Anion mediated activation of guanidine rich small molecules

Abstract

Cell penetrating peptides (CPPs) and their synthetic analogs are of widespread interest. Here we report that guanidine rich small molecules can be potential membrane transporters in the presence of hydrophobic counteranion activators. To our knowledge, this is the first example of small molecules that mimic the anion-activated transport function of CPP.

---

Cell penetrating peptides (CPPs) and their synthetic analogs are of widespread interest as they show the remarkable ability to transverse membranes despite their high charge density.¹ They also show promise as unique delivery vehicles. Most CPPs are rich in arginines; arginines are well known to exist with tightly bound but exchangeable counteranions. According to Matile and co-workers this counteranion scavenging may be how guanidine rich molecules minimize intramolecular charge repulsion.² The weak acidity of the guanidinium group hinders partial deprotonation unlike free amine groups under physiological condition.^1g Their transduction activity has also been linked to this counterion scavenging.^1g It has been proposed that guanidine rich CPPs can alter their solubility upon counteranion exchange with more or less polar anions allowing them to adapt to their environment and cross the lipid bilayer.³ It appears that neutralization of the charge is not essential assuming the guanidine rich molecules have other hydrophobes in their structure. This ability to incorporate hydrophobic activators into the molecular structure was termed ‘self-activation’.^1o

Early on Rothbard and co-workers showed that polyamides, mostly peptides, entered cells in a length dependent fashion. Trimers, tetramers, and pentamers hardly entered Jurkat cells while hexamers showed significantly more internalization.^1i,1kIntracellular staining increased from hexamers to nonamers. Similar results have been observed for other CPPs including HIV-TAT derivatives; the nona-peptide TAT_49–57 efficiently translocated into Jurkat cells while shorter (truncated) analogs were much less active.^1i,4 Small biphenyl-based molecules with two and four guanidines were shown to enter human U2OS osteosarcoma cells, but the derivative with four guanidines worked better than the derivative with two.⁵ More recently, a number of synthetic polymers were shown to effectively mimic CPPs.^1l–o,6 In general, these papers reported molecules with more than eight repeat units, except for one case in which a pentamer was studied but this molecule contained ten guanidine units.^1l–o,4b,6 Given the widespread interest in CPP derivatives, the ability to design small guanidine rich molecules capable of transversing the plasma membrane would be exciting. Here, two guanidine rich molecules, CPPM1 and CPPM2, were designed to evaluate their transport activity (see Fig. 1).

Fig. 1 Cell penetrating peptide mimics (CPPMs) and the hydrophobic counteranion activators used in this study. The possible hydrogen bonding structures are shown between CPPM's guanidines and the activator's anions.

Model membrane studies were specifically chosen for two reasons. First, to study cellular internalization it is common to covalently attach a dye to the transporter. For these small molecules, a dye molecule would be of similar size and expected to have a significant influence on its uptake properties. Second, uptake into cells can involve a number of mechanistic pathways and we remain fundamentally interested in ‘passive diffusion’ or non-energy driven mechanisms. Of course cellular uptake studies are critical if delivery of therapeutics is to be achieved but that was not the goal of the current study. This study aimed to understand whether these two small oligomers effectively transverse plasma membranes by themselves or required assistance from hydrophobic counteranions.

CPPM1 consists of three guanidine groups and two amine groups in the side chains while CPPM2 contains five guanidine groups. This design provided the direct study of guanidine density on transport. At the same time, the role of the primary amine groups on anion mediated activation in these derivatives could be studied. CPPM1 and CPPM2 were synthesized in a few routine steps from commercially available starting materials as reported previously.⁷ Five amphiphilic anions were selected to evaluate their ability to activate these CPPMs (see Fig. 1). Pyrenebutyrate (PB) and pyreneacetate (PA) were chosen as they are the most commonly used aromatic activators for polyarginines (pR).⁸ In fact, PB was previously found to be the best activator for CPP mediated delivery.^8b,9 To directly investigate the effect of various anions in activation, three aliphatic anions containing an equal number of carbon atoms, sodium laurate (SL), sodium dodecyl phosphate (SDP), and sodium dodecyl sulfate (SDS) were picked from the known collection of aliphatic counteranion activators.^2,8a

In order to assess the ability of these CPPMs to transverse mammalian-like membranes, neutral phosphatidylcholine large unilamellar vesicles (EYPC-LUVs) were prepared and used in the classical 5(6)-carboxyfluorescein (CF) assay, well accepted in the CPP field.^1e,1m,8a,10 In this assay the release of self quenched CF from EYPC-LUVs⊃CF is monitored continuously as an increase of fluorescence intensity against time. As shown in Fig. 2a and 2d (red curves), neither CPPM showed significant activity in the absence of hydrophobic counteranions. However, addition of these counteranions had the expected activating effect as measured by increased fluorescence intensity.

Fig. 2 Changes in CF emission (I_F) (λ_ex 492 nm, λ_em 517 nm) as a function of time with increasing activator concentration [0 (red curve), 0.1, 1, 5, 10, 20, 25, 50, 100 μM PB in A; 1, 5, 10, 30, 60, 90, 180 μM SL in B; 1, 10, 30, 60, 90, 120 μM SDP in C; 0 (red curve), 0.1, 0.5, 5, 10, 20, 25, 50, 75, 100 μM PB in D; 5, 10, 30, 60, 90, 120, 180, 300 μM SL in E; 0.1, 1, 5, 10, 20, 30, 60, 120 μM SDP in F] during addition of CPPM1 (2.5 μM, final concentration) or CPPM2 (2.5 μM, final concentration) at t = 100 s to EYPC-LUVs⊃CF (50 μM EYPC), calibrated by final analysis (I_F = 1.0, with 40 μL 1.2% aqueous triton X-100).

The activity of the CPPMs increased with increasing activator concentration at a constant CPPM concentration and constant lipid concentration, yielding plots of fractional fluorescence intensity (I_F) versus time for different activator concentrations (Fig. 2). Fitting the Hill equation, Y ∝ (c/EC₅₀)ⁿ, for each individual activator revealed a nonlinear dependence of the fractional activity, Y, on the activator concentration, c. This analysis gave Y_max (maximal CF release relative to complete release by Triton X-100), EC₅₀ (effective activator concentration needed to reach Y_max/2), and n (the Hill coefficient).

Fig. 3 and Table 1 collects the EC₅₀ and Y_max values for these two CPPMs with the various activators at a constant CPPM/lipid (P/L) ratio 0.05 (guanidine to lipid ratio, 0.15 and 0.25 for CPPM1 and CPPM2 respectively). Consistent with previous reports of pRactivation,^2,8a the two aromatic activators (PB and PA) gave the highest Y_max values for these CPPMs. PB had an EC₅₀ approximately 3.5 times lower than PA for both CPPM1 and CPPM2 indicating this hydrophobic, aromatic anion activator is more effective. The three aliphatic activators had considerably lower Y_max values. From Table 1, it is clear that these activators fall into two categories based on their Y_max values in which the aromatic anions are more effective at inducing CF transport out of the vesicles (also see Fig. 3 and Supplementary Information, SI, Figure S2†). This highlights a problem with how EC₅₀ values should be compared. When all activators give similar Y_max valves, the comparison of EC₅₀ is straightforward and meaningful. Therefore, from Table 1, it is easy to compare the EC₅₀'s of PB to PA and SL to SDS to SDP.

Fig. 3 Dose response curves for PB, PA, SL, SDP, and SDS with (A) CPPM1 (2.5 μM) and (B) CPPM2 (2.5 μM) against EYPC-LUVs⊃CF vesicles (50 μM EYPC), with curve fit to Hill equation.

Table 1 EC₅₀, Y_max, and E values for anion activators in EYPC-LUVs⊃CF at constant lipid concentration (50 μM) and CPPM concentration (2.5 μM)

Activator	CPPM 1			CPPM 2			pR^a
	EC 50 (μM)	Y max	E	EC 50 (μM)	Y max	E	EC 50 (μM)	Y max	E
Y max (maximal CF release relative to complete release by Triton X-100); EC50 (effective polymer concentration needed to reach Ymax/2); E, activator efficiency. Each data point was collected in three independent experiments.a pR activation data incorporated from refs. 2, 8a.
PB	20 ± 1.3	1.0 ± 0.04	8.2	24 ± 0.8	1.0 ± 0.02	7.9	44 ± 2.0	0.78 ± 0.02	5.1
PA	70 ± 2.4	0.85 ± 0.01	4.8	88 ± 3.6	0.96 ± 0.03	4.9	86 ± 3.0	0.80 ± 0.03	4.1
SL	15 ± 2.0	0.17 ± 0.02	1.5	76 ± 4.3	0.17 ± 0.006	0.9	34 ± 1.0	0.10 ± 0.01	0.7
SDP	25 ± 3.1	0.15 ± 0.008	1.2	16 ± 0.4	0.3 ± 0.004	2.6	19 ± 1.0	0.61 ± 0.03	5.1
SDS	10 ± 1.0	0.10 ± 0.003	1.0	103 ± 0.02	0.14 ± 0.02	0.7	16 ± 1.0	0.27 ± 0.01	2.4

---

However, when the Y_max values differ significantly meaningful comparisons of EC₅₀ are less obvious. A perfect activator would have a maximum Y_max and low EC₅₀.² Previous literature with a classical CPP, pR, at the P/L 0.02 (guanidine to lipid ratio 1.4) showed a general trend of decreasing EC₅₀ accompanied by decreasing Y_max for a group of activators; not very promising for potent activator selection.^2,8a Fortunately, Matile and co-workers proposed a solution to this activator comparison dilemma by creating a term called activator efficiency, E, based on the exponential relationship between Y_max and EC₅₀, with a scale between 0 and 10.² This term is also helpful since it allows experiments to be compared when the concentrations of the activators and/or transporters are different between experiments.

Table 1 also shows E values for each activator with CPPM1, CPPM2, and literature values of pR. Not surprisingly, PB is the best activator for all three of these guanidinium-rich transporters. It even activates the two small oligomers better than pR. The phosphate containing activator, SDP, also performed well for pR and CPPM2. Both of these molecules contain only guanidinium cations which are known to form strong and unique interactions with phosphate anions (see Fig. 1). In fact, comparing activation of CPPM1 and CPPM2 with SDP shows that the two extra guanidines present in CPPM2 yield a lower EC₅₀ and higher Y_max (see SI, Figure S2† for statistical analyses), or overall better activation (E = 2.6 vs. 1.2). This is consistent with the ability of phosphate ions to form specific interactions with guanidinium over amines.^1d,3 Further, aromatic counteranion activation of both CPPMs in EYPC vesicles is more selective than pR as shown by comparing the E values of PB vs. PA.

Comparing the two non-phosphate aliphatic anions shows that these two activators, SL and SDS, are more effective for CPPM1 than CPPM2 (Table 1). For example, SL gives identical Y_max values for both CPPMs but the EC₅₀ of CPPM1-SL is five times lower than CPPM2-SL (also see SI, Figure S2† for statistical analyses). This suggests an important role for the amines present in CPPM1. Lehn et al. reported that the less-delocalized, harder amine cation binds anions more strongly even though guanidinium had a prefer geometry to form bi-dentate interactions.¹¹ It thus seems likely that the similarly less delocalized carboxylate can interact with the amines of CPPM1.

In summary, this paper demonstrates the ability of common hydrophobic anions to serve as activators for two guanidinium-rich small molecules. Although the detailed mechanism is unknown and beyond the scope of this communication,¹² the results confirm the ability of these small molecules to serve as transporters through model membranes. It re-confirms the special role aromatic anions, like pyrenebutyrate, have in facilitating transport. Comparing activator efficiencies to the well-studied polyarginine shows that PB activates these small molecules better. Comparing CPPM1 to CPPM2 showed that the amine groups allowed better activation by the aliphatic carboxylate anion. This strategy can facilitate guanidine rich small molecule mediated cargo delivery and suggests small molecules may present a rich new opportunity to design novel CPPMs.

Acknowledgements

This work was supported by a NSF grant (CHE-0910963) to GNT.

Notes and references

1. (a) I. Nakase, T. Takeuchi, G. Tanaka and S. Futaki, Adv. Drug Delivery Rev., 2008, 60, 598–607; (b) P. A. Wender, W. C. Galliher, E. A. Goun, L. R. Jones and T. H. Pillow, Adv. Drug Delivery Rev., 2008, 60, 452–472; (c) A. Ziegler, Adv. Drug Delivery Rev., 2008, 60, 580–597; (d) A. Pantos, I. Tsogas and C. A. Paleos, Biochim. Biophys. Acta, Biomembr., 2008, 1778, 811–823; (e) S. M. Fuchs and R. T. Raines, Biochemistry, 2004, 43, 2438–2444; (f) N. Sakai, S. Futaki and S. Matile, Soft Matter, 2006, 2, 636–641; (g) N. Sakai and S. Matile, J. Am. Chem. Soc., 2003, 125, 14348–14356; (h) T. Hitz, R. Iten, J. Gardiner, K. Namoto, P. Walde and D. Seebach, Biochemistry, 2006, 45, 5817–5829; (i) E. A. Goun, T. H. Pillow, L. R. Jones, J. B. Rothbard and P. A. Wender, ChemBioChem, 2006, 7, 1497–1515; (j) K. Melikov and L. Chernomordik, Cell. Mol. Life Sci., 2005, 62, 2739–2749; (k) D. J. Mitchell, D. T. Kim, L. Steinman, C. G. Fathman and J. B. Rothbard, J. Pept. Res., 2000, 56, 318–325; (l) C. B. Cooley, B. M. Trantow, F. Nederberg, M. K. Kiesewetter, J. L. Hedrick, R. M. Waymouth and P. A. Wender, J. Am. Chem. Soc., 2009, 131, 16401–16403; (m) A. Hennig, G. J. Gabriel, G. N. Tew and S. Matile, J. Am. Chem. Soc., 2008, 130, 10338–10344; (n) E. M. Kolonko and L. L. Kiessling, J. Am. Chem. Soc., 2008, 130, 5626–5627; (o) A. Som, A. O. Tezgel, G. J. Gabriel and G. N. Tew, Angew. Chem., Int. Ed., 2011, 50, 6147–6150.
2. M. Nishihara, F. Perret, T. Takeuchi, S. Futaki, A. N. Lazar, A. W. Coleman, N. Sakai and S. Matile, Org. Biomol. Chem., 2005, 3, 1659–1669.
3. N. Sakai, T. Takeuchi, S. Futaki and S. Matile, ChemBioChem, 2005, 6, 114–122.
4. (a) E. Vives, P. Brodin and B. Lebleu, J. Biol. Chem., 1997, 272, 16010–16017; (b) P. A. Wender, D. J. Mitchell, K. Pattabiraman, E. T. Pelkey, L. Steinman and J. B. Rothbard, Proc. Natl. Acad. Sci. U. S. A., 2000, 97, 13003–13008.
5. M. Okuyama, H. Laman, S. R. Kingsbury, C. Visintin, E. Leo, K. L. Eward, K. Stoeber, C. Boshoff, G. H. Williams and D. L. Selwood, Nat. Methods, 2007, 4, 153–159.
6. (a) E. M. Kolonko, J. K. Pontrello, S. L. Mangold and L. L. Kiessling, J. Am. Chem. Soc., 2009, 131, 7327–7333; (b) A. O. Tezgel, J. C. Telfer and G. N. Tew, Biomacromolecules, 2011, 12, 3078–3083.
7. S. Choi, A. Isaacs, D. Clements, D. H. Liu, H. Kim, R. W. Scott, J. D. Winkler and W. F. DeGrado, Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 6968–6973.
8. (a) F. Perret, M. Nishihara, T. Takeuchi, S. Futaki, A. N. Lazar, A. W. Coleman, N. Sakai and S. Matile, J. Am. Chem. Soc., 2005, 127, 1114–1115; (b) T. Takeuchi, M. Kosuge, A. Tadokoro, Y. Sugiura, M. Nishi, M. Kawata, N. Sakai, S. Matile and S. Futaki, ACS Chem. Biol., 2006, 1, 299–303.
9. P. Guterstam, F. Madani, H. Hirose, T. Takeuchi, S. Futaki, S. El Andaloussi, A. Graslund and U. Langel, Biochim. Biophys. Acta, Biomembr., 2009, 1788, 2509–2517.
10. S. Matile, T. Miyatake and M. Nishihara, J. Am. Chem. Soc., 2006, 128, 12420–12421.
11. B. Dietrich, D. L. Fyles, T. M. Fyles and J. M. Lehn, Helv. Chim. Acta, 1979, 62, 2763–2787.
12. (a) P. F. Almeida and A. Pokorny, Biochemistry, 2009, 48, 8083–8093; (b) C. A. Paleos, A. Pantos and I. Tsogas, Biochim. Biophys. Acta, Biomembr., 2008, 1778, 811–823; (c) A. Ziegler, X. L. Blatter, A. Seelig and J. Seelig, Biochemistry, 2003, 42, 9185–9194; (d) R. Rathinakumar and W. C. Wimley, J. Am. Chem. Soc., 2008, 130, 9849–9858.

---

Footnote

† Electronic supplementary information (ESI) available: Liposome preparation, fluorescence assay, transporter activity with counteranions, statistical analyses. See DOI: 10.1039/c1ob06373a

---