi.​org/​10.​1016/j.​tibte​ch.​2011.​09.​003. Reduction of disulphide bonds unmasks potent antimicrobial activity of 49. Roca-Pinilla R, López-Cano A, Saubi C, Garcia-Fruitós E, Arís A. A new genera‑ human β-defensin 1. Nature. 2011;469(7330):419–23. https://​doi.​org/​10.​ tion of recombinant polypeptides combines multiple protein domains for 1038/​natur​e09674. effective antimicrobial activity. Microb Cell Fact. 2020;19(1):122. https://​doi.​ 30. Azari M, Asad S, Mehrnia MR. Heterologous production of porcine derived org/​10.​1186/​s12934-​020-​01380-7. antimicrobial peptide PR-39 in Escherichia coli using SUMO and intein fusion 50. Jenssen H, Hamill P, Hancock RE. Peptide antimicrobial agents. Clin Microbiol systems. Protein Expr Purif. 2020;169: 105568. https://​doi.​org/​10.​1016/j.​pep.​ Rev. 2006;19(3):491–511. https://​doi.​org/​10.​1128/​CMR.​00056-​05. 2020.​105568. 51. Afacan NJ, Yeung AT, Pena OM, Hancock RE. Therapeutic potential of 31. Jin F, Xu X, Wang L, Zhang W, Gu D. Expression of recombinant hybrid host defense peptides in antibiotic-resistant infections. Curr Pharm Des. peptide cecropinA(1–8)-magainin2(1–12) in Pichia pastoris: purification and 2012;18(6):807–19. https://​doi.​org/​10.​2174/​13816​12127​99277​617. characterization. Protein Expr Purif. 2006;50(2):147–56. https://​doi.​org/​10.​ 52. Wilmes M, Cammue BP, Sahl HG, Thevissen K. Antibiotic activities of host 1016/j.​pep.​2006.​05.​023. defense peptides: more to it than lipid bilayer perturbation. Nat Prod Rep. 32. Corrales-Garcia LL, Possani LD, Corzo G. Expression systems of human 2011;28(8):1350–8. https://​doi.​org/​10.​1039/​c1np0​0022e. β-defensins: vectors, purification and biological activities. Amino Acids. 53. Stolzenberg ED, Anderson GM, Ackermann MR, Whitlock RH, Zasloff M. 2011;40(1):5–13. https://​doi.​org/​10.​1007/​s00726-​010-​0493-7. Epithelial antibiotic induced in states of disease. Proc Natl Acad Sci USA. 33. Clement H, Flores V, Diego-Garcia E, Corrales-Garcia L, Villegas E, Corzo G. A 1997;94(16):8686–90. https://​doi.​org/​10.​1073/​pnas.​94.​16.​8686. comparison between the recombinant expression and chemical synthesis 54. Hancock RE, Sahl HG. Antimicrobial and host-defense peptides as new anti- of a short cysteine-rich insecticidal spider peptide. J Venom Anim Toxins Incl infective therapeutic strategies. Nat Biotechnol. 2006;24(12):1551–7. https://​ Trop Dis. 2015;21:19. https://​doi.​org/​10.​1186/​s40409-​015-​0018-7. doi.​org/​10.​1038/​nbt12​67. 34. Ongey EL, Neubauer P. Lanthipeptides: chemical synthesis versus in vivo 55. Wanniarachchi YA, Kaczmarek P, Wan A, Nolan EM. Human defensin 5 biosynthesis as tools for pharmaceutical production. Microb Cell Fact. disulfide array mutants: disulfide bond deletion attenuates antibacterial 2016;15:97. https://​doi.​org/​10.​1186/​s12934-​016-​0502-y. activity against Staphylococcus aureus. Biochemistry. 2011;50(37):8005–17. 35. Wegmuller S, Schmid S. Recombinant peptide production in microbial cells. https://​doi.​org/​10.​1021/​bi201​043j. Curr Org Chem. 2014. https://​doi.​org/​10.​2174/​13852​72818​08140​61616​ 56. Maemoto A, Qu X, Rosengren KJ, Tanabe H, Henschen-Edman A, Craik DJ, 0013. et al. Functional analysis of the alpha-defensin disulfide array in mouse 36. Klüver E, Adermann K, Schulz A. Synthesis and structure-activity relationship cryptdin-4. J Biol Chem. 2004;279(42):44188–96. https://​doi.​org/​10.​1074/​jbc.​ of beta-defensins, multi-functional peptides of the immune system. J Pept M4061​54200. Sci. 2006;12(4):243–57. https://​doi.​org/​10.​1002/​psc.​749. 57. Tanabe H, Ayabe T, Maemoto A, Ishikawa C, Inaba Y, Sato R, et al. Dena‑ 37. Li Y. Carrier proteins for fusion expression of antimicrobial peptides in tured human alpha-defensin attenuates the bactericidal activity and the Escherichia coli. Biotechnol Appl Biochem. 2009;54(1):1–9. https://​doi.​org/​10.​ stability against enzymatic digestion. Biochem Biophys Res Commun. 1042/​BA200​90087. 2007;358(1):349–55. https://​doi.​org/​10.​1016/j.​bbrc.​2007.​04.​132. 38. de Oliveira KBS, Leite ML, Rodrigues GR, Duque HM, da Costa RA, Cunha VA, 58. Zhang Y, Cougnon FB, Wanniarachchi YA, Hayden JA, Nolan EM. Reduction et al. Strategies for recombinant production of antimicrobial peptides with of human defensin 5 affords a high-affinity zinc-chelating peptide. ACS pharmacological potential. Expert Rev Clin Pharmacol. 2020;13(4):367–90. Chem Biol. 2013;8(9):1907–11. https://​doi.​org/​10.​1021/​cb400​340k. https://​doi.​org/​10.​1080/​17512​433.​2020.​17643​47. 59. González-Montalbán N, García-Fruitós E, Villaverde A. Recombinant protein 39. Ingham AB, Moore RJ. Recombinant production of antimicrobial peptides solubility—does more mean better? Nat Biotechnol. 2007;25(7):718–20. in heterologous microbial systems. Biotechnol Appl Biochem. 2007;47(Pt https://​doi.​org/​10.​1038/​nbt07​07-​718. 1):1–9. https://​doi.​org/​10.​1042/​BA200​60207. 60. Silva ON, Mulder KC, Barbosa AE, Otero-Gonzalez AJ, Lopez-Abarrategui C, 40. Bommarius B, Jenssen H, Elliott M, Kindrachuk J, Pasupuleti M, Gieren H, Rezende TM, et al. Exploring the pharmacological potential of promiscuous et al. Cost-effective expression and purification of antimicrobial and host host-defense peptides: from natural screenings to biotechnological applica‑ defense peptides in Escherichia coli. Peptides. 2010;31(11):1957–65. https://​ tions. Front Microbiol. 2011;2:232. https://​doi.​org/​10.​3389/​fmicb.​2011.​00232. doi.​org/​10.​1016/j.​pepti​des.​2010.​08.​008. 61. Roca-Pinilla R, Fortuna S, Natalello A, Sánchez-Chardi A, Ami D, Arís A, et al. 41. Waegeman H, Soetaert W. Increasing recombinant protein production Exploring the use of leucine zippers for the generation of a new class of in Escherichia coli through metabolic and genetic engineering. J Ind inclusion bodies for pharma and biotechnological applications. Microb Cell Microbiol Biotechnol. 2011;38(12):1891–910. https://​doi.​org/​10.​1007/​ Fact. 2020;19(1):175. https://​doi.​org/​10.​1186/​s12934-​020-​01425-x. s10295-​011-​1034-4. 62. Riener CK, Kada G, Gruber HJ. Quick measurement of protein sulfhydryls 42. de Marco A. Strategies for successful recombinant expression of disulfide with Ellman’s reagent and with 4,4′-dithiodipyridine. Anal Bioanal Chem. bond-dependent proteins in Escherichia coli. Microb Cell Fact. 2009;8:26. 2002;373(4–5):266–76. https://​doi.​org/​10.​1007/​s00216-​002-​1347-2. https://​doi.​org/​10.​1186/​1475-​2859-8-​26. 43. Wang A, Su Y, Wang S, Shen M, Chen F, Chen M, et al. High efficiency preparation of bioactive human alpha-defensin 6 in Escherichia coli Publisher’s Note origami(DE3)pLysS by soluble fusion expression. Appl Microbiol Biotechnol. Springer Nature remains neutral with regard to jurisdictional claims in pub‑ 2010;87(5):1935–42. https://​doi.​org/​10.​1007/​s00253-​010-​2688-y. lished maps and institutional affiliations. 44. Su G, Xie K, Chen D, Yu B, Huang Z, Luo Y, et al. Differential expres‑ sion, molecular cloning, and characterization of porcine beta