Introduction

The worldwide emergence of multi-drug resistant bacteria to antibiotics is a significant problem at present. These bacteria cause fatal incurable diseases (World Health Organization 2021). Various new therapies are currently being developed, and such as enzyme therapy (Ghosh et al. 2019; Hojckova et al. 2013), phage therapy (Kakasis and Panista 2019), antibacterial light therapy (Yuan et al. 2022) and therapy by nanoparticles (Franci et al. 2015; Ribeiro et al. 2022). One possible alternative to antibiotics are antimicrobial peptides (AMPs), which can effectively kill even multi-drug resistant bacteria. Antimicrobial peptides are found in a wide range of organisms, where they have an important function in defence of the organism against foreign pathogens as part of innate immunity (Kang et al. 2017). AMPs manifest natural antimicrobial activity; they act fast and are characterised by a wide range of action both Gram-negative and Gram-positive bacteria, viruses, yeasts. In addition, the emergence of resistance to AMP is less frequent (Baltzer and Brown 2011). The limiting factor of antimicrobial peptides is their susceptibility to proteolytic degradation and cytotoxicity at higher concentrations (Han et al. 2021). A viable alternative to antibiotics may be the human protein hCAP18 – up until the present, the only cathelicidin discovered in humans (Majewska et al. 2021).

The product of the gene hCAP18 is 18 kDa pre-pro-protein hCAP18, which contains a signal peptide, a conserved cathelin domain, and the antimicrobial peptide LL-37 on the C-terminus of protein (Dürr et al. 2006). Following the cleavage of the signal peptide, a pro-protein is created which is stored in epithelial cells and neutrophil granules (Burton and Steel 2009). The pro-protein is subsequently split with human serine proteinase 3 (PRTN3) in the case of neutrophils, or kallikrein (KLK) in the case of keratinocytes, into the cathelin domain and the active peptide LL-37 (Vandamme et al. 2012). This peptide may be found in different concentrations in many various types of bodily fluids, tissues and cells. LL-37 may be found in bone marrow, nails, and fluids, such as saliva, breast milk, sweat and seminal plasma (Dürr et al. 2006). LL-37 is produced mainly by epithelial cells, cells of the respiratory tract, skin, intestines, eyes and genitals. This peptide is also produced by immune cells such as neutrophils, mast cells and NK cells (Vandamme et al. 2012).

LL-37 consists of 37 amino acids; the first two amino acids are leucines. Under physiological pH, the total charge is + 6, and 16 out of 37 amino acids are charged (Dürr et al. 2006). Its molecule contains 12 hydrophobic amino acids (DRAMP 2023). In a pure aqueous solution, LL-37 assumes a random-coil conformation. However, in the vicinity of lipid micelles and negatively charged bacterial membranes, it takes the shape of α-helix with an amphipathic structure. Thanks to its characteristics, such as amphipathicity, the ability to form α-helix, and the positive charge of the molecule, the peptide LL-37 can act as an antimicrobial agent (Porcelli et al. 2008). Except α-helical peptides there are three other structure classes: β-sheet peptides, linear extension structure and both α-helix and β-sheet peptides (Huan et al. 2020).

LL-37 manifests antimicrobial activity against certain viruses, the yeast Candida albicans and Gram-positive and Gram-negative bacteria. In addition, LL-37 inhibits the growth of bacterial biofilms (Vandamme et al. 2012; Memariani and Memariani 2023). LL-37 demonstrated antiviral activity e.g. against HSV-1 or human adenovirus 19 (Gordon et al. 2005). As measured by Tsai et al. 2014, a significant growth inhibition of the yeast C. albicans began with the concentration of the peptide 20 µg/mL and LC50 was 30 µg/mL of LL-37. The bacteria most sensitive to the effect of peptide are Escherichia coli (Luo et al. 2017), Helicobacter pylori (Leszczynska et al. 2013), Acinetobacter baumannii (Zarei-Mehrvarz et al. 2022), Salmonella Typhimurium (Nan et al. 2012), Enterococcus faecalis (Leszczynska et al. 2013) and Mycobacterium tuberculosis (Rivas-Santiago et al. 2013), with MIC values ranging from 5 µg/mL to 32 µg/mL. Furthermore, LL-37 is considerably active against multi-drug resistant Pseudomonas aeruginosa – MIC = 32–64 µg/mL (Geitani et al. 2019) and Staphylococcus epidermidis (Leszczynska et al. 2013). The bacteria most resistant to LL-37 are Klebsiella pneumoniae – MIC = 224.7 µg/mL (Smeianov et al. 2000) and methicillin-resistant Staphylococcus aureus – MIC ˃ 128 µg/mL (Geitani et al. 2019).

LL-37 has a substantial area of application in wound healing because even 1 µM (4.5 µg/mL) of the peptide can induce cell migration and promote proliferation. The final consequence is in all probability a faster closing of the wound (Wang Guangshun 2014). Through interaction with cell receptors, LL-37 exerts its immunomodulatory effects – chemotactic activity and regulation of inflammatory response. Additional observations include the influence of the peptide LL-37 on apoptosis, angiogenesis and tumour proliferation (Burton and Steel 2009). High amounts of this peptide in inflamed and infected tissue prevent septic shock (Dürr et al. 2006). The greatest obstacles to clinical application of LL-37 are its cytotoxicity, particularly towards human erythrocytes and keratinocytes (Burton and Steel 2009), as well as a decrease in the antimicrobial activity of LL-37 due to environmental effects, such as solutions with the high content of salts (Dürr et al. 2006) and human serum (Johansson et al. 1998). In general, other constraints for application of AMP may include proteolytic degradation and the high financial costs of industrial production of peptides (Ghosh et al. 2019, Wang Jiajun et al. 2019).

In this study, a synthetic gene encoding the separate active peptide LL-37 was inserted into the laboratory made express vector pUbEx100, with is a high copy number plasmid containing a fusion gene for ubiquitin. Regarding the gene LL-37, an optimization of frequency and utilization of codons for heterologous expression in E. coli was carried out. The resulting product of expression in E. coli was a fusion protein containing ubiquitin, a His-tag, TEV recognition sequence and the peptide LL-37. The presence of ubiquitin prevented the toxic effect of the peptide on the E. coli bacteria producing it. In order to cleave LL-37 from the fusion protein, TEV protease was used. The fusion protein and the cleaved GLL-37 were purified using immobilized metal affinity chromatography (IMAC). In the final stage of the study, the antibacterial activity of the prepared GLL-37 and synthetic peptide LL-37 against selected pathogens was measured and compared. The secondary structure of both peptides and their behaviour in the membrane environment were analysed using CD spectroscopy.

Materials and methods

Materials

The gene hCAP18 was synthesized by GenScript, USA, with an optimization of frequency and utilization of codons for expression in E. coli. Express vector pUbEx100 and TEV protease were in-house produced at the Veterinary Research Institute, Brno. Host strains of Escherichia coli DH10β and Escherichia coli BL21 (DE3) were purchased from ThermoFisher Scientific, USA. Restriction enzymes, cut buffer, ligation buffer, calf intestinal alkaline phosphatase, and T4 DNA ligase were bought from New England Biolabs. Agarose for PCR was purchased from Top-Bio, Vestec. LB Broth with agar (Luria low salt) was obtained from Sigma-Aldrich. LB Broth Base Invitrogen (Lennox) was purchased from ThermoFisher Scientific, USA. Kit NucleoSpin Gel and PCR Clean-up was bought from Machery-Nagel, Germany and E.Z.N.A. Plasmid DNA Mini Kit I was bought from Omega Bio-Tek, USA. Ni affinity column 5 ml HisTrap™ HP and Sephadex G-25 HiPrep 26/10 Desalting column were purchased from GE Healthcare, USA. 40% acrylamide/bis was obtained from SERVA – USA, tricine from Carl Roth, Germany and protein standards from Bio-Rad, USA. Isopropyl β-d-1-thiogalactopyranoside (IPTG) was bought from UBPbio, USA. The synthetic peptide LL-37 was purchased from sb-PEPTIDE, France.

Expression vector contruction: pUbEx100-GLL37

The synthetized gene hCAP18 stored in plasmid pUC57 was digested by KpnI and NotI to obtain DNA region corresponding to active peptide LL-37. The recipient expression vector pUbEx100 (Janda 2022) was linearized by KpnI and NotI for the ligation of the active peptide LL-37. Both linearized expression vector pUbEx100 and the product of KpnI/NotI digestion of hCAP18 (LL-37) were recovered from 2% agarose gel and subsequently ligated and transformed to E. coli strain DH10β. The recombinant plasmid pUbEx100-GLL37 was isolated from E. coli using the isolation kit E.Z.N.A. Plasmid DNA Mini Kit I. The presence of the peptide LL-37 in expression vector was verified by digestion of restriction enzymes (agarose gel electrophoresis) and DNA plasmid sequencing (carried out at SEQme, Dobris). Lastly, the created recombinant vector pUbEx100 with LL-37 was used to transform E. coli BL21 (DE3) competent cells for later expression of fusion protein. Freshly transformed E. coli were grew overnight, centrifugated at 4000 g for 20 min at 4 °C. The pellet was resuspended in 10% sterile glycerol (1:1 v/v ratio) and shocke-freezed for storage. These bacterial stocks were used for inoculation.

Expression and purification of fusion protein UBQ-GLL-37

A volume of 1 mL of E. coli BL21 (DE3) bacterial stock containing recombinant vector pUbEx100 with LL-37 was used to inoculate 1 L of TB medium (with 1.25% glucose and 50 µg/mL kanamycin). Bacterial cells were grown at 37 oC with shaking at 280 rpm, until the culture in TB medium reached O.D.600 = 1.0; expression of fusion protein UBQ-GLL-37 (Kozlov et al. 2008) was induced by IPTG (final concentration in the medium was 0.5 mM). Subsequently, after a 6-hour of the cultivation (37 oC, 280 rpm) the cells were harvested by centrifugation (15 000 g, 10 min, 4 oC). The bacterial pellet was resuspended in 40 ml of equilibration buffer (25 mM TRIS, 100 mM NaCl, 10 mM imidazole, pH 8.0) and bacterial cells in the suspension were lysed in a homogeniser (Pressure cell homogeniser SPCH-EP – Homogenising systems LTD., UK) using 1 bar pressure. After that, the lysate was centrifuged (15 000 g, 30 min, 4 oC) and the supernatant was loaded onto a purification column.

The fusion protein UBQ-GLL-37 was purified by FPLC (Fast Protein Liquid Chromatography) Azura ASM 2.1 L (Knauer, Germany) using IMAC column HisTrap™ HP. First, the IMAC column was equilibrated five times by equilibration buffer. After loading supernatant on the column, the column was washed five times by equilibration buffer containing 48 mM imidazole. The fusion protein UBQ-LL-37 bound to the column was eluted by elution buffer (25 mM TRIS, 100 mM NaCl, 300 mM imidazole, pH 8.0). Further, the buffer exchang into TEV buffer (25 mM TRIS, 25 mM NaCl, 2 mM DTT, pH 8.0) was performed, using a Sephadex G-25 HiPrep 26/10 Desalting column for the following digestion. The concentration of the prepared fusion protein was measured by BCA protein assay (ThermoFisher Scientific, USA).

Release and purification of peptide GLL-37

To cleave the peptide GLL-37 from UBQ-His tag, the fusion protein was treated with TEV protease at 25 oC for 24 h, using a mass ratio of 1:20 (1 mg TEV protease and 20 mg protein). After cleavage, the cleaved peptide GLL-37 was purified, removing UBQ-His tag and TEV protease (which also possesses His-tag) using IMAC as described in subsection 2.3., but with slight modifications. During this purification process, an equilibration buffer without 10 mM imidazole was used, and the peptide GLL-37 was eluted in the flow throught fraction (as separate LL-37 does not bind to the column). The eluted peptide was lyophilized for further use. The quality of the prepared peptide LL-37 was analyzed by MALDI-TOF mass spectrometry and Tris-Tricine-SDS-PAGE (visualization by silver staining). The concentration of the prepared and sLL-37 was measured by Bradford assay (Carl Roth, Germany).

AMIT (antimicrobial inhibition test)

A bacterial inoculum was prepared from a 24-hour bacterial culture by repeated steps of centrifugation and cell pellet resuspension in PBS. Working bacterial suspension was prepared by dilution of bacterial inoculum to the final concentration of 1 × 106 CFU/mL using PBS. A volume of 5 µL of this working suspension was mixed with 5 µL of antimicrobial peptides. This suspension was incubated at 37 oC for 1 h. From the incubated mixture, 5 µL was taken and mixed with 95 µL of PBS. This mixture was seeded on plates with Mueller-Hinton Agar (Sigma-Aldrich, Inc., USA). The plates were incubated for 24 hod at 37 oC. The following day, the number of colonies was evaluated to calculate CFU/mL and compared to the control. Peptide activity was measured against the following multiresistant bacterial strains: Acinetobacter baumannii ANC 4097 – AMI, CTZ, GEN, IMI, MEM, OFL, PIP, TZP, AMS, SMX (Krizova et al. 2012; Nemec and Krizova 2012); Acinetobacter baumannii ANC 6007 – AMI, CTZ, COL (MIC = 16 mg/l), DOX, GEN, IMI, MEM, NET, OFL, PIP, TZP, AMS, TOB; Acinetobacter baumannii ANC 7347 – AMI, CTZ, DOX, GEN, IMI, MEM, NET, OFL, PIP, TZP, AMS, SMX, TOB; Pseudomonas aeruginosa ANC 3488 – IMI, MEM, CTZ, CPM, PIP, CIP, GEN, TOB, AMI; Pseudomonas aeruginosa ANC 3157 – IMI, MEM, CTZ, CPM, PIP, CIP, GEN, TOB; Pseudomonas aeruginosa ANC 3271 – IMI, MEM, CTZ, CPM, PIP, CIP, GEN, TOB (Nemec et al. 2010).

AMI, amikacin; AMS, ampicilin-sulbactam; CTZ, ceftazidime; CIP, ciprofloxacin; COL, colistin; CPM, cefepime; DOX, doxycycline; GEN, gentamicin; IMI, imipenem; MEM, meropenem; NET, netilmicin; OFL, ofloxacin; PIP, piperacilin; TZP, piperacilin-tazobactam; SMX, sulfamethoxazole; TOB, tobramycin.

CD spectroscopy

The Chirascan™ V100 CD Spectrometer (Applied Photophysics Limited, UK) was used to measure CD spectra. The instrument was set to the wavelength range of 185–260 nm with measurements taken at 1 nm intervals. Water was used as the blank. The CD spectra of the antimicrobial peptides were measured and values of mean residue ellipticity were calculated for each wavelength.

The effect of two reagents, urea and trifluorethanol, on the CD spectra was measured. After the addition 2,2,2-trifluoroethanol/water (25% v/v) solution (25% TFE/water), the peptides were incubated for 10 min, and CD spectra were measured. The procedure for denaturing the GLL-37 peptide with urea involved adding different concentrations of urea (1–10 M) to the GLL-37 peptide, followed by overnight incubation at 4 oC. The following day, spectra were measured at wavelengths ranging from 210 to 260 nm. A graph was generated to show the effect of urea on the denaturation of the GLL-37 peptide.

Results

Expression vector pUbEx100-GLL37

In this study, we inserted the separate gene LL-37 (without the cathelin domain) into expression vector pUbEx100. Expression of the fusion protein UBQ-GLL37 is controlled by the T7 promotor. The fusion protein contains ubiquitin, which is followed by a His-tag. Between the fusion tag and the separate GLL-37, there is a recognition site for TEV protease (ENLYFQ/G). This is then followed by T7 terminator. Vector pUbEx100 contains the gene for kanamycin resistance.

Production and purification fusion protein UBQ-GLL37

The expressed fusion protein UBQ-GLL37 consists of three parts: ubiquitin protein, histidine tag and antimicrobial peptide LL-37 (Fig. 1.). There is a cleavage site for TEV protease between the histidine tag and the peptide LL-37. The fusion protein is composed of ubiquitin because it increases the yield of expression and prevents the peptide from becoming cytotoxic to bacteria.

Fig. 1
figure 1

Structure of fusion protein UBQ-GLL37

Full size image
Fig. 2
figure 2

Expression and purification of UBQ-GLL37 fusion protein – 15% SDS-PAGE. Lane 1, cell lysate supernatant; Lane 2, flow-through eluate; Lane 3, wash eluate; Lane 4, imidazole eluate; Lane 5, desalted fusion protein UBQ-GLL37; kDa, protein standard

Full size image

The fusion protein was expressed at 37 oC for 6 h, and induction was performed at OD600 = 1.0. It was purified by affinity chromatography using washing buffer containing 48 mM imidazole. The protein was desalted and converted into TEV protease digestion buffer. The composition and purity of the individual fractions from both chromatographies are presented in Fig. 2. where we can also see the whole fusion protein UBQ-GLL37 after desalting.

Release and purification of peptide GLL-37

The cleavage conditions for TEV protease were optimized (25 oC, 24 h, protein:protease mass ratio 20:1). The digest mixture was applied to the affinity column and a second purification step was performed to remove the ubiquitin and obtain the purified peptide GLL-37. In Fig. 3 we can see the comparison between the purity of synthetically produced LL-37 and recombinantly produced GLL-37. Regarding recombinant GLL-37, this is the final purity of the lyophilized peptide (flow-throught fraction from the second purification).

Fig. 3
figure 3

A comparison between two peptides sLL-37 and recombinant GLL-37, using 18% Tricine SDS-PAGE. Lane 1, sLL-37; Lane 2, recombinant GLL-37; kDa, protein standard

Full size image

The average yields for purified fusion protein UBQ–GLL37 and GLL-37 from one litre of culture were 69.5 and 18.6 mg, respectively. The final yield of a whole peptide GLL-37 depends on the efficiency of TEV protease cleavage.

Comparison of antibacterial activity between synthetic and recombinant peptide

The synthetic peptide showed activity against multidrug-resistant pathogens and reduced their numbers by 4 to 5 orders of magnitude compared to the growth control at a concentration of 250 µg/mL. The recombinant peptide GLL-37 did not cause statistically significant reduction in bacterial numbers and did not show antimicrobial activity (Fig. 4).

Fig. 4
figure 4

Comparison of antibacterial activity between recombinant peptide GLL-37 and sLL-37 (c = 250 µg/mL) against multidrug resistant bacteria (a) Acinetobacter baumannii and (b) Pseudomonas aeruginosa (AMIT method)

Full size image

Secondary structure of sLL-37 and GLL-37

CD spectra of sLL-37 and recombinant GLL-37 were measured to determine whether the problem of bacterial inactivity of GLL-37 is due to an incorrect secondary structure. As shown in Fig. 5, the peptide GLL-37 is in the α-helix form. Figure 5 also shows the CD spectrum of sLL-37, which has a random-coil form.

Fig. 5
figure 5

(a) CD spectra of sLL-37 (at pH 7 and 25 oC in dH2O) and (b) recombinant peptide GLL-37 at pH 8 and 25 oC in 25 mM TRIS, 25 mM NaCl, and 2 mM DTT

Full size image

GLL-37 and sLL-37 in contact with membrane

Synthetic LL-37 forms a random-coil conformation in an aqueous solution. After incubation in 2,2,2-trifluoroethanol/water (25% v/v) solution (mimicking the membrane environment – Chaloin at al. 1997), the secondary structure was changed. In Fig. 6, we can see that there is a change in the sLL-37 spectrum towards an α-helix form (green curve). For the recombinant peptide GLL-37, no change in conformation was observed after the addition of trifluoroethanol; the peptide remained in the α-helix conformation.

Fig. 6
figure 6

CD spectra of peptides GLL-37 and LL-37 in buffers containing trifluorethanol (TFE). CD spectra were recorded after exposing the samples to 5% and 25% (TFE/water) for 10 min. Helical content: GLL-37: 64.0%, LL-37: 1.6%, sLL-37 25% TFE: 21.0% evaluated by the bioinformatics tool BESTSEL

Full size image

GLL-37 denaturation by urea

Figure 7. shows the effect of urea on the denaturation of the GLL-37 peptide. Synthetic LL-37 (displaying a random-coil spectrum) was used as a control. The spectrum of GLL-37 without the addition of urea forms α-helix. As the concentration of urea increases, the GLL-37 peptide undergoes a transition from an α-helix conformation to a random-coil form. Even at the highest concentration of urea (10 M), the peptide is not completely denatured.

Fig. 7
figure 7

CD spectra of GLL-37 peptide denaturation obtained at different urea concentrations (0–10 M). Incubation with urea was performed at RT overnight

Full size image

Discussion

The aim of this work was to try to produce in E. coli the active antibacterial peptide LL-37 in high yields. Antimicrobial peptides are toxic to bacterial cells (Büyükkiraz & Kesmen 2022). Therefore, it is necessary to produce these peptides in fusion with protein tag to avoid toxicity, increase the yield, and improve stability (Deo et al. 2022). We used a custom expression vector containing the ubiquitin fusion protein. We also optimized the LL-37 gene sequence (codon usage) for heterologous expression in E. coli. Higher yields of LL-37 peptide production may bring it closer to clinical application. Recombinant production of the peptide will be cheaper than its production by chemical synthesis.

The recombinant production of the UBQ-LL37 fusion protein results in two forms as shown in Fig. 2, although the MS results indicate (data not shown) that these forms are identical proteins. Figure 3 displays the molecular weight comparison between recombinant GLL-37 and sLL-37. Additionally, it illustrates the purity of GLL-37 after undergoing two purification steps. For GLL-37 produced in this way, the secondary structure can be accurately measured using CD spectroscopy. So far, a total of eight different methods of producing LL-37 peptide in E. coli have been published. In most of these publications (Moon et al. 2006, 2011; Krahulec et al. 2010; Li 2012, 2013), the yield of fusion protein with LL-37 ranged from 31.5 to 37.8 mg per 1 L of culture, except for the publication Krahulec et al. 2010. Our work achieved up to twice the yield of fusion protein compared to these publications. Krahulec et al. 2010 achieved a fusion protein yield of up to 1 g/L of culture, and after cleavage with enterokinase and purification, they obtained 40 mg of pure LL-37. However, it should be noted that they used a fermentor for cultivation.

Recombinant GLL-37 produced by us did not show antimicrobial activity compared to sLL-37. The presence of an extra amino acid (glycine) will not be a problem, as Yang et al. 2004 measured the activity of a peptide extended by two amino acids (GSLL-37). Also, Shi et al. 2008 observed the antibacterial activity of the GLL-37 peptide despite the extra glycine upstream of the peptide sequence. We assumed that the problem might lie in the secondary structure on which the LL-37 peptide mechanism depends. Synthetic (random-coil) and recombinant (α-helix) peptide LL-37 formed different structures in aqueous medium. Porcelli et al. 2008 hypothesized, based on NMR experiments (Henzler-Wildman et al. 2003, 2004), that the LL-37 peptide in its random coil form interacts with the membrane and then transforms into an α-helix. This structure causes the peptide to burrow into the membrane and disrupt its integrity (carpet model mechanism).

This NMR observation is confirmed by our CD spectra (Fig. 6) where the sLL-37 undergoes a structural transition to the helical conformation upon the addition of TFE. This model represents the situation when the peptide contacts the bacterial membrane. The recombinant peptide GLL-37 lacks antibacterial activity because as an α-helix form it cannot approach the membrane and change its structure (as shown in Fig. 6). It is likely that the recombinant peptide GLL-37 folded during expression while being fused with ubiquitin. For the other expression systems, the authors were more successful and managed to express the antimicrobially active peptide LL-37 (see above). These authors used other expression systems that did not cause the peptide to fold into α-helix form.

A possible solution to change the secondary structure of the peptide to a random-coil is through denaturation using urea. However, even at a 10 M concentration of urea, the GLL-37 peptide could not be completely denatured (only 50% denaturation). In addition, this step introduces another procedure into the production of the GLL-37 peptide, namely the removal of urea from the peptide solution, which is not desirable. Measurement the cytotoxicity of both peptides synthetic and recombinant is planned in future studies together with the measurement of imunomodulatory activity.

Conclusion

LL-37 is the most studied human antimicrobial peptide that effectively inhibits bacterial growth through direct disruption of the bacterial membrane, but also has a stimulatory effect on the immune system. In this study, recombinant peptide GLL-37 was produced in high amounts by own expression system, but did not show antibacterial activity unlike synthetic LL-37. The reason for the inactivity is the incorrect conformation of GLL-37, which is in the form of an α-helix. In this work, we present evidence that for a peptide to disrupt bacterial membranes, it must be in the form of a random coil.