Designed Antimicrobial and Antitumor Peptides with High Selectivity

Communication pubs.acs.org/Biomac

Designed Antimicrobial and Antitumor Peptides with High Selectivity Jing Hu,†,⊥ Cuixia Chen,†,⊥ Shengzhong Zhang,† Xichen Zhao,‡ Hai Xu,*,† Xiubo Zhao,§ and Jian R. Lu*,§ †

State Key Laboratory of Heavy Oil Processing and the Centre for Bioengineering and Biotechnology, China University of Petroleum (East China), Qingdao 266555, P. R. China ‡ The First People’s Hospital, Qingdao Economic & Technological Development Zone, Qingdao 266555, P. R. China § Biological Physics Group, School of Physics and Astronomy, University of Manchester, Schuster Building, Oxford Road, Manchester M13 9PL, United Kingdom S Supporting Information *

ABSTRACT: We report a new class of cationic amphiphilic peptides with short sequences, G(IIKK)nI-NH2 (n = 1−4), that can kill Gram-positive and Gram-negative bacteria as effectively as several well-known antimicrobial peptides and antibiotics. In addition, some of these peptides possess potent antitumor activities against cancer cell lines. Moreover, their hemolytic activities against human red blood cells (hRBCs) remain remarkably low even at some 10-fold bactericidal minimum inhibitory concentrations (MICs). When bacteria or tumor cells are cocultured with NIH 3T3 fibroblast cells, G(IIKK)3I-NH2 showed fast and strong selectivity against microbial or tumor cells, without any adverse effect on NIH 3T3 cells. The high selectivity and associated features are attributed to two design tactics: the use of Ile residues rather than Leu and the perturbation of the hydrophobic face of the helical structure with the insertion of a positively charged Lys residue. This class of simple peptides hence offers new opportunities in the development of cost-effective and highly selective antimicrobial and antitumor peptide-based treatments.

A

natural defense, thereby imposing possible threats to public health.12 To circumvent the hurdles limiting their applications and also to elucidate the structure−activity relationship, tremendous efforts have been put into sequence optimization and modification based on natural peptide sequences and traits, including the use of unnatural building blocks such as D-amino acids, β-amino acids, and fluorinated amino acids.13−27 Peptide charges, amphipathicity, and the size of hydrophobic/hydrophilic domain have been shown to affect the antimicrobial activity and selectivity.18,20 Natural magainins with high antibacterial activities but little hemolytic effect provide a good basis for selectivity optimization, in which the cationic charges are spread over the whole peptide chain and hydrophobic distributions (amphipathicity) are moderated.18,28,29 Furthermore, the shorter the peptide sequence, the less the manufacturing cost and the easier the investigation of the structure−activity relationship. Taking the above issues into consideration, we have focused our effort on optimizing short, cationic amphiphilic peptides.

ntibiotic resistance is a natural evolutionary phenomenon that is becoming increasingly difficult to tame. The recent emergence and spread of NDM-1 (New Delhi metallo-βlactamase-1) producing bacteria, resistant to many groups of antibiotics including the powerful β-lactams (e.g., carbapenems), has led to infections across many countries. The global panic has again fueled the urgent need to discover novel bactericidal agents.1−3 Antimicrobial peptides (AMPs) are the evolutionarily conserved effectors in innate immunity. Their broad-spectral bactericidal activity, rapid killing rate, and the distinctive mode of action (i.e., targeting the bacterial cell membrane itself rather than specific receptors such as proteins and DNA) have made them promising candidates for the development of alternative agents to cope with the widespread challenges of bacterial resistance.4−8 Although over 1000 AMPs have been isolated and characterized from different sources, only limited successes have so far been achieved in clinical trials. The major barriers for converting them into drugs lie in the high cost of production on a large scale, toxicity to host cells, and susceptibility to proteolytic degradation. 9−11 Furthermore, concerns have recently emerged from the clinical use of AMPs with sequences that are too close to those of natural human AMPs relating to the inevitable compromise of human © 2011 American Chemical Society

Received: August 8, 2011 Revised: September 14, 2011 Published: September 28, 2011 3839

dx.doi.org/10.1021/bm201098j | Biomacromolecules 2011, 12, 3839−3843

Biomacromolecules

Communication

Figure 1. CD spectra of the G(IIKK)nI-NH2 (n = 1−4) peptides in water (A), 0.5 mg/mL DPPC small unilamellar vesicles (SUVs) in 10 mM Tris buffer (B), and 0.5 mg/mL DPPG SUVs in 10 mM Tris buffer (C). The concentrations of peptides were maintained at 0.1 mM.

Table 1. Bioactivities and Cytotoxicities of G(IIKK)nI-NH2 (n = 1−4) Peptides, with Two Natural Antibacterial Peptides, Antibiotic Ampicillin and Anticancer Drug Cisplatin Cited for Comparison a MIC (μM)

IC50 (μM)

peptides

E. coli

B. subtilis

G(IIKK)I G(IIKK)2I G(IIKK)3I G(IIKK)4I magainin-2b melittinc ampicillind cisplatine

125 ± 5.6 8 ± 0.2 2 ± 0.5 38 3.9 ± 0.6 ∼11

30 ± 2.5 2 ± 0.5 0.5 ± 0.1 >80 2.0 ± 0.2 5.5

HeLa 160 ± 7.6 15 ± 5.0 4 ± 2.3 >60

HL60 >500 25 ± 5.0 10 ± 2.5

EC50 (μM)

>250 41 ± 2 430 3

1.34 ± 0.32

a

MIC: the lowest peptide/drug concentration to inhibit bacterial growth; IC 50: concentration causing 50% tumor cell growth inhibition; and EC 50: concentration to induce 50% lysis of erythrocytes. Note that the antimicrobial activities of natural AMPs and ampicillin were also determined by using the presently used microdilution method; the antitumor activity of cisplatin was assessed via the WST-1 colorimetric assay, similar to the present MTT method. bCited from refs 22, 30, and 31. cCited from ref 32. dCited from ref 33. eCited from ref 35.

Furthermore, the cytotoxicity of the peptides, determined on human erythrocytes cells, remains remarkably low at values up to 10-fold MICs. More importantly, G(IIKK)3I-NH2 showed a rapid and high selectivity to microbes or tumor cells (HL60 cancer cell line) with minimal cytotoxicity to model host cells (NIH 3T3 cell line) when cocultured in vitro. AMPs are highly diverse in terms of length, sequence, and structure, making it challenging to design short and efficient sequences. Most antibacterial peptides however share two common hallmarks, i.e., being cationic and amphipathic. In designing this series of cationic peptides, we have tested various combinations of amphiphilic sequences by aligning hydrophobic and cationic amino acids differently. When the designed peptides contain the simple repeat sequence of IIKK, the ratio of hydrophobic isoleucine to cationic lysine residues remains at 1:1, consistent with the requirements for cationicity and amphipathicity. All of them have been capped with Gly at the N-terminus and amidation at the C-terminus which were believed to not only provide resistance to peptidases but also favor α-helical hydrogen bonding.17,20 The peptides were synthesized by solid-phase peptide synthesis on Rink amide MBHA resin and were all purified to >95% homogeneity. The mass of each peptide was confirmed by MALDI-TOF MS (Figure S1 and Table S1 in the Supporting Information). Circular dichroism (CD) spectroscopy was used to assess the folded structures of these peptides in aqueous solution and in membrane-mimetic lipid vesicles. All the peptides were

Figure 2. Hemolytic activities of G(IIKK)nI-NH2 (n = 1−4) series. Human red blood cells (hRBCs) were incubated in PBS with different concentrations of peptides for 1 h at 37 °C, followed by the monitoring of hemoglobin release at 540 nm.

We here describe the design and the biological evaluation of a novel class of short peptides containing simple sequence repeats, G(IIKK)nI-NH2 (n = 1−4). Our studies have demonstrated that the G(IIKK)nI-NH2 (n = 3−4) peptides not only displayed potent bactericidal activity against both Gram-positive and Gram-negative bacteria with minimum inhibitory concentrations (MICs) around 0.5−8 μM but are also effective against two selected cancer cell lines with 50% growth inhibition concentrations (IC50) around 4−25 μM. 3840

dx.doi.org/10.1021/bm201098j | Biomacromolecules 2011, 12, 3839−3843

Biomacromolecules

Communication

Figure 3. Fluorescent images of the distribution of FITC-labeled G(IIKK)3I-NH2 in two coculture systems containing NIH 3T3 cells (model mammalian host) with bacteria B. subtilis (upper panel) or HL60 cancer cells (bottom panel). (A) and (D) are the images observed under the bright field; (B) and (E) are the images acquired under blue exciting light; (C) and (F) are the overlapping images of (A) and (B), and (D) and (E), respectively. In these images, adherent cells are NIH 3T3, rodlike cells are B. subtilis, and round suspension cells are HL 60 as marked by black arrows.

Figure 4. Hemolytic activities of G(IIKK)3I-NH2, GKI(KKII)2KIINH2, and G(LLKK)3L-NH2. Human red blood cells (hRBCs) were incubated in PBS with different concentrations of peptides for 1 h at 37 °C, followed by the monitoring of hemoglobin release at 540 nm.

Figure 5. The Schiffer−Edmundson wheel projection of the G(IIKK)3I-NH2 (A) and GKI(KKII)2KII-NH2 (B). Yellow circles indicate hydrophobic amino acids, and blue circles represent hydrophilic amino acids (drawn by using the online tool of the Web site http://heliquest.ipmc.cnrs.fr/cgi-bin/ComputParams.py). In the figure, m represents the number of coils of α-helix. The horizontal line in (B) separates the hydrophobic and the hydrophilic surface sections.

unfolded in the aqueous solution (Figure 1A) due to the considerable intermolecular hydrogen bonding between peptide backbones and the solvent molecules and also the intramolecular cationic repulsion between the neighboring Lys residues. In the negatively charged DPPG (1,2-dipalmitoylsn-glycero-3-phosphoglycerol) SUVs (mimicking bacterial and tumor cell membranes), all the peptides except G(IIKK)I-NH 2 adopted the typical α-helix structure, with minimal mean residual molar ellipticity values at 208 and 222 nm (Figure 1C). In the zwitterionic DPPC (1,2-dipalmitoyl-sn-glycero-3-phosphocholine) SUVs (mimicking normal mammalian host cell membranes), all of them remained unstructured (Figure 1B). Thus, the CD measurements not only unraveled the delicate interplay between hydrophobic effect, charge interaction, and hydrogen bonding crucial for the helix formation but also

suggested the molecular basis of peptide selectivity between different cell membranes. The antimicrobial activities of the peptides were measured against Gram-negative Escherichia coli (E. coli DH 5 α) and Gram-positive Bacillus subtilis (B. subtilis 168), with the results shown in Table 1. G(IIKK)I-NH2 has the weakest activity, and it did not show much antimicrobial and antitumor effect even up to the concentration of 1 mM, consistent with its short length (n = 1) and lack of structural responses. The bactericidal activity increased significantly with increasing peptide length or repeat unit, consistent with the enhanced helical structuring, amphipathicity, and cationicity. G(IIKK) 3I-NH2 and G(IIKK)4I-NH2 are particularly effective against both bacteria, with MICs in the range of 0.5−8 μM. These MIC values are 3841

dx.doi.org/10.1021/bm201098j | Biomacromolecules 2011, 12, 3839−3843

Biomacromolecules

Communication

highly comparable to those of melittin and ampicillin with MICs around 2−10 μM, making them as potent as many natural AMPs and commercial antibiotics.22,30−33 Note that the designed peptides are more effective against Gram-positive B. subtilis than Gram-negative E. coli, also similar to melittin and ampicillin. Similar to bacteria, tumor cell membranes carry net negative charges, a feature different from normal mammalian cells. Only a small number of cationic AMPs display antitumor effects. 34 Interestingly, the peptides used in this work also show high activities against tumor cells. The antitumor results assessed from HeLa and HL60 cancer cell lines show similar trend to antibacterial activities with the strength increasing with n (Table 1). G(IIKK)3I-NH2 and G(IIKK)4I-NH2 show high potency against both types of cancer cell lines, with IC 50 values (concentration to induce 50% tumor cell inhibition) in the 4− 25 μM range (Table 1). The activity of G(IIKK)4I-NH2 against HeLa cells is comparable to cisplatin, one of the most widely used and potent chemotherapeutic agents against cancers.35 For a peptide to be of pharmaceutical interest, high antimicrobial and antitumor activities must be combined with low toxicity against normal host cells. We first determined their hemolytic activity using fresh human red blood cells (hRBCs). Dose−response curves for the hemolytic activities of the peptides are shown in Figure 2. The broad trend is similar to their antimicrobial and antitumor activities, that is, potency increases with increasing n. Importantly, however, over the active concentration ranges against bacteria and cancer cells, there is little activity or toxicity against hRBCs. This feature is crucial to their therapeutic applications. Specifically, the two short peptides G(IIKK)I-NH2 and G(IIKK)2I-NH2 are inactive up to 250 μM. Although peptide G(IIKK)3I-NH2 showed some hemolytic activity, it only caused less than 10% hemolysis of erythrocytes at 10-fold MIC (80 μM) against E. coli and very little hemolysis at 10-fold MIC (20 μM) against B. subtilis. G(IIKK)4I-NH2 had the greatest hemolytic activity in the series studied, producing ∼25% lysis of erythrocytes at the 10-fold MIC (20 μM) against E. coli. Overall, G(IIKK)3I-NH2 offers an optimal selectivity against bacteria and cancer cells while its toxicity to hRBCs remains low, reflecting the balance of the effects of peptide length. To further confirm the selective responses and cytotoxicity of these peptides, we synthesized the FITC-labeled G(IIKK)3INH2 according to the mothod desribed previously by Jullian et al.36 The labeled peptide was added into two coculture systems containing both bacteria (B. subtilis 168) or HL60 cancer cells and model host mammalian cells (NIH 3T3). After incubation for 1 h, the distribution of FITC-G(IIKK)3I-NH2 was observed using fluorescence microscopy, and the results are shown in Figure 3. The green fluorescence of FITC-G(IIKK)3I-NH2 is concentrated in the bacterial clusters or the HL60 cells but not in any part of the NIH 3T3 cells. This further confirms that G(IIKK)3I-NH2 shows fast responses and highly effective recognition of pathogenic bacteria and the tumor cells in the coculture environment, with no binding or association to the model host cells. Such a coculture system can mimic the in vivo infected environment better than commonly used parallel assays. To our best knowledge, this is the first in vitro selectivity assay ever reported for such studies. The prime feature of the current peptides is their selectivity against bacteria and tumor cells while remaining benign to model mammalian cells. We attribute this characteristic behavior to two aspects of peptide structural design. First, we

used Ile instead of Leu as constituent hydrophobic residues. Given that Ile generally displays great β-sheet forming potential and Leu has strong helix-forming propensity and is highly abundant in many AMPs, the use of Ile instead of Leu in helix conserved sequences works to weaken their helical forming propensity. We observed that the designed peptides (G(IIKK)nI-NH2 (n = 2−3)) formed amphipathic α-helical structures in the bacterial membrane-mimicking environment but were unstructured in the normal mammalian membranemimicking environment. Correspondingly, the peptides showed greatly enhanced selectivity against bacteria and tumor cells with little toxicity to normal mammalian cells. In contrast, the matching G(LLKK)nL-NH2 peptides had good antibacterial activity, but their toxicity against host mammalian cells also became significant. For example, G(IIKK)3I-NH2 and G(LLKK)3L-NH2 had almost identical killing activity against bacteria (Figure S2 in the Supporting Information), but the latter caused ∼80% hemolysis of erythrocytes at even 10-fold MIC (80 μM) against E. coli (Figure 4). Second, we introduced positively charged Lys residues into the hydrophobic face of the α-helical structure. For example, GKI(KKII)2KII-NH2 has a well-defined helical structure with clear hydrophobic and hydrophilic faces (Figure 5B), In contrast, G(IIKK)3I-NH2 has one Lys inserted into the hydrophobic Ile face (Figure 5A), making it displaying far lower hemolytic activity. Thus, such a perturbation to the hydrophobic face contributes to the significant reduction of hemolytic activity without deteriorating antimicrobial activity.27,37 Blondelle et al. have indicated that an uninterrupted sector of 5 hydrophobic residues (as seen on a helical wheel projection) is sufficient for good antimicrobial activity with reduced hemolysis.14,20 The current observation thus supports their proposition in the link between moderate destabilization of the helical wheel and improvement in selectivity. In summary, for potential clinical applications, it is important to develop potent AMPs with low toxicity to host cells and with small size and simple composition that favor low manufacturing cost and easy quality control. To this end, we have developed a novel class of cationic peptides containing simple sequence repeats, G(IIKK)nI-NH2 (where n equals 1, 2, 3, and 4). Peptide (G(IIKK)3I-NH2) exhibits potent antimicrobial activity against both Gram-positive and Gram-negative bacteria and antitumor activity against two cancer cell lines, with little cytotoxicity to model mammalian host cells. Further, the coculturing developed in this work offers selectivity assays mimicking the in vivo infected environment, further demonstrating the fast and high selective actions of the FITC labeled G(IIKK)3I-NH2. Studies are in progress to examine their antimicrobial and antitumor performance against more bacterial strains and cancer cell lines and their protease stability. Further work is also under way to assess their activities in a formulated composition containing common primary cells and to undertake in vitro immunogenicity tests.

■

ASSOCIATED CONTENT

S Supporting Information *

Detailed experimental procedures and supplementary results. This material is available free of charge via the Internet at http://pubs.acs.org. 3842

dx.doi.org/10.1021/bm201098j | Biomacromolecules 2011, 12, 3839−3843

Biomacromolecules

Communication

■

(29) Zasloff, M.; Martin, B.; Chem, H. C. Proc. Natl. Acad. Sci. U. S. A. 1988, 85, 910−913. (30) Cruciani, R. A.; Barker, J. L.; Zasloff, M.; Chen, H. C.; Colamonici, O. Proc. Natl. Acad. Sci. U. S. A. 1991, 88, 3792−3796. (31) Wieprecht, T.; Dathe, M.; Beyermann, M.; Krause, E.; Maloy, W. L.; MacDonald, D. L.; Bienert, M. Biochemistry 1997, 36, 6124− 6132. (32) Asthana, N.; Yadav, S. P.; Ghosh, J. K. J. Biol. Chem. 2004, 279, 55042−55050. (33) Greenwood, D.; O’Grady, F. J. Infect. Dis. 1970, 122, 465−471. (34) Hoskin, D. W.; Ramamoorthy, A. Biochim. Biophys. Acta 2008, 1778, 357−375. (35) Takara, K.; Obata, Y.; Yoshikawa, E.; Kitada, N.; Sakaeda, T.; Ohnishi, N.; Yokoyama, T. Cancer Chemother. Pharmacol. 2006, 58, 785−793. (36) Jullian, M.; Hernandez, A.; Maurras, A.; Puget, K.; Amblard, M.; Martinez, J.; Subra, G. Tetrahedron Lett. 2009, 50, 260−263. (37) Hawrani, A.; Howe, R. A.; Walsh, T. R.; Dempsey, C. E. J. Biol. Chem. 2008, 283, 18636−18645.

AUTHOR INFORMATION Corresponding Author *Tel (+86)532-8698-1569, e-mail [email protected] (H.X.); Tel (+44)-161-306-3926, e-mail [email protected] (J.R.L.). Author Contributions ⊥ J.H. and C.C. contributed equally to this work.

■

ACKNOWLEDGMENTS

■

REFERENCES

This work was supported by the National Natural Science Foundation of China (Grant 30900765) and the Natural Science Foundation of Shandong Province (ZR2009DQ001 and JQ201105). We thank UK Engineering and Physical Sciences Research Council (EPSRC) for support.

(1) Kumarasamy, K. K.; Toleman, M. A.; Walsh, T. R.; Bagaria, J.; Butt, F.; Balakrishnan, R.; Chaudhary, U.; Doumith, M.; Giske, C. G.; Irfan, S.; Krishnan, P.; Kumar, A. V.; Maharjan, S.; Mushtaq, S.; Noorie, T.; Paterson, D.; Pearson, A.; Perry, C.; Pike, R.; Rao, B.; Ray, U.; Sarma, J. B.; Sharma, M.; Sheridan, E.; Thirunarayan, M. A.; Turton, J.; Upadhyay, S.; Warner, M.; Welfare, W.; Livermore, D. M.; Woodford, N. Lancet Infect. Dis. 2010, 10, 597−602. (2) Moellering, R. C. Jr. N. Engl. J. Med. 2010, 363, 2377−2379. (3) Nordmann, P.; Poirel, L.; Toleman, M. A.; Walsh, T. R. J. Antimicrob. Chemother. 2011, 66, 689−692. (4) Boman, H. G. Annu. Rev. Immunol. 1995, 13, 61−69. (5) Shai, Y. Biopolymers 2002, 66, 236−248. (6) Zasloff, M. Nature 2002, 415, 389−395. (7) Yeaman, M. R.; Yount, N. Y. Pharmacol. Rev. 2003, 55, 27−55. (8) Hancock, R. E. W.; Sahl, H. Nature Biotechnol. 2006, 24, 1551− 1557. (9) Hancock, R. E. W. Lancet Infect. Dis. 2001, 1, 156−165. (10) Marr, A. K.; Gooderham, W. J.; Hancock, R. E. W. Curr. Opin. Pharmacol. 2006, 6, 468−472. (11) Peters, B. M.; Shirtliff, M. E.; Jabra-Rizk, M. A. PLoS Pathog. 2010, 6, e1001067. (12) Bell, G.; Gouyon, P. H. Microbiology 2003, 149, 1367−1375. (13) Boman, H. G.; Wade, D.; Boman, I. A.; Wåhlin, B.; Merrifield, R. B. FEBS Lett. 1989, 259, 103−106. (14) Blondelle, S. E.; Houghten, R. A. Biochemistry 1992, 31, 12688− 12694. (15) Kiyota, T.; Lee, S.; Sugihara, G. Biochemistry 1996, 35, 13196− 13204. (16) Oren, Z.; Shai, Y. Biochemistry 1997, 36, 1826−1835. (17) Tossi, A.; Tarantino, C.; Romeo, D. Eur. J. Biochem. 1997, 250, 549−558. (18) Dathe, M.; Wieprecht, T. Biochim. Biophys. Acta 1999, 1462, 71−87. (19) Hamuro, Y.; Schneider, J. P.; DeGrado, W. F. J. Am. Chem. Soc. 1999, 121, 12200−12201. (20) Tossi, A.; Sandri, L.; Giangaspero, A. Bioplymers 2000, 55, 4− 30. (21) Porter, E. A.; Wang, X.; Lee, H. S.; Weisblum, B.; Gellman, S. H. Nature 2000, 404, 565−565. (22) Dathe, M.; Nikolenko, H.; Meyer, J.; Beyermann, M.; Bienert, M. FEBS Lett. 2001, 501, 146−150. (23) Asthana, N.; Yadav, S. P.; Ghosh, J. K. J. Biol. Chem. 2004, 279, 55042−55050. (24) Hilpert, K.; Volkmer-Engert, R.; Walter, T.; Hancock, R. E.W. Nature Biotechnol. 2005, 23, 1008−1012. (25) Porter, E. A.; Weisblum, B.; Gellman, S. H. J. Am. Chem. Soc. 2005, 127, 11516−11529. (26) Meng, H.; Kuma, K. J. Am. Chem. Soc. 2007, 129, 15615− 15622. (27) Matsuzaki, K. Biochim. Biophys. Acta, Biomembr. 2009, 1788, 1687−1692. (28) Zasloff, M. Proc. Natl. Acad. Sci. U. S. A. 1987, 84, 5449−5453. 3843

dx.doi.org/10.1021/bm201098j | Biomacromolecules 2011, 12, 3839−3843