Intein-mediated backbone cyclization of entolimod confers enhanced radioprotective activity in mouse models

Cloning and construction of lin- and circ-entolimod expression plasmids

The full-length flagellin gene was amplified by forward primer 1 carrying the NdeI restriction site and StrepII-tag (Table 1, bold font) and reverse primer 1 with the HindIII site and His₆-tag (Table 1, bold font) from Salmonella enterica serovar Dublin genome, as previously described (Burdelya et al., 2008). We used the flagellin gene as a template and amplified the entolimod N-terminal using forward primer 1 and reverse primer 2, and the C-terminal using forward primer 2 and reverse primer 1. The PCR products of the entolimod N-terminal and C-terminal were mixed and used as templates to amplify entolimod using forward primer 1 and reverse primer 1. The sequence 5′-TCCCCGGGAATTTCCGGTGGTGGTGGTGGAATTCTAGACTCCATGGGT-3′ was the linker between entolimod N-terminal and C-terminal. The resulting DNA segment was ligated into pET-28a(+) (Novagen #69864-3), resulting in pET/entolimod-28a(+). We then used the same strategies to clone Ssp DnaE intein and Npu DnaE intein into pET-28a(+). We amplified Ssp I_C using forward primer 3 carrying an NcoI restriction site and reverse primer 3 carrying a BamHI site from pSFBAD09 (Addgene # 11963, Züger & Iwai, 2005), the product was ligated into pET-28a(+), to generate pET/Ssp I_C-28a(+), the italicized base in reverse primer 3 contain a linker sequence with KpnI, BamHI, and NdeI cloning sites and three different termination codon frames TAA, TGA, and TAG. Ssp I_N was amplified using forward primer 4 with a BamHI site and reverse primer 4 with a HindIII site from pJJDuet30 (Addgene # 11962, Züger & Iwai, 2005), and then ligated into pET/Ssp I_C-28a(+) to produce pET/Ssp DnaE-28a(+). We obtained pET/Npu I_C-28a(+) using forward primer 5 and reverse primer 5 with the same restriction enzymes as those used for pET/Ssp I_C-28a(+) from pSKBAD02 (Addgene # 15335, Iwai et al., 2006), and pET/Npu DnaE-28a(+) was constructed using forward primer 6 and reverse primer 6 with the same restriction enzymes as pET/Ssp DnaE-28a(+) from pSKDuet01 (Addgene # 12172, Iwai et al., 2006). Lastly, we used forward primer 1 and primer 7 carrying a BamHI site, His₆-tag (Table 1, bold font) to amplify entolimod from pET/entolimod-28a(+), and then produce pET/Ssp DnaE-28a(+) and pET/Npu DnaE-28a(+) to generate the final circ-entolimod expression plasmids pET/Ssp DnaE/entolimod-28a(+) and pET/Npu DnaE/entolimod-28a(+).

Expression of recombinant entolimod proteins and cell lysis

The expression plasmids pET/entolimod-28a(+), pET/Ssp DnaE/entolimod-28a(+) and pET/Npu DnaE/entolimod-28a(+) were transformed into E. coli BL21(DE3) (Tiangen Biotech, Beijing, China). The cells were grown overnight at 37 °C in LB medium supplemented with 50 μg/mL kanamycin. Protein overexpression was induced by adding isopropyl β-D-1-thiogalactopyranoside (IPTG) to a final concentration of 0.5 mM in the bacterial culture (when optical density at 600 nm reached ∼0.5) and the incubation was extended for additional 3–5 h at 37 °C. The cells containing the expression plasmids pET/entolimod-28a(+), pET/Ssp DnaE/entolimod-28a(+), and pET/Npu DnaE/entolimod-28a(+) were subsequently harvested at 6,000×g for 25 min. After centrifugation, the three cell pellets were respectively sonicated (30% amplitude for 30 min, 5 s on, 5 s off) in lysis buffer: 20 mM Tris, 500 mM NaCl, 10 mM Imidazole, pH 8.0. The target proteins expressed as inclusion bodies (∼80% of total) were confirmed by SDS–PAGE analysis.

Purification of lin- and circ-entolimod proteins

The lin-entolimod was purified described previously (Burdelya et al., 2008). First, the lin-entolimod-expressing cells were suspended in lysis buffer comprising 20 mM Tris, 500 mM NaCl, and 10 mM Imidazole, pH 8.0, followed by sonication in an ice water bath to minimize thermal damage to proteins. The supernatant and pellet were then collected. The pellet was used for further purification and could be completely solubilized in Buffer A (20 mM Tris, 500 mM NaCl, 2 M Urea, pH 8.0). Then, the lin-entolimod was further purified by Ni-chelating affinity column chromatography (GE Healthcare, Little Chalfont, UK) and could be readily eluted in Buffer B (20 mM Tris, 500 mM NaCl, 250 mM Imidazole, pH 8.0). We further purified lin-entolimod according to the procedure described previously, and performed desalting followed by size-exclusion chromatography (GE Healthcare, Little Chalfont, UK) (Burdelya et al., 2008). However, the same purification methods were not suitable for circ-entolimod. Like lin-entolimod, circ-entolimod was also completely solubilized in Buffer A as described above. The first one-step elution of lin-entolimod (Ni-chelating affinity column chromatography) was not suitable for circ-entolimod purification. Therefore, we increased the concentration of imidazole (0–500 mM) in a step-wise manner to elute circ-entolimod (three imidazole gradients: 50, 200, and 500 mM). In this step, an alternative strategy for purifying circ-entolimod (or lin-entolimod) is Strep-Tactin affinity purification system (IBA Lifesciences, Goettingen, Germany) (according to manufacture’s instructions). After purification by Ni-chelating affinity chromatography, the total sample was desalted (20 mM phosphate buffer (PB), pH 7.4) using a G25 column (GE Healthcare, Little Chalfont, UK). After desalting, ion-exchange chromatography (IEC) was performed on a column of HiTrap Q HP (anion exchange) resin (GE Healthcare, Little Chalfont, UK). We used 20 mM PB, pH 7.4 as the ion-exchange binding buffer. The circ-entolimod was then eluted using an increasing salt (10 mM–1 M NaCl) gradient. Protein concentration was measured by a BCA Protein Assay Kit (Tiangen Biotech, Beijing, China) and protein fractions were analyzed by SDS–PAGE stained with Coomassie blue. Endotoxin was measured by limulus lysate agent. The buffers of purified lin-entolimod and circ-entolimod were changed to phosphate-buffered saline (PBS) and the proteins were stored at −80 °C until used.

MALDI-TOF imaging mass spectrometry

Mass spectrometer (MS) was performed on an ultrafleXtreme MALDI-TOF/TOF MS instrument (Bruker Daltonics, Billerica, MA, USA). The sample solution is mixed in equal volumes with matrix solution (containing 4 mg/ml CHCA, 80%CAN and 0.1%TFA), and then take 1 μl of the mixed sample point to the target for dry detection. The MS parameters were as follows, mode: LP 5–50 KDa, m/z range 10 to 60 K, 500 laser shots per spectrum, ion source voltage 1 set at 20 kV, ion source voltage 2 set at 18.40 kV, lens voltage set at 6.2 kV.

NF-κB-dependent dual-luciferase reporter assays

Human embryonic kidney cells 293 (HEK293) (ATCC # CRL-1573) at 70–90% confluence were transfected with plasmids pNF-κB-Luc (#219078; Agilent, Santa Clara, CA, USA) and pRL-SV40 (#E2231; Promega, Madison, WI, USA) using Lipofectamine^™ 2000 (#11668019; ThermoFisher Scientific, Waltham, MA, USA) as per the manufacturer’s protocol. The lin- and circ-entolimod were added to the cells at different concentrations (0.00001, 0.001, and 0.1 μM) at 24 h after transfection. Firefly and Renilla luciferase assays were performed 6 h after addition of lin- and circ-entolimod to the culture medium using the dual-luciferase reporter assay system (#E1980; Promega, Madison, WI, USA).

Irradiation and radioprotection of mice

C57BL/6 mice were obtained from Vital River Laboratories (VRL) (Beijing, China). All animal experiments were performed according to the Guide for the Care and Use of Medical Laboratory Animals (Ministry of Health, China, 1998) and with the ethical approval of the Beijing Institute of Biotechnology. The approval documentation number is No.545 ([2001]). Male mice, 6–8 weeks old, weighing 18–22 g were used for the experiment (12 animals per group, six groups, a total of 72 animals). The mice were divided into three groups with similar weight distribution: PBS (control) group, lin-entolimod treatment group, and circ-entolimod treatment group. The mice were subjected to total body irradiation (TBI) with ⁶⁰Co-gamma to a total dose of 9 and 14 Gy at a dose rate of 153.33 cGy/min (Institute of Radiation Medicine, Beijing, China). Mice were irradiated on a rotating platform to ensure even dose delivery to all tissues. The lin- and circ-entolimod were injected subcutaneously (s.c.) 30 min before TBI at a dose of 0.2 mg/kg. The PBS group was injected an equivalent volume of PBS instead. The survival rate of mice was observed 30 days after TBI.

Statistical analysis

Each column is presented as means ± SD of three independent experiments. The results were statistically evaluated for significance using the Student’s t-test. The survival rate was statistically evaluated for significance using the log-rank test. P < 0.05 was considered statistically significant.

Npu DnaE split intein-mediated entolimod cyclization

According to the principle of PTS, the two fragments of intein, I_N and I_C, interact with each other to form an active intein that splices to cyclize the proteins or peptides placed in between. In this case, the entire entolimod gene was sandwiched between the I_C and I_N of Ssp or Npu intein with CFN residues at the C terminus from the native C-terminal extein sequence of DnaE intein, and HM and GS residues at its C terminus and N terminus due to cloning (Figs. 1A–1C).

The circ-entolimod was readily apparent by SDS–PAGE upon IPTG induction under the expression plasmids pET/Npu DnaE/entolimod-28a(+). Figure 2A shows the expression and the expected splice product circ-entolimod (∼33.7 kDa); the intein (I_N and I_C) band was clearly observed in two fragments, I_N (∼11.9 kDa) and I_C (∼4.1 kDa), after splicing. The circ-entolimod migrated more rapidly in SDS–PAGE analyses than did lin-entolimod (Fig. S1), implying an additional topological constraint (Iwai & Plückthun, 1999; Zhao et al., 2010; Qi & Xiong, 2017). As reported previously, lin-entolimod is mainly expressed in the form of inclusion bodies (∼80%), but is resistant to thermal denaturation at 90 °C for 20 min and does not need refolding (Burdelya et al., 2008). In our study, after sonication, circ-entolimod also predominantly formed inclusion bodies (Fig. 2A). We then tested whether entolimod cyclization could be mediated by Ssp intein. Unfortunately, the results indicated that pET/Ssp DnaE/entolimod-28a(+) failed to induce the expression of circ-entolimod, and the precursor (Ssp DnaE- entolimod) was clearly observed in the gel (Fig. 2B). Therefore, Npu intein was considered a better choice for entolimod cyclization. Previous studies have shown that Npu intein is more tolerant of amino acid substitutions in the C-terminal extein sequence than Ssp intein (Townend & Tavassoli, 2016), and also has exhibited the highest efficiency for the PTS reaction so far (t_1/2 of ∼60 s), which was 33–170-fold higher than that of Ssp intein (Zettler, Schütz & Mootz, 2009). Thus, we could successfully express circ-entolimod by Npu intein-mediated its cyclization.

Purification and thermo stability of lin- and circ-entolimod

In this step, we chosed different strategies to purify lin- and circ-entolimod, respectively. After analyzing the purity of lin-entolimod (Ni-chelating affinity) by SDS–PAGE, we only found a few bands below lin-entolimod (Fig. 3A). However, the one-step elution was not suitable for circ-entolimod purification. Therefore, we increased the concentration of imidazole (0–500 mM) in a step-wise manner to elute circ-entolimod. Figure 3B shows the result of using three imidazole gradients (50, 200, and 500 mM) to elute circ-entolimod. Unfortunately, it could not yield the same purity as lin-entolimod. We further purified lin-entolimod according to the procedure described previously, and performed desalting followed by size-exclusion chromatography (Burdelya et al., 2008). Through these two purification steps, lin-entolimod was purified to >95% (Fig. 3C). However, size-exclusion chromatography to purify circ-entolimod yielded results similar to those obtained using Ni-chelating affinity purification. As predicted by an online protein isoelectric point calculator tool, the isoelectric point of circ-entolimod was approximately 5.6. We, therefore, decided to purify circ-entolimod by anion exchange chromatography using a HiTrap Q HP resin. As shown in Fig. 3D, circ-entolimod with high purity was obtained with the elution buffer containing 10 and 20 mM NaCl (lane 3 and 4). These two were then combined as the final circ-entolimod purification (Fig. 3E). To further confirm the purified product was circ-entolimod, we used MALDI mass spectrometry and Q Exactive MS to analyze its molecular weight and sequence, the results indicated that both size and sequence were right as expected (Fig. 4 and Fig. S2). In our study, at various temperatures (from 90 to 100 °C), the linear and circular form of entolimod were both performed the enhanced thermostability (Fig. S3).

Effect of circ-entolimod on NF-κB pathway activation

To investigate whether circ-entolimod has in vitro biological activity, we added both the lin-entolimod and circ-entolimod to the dual-luciferase reporter plasmids transfected HEK293 cells. Then the luciferase activity was measured to check/determine the in vitro activity that the both products activate the NF-κB-dependent pathway by inducing production of numerous bioactive factors including anti-apoptotic proteins, reactive oxygen species (ROS) scavengers, cytokines, and anti-inflammatory agents (Burdelya et al., 2013; Lockless & Muir, 2009).

After purification, we measured and controlled the contamination levels of less than 5 endotoxin units, which was approximately equivalent to 0.5 ng of E. coli Lipopolysaccharides (LPS) per microgram of recombinant protein. Based on this level, LPS had minimal impact on the following detection, and then we compared the biological activity of lin-entolimod and circ-entolimod at three different concentrations (0.00001, 0.001, and 0.1 μM). A total of three independent replicates were set up for each concentration and PBS control. The results shown that circ-entolimod exhibited an excellent performance at all three concentration levels. The differences between the corresponding lin-entolimod and circ-entolimod groups were statistically significant (Fig. 5, **P < 0.01). Additionally, even at the very low concentration (0.00001 μM), circ-entolimod still activated NF-κB signaling at approximately 40% higher than that by lin-entolimod (Fig. 5). Cyclization of proteins has been previously reported to enhance biological activity (Iwai & Plückthun, 1999; Zhao et al., 2010; Qi & Xiong, 2017). In the present work, our novel protein circ-entolimod also showed a similar outcome. These results thus indicate that circ-entolimod has a significant biological activity in NF-κB pathway activation.

Effect of circ-entolimod on mouse survival after irradiation

As an anti-radiation drug, lin-entolimod showed a significant radioprotective effect on mouse and primate models. To address whether circ-entolimod has a in vivo protective effect same as or even better than that of lin-entolimod, we injected both agents s.c. into C57BL/6 mice at 0.2 mg/kg of body weight. We administered lin-entolimod and circ-entolimod 30 min prior to irradiation at two different doses (9 and 14 Gy), and compared them with control mice that received PBS (12 animals per group, in total 72 mice). We determined the survival rate of mice at 30 days after TBI at radiation doses of 9 and 14 Gy. At the lower radiation dose of 9 Gy, which caused lethality to all the PBS-treated mice, there was no significant difference in survival rate between circ-entolimod and lin-entolimod-treated mice (Fig. 6A, P = 0.3173). However, in the higher radiation dose of 14 Gy group, 40% of the circ-entolimod-treated mice survived whereas all lin-entolimod-treated mice were caused lethality by the ninth day (Fig. 6B, **P < 0.01). These results indicate that circ-entolimod has a better radioprotective effect than lin-entolimod at higher radiation doses. Thus, circ-entolimod is a potential novel and effective drug with radioprotective application.