Synthesis, characterization and biological properties of new hybrid carbosilane–viologen–phosphorus dendrimers

Abstract

A series of hybrid carbosilane–viologen–phosphorus dendrimers was prepared, as a new example of the synthetic “onion peel” approach. This is based on a convergent strategy by combination of double alkylation of 4,4-bipyridine units with two different halogenated reagents, one of them as a carbosilane dendron, and their subsequent ligation to a hexafunctionalized phosphorus core through amine–aldehyde condensation reactions. In these systems two kinds of cationic groups were included: those located at the branches due to viologen quaternized units and those related to the ammonium groups at the surface of carbosilane wedges. This feature constitutes a novel situation to be explored in the search for new physical–chemical and biological properties, respecting traditional dendritic architectures. The biological properties of two of these hybrid molecules have been studied, focusing the investigation on their interactions with plasma proteins like human serum albumin (HSA), cytotoxicity and hemotoxicity experiments. Although the observed biological behaviors were mainly related to the presence of outer positive charges, in some cases the inner positive charges acted as fine tuning factors.

---

Introduction

The properties of dendrimers, such as a well-defined size and structure, flexibility, monodispersity and multivalent molecular surface, have attracted attention for biomedical applications.^1–3 In addition, multi-charged dendrimers are particularly interesting, since they are generally soluble in water, opening the way to biological uses, like drug delivery or gene therapy. In this regard, several types of cationic dendrimers containing different skeletons have been explored for gene therapy.^4–13 Depending on the chemical structure of the dendrimers, the positive charges can be located on the terminal groups, in the branching points or event at the core. In the case of phosphorus-containing dendrimers, an example with phosphonium salt at the branching point and core has been reported elsewhere,¹⁴ while phosphorhydrazone dendrimers show neutral skeleton the charges being introduced at the periphery.¹⁵ These homodendrimers have shown important biological properties and play a key role in different biomedical applications, as well as in other fields such as catalysis, fluorescence or materials.¹⁶ Dendrimers based on viologen as building blocks for the design of polycationic dendrimers have been also reported showing antiviral activity mainly against human immunodeficiency virus (HIV-1).¹⁷ Recently, hybrid viologen–phosphorus dendrimers have been synthesized containing charged viologen units at the branches, with neutral groups at the periphery and their biological properties studied.^18,19 Another interesting dendritic typology is based on water-soluble cationic carbosilane dendrimers. Here, the strong amphiphilic behavior with positive charges at the surface and a very inert skeleton, differentiates them from most of the other kinds of dendrimers with a more hydrophilic skeleton. In the same way, cationic carbosilane homodendrimers have proven to be useful as non-viral vectors for gene therapy^20–23 or as antibacterial agents.^24,25

Very recently, a novel “onion peel” strategy for the divergent construction of glycodendrimers using different building blocks at each layer of the dendritic growth has been reported by Roy et al.^26–30 A combination of successive highly efficient, and versatile chemical reactions, namely thiol–ene or thiol–yne, esterification, and azide–alkyne click chemistry, have been used to form heterogeneous layers. The dendrimers prepared using this strategy, are fundamentally different from conventional dendrimers that are usually built from repetitive building nanosynthons with limited surface groups. Similarly, phosphorus dendrimers constituted with heterogeneous layers have been also reported.^31,32

In this paper, we want to show another example of this onion peel strategy, grounded on a convergent strategy by combination of double alkylation of 4,4-bipyridine units with two different halogenated reagents, one of them based on a carbosilane dendron, and the subsequent ligation to a hexafunctionalized phosphorus core through an amine–aldehyde condensation reaction. This strategy allows combining the structural diversity of the three building blocks based on the lipophilicity of the carbosilane scaffold, the polarity imposed by the viologen branches scaffolds and the rigidity of the phosphorus based cored, along with their individually potential biological uses described above. Namely, hybrid systems containing carbosilane, viologen and phosphorus dendritic scaffolds into one single molecule have been designed. Starting from the hexafunctionalized core [(P₃N₃)NMeNH₂)₆],³³ viologen units were incorporated into the branching points, decorating their surface by inclusion of cationic carbosilane dendrons. The obtained dendrimers contains two kinds of cations. One is formed by viologen quaternized units located at inner branches while the other is ammonium groups at the surface of the carbosilane wedges. This situation makes them unique and interesting, to be explored from the biological point of view. The biological properties of two of these hybrid molecules have been studied, focusing the investigation on the interactions with plasma proteins like human serum albumin (HSA), and toxicity experiments.

Results and discussion

Synthesis of dendrimers

Two kinds of topologies of dendrimers were prepared: dendrons, which are cone-shaped molecules with two different functions, one at the periphery and another at the focal point, and spherical dendrimers, which usually show a symmetric distribution around a core.

A variety of cationic carbosilane dendrons with ammonium groups at the periphery containing viologen units at the focal point were synthesized. It starts from carbosilane dendrons (G1 (I) and G2 (II)), with a bromine atom at the focal point and NHBoc groups on the periphery.³⁴

Dendrons with viologen at the focal point were prepared by two subsequent quaternization steps on bipyridine ligand. The first one, using carbosilane dendrons I or II affords compounds 1 and 2 in an excess of bipyridine to facilitate the monosubstitution (Scheme 1). The reaction was followed by ¹H NMR for the disappearance of CH₂Br signal at 3.40 ppm. These products were purified by size exclusion chromatography obtaining the dendritic wedges as yellow oils, soluble in organic solvents. The NMR data confirm the presence of the new CH₂N– bond at 4.96 ppm in the ¹H NMR spectrum and at 61.5 ppm in the ¹³C NMR spectrum.

Scheme 1 Synthesis of amine and ammonium terminated dendrons 1–8. Conditions: (i) 4,4′-bipyridine, MeCN, 80 °C, 48 h; (ii) 4-(bromomethyl)benzaldehyde, MeCN, 80 °C, 24 h; (iii) TFA/DCM; ^tBu₄NCl; (iv) CHCl₃/H₂O, NH₄PF₆.

For the second, quaternization, 4-(bromomethyl)benzaldehyde was used to obtain dendrons with viologen at the focal point, 3 and 4 (Scheme 1). These dendrones have an aldehyde group necessary to prepare spherical phosphorus cored dendrimers through condensation reactions. This quaternization was performed using the same conditions to that of compounds 1 and 2. Generally, compounds with viologen and bromide as counterion are soluble in polar solvents, however, in this case the compounds were slightly soluble in all solvents, precluding simple purifications. Their purifications were accomplished by extraction with dichloromethane–water, in very low yields. For this reason, the compounds were isolated only for characterization purposes and used in situ for posterior reactions. The NMR data confirm the incorporation of bromobenzaldehyde through the appearance of signals at 10.08 ppm attributed to CHO group, and the new CH₂N group at 6.09 ppm in the ¹H NMR spectrum and at 192.21 and 63.49 ppm respectively in the ¹³C NMR spectrum.

On one side, the compounds 3 and 4 were deprotected with TFA and their counterions changed by chloride anions adding ^tBu₄NCl, obtaining dendrons 5 and 6 with viologen at the focal point and ammonium groups at the periphery in high yield after the corresponding purification process. This reaction was completed when the characteristic resonance of the Boc group at 1.47 ppm in ¹H NMR disappeared. The presence of ammonium groups at the periphery in these dendrons was clearly identified by NMR spectroscopy, showing a triplet at 2.84 ppm in the ¹H NMR and in the ¹³C NMR spectra at 42.32 ppm for all CH₂NH₃Cl groups.

However, to obtain spherical phosphorus cored dendrimers, dendrons protected with NHBoc and soluble in organic solvent are needed. Then, for dendrons 3 and 4, bromide counterion was changed by adding NH₄PF₆ in order to obtain better solubilities in organic solvent. These products were purified by size exclusion chromatography thus allowing to get compounds 7 and 8 as yellow solids.

The NMR spectra are similar to those of compounds 3 and 4, with only small differences in the bipyridine aromatic framework due to the counterion exchange, evidenced by ³¹P NMR (Fig. 1).

Fig. 1 ¹H (A), ¹³C (B) and ³¹P NMR (C) spectra of derivative 7 in CD₃CN. *denotes residual acetonitrile as solvent.

For the synthesis of hybrid spherical dendrimers, G1 and G2 dendrons, 7 and 8, with an aldehyde group at the focal point were linked to the N₃P₃[N(Me)NH₂]₆ core.³³ New hybrid systems with a phosphorus based core, six viologen units at the branches and carbosilane wedges of different generations on the surface were prepared (Scheme 2).

Scheme 2 Synthesis of hybrid spherical dendrimers 9–14. Conditions: (i) MeCN, 24 h; (ii) TFA/DCM, ^tBu₄NCl; (iii) CHCl₃/H₂O, NH₄PF₆.

The condensation between the phosphorus core and the compounds 7 or 8 was carried out in acetonitrile at room temperature affording the dendrimers 9 and 10 (Scheme 2) as yellow solids soluble in organic solvents. The progress of the condensation was monitored by ¹H and ³¹P NMR spectroscopy and this reaction was completed when the characteristic resonance of the CHO group disappeared in the ¹H NMR and by shifting the phosphorus core resonance from 22.21 to 17.14 ppm in the ³¹P NMR (Fig. 2). The condensation reactions were also confirmed by ¹H NMR through the appearance of the methyl group of the core at 3.30 ppm and the CH imine group at 7.73 ppm, partially overlapped with aromatic protons.

Fig. 2 ³¹P NMR spectrum of dendrimer 10 in CD₃CN.

In order to provide water-solubility to all these dendritic structures, deprotection of NHBoc groups and their counterion exchanges by chlorides, was performed using the same conditions used for the synthesis of 5 and 6. These procedures allow the preparation of dendrimers 11 and 12 (Scheme 2) with two kinds of cationic groups: those located at the branches due to viologen quaternized units and those related to the ammonium groups at the surface of carbosilane wedges. The cationic dendrimers, thus obtained, were water-soluble, which is an important requirement for biomedical applications. However for improving NMR spectral quality, NMR data was recorded in deuterated methanol (see Fig. 3C and ESI†) or dendrimers with PF₆⁻ as counterion were formed by exchanging counteranion once again. In the later case, dendrimers 13 and 14 were prepared (Scheme 2) which were soluble in organic solvents such as acetonitrile. The presence of ammonium groups in these dendrons was clearly identified by NMR spectroscopy, showing a triplet at 2.93 ppm for all CH₂NH₃⁺PF₆⁻ in the ¹H NMR spectra along with a signal at 42.99 ppm in the ¹³C NMR (Fig. 3). Chemical structures of final dendrimers 11 and 12 are depicted in Fig. 4.

Fig. 3 ¹H (A) and ³¹P (B) NMR spectra of dendrimer 11 in CD₃OD and ¹H (C) and ¹³C (D) NMR spectra of dendrimer 13 in CD₃CN. See ref. 34 for carbosilane part and ref. 18 for viologen and cyclophosphazene frameworks for an accurate assignment.

Fig. 4 Chemical structures of dendrimers 11 and 12.

Biophysical properties and toxicity of dendrimers

The physical–chemical characterization and biological properties of hybrid dendrimers 11 and 12 have been examined, while dendrimers 13 and 14 containing PF₆⁻ anions were discarded because they were not soluble in water.

The first stage of the study was to investigate the zeta potential of the dendrimers and their effect on the zeta potential of HSA (human serum albumin). It has been observed that the zeta potential value for dendrimer 12 dissolved in water was higher than that for 11, being practically of the same size when they are dissolved in phosphate buffer within the standard deviation. In general, the values for zeta potential are significantly lower in the phosphate buffer than in water solutions, probably due to the modification of the number of terminal ammonium units in buffer, although interactions with the phosphate counteranion by ionic pair formation cannot be ruled out.

The interaction of dendrimers with HSA has also been analyzed. The zeta potential values plotted versus the dendrimer/HSA molar ratio allowed to determine the possible number of dendrimer molecules attached to one protein molecule. For 11, the initially negative zeta potential of albumin (−12.6 mV) changed to positive on increasing concentration of dendrimer (Fig. 5A). Based on these results, it can be calculated that a single HSA molecule binds from 11 to 12 molecules of dendrimer 11. Similarly, the increasing concentration of 12 added to the HSA solution caused the change of the zeta potential from negative to positive up to 9.1 mV (Fig. 5B). It means that one albumin molecule can bind from 2 to 3 molecules of the dendrimer 12. These differences have to be ascribed to the increase of size and number of positive charges on going from 11 to 12 (Table 1).

Fig. 5 Changes of HSA zeta potential after addition of dendrimers: (A) 11; (B) 12 (x = 12, *p < 0.05; **p < 0.01; ***p < 0.001, HSA concentration 0.2 μM), x is the number of repetitions.

Table 1 Zeta potential values of hybrid dendrimers dissolved in water or in phosphate buffer

Dendrimer	Zeta potential [mV]
	In water	In buffer
11	17.0 ± 2.4	4.1 ± 0.7
12	22.3 ± 0.7	2.8 ± 1.5

---

It is known that cationic dendrimers (e.g. polyamidoamine, viologen–phosphorus) can interact with human serum albumin. HSA has five nonspecific anionic regions and binds positively charged dendrimer due to electrostatic interactions.^35–38 Giri et al.³⁹ showed that the interactions between polyamidoamine (PAMAM) dendrimers and human serum albumin depend strongly on size and terminal group chemistry. They suggested several mechanisms of interactions between PAMAM dendrimers and HSA proteins including (i) electrostatic interactions between charged dendrimer terminal groups and protein residues, (ii) hydrogen bonding between dendrimer internal groups and protein amino acid residues, (iii) hydrophobic interactions between the nonpolar dendrimer and HSA groups, and (iv) specific interactions between dendrimer carboxylic groups and protein aliphatic acid binding sites.

The effect of the dendrimers on the secondary structure of HSA molecule was evaluated. This was achieved by measuring the circular dichroism (CD) spectra of HSA after addition of dendrimers 11 or 12. When dendrimer 11 was added to the protein solution at low concentrations of 2.5 or 5 μM (HSA/11 molar ratio 1 : 1 or 1 : 2), the changes in the spectrum shape could be regarded as negligible because they were lower than 10%. Higher concentrations of 11 (HSA/11 molar ratio 1 : 5 and 1 : 10) significantly altered the shape of the albumin CD spectrum, indicating changes in the albumin secondary structure (Fig. 6A).

Fig. 6 Circular dichroism spectra of HSA (c = 2.5 μM) in the presence of dendrimers: (A) 11; (B) 12; (x = 6). x is the number of repetitions, i.e. each spectrum is the average of six replicates.

The addition of 12 to HSA resulted in significant changes in the HSA CD spectrum intensity already at low concentrations (Fig. 6B). The biggest modification has been observed for addition of the highest concentration (25 μM) of dendrimer 12 (HSA/12 = 1 : 10). The results indicate that this dendrimer affects the HSA secondary structure at least beyond 2.0 μM concentration. Base on the figures we can conclude that the highest concentration of both used dendrimers strongly changed the secondary structure of HSA, but we cannot clearly state which dendrimer alters the structure of proteins more significantly.

The HSA fluorescence intensity measurements were conducted to corroborate whether dendrimers influence the protein conformation. The fluorescent spectra of albumin after addition of the increasing concentrations of dendrimers 11 and 12 were plotted (Fig. 7). The decrease in fluorescence intensity with increasing dendrimer concentration has been observed indicating HSA fluorescence quenching by both dendrimers. These changes suggest that dendrimers modify the secondary structure of protein, quenching being stronger for dendrimer 12.

Fig. 7 Fluorescence spectra of the HSA (c = 2.5 μM) in the presence of dendrimers: (A) 11; (B) 12. (x = 6). x is the number of repetitions.

The Stern–Volmer plots for the quenching of HSA with both used dendrimers (quenchers) are shown in Fig. 8. The Stern–Volmer plots for both dendrimers are linear indicating that the quenching mechanism is dynamic and the Stern–Volmer quenching constants, K_SV, are K_SV = 0.65 × 10⁵ ± 0.6 × 10⁴ M⁻¹ for 11 and K_SV = 3.73 × 10⁵ ± 0.8 × 10⁵ M⁻¹ for 12.

Fig. 8 Stern–Volmer graphs. Quenching of HSA fluorescence by dendrimers: (A) 11; (B) 12. (x = 6).

Finally, albumin fluorescence quenching was not accompanied by any red or blue shift of the emission maximum for both dendrimers. Quenching of fluorescence means that the chromophore in HSA (tryptophan residue) is accessible to the quencher (dendrimer). The shift of λ_max during fluorescence quenching would suggest that the tryptophan residue was exposed to either more polar (red shift) or less polar (blue shift) environment. However, this was not the case in our studies. This may mean that although the tryptophan (Trp) residue became more accessible to the quencher it was not moved to an environment of different polarity. In another words, there are some structural rearrangements going on but not linked with the change in the polarity in the immediate vicinity of the chromophore.

The interaction between a series of viologen–phosphorus based dendrimers and HSA has been studied elsewhere³⁵ by fluorescence and circular dichroism measurements, as an example of the interaction with proteins. Compounds having very reactive aldehyde terminal groups (see Fig. 9C) quenched HSA fluorescence and changed the secondary structure of albumin. However, compounds with phosphonate terminal groups (see Fig. 9D and E) or polyethyleneglycol (PEG) terminal groups strongly quenched HSA fluorescence, but did not affect protein conformation. In the case of hybrid carbosilane–viologen–phosphorus dendrimers, the presence of terminal NH₃⁺ groups enables interaction with the protein by hydrogen bonding explaining the changes in secondary structures observed. For both used hybrid dendrimers, 11 and 12, the Stern–Volmer constant is in the same range to those observed for viologen–phosphorus dendrimers: the highest value detected for a system having viologen building block units with 36 inner positive charges (Fig. 9E) was 13.1 × 10⁵ M⁻¹ and the lowest (0.75 × 10⁵) was found for compounds with 6 inner positive charges. However, dendrimer 12, possessing 12 inner and 24 outer charges, quenched the HSA fluorescence stronger than dendrimer 11 with 24 total positive charges. This may suggest that the strongest interaction between protein and dendrimer, is a consequence of the presence of outer positive charges, as expected.

Fig. 9 Representative structures of third generation (A) phosphorus and (B) carbosilane homodendrimers and hybrids viologen–phophorus dendrimers of first (C and D) and second (E) generations.

Interestingly, it has been mentioned elsewhere³⁵ that the number of molecules bound to HSA for a series of viologen phosphorus-based dendrimers measured by fluorescence quenching of tryptophan residues in HSA showed values ranged from 0.9–1.4. These values are very low compared to those observed for 11 and 12. In the present study, the number of hybrid molecules bound to HSA, determined by measuring the zeta potential, was higher. Probably, the use of different techniques is a factor to take into account for the high discrepancy of the values found.

Another important difference concerns the changes in the protein secondary structure composition. For small viologen–phosphorus dendrimers having low number of inner positive charges but reactive CHO groups at the periphery (see Fig. 9C), changes from α-helix to β-structures were observed after addition of dendrimers. For other viologen dendrimers having bigger sizes and charges but non-reactive groups at the periphery (see Fig. 9D and E), the increase of their concentration caused only small changes in the ellipticity of protein CD spectra, not altering the spectra shape. In the case of 11 and 12, the presence of terminal ammonium groups could be responsible for inducing critical changes on protein structure.

Biocompatibility studies of dendrimers 11 and 12 have also been performed. In order to assess the effect of hybrid dendrimers on eukaryotic cells, Chinese hamster fibroblasts (B14 cell line) were chosen and MTT assay was performed for this study. The results show that 11 at the highest concentration (10 μM) decreases B14 cells viability up to 70%, with CC₅₀ > 10 μM, whereas, incubation with 12 at this concentration decreases the viability up to 8% with CC₅₀ ca. 1 μM (Fig. 10). A value of 80% considered as the cut-off for toxicity showed B14 cells viability at 5 μM and 0.1 μM for hybrid systems 11 and 12, respectively.

Fig. 10 Viability of the B14 cells treated with dendrimers 11 and 12 measured by MTT (x = 10, *p < 0.05; **p < 0.01; ***p < 0.001).

Statistically significant differences were obtained for the concentrations 1, 5 and 10 μM of dendrimer 11 and for all used concentrations of 12. Statistical significance was determined relative to the control cells in the absence of dendrimers. Cell not treated with dendrimers were taken as 100% viability.

The morphology of Chinese hamster fibroblasts exposed to hybrid dendrimers of different concentrations was examined (Fig. 11). Dendrimer 11 at a concentration of 0.1 μM does not cause changes in cell morphology. However, higher concentrations of this dendrimer and all concentrations of dendrimer 12 contribute to significant changes in cell morphology, and cell fragments are visible at the photographs, indicating again the differences between the two dendrimers.

Fig. 11 Photographs of control B14 cells and cells treated with 11 or 12 at concentrations of 0.1, 1 and 10 μM (magnification, 400×).

Although dendrimers of the same or different nature may behave differently from one cell line to another, the comparison with the values obtained from the literature may give an idea of the order of magnitude of the cell viability shown. A comparative analysis with homophosphorus, homocarbosilane and hybrids viologen–phosphorus dendrimers has been carried out (see Fig. 9 and Table 2 for key features of dendritic structures and toxicity values). From data of Table 2, the hybrid system 11 resulted to be more biocompatible than homophosphorus⁴⁰ (as an example of the structure see Fig. 9A) and homocarbosilane dendrimers^41,42 (as an example of the structure see Fig. 9B) probably due to the lower number of outer positive charges. Comparing to hybrid viologen–phosphorus based dendrimers capped with neutral phosphonate groups at the periphery, (see Fig. 9D and E)^18,19 dendrimer 11 exerted similar biocompatibility, indicating certain degree of contribution of the inner positive charges on toxicity. For the hybrid system 12, the total number of positive charges seems to dominate the toxicity when compared with carbosilane dendrimers^41,42 and hybrid viologen–phosphorus dendrimers.^18,19 Taking into account that we are comparing different cell lines, the analysis suggests that the number of outer positive charges is an important factor in the toxicity process, as expected, but noticeable contribution of the inner positive charges is observed. A straightforward toxicity correlation is not deduced respecting the dendritic nature of the different building blocks.

Table 2 Key features of dendritic structures and toxicity values of some homophosphorus (PD), homocarbosilane (CD), and hybrid viologen–phosphorus (VPD) dendrimers along with hybrid dendrimers 11 and 12

Dendrimer	Generation	Surface groups	Outer charge	Inner charge	Toxicity (μM)	Cell line	Ref.
11	1	–NH3^+	12	12	CC50 > 10, CC80 = 5	B14	This work
12	2	–NH3^+	24	12	CC50 ≈ 1, CC80 = 0.1	B14	This work
PD	2	–NHEt2^+	24	0	CC50 = 1.84	N2a	40
PD (Fig. 9A)	3	–NHEt2^+	24	0	CC50 = 1.74	N2a	40
CD	3	–NH3^+	24	0	CC50 = 2.0	HeLa	41
CD (Fig. 9B)	3	–NMe3^+	24	0	CC80 = 1.0	PBMC	42
VPD (Fig. 9C)	0	–CHO	0	12	CC80 = 1.0	B14	18
VPD (Fig. 9D)	0	–P(O) (OEt)2	0	12	CC80 = 2.5	B14	18
VPD (Fig. 9D)	1	–P(O) (OEt)2	0	36	CC80 = 10	B14	18

---

In order to assess whether the hybrid dendrimers damage red blood cells their hemolytic activity was examined after 3 hour incubation of cells with different concentrations of dendrimers. Both dendrimers 11 and 12 caused statistically significant increase in the degree of hemolysis in the treated erythrocytes. However, at a concentration range of 0.01–0.1 μM, these changes were small – from 2.5% (control value) to about 7.5% (0.1 μM). For higher concentrations, significantly more erythrocytes underwent hemolysis. Dendrimer 11 at the concentration of 10 μM caused 78% hemolysis while for dendrimer 12 at the same concentration hemolysis was equal to 84% (Fig. 12). Basing on these results, we can conclude that studied dendrimers are hemotoxic at the concentration above 0.1 μM.

Fig. 12 Hemolysis of erythrocytes induced by dendrimers 11 and 12 (x = 12, ***p < 0.001).

Microscopic analysis of red blood cells shows that hybrid dendrimers cause the loss of erythrocytes normal morphology. After 3 h of incubation with the dendrimers at the concentrations of 0.1 or 1 μM, the appearance of the cytoplasmic projections on the surface and the conversion of normal erythrocytes to echinocytes were noticed. At the highest used concentration (10 μM) many residues resulting from the escape of hemoglobin were observed. From Fig. 13, dendrimer 11 showed hemotoxicity at 1 μM while for 12 the same is true at all concentrations studied.

Fig. 13 Photographs of control erythrocytes and erythrocytes treated with hybrid dendrimers 11 and 12 (magnification, 400×).

The hemotoxicity observed for 11 and 12 (at 1 μM lysed ca. 30% for both cases) is lower than that observed for homodendrimers based on carbosilane structure (at 1 μM lysed ca. 87%, see Fig. 9B)⁴² and hybrid systems based on viologen–phosphorus dendrimer of second generation (at 1 μM lysed ca. 40%, see Fig. 9E).¹⁸ However, the hemotoxicity is significantly reduced for first generation viologen–phosphorus dendrimers (at 1 μM lysed ca. 5–10%, see Fig. 9C and D).¹⁸ These facts suggest that the toxicity towards red blood cells is not following the same parameters that in non-red blood eukaryotic cell. No simple correlation with the number of positive charges can be inferred in this case, where other factors like size, topology, nature of the ammonium groups, local charge density, typology, or a set of them, cannot be ruled out.

Conclusions

A family of hybrid carbosilane–viologen–phosphorus dendrimers was prepared, constituting an example of the synthetic “onion peel” approach. This is based on a convergent strategy by combination of two reactions: (i) a double alkylation of 4,4-bipyridine units with two different halogenated reagents, one of them as carbosilane dendron, and (ii) the subsequent ligation to a hexafunctionalized phosphorus core through amine–aldehyde condensation. In these systems two kinds of cationic groups were included: (i) those located at the branches due to viologen quaternized units and (ii) those related to the ammonium groups at the surface of carbosilane wedges. The different interaction observed for water-soluble hybrid dendrimers 11 or 12 with serum protein like HSA seems to be ascribed mainly to the size, instead of the increase of theoretical number of positive charges according to zeta-potential data. Dendrimer 12 modified the secondary structure of HSA more deeply than 11 as the HSA fluorescence intensity measurements and the Stern–Volmer quenching constants, K_SV, revealed. In this case, this behavior is probably due to the presence of higher number of positive charges at the periphery, being the inner positive charges associated to viologen units less important for this effect. Cell viability of dendrimers 11 and 12 measured on the eukaryotic cells, Chinese hamster fibroblasts (B14 cell line) and compared with analogous viologen–phosphorus dendrimers tentatively suggested that toxicity was related to the outer positive charges, with significant effect caused by the internal ones. However, respecting hemotoxicity, no correlation due to the number of outer positive charges and the different charge location was observed. Additional studies are underway for obtaining a better understanding on their antibacterial activity or as non-viral vectors for nucleic acids as well as in the design of more biocompatible systems by capping carbosilane dendrons with neutral groups.

Experimental section

Materials

Dimethyl sulfoxide (DMSO), phosphate buffered saline (PBS) tablets, foetal bovine serum, trypsin were purchased from Sigma-Aldrich (USA). All other chemicals used were of analytical grade. All solutions were prepared using water purified by the Milli-Q system.

Carbosilane–viologen–phosphorus dendrimers

Synthesis of dendron (viologen)G₁CBS(NHBoc)₂ (1). A mixture of carbosilane dendron (I)³⁴ (0.350 g, 0.503 mmol) and 4,4′-bipyridine (0.235 g, 1.51 mmol) in 4 mL of acetonitrile was heated at 80 °C for 48 h. The solvent was removed in vacuum. The residue can be purified by sublimation (T = 100 °C, 48 h) to remove excess of bipyridine used to promote monosubstitution. To eliminate small amounts of the disubstitution product observed, size exclusion chromatography was used. This was performed with a polystyrene stationary phase (Bio-Beads SX-1 by Bio-Rad) and THF as solvent, which gave the compound 1 as brown oil (390 mg, 90%).

¹H NMR (300 MHz, CDCl₃): δ −0.09 (s, 3H, SiMe), 0.06 (s, 12H, SiMe₂), 0.45 (t, ³J_HH = 9.0 Hz, 4H, SiCH₂CH₂CH₂NH), 0.50–0.56 (m, 10H, SiCH₂), 1.21–1.32 (m, 4H, CH₂), 1.39–1.58 (m, 24H, CH₂ and –C(CH₃)₃), 2.09 (t, ³J_HH = 9.0 Hz, 2H, N⁺CH₂CH₂), 3.04–3.10 (m, 4H, CH₂NH), 4.69 (br s, 2H, NH), 4.96 (t, ³J_HH = 6.0 Hz, 2H, N⁺CH₂), 7.73 (d, ³J_HH = 6.2 Hz, 2H, H²), 8.43 (d, ³J_HH = 6.8 Hz, 2H, H⁵), 8.86 (d, ³J_HH = 6.1 Hz, 2H, H¹), 9.58 (d, ³J_HH = 6.9 Hz, 2H, H⁶). ¹³C {¹H} NMR (75 MHz, CDCl₃): δ −5.17 (SiMe), −3.35 (SiMe₂), 12.35 (t, 4H, SiCH₂CH₂CH₂NH), 13.77–24.61 (SiCH₂), 28.42 (C(CH₃)₃), 35.72 (N⁺CH₂CH₂), 43.61 (CH₂NH), 61.54 (N⁺CH₂), 78.90 (C(CH₃)₃), 121.48 (C²), 125.94 (C⁵), 140.83 (C³), 145.73 (C⁶), 151.42 (C¹), 153.57 (C⁴), 156.00 (CO, Boc). Anal. calcd for C₄₁H₇₅BrN₄O₄Si₃ (852.22 g mol⁻¹): C, 57.78; H, 8.87; N, 6.57. Found: C, 57.54; H, 8.84; N, 6.56. MS: [M − Br]⁺ = 771.5089 uma (calcd = 771.5096 uma).

Synthesis of dendron (viologen)G₂CBS(NHBoc)₄ (2). The reaction of carbosilane dendron (II)³⁴ (0.600 g, 0.434 mmol) with 4,4′-bipyridine (0.203 g, 1.30 mmol) was carried out in the similar way than the synthesis of 1. The compound 2 was obtained as brown oil (460 mg, 70%) after using the same purification process shown in 1.

¹H NMR (300 MHz, CDCl₃): δ −0.017 (s, 33H, SiMe and SiMe₂), 0.48 (t, ³J_HH = 8.4 Hz, 8H, SiCH₂CH₂CH₂NH), 0.53–0.59 (m, 26H, SiCH₂), 1.27–1.35 (m, 12H, CH₂), 1.43–1.50 (m, 46H, CH₂ and C(CH₃)₃), 2.12 (t, ³J_HH = 7.2 Hz, 2H, N⁺CH₂CH₂), 3.02–3.16 (m, 8H, CH₂NH), 4.64 (br s, 4H, NH), 4.99 (t, ³J_HH = 6.8 Hz, 2H, N⁺CH₂), 7.71 (d, ³J_HH = 5.6 Hz, 2H, H²), 8.38 (d, ³J_HH = 6.4 Hz, 2H, H⁵), 8.90 (d, ³J_HH = 5.2 Hz, 2H, H¹), 9.55 (d, ³J_HH = 6.8 Hz, 2H, H⁶). ¹³C {¹H} NMR (300 MHz, CDCl₃): δ −4.97 (SiMe), −3.22 (SiMe₂), 12.36 (SiCH₂CH₂CH₂NH), 13.95–24.75 (SiCH₂), 28.38 (C(CH₃)₃), 36.05 (N⁺CH₂CH₂), 43.94 (CH₂NH), 61.85 (N⁺CH₂), 79.03 (C(CH₃)₃), 121.44 (C²), 125.84 (C⁵), 140.83 (C³), 145.66 (C⁶), 151.16 (C¹), 154.00 (C⁴), 156.00 (CO, Boc). Anal. calcd for: C₇₅H₁₄₉N₆O₈SiBr (1539.52 g mol⁻¹): C, 58.51; H, 9.76; N, 5.46. Found: C, 58.15; H, 9.40; N, 5.00. MS: [M − Br]⁺ = 1457.98 D (calcd = 1457.98 D).

Synthesis of dendron [CHO(viologen)G₁CBS(NHBoc)₂]Br₂ (3). A solution of compound 1 (0.390 g, 0.458 mmol) and 4-(bromomethyl)benzaldehyde (0.103 g, 0.514 mmol) in 4 mL of acetonitrile was heated at 80 °C for 24 h. The solvent was removed in vacuum and the crude was dissolved in a mixture CH₂Cl₂/H₂O (1 : 1). The product was extracted many times with water, the solvent was removed in vacuum to give a pale yellow solid with low yields, and this compound was isolated just for its characterization.

¹H NMR (300 MHz, CD₃CN): δ −0.017 (s, 15H, SiMe and SiMe₂), 0.48 (t, 4H, SiCH₂CH₂CH₂NH), 0.58–0.64 (m, 10H, SiCH₂) 1.31–1.48 (m, 28H, CH₂ and C(CH₃)₃), 2.00 (t, 2H, N⁺CH₂CH₂), 2.96–3.03 (m, 4H, CH₂NH) 4.73 (t, ³J_HH = 7.5 Hz, 2H, N⁺CH₂), 5.40 (br s, 2H, NH), 6.19 (s, 2H, CH₂), 7.85 (d, ³J_HH = 8.2 Hz, 2H, H²), 8.02 (d, ³J_HH = 8.2 Hz, 2H, H³), 8.66–8.68 (m, 4H, H⁶ and H⁹), 9.17 (d, ³J_HH = 5.0 Hz, 2H, H⁵), 9.44 (d, ³J_HH = 6.8 Hz, 2H, H¹⁰), 10.07 (s, 1H, CHO). ¹³C {¹H} NMR (300 MHz, CD₃CN): δ −0.47 (SiMe), −0.20 (SiMe₂), 11.93 (SiCH₂CH₂CH₂NH), 13.20–24.34 (SiCH₂), 27.71 (C(CH₃)₃), 34.94 (N⁺CH₂CH₂), 43.42 (CH₂NH), 61.71 (N⁺CH₂), 63.49 (CH₂), 77.94 (C(CH₃)₃), 127.37 (C⁹), 127.63 (C⁶), 130.08 (C²), 130.27 (C³), 137.31 (C¹), 139.07 (C⁴), 145.76 (C¹⁰), 149.19 (C⁵), 149.49 (C⁸), 150.22 (C⁷), 155.99 (CO, Boc), 192.21 (CHO).

Synthesis of dendron [CHO(viologen)G₂CBS(NHBoc)₄]Br₂ (4). The reaction of compound 2 (0.400 g, 0.260 mmol) with 4-(bromomethyl)benzaldehyde (0.068 g, 0.338 mmol) was carried out in the similar way than the synthesis of 3. The compound 4 was obtained as yellow solid.

¹H NMR (300 MHz, CD₃CN): δ −0.046 (s, 9H, SiMe), −0.013 (s, 24H, SiMe₂), 0.48 (t, 8H, SiCH₂CH₂CH₂NH), 0.58–0.63 (m, 28H, SiCH₂) 1.39–1.42 (m, 56H, CH₂ and C(CH₃)₃), 2.00 (t, 2H, N⁺CH₂CH₂), 2.96–3.01 (m, 8H, CH₂NH), 4.67 (t, 2H, N⁺CH₂), 5.34 (br s, 4H, NH), 6.09 (s, 2H, CH₂), 7.78 (d, ³J_HH = 8.1 Hz, 2H, H²), 8.04 (d, ³J_HH = 8.4 Hz, 2H, H³), 8.53–8.58 (m, 4H, H⁶ and H⁹), 9.04 (d, ³J_HH = 6.6 Hz, 2H, H⁵), 9.28 (d, ³J_HH = 6.6 Hz, 2H, H¹⁰), 10.08 (s, 1H, CHO). ¹³C {¹H} NMR (300 MHz, CD₃CN): δ −0.47 (SiMe), −0.20 (SiMe₂), 11.93 (SiCH₂CH₂CH₂NH), 13.20–24.34 (SiCH₂), 27.71 (C(CH₃)₃), 34.94 (N⁺CH₂CH₂), 43.42 (CH₂NH), 61.71 (N⁺CH₂), 63.49 (CH₂), 77.94 (C(CH₃)₃), 127.37 (C⁹), 127.63 (C⁵), 130.08 (C²), 130.27 (C³), 137.31 (C¹), 139.07 (C⁴), 145.76 (C¹⁰), 149.19 (C⁵), 149.49 (C⁸), 150.22 (C⁷), 155.99 (CO, Boc), 192.21 (CHO).

Synthesis of dendron [CHO(viologen)G₁CBS(NH₃)₂]Cl₄ (5). Dendron 3 (0.100 g, 0.085 mmol) was deprotected with 3 mL of a 20% solution of TFA in anhydrous dichloromethane, the mixture was stirred for 1 h. The solvent was removed in vacuum. After, the residue was dissolved in acetonitrile (20 mL) and 1.5 mL of a saturated aqueous solution of ^tBu₄NCl was added carefully, provoking the precipitation of the desired product. The resulting solid was filtered and washed several times with acetonitrile, for obtaining the compound 5 as yellow solid (0.060 mg, 77%). This purification process is perfectly enough to removed byproducts.

¹H NMR (300 MHz, D₂O): δ −0.18 (s, 15H, SiMe and SiMe₂), 0.42 (m, 14H, CH₂Si), 1.23 (m, 6H, CH₂CH₂Si), 1.52 (m, 4H, CH₂CH₂NH₃Cl), 1.95 (m, 2H, N⁺CH₂CH₂), 2.84 (m, 4H, CH₂NH₃Cl), 4.64 (m, 2H, N⁺CH₂ overlapped with water), 5.96 (s, 2H, CH₂), 7.59 (d, ³J_HH = 8.4 Hz, 2H, H²), 7.95 (d, ³J_HH = 8.4 Hz, 2H, H³), 8.45–8.51 (m, 4H, H⁶ and H⁹), 9.02 (d, 2H, H⁵), 9.13 (d, 2H, H⁵), 9.88 (s, 1H, CHO). ¹³C {¹H} NMR (300 MHz, D₂O): δ −5.89 (MeSi), −4.39 (SiMe₂), 11.19–21.55 (SiCH₂), 34.17 (N⁺CH₂CH₂), 42.32 (CH₂NH₃Cl), 61.93 (NCH₂), 64.13 (CH₂), 126.90 (C⁹), 127.20 (C⁶), 129.61 (C²), 130.91 (C³), 136.57 (C¹), 138.82 (C⁴), 145.46 (C¹⁰), 145.83 (C⁵), 149.69 (C⁸), 150.67 (C⁷), 195.51 (CHO).

Synthesis of dendron [CHO(viologen)G₂CBS(NH₃)₄]Cl₆ (6). The deprotection of compound 4 (0.100, 0.053 mmol) with TFA (5 mL) was carried out in the similar way than the synthesis of 5. The compound 6 was obtained as yellow solid (0.058 mg, 79%).

¹H NMR (300 MHz, D₂O): δ −0.015 (s, 33H, SiMe and SiMe₂), 0.39 (m, 34H, SiCH₂), 1.08 (m, 14H, CH₂CH₂Si) 1.55 (m, 8H, CH₂CH₂NH₃Cl), 1.95 (m, 2H, N⁺CH₂CH₂), 2.84 (m, 8H, CH₂NH₃Cl), 4.64 (m, 2H, N⁺CH₂ overlapped with water), 5.98 (s, 2H, CH₂), 7.65 (d, ³J_HH = 8.4 Hz, 2H, H²), 7.92 (d, ³J_HH = 8.4 Hz, 2H, H³), 8.51 (m, 4H, H⁶ and H⁹), 9.10 (d, 2H, H⁵), 9.19 (d, 2H, H¹⁰), 9.87 (s, 1H, CHO). ¹³C {¹H} NMR (300 MHz, D₂O): δ −5.89 (MeSi), −3.94 (SiMe₂), 11.66–21.60 (SiCH₂), 34.78 (N⁺CH₂CH₂), 42.37 (CH₂NH₃Cl), 61.93 (NCH₂), 64.13 (CH₂), 126.98 (C⁹), 127.14 (C⁶), 129.28 (C²), 130.80 (C³), 136.58 (C¹), 140.37 (C⁴), 145.46 (C¹⁰), 145.70 (C⁵), 149.69 (C⁸), 150.67 (C⁷), 201.7 (CHO).

Synthesis of dendron [CHO(viologen)G₁CBS(NHBoc)₂](PF₆)₂ (7). The compound 3 was dissolved in a mixture CHCl₃/H₂O (1 : 3), and 2 mL of a saturated aqueous solution of NH₄PF₆ was added dropwise. The mixture was extracted two times with CHCl₃ (2 × 20 mL), and the combined organic phase was dried with MgSO₄, and the solvent was removed in vacuum. The residue was purified by size exclusion chromatography in THF (see compound 1 for process), which gave the compound 7 as brown oil (0.420 g, 78%).

¹H NMR (400 MHz, CD₃CN): δ −0.020 (s, 15H, SiMe and SiMe₂), 0.48 (t, ³J_HH = 8.8 Hz, 4H, SiCH₂CH₂CH₂NH), 0.60–0.62 (m, 10H, SiCH₂), 1.33–1.49 (m, 28H, CH₂ and C(CH₃)₃), 2.05 (t, ³J_HH = 7.2 Hz, 2H, N⁺CH₂CH₂), 2.96–3.01 (m, 4H, CH₂NH), 4.64 (t, ³J_HH = 7.6 Hz, 2H, N⁺CH₂), 5.31 (br s, 2H, NH), 5.95 (s, 2H, CH₂), 7.69 (d, ³J_HH = 8.4 Hz, 2H, H²), 8.03 (d, ³J_HH = 8.4 Hz, 2H, H³), 8.41–8.45 (m, 4H, H⁶ and H⁹), 8.94 (d, ³J_HH = 6.8 Hz, 2H, H⁵), 9.02 (d, ³J_HH = 7.2 Hz, 2H, H¹⁰), 10.07 (s, 1H, CHO). ¹³C {¹H} NMR (400 MHz, CD₃CN): δ −5.85 (SiMe), −4.12 (SiMe₂), 11.93 (SiCH₂CH₂CH₂NH), 13.18–24.34 (SiCH₂), 27.71 (C(CH₃)₃), 34.87 (N⁺CH₂CH₂), 43.43 (CH₂NH), 61.94 (NCH₂), 64.10 (CH₂), 77.91 (C(CH₃)₃), 127.25 (C⁹), 127.58 (C⁶), 129.84 (C²), 130.33 (C³), 137.43 (C¹), 138.51 (C⁴), 145.57 (C¹⁰), 145.91 (C⁵), 149.69 (C⁸), 150.67 (C⁷), 155.94 (CO, Boc), 192.10 (CHO). ³¹P NMR (300 MHz, CD₃CN): δ −144.62 (h, PF₆). ²⁹Si NMR (300 MHz, CD₃CN): δ 1.77 (SiMe), 1.81 (SiMe₂). Anal. calcd for: C₄₉H₈₂F₁₂N₄O₅P₂Si₃ (1181.38 g mol⁻¹): C, 49.82; H, 7.00; N, 4.74. Found: C, 49.46; H, 7.21; N, 4.36. MS: [M − PF₆]⁺ = 1035.52 D (calcd = 1035.47 D).

Synthesis of dendron [CHO(viologen)G₂CBS(NHBoc)₄](PF₆)₂ (8). The change of counterion of compound 4 was carried out in the similar way than the synthesis of 7. The compound 8 was obtained as a pale yellow solid (0.430 g, 88%) after a purification process identical to 7.

¹H NMR (400 MHz, CD₃CN): δ −0.045 (s, 9H, SiMe), −0.018 (s, 24H, SiMe₂), 0.48 (t, ³J_HH = 8.4 Hz, 8H, SiCH₂CH₂CH₂NH), 0.57–0.62 (m, 28H, SiCH₂) 1.34–1.43 (m, 56H, CH₂ and C(CH₃)₃), 2.04 (t, ³J_HH = 7.6 Hz, 2H, N⁺CH₂CH₂), 2.97–3.01 (m, 8H, CH₂NH) 4.63 (t, ³J_HH = 7.6 Hz, 2H, N⁺CH₂), 5.31 (br s, 4H, NH), 5.95 (s, 2H, CH₂), 7.69 (d, ³J_HH = 8.0 Hz, 2H, H²), 8.03 (d, ³J_HH = 8.4 Hz, 2H, H³), 8.40–8.44 (m, 4H, H⁶ and H⁹), 8.93 (d, ³J_HH = 6.8 Hz, 2H, H⁵), 9.02 (d, ³J_HH = 6.4 Hz, 2H, H¹⁰) 10.08 (s, 1H, CHO). ¹³C {¹H} NMR (400 MHz, CD₃CN): δ −5.47 (SiMe), −4.04 (SiMe₂), 11.99 (SiCH₂CH₂CH₂NH), 13.38–24.36 (SiCH₂), 27.74 (C(CH₃)₃), 34.99 (N⁺CH₂CH₂), 43.47 (CH₂NH), 61.93 (N⁺CH₂), 64.12 (CH₂), 77.92 (C(CH₃)₃), 127.28 (C⁹), 127.59 (C⁶), 129.84 (C²), 130.34 (C³), 137.47 (C¹), 138.50 (C⁴), 145.57 (C¹⁰), 145.94 (C⁵), 149.70 (C⁸), 150.64 (C⁷), 155.93 (CO, Boc), 192.08 (CHO). ³¹P NMR (400 MHz, CD₃CN): δ −144.61 (h, PF₆). ²⁹Si NMR (300 MHz, CD₃CN): δ 0.98 (MeSi), 1.77 (MeSi), −1.82 (SiMe₂ and SiMe). Anal. calcd for: C₈₃H₁₅₆F₁₂N₆O₉P₂Si₇ (1868.69 g mol⁻¹): 53.35; H, 8.41; N, 4.50. Found: C, 53.46; H, 8.18, N, 4.05. MS: [M − PF₆]⁺ = 1722.00 D (calcd = 1722.78 D).

Synthesis of dendrimer {N₃P₃[(viologen)G₁CBS(NHBoc)₂]₆}(PF₆)₁₂ (9). A solution of dendron 7 (350 mg, 0.296 mmol) in acetonitrile (4 mL) was added a solution of N₃P₃ [N(Me)NH₂] (III) (0.020 g, 0.048 mmol) in acetonitrile. This mixture was stirred 48 h. The solvent was removed in vacuum to yield the compound 9 as an orange powder (350 mg, 98%). No further purification was needed.

¹H NMR (300 MHz, CD₃CN): δ −0.02 (s, 90H, SiMe and SiMe₂), 0.50 (t, ³J_HH = 8.7 Hz, 24H, SiCH₂CH₂CH₂NH), 0.58–0.64 (m, 60H, SiCH₂) 1.33–1.49 (m, 168H, CH₂ and C(CH₃)₃), 2.04 (t, ³J_HH = 7.8 Hz, 12H, N⁺CH₂CH₂), 2.95–3.02 (m, 24H, CH₂NH), 3.30 (s, 18H, CH₃), 4.63 (t, ³J_HH = 7.8 Hz, 12H, N⁺CH₂), 5.31 (br s, 12H, NH), 5.82 (s, 12H, CH₂), 7.47 (d, ³J_HH = 8.1 Hz, 12H, H³), 7.72 (d, ³J_HH = 8.1 Hz, 12H, H²), 7.72 (s, 6H, CH=N), 8.40 (d, ³J_HH = 6.0 Hz, 24H, H⁶ and H⁹), 8.93 (d, ³J_HH = 6.9 Hz, 12H, H⁵), 8.98 (d, ³J_HH = 6.9 Hz, 12H, H¹⁰). ¹³C {¹H} NMR (300 MHz, CD₃CN): δ −5.84 (SiMe), −4.12 (SiMe₂), 11.93 (SiCH₂CH₂CH₂NH), 13.22–24.35 (SiCH₂), 27.72 (C(CH₃)₃), 31.92, 31.97 (2d, ²J_PC = 5.5 Hz, N–CH₃), 34.94 (N⁺CH₂CH₂), 43.43 (CH₂NH), 61.91 (N⁺CH₂), 64.41 (CH₂), 77.93 (C(CH₃)₃), 127.09 (C²), 127.21, 127.36 (C⁹ and C⁶), 129.86 (C³), 136.25, 136.32 (2d, ³J_PC = 10.0 Hz, CH=N), 137.88 (C¹), 139.07 (C⁴), 145.58 (C¹⁰ and C⁵), 149.74, 150.22 (C⁷ and C⁸), 155.95 (CO, Boc). ³¹P NMR (300 MHz, CD₃CN): δ −144.62 (h, PF₆), 17.14 (N₃P₃). ²⁹Si NMR (300 MHz, CD₃CN): δ 1.76 (SiMe), 1.81 (SiMe₂). Anal. calcd for: C₃₀₀H₅₁₀F₇₂N₃₉O₂₄P₁₅Si₁₈ (7385.54 g mol⁻¹): C, 48.79; H, 6.96; N, 7.40. Found: C, 48.32; H, 6.68; N, 7.56.

Synthesis of dendrimer {N₃P₃[(viologen)G₂CBS(NHBoc)₄]₆}(PF₆)₁₂ (10). The reaction of compound 8 (0.353 g, 0.189 mmol) with N₃P₃ [N(Me)NH₂] (0.012 g, 0.031 mmol) was carried out in the similar way than the synthesis of 9. The compound 10 was obtained as an orange powder (0.350 g, 98%). No further purification was needed.

¹H NMR (300 MHz, CD₃CN): δ −0.049, −0.022 (s, 198H, SiMe and SiMe₂), 0.47 (t, ³J_HH = 8.4 Hz, 48H, SiCH₂CH₂CH₂NH), 0.57–0.62 (m, 168H, SiCH₂), 1.31–1.47 (m, 336H, CH₂ and C(CH₃)₃), 2.04 (t, ³J_HH = 7.2 Hz, 12H, N⁺CH₂CH₂), 3.30 (s, 18H, CH₃), 2.95–3.02 (m, 48H, CH₂NH), 4.62 (t, ³J_HH = 7.2 Hz, 12H, N⁺CH₂), 5.34 (br s, 24H, NH), 5.83 (s, 12H, CH₂), 7.48 (d, ³J_HH = 6.0 Hz, 12H, H³), 7.72 (d, ³J_HH = 8.1 Hz, 12H, H²), 7.72 (s, 6H, CH=N), 8.41 (d, ³J_HH = 4.2 Hz, 24H, H⁶ and H⁹), 8.93 (d, ³J_HH = 6.0 Hz, 12H, H⁵), 8.99 (d, ³J_HH = 6.0 Hz, 12H, H¹⁰). ¹³C {¹H} NMR (300 MHz, CD₃CN): δ −0.049 (SiMe), −0.022 (SiMe₂), 11.93 (SiCH₂CH₂CH₂NH), 13.20–24.34 (SiCH₂), 27.71 (C(CH₃)₃), 34.94 (N⁺CH₂CH₂), 43.42 (CH₂NH), 61.71 (N⁺CH₂), 63.49 (CH₂), 77.94 (C(CH₃)₃), 127.21, 127.37 (C⁶ and C⁹), 130.08 (C³), 130.27 (C²), 136.25, 136.32 (2d, ³J_PC = 10.0 Hz, CH=N), 137.31 (C¹), 139.07 (C⁴), 145.76 (C¹⁰), 149.19 (C⁵), 149.49 (C⁸), 150.22 (C⁷), 155.99 (CO, Boc), 192.21 (CHO). ³¹P NMR (300 MHz, CD₃CN): δ 17.17 (N₃P₃), −144.60 (h, PF₆). ²⁹Si NMR (300 MHz, CD₃CN): δ 0.97 (SiMe), 1.82 (SiMe₂ and SiMe). Anal. calcd for: C₅₀₄H₉₅₄F₇₂N₅₁O₄₈P₁₅Si₄₂ (11521.2 g mol⁻¹): C, 52.60; H, 8.35; N, 6.21. Found: C, 52.30; H, 7.95 N, 6.18.

Synthesis of dendrimer {N₃P₃[(viologen)G₁CBS(NH₃)₂]₆}(Cl)₂₄ (11). Dendrimer 9 (360 mg, 0.049 mmol) was deprotected with 5 mL of a solution of TFA al 20% in anhydrous dichloromethane, and the mixture was stirred for 30 min. The solvent was removed in vacuum. After, the crude was dissolved in acetonitrile (20 mL) and 3 mL of a saturated aqueous solution of ^tBu₄NCl added carefully, provoking the precipitation of the desired product. The resulting solid was filtered and washed several times with acetonitrile, obtaining the compound 11 as yellow solid (200 mg, 78%). This purification process is perfectly enough to removed byproducts.

¹H NMR (300 MHz, D₂O): δ 0.10 (s, 198H, SiMe and SiMe₂), 0.40 (m 196H, SiCH₂CH₂CH₂NH and SiCH₂), 1.22 (m, 92H, CH₂), 1.51 (m, 48H, CH₂), 1.97 (m, 12H, N⁺CH₂CH₂), 2.82 (m, 48H, CH₂NH), 3.07 (s, 18H, CH₃), 4.60 (12H, N⁺CH₂ overlapped with water), 5.82 (s, 12H, CH₂), 7.51 (m, 30H, H³, H² and CH=N), 8.46 (m, 24H, H⁶ and H⁹), 9.02 (m, 12H, H⁵ and 12H, H¹⁰). ³¹P NMR (300 MHz, CD₃OD): δ 18.12 (N₃P₃). Anal. calcd for: C₂₄₀Cl₂₄H₄₂₆N₃₉P₃Si₁₈ (5307.54 g mol⁻¹): C, 54.31; H, 8.09; N, 10.29. Found: C, 53.78; H, 8.17 N, 10.15.

Synthesis of dendrimer {N₃P₃[(viologen)G₂CBS(NH₃)₄]₆}(Cl)₃₆ (12). The deprotection of dendrimer 10 (0.350, 0.030 mmol) with TFA (5 mL) was carried out in the similar way than the synthesis of 11. The compound 12 was obtained as yellow solid (0.210 mg, 80%).

¹H NMR (300 MHz, CD₃CN): δ −0.10 (s, 198H, SiMe and SiMe₂), 0.45 (m, 196H, SiCH₂CH₂CH₂NH and SiCH₂), 1.20 (m, 92H, CH₂), 1.58 (m, 48H, CH₂), 2.0 (m, 12H, N⁺CH₂CH₂), 2.87 (m, 48H, CH₂NH), 3.11 (s, 18H, CH₃), 4.60 (, 12H, N⁺CH₂ overlapped with water), 5.93 (s, 12H, CH₂), 7.50 (m, 30H, H³, H² and CH=N), 8.70 (m, 24H, H⁶ and H⁹), 9.19 (m, 24H, H⁵ and H¹⁰). ³¹P NMR (300 MHz, CD₃OD): δ 16.42 (N₃P₃). Anal. calcd for: C₃₈₄H₇₈₅Cl₃₆N₅₁P₃Si₄₂ (8666.50 g mol⁻¹): C, 53.22; H, 9.13; N, 8.24. Found: C, 52.93; H, 9.42; N, 7.94.

Synthesis of dendrimer {N₃P₃[(viologen)G₁CBS(NH₃)₂]₆}(PF₆)₂₄ (13). The dendrimer 11 (0.040 g, 0.008 mmol) was dissolved in 10 mL of water and 2.5 mL of a saturated aqueous solution of NH₄PF₆ was added dropwise. The resulting solid was filtered, and washed with water and diethyl ether, obtaining the compound 13 as a yellow solid (0.052 g, 86%).

¹H NMR (300 MHz, CD₃CN): δ −0.01 (s, 90H, SiMe and SiMe₂), 0.52 (t, ³J_HH = 9.0 Hz, 24H, SiCH₂CH₂CH₂NH), 0.59–0.65 (m, 60H, SiCH₂), 1.26–1.44 (m, 36H, CH₂), 1.55–1.66 (m, 24H, CH₂), 2.04 (t, ³J_HH = 8.0 Hz, 12H, N⁺CH₂CH₂), 2.91–2.96 (m, 24H, CH₂NH), 3.31 (s, 18H, CH₃), 4.63 (t, ³J_HH = 8.0 Hz, 12H, N⁺CH₂), 5.81 (s, 12H, CH₂), 7.46 (d, ³J_HH = 8.2 Hz, 12H, H³), 7.72 (d, ³J_HH = 8.5 Hz, 12H, H²), 7.73 (s, 6H, CH=N), 8.40 (d, ³J_HH = 5.5 Hz, 24H, H⁶ and H⁹), 8.91 (d, ³J_HH = 6.7 Hz, 12H, H⁵), 8.96 (d, ³J_HH = 6.6 Hz, 12H, H¹⁰). ¹³C {¹H} NMR (75 MHz, CD₃CN): δ −6.11 (SiMe), −4.47 (SiMe₂), 11.39 (SiCH₂CH₂CH₂NH), 13.00–21.42 (SiCH₂), 34.96 (N⁺CH₂CH₂), 42.99 (CH₂NH), 61.82 (N⁺CH₂), 64.27 (CH₂), 127.08 (C²), 127.21, 127.33 (C⁹ and C⁶), 129.76 (C³), 132.11 (C¹), 136.25, 136.32 (2d, ³J_PC = 10.0 Hz, CH=N), 137.75 (C⁴), 145.45, 145.50 (C¹⁰ and C⁵), 149.73, 150.38 (C⁷ and C⁸). ³¹P NMR (300 MHz, CD₃CN): δ −144.62 (h, PF₆), 17.19 (N₃P₃). ²⁹Si NMR (300 MHz, CD₃CN): δ 1.75 (SiMe), 2.09 (SiMe₂).

Synthesis of dendrimer {N₃P₃[(viologen)G₂CBS(NH₃)₄]₆}(PF₆)₃₆ (14). The change of counterion of compound 12 (0.050 g, 0.006 mmol) was carried out in the similar way than the synthesis of 13. The compound 14 was obtained as a pale yellow solid (0.063 mg, 83%).

¹H NMR (400 MHz, CD₃CN): δ −0.03 (s, 198H, SiMe and SiMe₂), 0.52 (t, ³J_HH = 8.0 Hz, 48H, SiCH₂CH₂CH₂NH), 0.58–0.62 (m, 148H, SiCH₂), 1.31–1.46 (m, 92H, CH₂), 1.58–1.66 (m, 48H, CH₂), 2.04 (t, ³J_HH = 8.0 Hz, 12H, N⁺CH₂CH₂), 2.92–2.96 (m, 48H, CH₂NH), 3.31 (s, 18H, CH₃), 4.60 (t, ³J_HH = 8.0 Hz, 12H, N⁺CH₂), 5.80 (s, 12H, CH₂), 7.46 (d, ³J_HH = 9.6 Hz, 12H, H³), 7.73 (d, ³J_HH = 7.4 Hz, 12H, H²), 7.73 (s, 6H, CH=N), 8.40 (d, ³J_HH = 5.4 Hz, 24H, H⁶ and H⁹), 8.91 (d, ³J_HH = 5.8 Hz, 12H, H⁵), 8.96 (d, ³J_HH = 4.9 Hz, 12H, H¹⁰). ¹³C {¹H} NMR (75 MHz, CD₃CN): δ −5.68 (SiMe), −4.40 (SiMe₂), 11.41 (SiCH₂CH₂CH₂NH), 13.40–21.40 (SiCH₂), 35.07 (N⁺CH₂CH₂), 43.16 (CH₂NH), 61.86 (N⁺CH₂), 64.21 (CH₂), 127.02 (C²), 127.34, 127.20 (C⁹ and C⁶), 129.74 (C³), 132.11 (C¹), 136.25, 136.32 (2d, ³J_PC = 10.0 Hz, CH=N), 137.75 (C⁴), 145.45, 145.56 (C¹⁰ and C⁵), 149.75, 150.46 (C⁷ and C⁸). ³¹P NMR (300 MHz, CD₃CN): δ −144.62 (h, PF₆), 17.19 (N₃P₃). ²⁹Si NMR (300 MHz, CD₃CN): δ 0.97 (SiMe), 2.10 (SiMe₂).

Zeta potential measurements

The measurements of the zeta potential (electrokinetic potential) were performed with Zetasizer Nano ZS from Malvern, which uses electrophoresis and LDV (Laser Doppler Velocimetry) techniques. Applying a combination of these two techniques allowed measuring the electrophoretic mobility of the molecules in the solution. The zeta potential value was calculated directly from the Helmholtz–Smoluchowski equation using the Malvern software.⁴³

The measurements of zeta potential of dendrimers alone were performed in water and phosphate buffer (10 mM, pH = 7.4).

The measurements of HSA zeta potential in the presence of dendrimers were performed in phosphate buffer at 37 °C. Increasing concentrations of the dendrimer in the range 0.1–10 μM were added to HSA of a concentration of 0.2 μM and zeta potential was measured.

Circular dichroism (CD) measurements

The CD spectra (195–260 nm) for HSA were recorded in the presence/absence of dendrimers on a Jasco J-815 CD spectropolarimeter, in 5 mm path length quartz cuvettes, with a wavelength step of 1 nm, a response time of 4 s and a scan rate of 20 nm min⁻¹, thermostatted at 37 °C. Each spectrum was the average of three repetitions. HSA was used at a concentration of 2 μM. The molar ratio of dendrimers/HSA ranged from 0.5 to 20. Viologen–carbosilane–phosphorus dendrimers alone at the concentrations used in the experiments produced no discernible features in their CD spectra and were used as blanks (controls) for HSA/dendrimers complex spectra.

Fluorescence quenching of HSA

Fluorescence measurements were performed using a Perkin-Elmer LS-55 spectrofluorometer with excitation at 290 nm, and emission spectra were recorded in the range of 305–435 nm. HSA at a concentration of 2.5 μM was contained in 5 mm quartz cuvette, and it was thermostatted at 37 °C. Next, increasing concentrations of dendrimers were added to HSA and fluorescence spectra were recorded. Before examining the fluorescence properties of the protein, it was checked that dendrimers were not excited by 290 nm wavelength and did not emit fluorescence.

Cytotoxicity assay

Chinese hamster fibroblasts (B14 cell line) were purchased from Child Health Center in Warsaw (Poland). Cells were grown as a monolayer in DMEM medium supplemented with 10% foetal bovine serum with 100 units per ml gentamycin. The cells were maintained at 37 °C in an atmosphere of 5% CO₂ and 95% air with more than 95% humidity. Cells were split for subcultures every 2 days.

Cell viability was measured by the MTT assay. The test is based on the reduction of the soluble yellow MTT tetrazolium salt to a blue, insoluble formazan produced by mitochondrial succinate dehydrogenase. The amount of formazan produced is proportional to the number of living cells. Cells were seeded at a density of 1.5 × 10⁵ per well into 96-well microtitrate plates using DMEM medium. They were treated with dendrimers at 37 °C in a 5% carbon dioxide–95% air atmosphere for 24 h, and recovered by gentle washing with PBS (pH = 7.4) twice. After incubation, 50 μl of 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT) solution was added to each well, followed by 3 h of incubation. Next, MTT-containing medium was removed, and 100 μl of DMSO was added to each well to dissolve formazan crystals. Absorbance of the converted dye was measured at 570 nm using a microplate spectrophotometer (BioTek). Cell viability was calculated as the percent ratio of absorbance of the samples to the referent control.

Morphological changes of B14 cell line

Morphology of B14 cells line and erythrocytes was evaluated on the basis of microscopic observation. The final concentrations of dendrimers in the sample were 0.1; 1 and 10 μM. Cells were incubated with dendrimers at 37 °C for 24 h. Then, cells were observed under Olympus CKX41 optical microscope at a magnification of 400×.

Hemotoxicity

Blood from healthy donors was obtained from Central Blood Bank in Lodz. Blood was anticoagulated with 3% sodium citrate. Erythrocytes were separated from blood plasma and leukocytes by centrifugation (4000g, 10 min) at 4 °C and washed three times with PBS (phosphate buffered saline; pH = 7.4). Erythrocytes were used immediately after isolation.

To study the effect of viologen–carbosilane–phosphorus dendrimers on erythrocytes, dendrimers were added in the concentration range of 0.01–10 μM to red blood cells of 2% hematocrit and incubated at 37 °C for 24 h. After incubation suspensions were centrifuged (1000g, 10 min). Hemolysis was determined by measuring the hemoglobin content in the supernatant at 540 nm.

Percent of hemolysis was calculated from the formula:

Hemolysis [%] = (A/A_c) × 100%

where, A is the absorbance of the sample, and A_c is the absorbance of the sample in water (100% hemolysis).

Statistical analysis

The results are presented as mean ± SD. Statistical analysis was performed with the use of the SigmaPlot 11.0 Systat Software. Statistical evaluation of the difference between control and treated group was performed using Student's t-test. P < 0.05 and below was accepted as statistically significant.

Acknowledgements

This work has been supported by grants from MINECO CTQ2011-23245, Consortium NANODENDMED ref S2011/BMD-2351 (CAM) and also supported by CIBER-BBN as an initiative funded by the VI National R & D & i Plan 2008–2011, Iniciativa Ingenio 2010, Consolider Program, CIBER Actions and financed by the Instituto de Salud Carlos III, with assistance from the European Regional Development Fund. National Science Centre of Poland, Project no. UMO-2012/04/M/NZ1/00059, the project Iuventus Plus (nr IP2011 041071). The CNRS (France) was also supporting this work.

References

1. J. M. J. Fréchet and D. A. Tomalia, Dendrimers and Other Dendritic Polymers, VCH, Weinheim, 2002.
2. (a) S. Mignani, S. El Kazzouli, M. Bousmina and J. P. Majoral, Prog. Polym. Sci., 2013, 38, 993–1008; (b) S. Mignani, S. El Kazzouli, M. Bousmina and J.-P. Majoral, Adv. Drug Delivery Rev., 2013, 65, 1316–1330.
3. P. Kesharwani, K. Jain and N. K. Jain, Prog. Polym. Sci., 2014, 39, 268–307.
4. M. A. Mintzer and E. E. Simanek, Chem. Rev., 2009, 109, 259–302.
5. X. Liu, P. Rocchi and L. Peng, New J. Chem., 2012, 9, 256–263.
6. P. M. H. Heegaard, U. Boas and N. S. Sorensen, Bioconjugate Chem., 2010, 21, 405–418.
7. D. G. Shcharbin, B. Klajnert and M. Bryszewska, Biochemistry, 2009, 74, 1070–1079.
8. H. Zhong, Z. G. He, Z. Li, G. Y. Li, S. R. Shen and X. L. Li, J. Biomater. Appl., 2008, 22, 527–544.
9. D. K. Smith, Curr. Top. Med. Chem., 2008, 8, 1187–1203.
10. Y. Gao, G. Gao, Y. He, T. L. Liu and R. Qi, Mini-Rev. Med. Chem., 2008, 8, 889–900.
11. H. S. Parekh, Curr. Pharm. Des., 2007, 13, 2837–2850.
12. M. Guillot-Nieckowski, S. Eisler and F. Diederich, New J. Chem., 2007, 31, 1111–1127.
13. C. Dùfes, I. F. Uchegbu and A. G. Schatzlein, Adv. Drug Delivery Rev., 2005, 57, 2177–2202.
14. K. Kegan and R. Engel, J. Chem. Soc., Chem. Commun., 1990, 1084–1085.
15. A. M. Caminade, C. O. Turrin, R. Laurent, C. Rebout and J. P. Majoral, Polym. Int., 2006, 55, 1155–1160.
16. A. M. Caminade and J. P. Majoral, New J. Chem., 2013, 37, 3358–3373.
17. S. Asaftei and E. De Clerck, J. Med. Chem., 2010, 53, 3480–3488.
18. K. Ciepluch, N. Katir, A. El Kadib, A. Felczak, K. Zawadzka, M. Weber, B. Klajnert, K. Lisowska, A. M. Caminade, M. Bousmina, M. Bryszewska and J. P. Majoral, Mol. Pharm., 2012, 9, 448–457.
19. J. Lazniewska, K. Milowska, N. Katir, A. El Kadib, M. Bryszewska, J. P. Majoral and T. Gabryelak, Cell. Mol. Biol. Lett., 2013, 18, 459–478.
20. T. Gonzalo, M. I. Clemente, L. Chonco, N. D. Weber, L. Díaz, M. J. Serramía, R. Gras, P. Ortega, F. J. de la Mata, R. Gómez, L. A. López-Fernández, M. A. Muñoz-Fernández and J. L. Jiménez, ChemMedChem, 2010, 5, 921–929.
21. I. Posadas, B. López-Hernández, M. I. Clemente, J. L. Jiménez, P. Ortega, F. J. de la Mata, R. Gómez, M. A. Muñoz-Fernández and V. Ceña, Pharm. Res., 2009, 26, 1181–1191.
22. N. Weber, P. Ortega, M. I. Clemente, D. Shcharbin, M. Bryszewska, F. J. de la Mata, R. Gómez and M. A. Muñoz-Fernández, J. Controlled Release, 2008, 132, 55–64.
23. J. F. Bermejo, P. Ortega, L. Chonco, R. Eritja, R. Samaniego, M. Mullner, E. de Jesús, F. J. de la Mata, J. C. Flores, R. Gómez and A. Muñoz-Fernández, Chem.–Eur. J., 2007, 13, 483–495.
24. B. Rasines, J. M. Hernández-Ros, N. de las Cuevas, J. L. Copa-Patiño, J. Soliveri, M. A. Muñoz-Fernández, R. Gómez and F. J. de la Mata, Dalton Trans., 2009, 8704–8713.
25. E. Fuentes-Paniagua, J. M. Hernandez-Ros, M. Sanchez-Milla, M. A. Camero, M. Maly, J. Perez-Serrano, J. L. Copa-Patiño, J. Sanchez-Nieves, J. Soliveri, R. Gomez and F. J. de la Mata, RSC Adv., 2014, 4, 1256–1265.
26. R. Sharma, N. Kottari, Y. M. Chabre, R. Rej, N. K. Saadeh and R. Roy, Polym. Chem., 2014, 5, 4321–4331.
27. N. Kottari, Y. M. Chabre, T. C. Shiao and R. Roy, Chem. Commun., 2014, 50, 1983–1985.
28. R. Sharma, N. Kottari, Y. M. Chabre, L. Abbassi, T. C. Shiao and R. Roy, Chem. Commun., 2014, 50, 13300–13303.
29. R. Sharma, I. Zhang, L. Abbassi, R. Rabindra, D. Maysinger and R. Roy, Polym. Chem., 2015, 6, 1436–1444.
30. R. Roy and T. C. Shiao, Chem. Soc. Rev., 2015 10.1039/c4cs00359d.
31. C. Galliot, A.-M. Caminade, F. Dahan and J. P. Majoral, Angew. Chem., Int. Ed. Engl., 1993, 32, 1477–1479.
32. V. Maraval, A.-M. Caminade, J.-P. Majoral and J. C. Blais, Angew. Chem., Int. Ed., 2003, 42, 1822–1826.
33. C. Galliot, A.-M. Caminade, F. Dahan and J.-P. Majoral, Angew. Chem., Int. Ed. Engl., 1993, 32, 1477–1479.
34. E. Fuentes-Paniagua, C. E. Peña-González, M. Galán, R. Gómez, F. J. de la Mata and J. Sánchez-Nieves, Organometallics, 2013, 32, 1789–1796.
35. K. Ciepluch, N. Katir, A. E. Kadib, M. Weber, A.-M. Caminade, M. Bousmina, J.-P. Majoral and M. Bryszewska, J. Lumin., 2012, 132, 1553–1563.
36. D. Shcharbin, B. Klajnert, V. Mazhul and M. Bryszewska, J. Fluoresc., 2005, 15, 21–28.
37. T. Peters Jr, All About Albumin: Biochemistry, Genetics and Medical Applications, Academic Press, 1995.
38. B. Klajnert, L. Stanisławska, M. Bryszewska and B. Pałecz, Biochim. Biophys. Acta, 2003, 1648, 115–126.
39. J. Giri, M. S. Diallo, A. J. Simpson, Y. Liu, W. A. Goddard, R. Kumar and G. C. Woods, ACS Nano, 2011, 5, 3456–3468.
40. J. Lazniewska, K. Milowska, M. Zablocka, S. Mignani, A.-M. Caminade, J.-P. Majoral, M. Bryszewska and T. Gabryelak, Mol. Pharm., 2013, 10, 3484–3496.
41. I. Heredero-Bermejo, J. L. Copa-Patiño, J. Soliveri, E. Fuentes-Paniagua, F. J. de la Mata, R. Gomez and J. Perez-Serrano, Parasitol. Res., 2015, 114, 473–486.
42. E. Pedziwiatr-Werbicka, E. Fuentes, V. Dzmitruk, J. Sánchez-Nieves, M. Sudas, E. Drozd, A. Shakhbazau, D. Shcharbin, F. J. de la Mata, R. Gómez-Ramírez, M. A. Muñoz-Fernández and M. Bryszewska, Colloids Surf., B, 2013, 109, 183–189.
43. A. Sze, D. Erickson, L. Ren and D. Li, J. Colloid Interface Sci., 2003, 261, 402–410.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra00960j

---