Synthesis of bifunctional integrin-binding peptides containing PEG spacers of defined length for non-viral gene delivery

FULL PAPER

DOI: 10.1002/ejoc.200701188

Synthesis of Bifunctional Integrin-Binding Peptides Containing PEG Spacers

of Defined Length for Non-Viral Gene Delivery

Michael A. Pilkington-Miksa,^[a]Supti Sarkar,^[b]Michele J. Writer,^[c]Susie E. Barker,^[c]

Parviz Ayazi Shamlou,^[b,d]Stephen L. Hart,^[c]Helen C. Hailes,^[a]and Alethea B. Tabor*^[a]

Keywords: Amino acids / Gene delivery / Peptides / Poly(ethylene glycol)

Improving the buffer and serum stability of non-viral gene

delivery vectors, and increasing their circulation time in vivo,

is an important focus of current research in gene therapy. The

most successful strategies to date have involved shielding the

complexes with large polydisperse PEG adducts. However,

this approach is accompanied by a fall in transfection effi-

ciency. In this paper we describe the solid-phase synthesis of

a series of bifunctional peptides bearing short PEG spacers

of defined structure as components of lipopolyplex gene de-

livery vectors. Short, high-yielding routes to a series of PEG-

amino acids are described: these PEG-amino acids can be

used in varying combinations to afford bifunctional peptides

with varying lengths of PEG spacers by using standard solid-

phase synthesis techniques. A series of lipopolyplexes were

formulated using these bifunctional peptides, and their trans-

fection properties assessed. Dynamic light scattering mea-

surements on the complex with the best transfection proper-

ties showed that in phosphate-buffered saline this complex

was considerably more stable, and aggregated more slowly,

than a complex formulated using a similar peptide lacking

the short PEG spacer.

Germany, 2008)

some lipoplex^[12–14]and polyplex^[15,16]systems such PEG

shielding works well, resulting in both serum stability and

efficient transfection. However, with many others^[17–20]the

vectors, although stable in the circulation, have greatly re-

duced transfection properties, either because of steric hin-

drance or failure of the DNA to be released from the cell

following internalisation. In some cases PEG conjugation is

even ineffective in stabilising the complex and protecting

the DNA from in vivo degradation.^[21]The effectiveness of

PEG conjugation appears to depend on the cell type and,

for modified cationic lipids, the length of the acyl

chain.^[22,23]Preparation of vectors with cleavable PEG

moieties frequently improves transfection,^[24,25]and, cru-

cially, if a cell targeting group is accessible on the outside

of the PEG coat, the transfection efficiencies are often

restored.^[26,27]It is notable, however, that all of these ap-

proaches have relied on the conjugation of large polymeric

PEG units, and, possibly because of the synthetic effort in-

volved, the role of vector components bearing short, de-

fined PEG units in shielding non-viral vectors has not pre-

viously been investigated.

Introduction

Non-viral gene delivery systems typically involve the con-

densation of plasmid DNA with either cationic polymers

(e.g., polylysine peptides) to form polyplexes,^[1]cationic

lipids to afford lipoplexes,^[2]or both lipids and peptides to

give lipopolyplexes.^[3–7]A recurrent problem with these pos-

itively charged complexes is that they are bound by plasma

proteins and therefore quickly cleared by the reticuloendo-

thelial system (RES), which limits their circulation time in

vivo.^[8]In addition, such formulations are unstable in buffer

and serum, forming aggregates and/or looser macromolecu-

lar complexes. Several approaches to improve the in vivo

stabilisation and serum shielding of non-viral vectors have

been attempted,^[9–11]including shielding the cationic com-

plexes with high molecular weight polymeric PEG. With

[a] Department of Chemistry, University College London, Christo-

pher Ingold Laboratories,

20, Gordon Street, London, WC1H 0AJ, UK

Fax: +44-207-7679-7463

E-mail: a.b.tabor@ucl.ac.uk

[b] Department of Biochemical Engineering, University College

London,

We have initiated a program of work to improve the sta-

bility, circulation time and transfection efficiencies of lipo-

polyplex vectors by synthesising the individual components

with short PEG spacers of defined length and structure as

an integral part of the component lipids and peptides. In

this paper we report the synthesis of a series of bifunctional

peptides that comprise a DNA-binding sequence (K₁₆) and

a disulfide-bridged sequence binding to the α₅β₁integrin

Torrington Place, London, WC1E 7JE, UK

[c] Wolfson Centre for Gene Therapy of Childhood Disease, Insti-

tute of Child Health, University College London

30, Guilford Street, London, WC1N 1EH, UK

[d] Present address: Process Engineering Center, Eli Lilly and Com-

pany, Lilly Corporate Center, DC 3127, Indianapolis, Indiana

46285, USA

Supporting information for this article is available on the

WWW under http://ww w .eurjoc.org/ or from the author.

2900

Eur. J. Org. Chem. 2008, 2900–2914

Non-Viral Gene Delivery

(CRRETAWAC) joined by a series of PEG-amino acids of and then oxidise it to give the carboxylic acid. For amino

defined length. As the optimum length of the PEG spacer acid 1, tetraethylene glycol was first monoprotected with

required to simultaneously target the complex to the inte- the trityl group to give 4 (Scheme 1). This compound was

grin and to shield the complex from unwanted interactions then mesylated, treated with sodium azide and reduced un-

with plasma proteins and serum was not known, we have der Staudinger conditions^[33]to give the free amine 5. This

also devised short, high-yielding routes to PEG-amino ac- was readily protected to give 6 and the trityl group then

ids 1, 2 and 3 (Figure 1), which could be used, singly or in removed to afford 7. Careful choice of reagents was then

combination, to produce a range of different PEG-linked necessary in order to effect oxidation to the carboxylic acid

peptides.

in the presence of both the Fmoc group and the PEG chain.

Previous reports^[34]had indicated that oxidation of PEG

alcohols with sodium hypochlorite simultaneously resulted

in extensive ether linkage cleavage, and we therefore carried

out a series of test reactions on monobenzylated tetraethyl-

ene glycol under a variety of oxidising conditions. Oxi-

dations with potassium dichromate,^[35]potassium perman-

ganate^[36]and silver(II) oxide^[37]were tested, but all caused

either PEG-chain cleavage (determined by mass spec-

troscopy and NMR) or afforded the desired product in low

yield. Oxidation with chromium trioxide in sulfuric and

acetic acids^[38]successfully converted the monobenzylated

tetraethylene glycol into the desired acid in 84% yield with-

out PEG-chain ether cleavage. However, when this method

was attempted on the Fmoc-protected amino alcohol 7,

poor yields of partially deprotected amino acid were reco-

vered. After further experimentation with various chro-

Figure 1. PEG-amino acids used in this study.

Results and Discussion

Synthesis of PEG-Amino Acids 1, 2 and 3

PEG-amino acids are valuable building blocks for the mium oxidation methods, the Jones oxidation^[39]proved sat-

synthesis of tools for biological chemistry as they enable isfactory, affording 1 in 68% yield. The overall yield of the

hydrophilic, flexible PEG spacers^[28]to be readily inserted synthesis, over six steps, was 51%, and the reactions could

into biologically active peptides using standard techniques be performed on a multigram scale.

of solid-phase peptide synthesis (SPPS). It is desirable that

synthetic approaches to these amino acids be high-yielding,

require few purification steps (as difficulties are often en-

countered in the purification of PEG-containing moieties)

and make use of readily available, low-cost starting materi-

als such as tri- and tetraethylene glycol where possible.

A synthetic route to Fmoc-protected PEG-amino acid 1

(Haa4) for peptide synthesis was first reported by Boumrah

et al.^[29]who utilised a homologation of 2-(chloroethoxy)-

ethoxyethanol with ethyl diazoacetate in the presence of

BF₃·OEt₂to build the desired tetraethylene glycol based es-

ter, and converted the free hydroxy group to the azide and

Scheme 1. Synthesis of Fmoc-Haa4 (1). Reagents: a) TrtCl, Et₃N,

thence to the Fmoc-protected amino acid. This approach

DCM (95%); b) (i) MsCl, Et₃N, DCM; (ii) NaN₃, DMF (91%); c)

was recently modified by Thumshirn et al.^[30]and Dekker

PPh₃, H₂O, THF (96%); d) Fmoc-Cl, NMM, DCM (95%); e) (i)

et al.^[31]to prepare the Fmoc-protected amino acid 2

TFA, DCM, Et₃SiH; (ii) NaHCO₃(aq.) (94%); f) CrO₃, 1.5 

H₂SO₄, acetone (68%).

(Fmoc-Haa7) from hexaethylene glycol. In all of these ap-

proaches, however, the overall reported yields were moder-

ate, and the routes to 2 use the more expensive hexaethylene

glycol. In addition, by following this approach, we encoun-

We were then able to generalise this approach to synthe-

tered similar problems to those reported in the literature,^[29]sise longer-chain PEG-amino acids. For amino acid 3

particularly with the final Fmoc protection of the amino (Fmoc-Haa9), as nonaethylene glycol is not available, the

acid, which has been reported in a number of systems to PEG framework was constructed from triethylene glycol.

give poor yields and competing di-and tripeptide forma- We first synthesised the hexaethylene glycol skeleton from

tion.^[32]

triethylene glycol as follows (Scheme 2). Monotritylation of

We therefore sought to develop a route to these amino an excess of triethylene glycol gave 8, which was then mesyl-

acids from low-cost starting materials, avoiding some of the ated to afford 9. Similarly, monobenzylation of an excess of

problems highlighted in the literature. To avoid the final triethylene glycol in the presence of tetrabutylammonium

Fmoc protection step, we aimed to synthesise the appropri- bromide afforded 10. Reaction of the mesylate 9 with the

ate PEG-amino alcohol, protect the amino functionality sodium salt of 10 gave the desired hexaethylene glycol unit

Eur. J. Org. Chem. 2008, 2900–2914

www.eurjoc.org

2901

A. B. Tabor et al.

FULL PAPER

11, which was then detritylated under standard conditions

We then adapted this approach to give an expedient and

to give 12. We initially coupled this to the dibenzylamine high-yielding synthesis of amino acid 2 (Fmoc-Haa7), start-

13. However, removal of the O-benzyl protecting group ing from the tri- and tetraethylene glycol derivatives pre-

from the resulting tribenzylated material proved impossible; pared previously. Thus, tritylated tetraethylene glycol 4 was

hydrogenation catalysts were poisoned by the primary mesylated and then treated with the sodium salt of benzyl-

amino group,^[40]whereas dissolving metal reduction, al- ated triethylene glycol 10 to give 19 (Scheme 3). This was

though successful in removing all three protecting groups, followed by a similar sequence of detritylation and conver-

was not satisfactory as it was then impossible to isolate and sion to the azide 20, with purification only necessary in the

purify the resulting highly water-soluble amino-PEG9 last step. Staudinger reduction of the azide was carried out

alcohol. Instead, mesylate 9 was coupled to the sodium salt as previously^[33]with purification being effected by filtration

of 12 to afford 14, which was detritylated (to 15) and con- and extraction to give 21. Debenzylation and immediate

verted into the azide 16 using the conditions previously de- Fmoc protection again worked successfully to afford 22,

scribed. Again, simultaneous azide hydrogenation and O- which was converted into the desired amino acid 2 by Jones

debenzylation using a variety of methods^[41–43]was unsuc- oxidation, giving an overall yield of 50%. All three syn-

cessful, with only partial removal of the O-benzyl group in thetic routes could be accomplished on a large scale with

all cases. Instead, Staudinger reduction to give the amine few chromatography steps being required.

17 was followed by careful debenzylation with sodium in

liquid ammonia,^[44]and the crude amino alcohol was then

taken up in dioxane/water for Fmoc protection. The re-

sulting alcohol 18 was then subjected to Jones oxidation, as

before, to give the protected amino acid 3.

Scheme 3. Synthesis of Fmoc-Haa7 (2). Reagents: a) MsCl, Et₃N,

DCM; b) NaH, 10, DMF; c) TFA, DCM, Et₃SiH; d) satd. aq.

Na₂CO₃; e) MsCl, Et₃N, DCM; f) NaN₃, DMF (76% from 4);

g) Ph₃P, THF, H₂O (97%); h) Na, liq. NH₃, –78 °C; i) Fmoc-Cl,

NaHCO₃, dioxane, water (97%); j) CrO₃, acetone, H₂SO₄(71%).

Bifunctional Peptides

Lipopolyplex gene delivery vectors, which combine the

advantages of polyplexes and lipoplexes, have recently been

described by a number of groups. One of the first of these

was the receptor-targeted, self-assembling LID vector com-

plex described by Hart et al.^[4]This consists of plasmid

DNA, a lipid component [a 1:1 mixture of a cationic lipid,

DOTMA, and a co-lipid, DOPE (Lipofectin, L)] and a bi-

functional, cationic peptide (Peptide A) with a K₁₆domain,

which binds and condenses DNA, and a cyclic domain

(CRRETAWAC), which specifically binds to the α₅β₁inte-

grin, allowing the complex to be internalised by receptor-

Scheme 2. Synthesis of Fmoc-Haa9 (3). Reagents: a) TrtCl, Et₃N,

mediated endocytosis. LID vectors have been found to dis-

play high transfection efficiency and low toxicity in vitro

and in vivo, to transfect non-dividing cells efficiently and

were well tolerated with low immunogenicity in vivo.^[4,45–47]

However, this formulation has not been developed for sys-

temic gene delivery.

DCM (100%); b) MsCl, Et₃N, DCM (99%); c) NaH, TBABr, ben-

zyl chloride (99%); d) NaH, DMF (74%) then 9; e) (i) TFA, Et_3-

SiH, DCM; (ii) satd. aq. Na₂CO₃(99%); f) NaH, DMF (75%);g)

(i) TFA, Et₃SiH, DCM; (ii) satd. aq. Na₂CO₃(100%); h) (i) MsCl,

Et₃N, DCM; (ii) NaN₃, DMF (85%); i) PPh₃, THF, H₂O (97%); j)

(i) Na, NH₃, THF, –78 °C; (ii) Fmoc-Cl, NaHCO₃, dioxane, water

(88%); k) CrO₃, H₂SO₄, acetone (70%).

2902

www.eurjoc.org

Eur. J. Org. Chem. 2008, 2900–2914

Non-Viral Gene Delivery

Figure 2. Peptides containing PEG residues, together with the peptide from the original LID vector.

We reasoned that insertion of a PEG spacer between the ammonium hydrogen carbonate^[48]or DMSO^[49]as solvent

integrin-binding domain and the K₁₆domain might both to speed up the cyclisation was also tried, but residual

stabilise the particle and position the integrin-binding do- buffer and solvent interfered with the subsequent HPLC

main further from the complex, making it more accessible purification, which resulted in low yields of purified pep-

to the cell surface. As a first step to designing targeted lipo- tides.

polyplex vector systems effective for in vivo gene transfer

after intravenous administration, we synthesised variants of

the bifunctional cationic peptide incorporating PEG-amino

acids 1, 2 and 3. Thus, peptides B, C and D (Figure 2) bear-

ing the CRRETAWAC integrin-binding domain, PEG-

amino acids 1, 2 and 3, respectively, and the lysine-rich

DNA-binding domain were designed. In order to compare

the effects of a single, longer PEG-amino acid with two

sequential shorter PEG-amino acids, peptide E, containing

a repeat of 1 (Haa4) was also synthesised.

Transfection of Lipopolyplex Complexes Containing PEG-

Linked Peptides

A series of lipopolyplexes LPD-A–E were formulated

from Lipofectin, plasmid DNA and peptides A–E, respec-

tively, using the previously described procedure.^[4]Addition-

ally, each lipopolyplex was formulated at three different

As the 12-residue C-terminus is common to all four pep- charge ratios: 0.5:1, 1:1 and 4:1.

tides, the head group 23 was first synthesised on a large

Transfections of HAEo-, N2A and AJ3.1 cells were then

scale using the acid-labile NovaSyn-TGT resin preloaded carried out with these lipopolyplexes under different condi-

with Fmoc-Gly using standard methods for automated so- tions. Figure 3 shows the transfection of HAEo- cells with

lid-phase synthesis. A portion of the fully protected pep- LPD-A (the original LID complex) and LPD-C under vari-

tide–resin was then transferred to a Merrifield bubbler. Pre- ous conditions. Although poor transfections were obtained

liminary experiments to determine the best coupling rea- with a 0.5:1 ratio for LPD-C, with ratios of 4:1, LPD-C

gents for the Fmoc-protected PEG-amino acid 2 were then and LPD-A show comparable transfection efficiencies. Fig-

carried out and it was found that the most efficient coupling ure 3 (b) illustrates that this cell line requires Lipofectin for

was effected with TBTU/HOBt/DIPEA. The synthesis of efficient transfection.

peptide B was then completed on the peptide synthesiser

Complexes LPD-A, LPD-C, LPD-D and LPD-E were

using the same reagents to add the remaining lysine resi- used to transfect HAEo- (Figure 4, a), N2a (Figure 4, b)

dues. Peptides C, D and E were synthesised by automated and AJ3.1 (Figure 4, c) cells. The results were extremely de-

solid-phase synthesis from the fully protected peptide–resin pendent on the cell line. The complex with the peptide with

23 using TBTU/HOBt/DIPEA with the Fmoc deprotection the shortest linker, LPD-B, showed good transfection in

time and the recycle time both extended by 10 minutes HAEo- cells but not in the other two cell lines, whereas the

when incorporating amino acids 1 and 3, respectively. After complex with the peptide with the longest linker, LPD-D,

standard resin cleavage and HPLC purification, the pep- showed good transfection in N2A cells but not in the other

tides were cyclised by aerial oxidation at high dilution in two cell lines. LPD-E performed poorly in all three cell

degassed water over a period of a week. The use of 0.01  lines.

Eur. J. Org. Chem. 2008, 2900–2914

www.eurjoc.org

2903

A. B. Tabor et al.

FULL PAPER

Figure 3. (a) Transfection of HAEo- cells with LPD-A and LPD-C at different charge ratios, with the DNA/Lipofectin weight ratio

constant at 1:0.75. (b) Transfection of HAEo- cells with LPD-A and LPD-C at different ratios, with and without Lipofectin. (c) Transfec-

tion of HAEo- cells with LPD-C with increasing amounts of serum and the DNA/Lipofectin weight ratio constant at 1:0.75.

Stability of Lipopolyplex Complex Containing PEG-Linked

Peptide C

In order to determine the size of the complex, dynamic

light scattering (DLS) measurements were taken^[50]and the

As a first step towards quantifiying the stability of com-

hydrodynamic radius or diameter of the particles calcu-

lated.^[51,52]The change in size distribution of the lipopol-

yplex containing peptide C over time is shown in Figure 5.

It can be clearly seen that, compared with the complex for-

mulated with the original peptide A (for which the particle

size increases rapidly, indicating extensive aggregation), the

PEG-spaced complex is much more stable and aggregates

more slowly.

plexes containing PEG-linked peptides, we investigated the

aggregation properties of the most promising lipopolyplex

complex, that containing peptide C. The lipopolyplex LPD-

C(4:1) was formulated from Lipofectin, plasmid DNA and

peptide C, as described previously,^[4]and the size of the re-

sulting particles measured at different concentrations of

phosphate-buffered saline (PBS) to mimic physiological

conditions.

2904

www.eurjoc.org

Eur. J. Org. Chem. 2008, 2900–2914

Non-Viral Gene Delivery

Conclusions

A series of PEG-amino acids, Fmoc-protected for solid-

phase peptide synthesis, have been prepared by a scalable

and high-yielding route from cheap and readily available

tri- and tetraethylene glycol. The strategy used avoided pre-

viously reported problems by first assembling the desired

PEG framework, then converting it into the desired Fmoc-

protected PEG-amino alcohol and finally oxidising it to the

PEG-amino acid in the last step. These amino acids were

incorporated into bifunctional peptides bearing integrin-

binding and DNA-condensing moieties using standard so-

lid-phase synthesis techniques. Although a range of PEG-

spaced amino acids have recently become commercially

available, none of them exactly corresponds to the amino

acids we have synthesised. We believe that our approach is

a valuable synthetic route to PEG-spaced amino acids,

which can be adapted to give a variety of linker lengths.

A series of lipopolyplex complexes was then formulated

using these PEG-containing bifunctional peptides. The

complex LPD-C, formulated with a peptide containing

Haa7, showed promising transfection results in preliminary

experiments. DLS measurements (Figure 5) showed that

this lipopolyplex had enhanced stability in PBS compared

with a lipopolyplex without a PEG spacer, LPD-A. How-

ever, this enhanced complex stability does not appear to

confer better transfection properties in serum as transfec-

tion is greatly reduced at even low serum concentrations

(Figure 3, c). Complexes LPD-B, formulated with a peptide

containing the shorter linker, Haa4, and LPD-D, formu-

lated with a peptide containing the longest linker, Haa9,

showed variable results, depending on the cell line investi-

gated. In HAEo- cells the complex with the shorter linker,

LPD-B, showed better transfection than the original vector,

LPD-A, whereas in N2A cells the complex with the longer

linker, LPD-D, showed reasonable transfection, although

not as good as the original vector, LPD-A. In all the cell

lines, the complex with two Haa4 amino acids, LPD-E, per-

formed poorly.

The reasons for the divergent transfection properties of

what appear to be rather similar PEG-containing complexes

are not at present clear. In particular, our results indicate

that the length of the PEG spacer is not the most important

factor, as complexes LPD-E and LPD-C, which have almost

the same length spacer, are so different in transfection effi-

ciency. Clearly in the case of LPD-E, the additional amide

bond in the centre of the linker must impart additional

structural rigidity to the linker, which may be unfavourable

for the formation of a stable complex or for its dissociation

within the endosome during transfection. A long linker be-

tween the integrin-binding and DNA-condensation por-

tions of the peptide does not, however, appear to be neces-

sary for successful transfection. We have recently carried

out a detailed biophysical study of the macromolecular

structure of the original LID vector, and have demonstrated

that the integrin-binding peptide is partly buried in a disor-

dered lipid layer.^[53]In light of these recent results, therefore,

it seems unlikely that simply increasing the length of the

Figure 4. (a) Transfection of HAEo- cells with LPD-A, LPD-B,

LPD-D and LPD-E at charge ratios of 1:1 and 4:1. (b) Transfection

of N2A cells with LPD-A, LPD-B, LPD-D and LPD-E at charge

ratios of 1:1 and 4:1. (c) Transfection of AJ3.1 cells with LPD-A,

LPD-B, LPD-D and LPD-E at charge ratios of 1:1 and 4:1. Lipo-

fectin was present throughout at 0.75.

Figure 5. Average aggregate sizes (y axis, nm) in PBS (pH 7.4) as a

function of time. Data refer to LID complexes containing peptide

A (LPD-A) (᭿) and peptide C (LPD-C) (ϫ), respectively.

Eur. J. Org. Chem. 2008, 2900–2914

www.eurjoc.org

2905

A. B. Tabor et al.

FULL PAPER

aq. NaHCO₃(2ϫ200 mL). The organic phase was then partitioned

with satd. aq. NaCl (200 mL), dried with anhydrous sodium sulfate

and concentrated in vacuo to give the mesylate as a brown oil. This

oil was dried by azeotropic distillation several times with toluene.

The brown oil was dissolved in dimethylformamide (150 mL) and

sodium azide (13.0 g, 200 mmol) was added. The reaction was left

to stir for 4 d after which time it was concentrated in vacuo and

then dissolved in dichloromethane (250 mL). This organic solution

was partitioned with water (2ϫ150 mL) followed by satd. aq. NaCl

(200 mL), dried with anhydrous sodium sulfate and concentrated

linker between the DNA-condensation moiety and the inte-

grin-binding sequence would have much impact. A more

subtle interplay between the different linker structures, the

lipid layer and the lipid composition of the different cell

lines is clearly operating and will affect the overall stability

of the particle and the ease of dissociation of the complex

in the endosome. Our preliminary biophysical results indi-

cate that the presence of the PEG spacer in LPD-C does

impart some degree of additional stability to the complex

when suspended in buffer solutions. Further detailed stud- in vacuo to yield a dark-brown oil. The product was isolated from

this oil by normal-phase silica gel chromatography, using a gradient

from chloroform/hexane (1:1) to chloroform only, to yield 1-azido-

11-(triphenylmethyloxy)-3,6,9-trioxaundecane (21.0 g, 91%) as a

ies of the transfection properties of the PEG-spacer pep-

tides in serum, alone and in combination with PEG-lipids

of defined length, are underway and will be reported sub-

sequently.

1

yellow oil. H NMR (300 MHz, CDCl₃): δ = 3.25 [t, J = 5.1 Hz, 2

H, (C₆H₅)₃COCH₂], 3.34 (t, J = 4.8 Hz, 2 H, CH₂N₃) 3.63–3.69

[m, 12 H, (CH₂OCH₂)₃CH₂N₃], 7.20–7.48 [m, 15 H, (C₆H₅)₃] ppm.

¹³C NMR (75 MHz, CDCl₃): δ = 51.1, 63.8, 70.4, 71.1, 71.16,

Experimental Section

71.22, 87.0, 127.3, 128.2, 129.2, 144.6 ppm. IR (oil): ν = 3059 (aryl-

˜

H), 2870 (C–H), 2108 (N₃), 1597 (C=C) cm^–1. MS (positive ion

FAB): m/z (%) = 484 (98) [M + Na]⁺, 243 (63) [(C₆H₅)₃C]⁺. HRMS

(positive ion FAB): calcd. for [M + Na]⁺484.2212; found 484.2203.

General: Solid-phase peptide synthesis was carried out either man-

ually on a Merrifield bubbler or automatically on a peptide syn-

thesiser module (MilliGen 9050Plus PepSynthesiser). All reagents

used for either manual or automatic synthesis were purchased from

commercial suppliers with the exception of the Fmoc-protected

amino acids 1, 2 and 3. HIPERSOLV^©DMF of HPLC grade was

used straight from the bottle, whereas dichloromethane and piperi-

dine were freshly distilled from calcium hydride before use. All pep-

tide syntheses were carried out under nitrogen. Preparative HPLC

was carried out on a Waters 600E instrument with a Waters 486

UV detector using a Vydac 218TP reversed-phase HPLC column

(25ϫ250 mm). Analytical HPLC was carried out on a Varian

ProStar 210 with a Varian ProStar 320 UV detector using a Vydac

218TP reversed-phase HPLC column (2.1ϫ250 mm). Water/aceto-

nitrile gradients were used as described, with both water and aceto-

nitrile containing 0.1% trifluoroacetic acid. Detection was per-

formed at 215 nm.

Triphenylphosphane (14.33 g, 54.6 mmol) was added to 1-azido-11-

(triphenylmethyloxy)-3,6,9-trioxaundecane (21.0 g, 45.5 mmol) in

tetrahydrofuran (100 mL). After 2 h, water (3 mL) was added to

the reaction. The reaction was then left to stir under argon at room

temperature for 24 h after which time it was concentrated in vacuo.

Diethyl ether (65 mL) was added to the remaining oil/solid and the

reaction mixture was then filtered and concentrated in vacuo. The

product was isolated by normal-phase silica gel chromatography

[chloroform, then changing to chloroform/methanol/triethylamine

(85:10:5)] to yield 5 (19.0 g, 96%) as a pale-yellow oil. ¹H NMR

(300 MHz, CDCl₃): δ = 2.83 (t, J = 5.2 Hz, 2 H, CH₂NH₂) 3.25

[t, J = 5.2 Hz, 2 H, (C₆H₅)₃COCH₂], 3.50 (t, J = 5.2 Hz, 2 H,

CH₂CH₂NH₂), 3.61–3.69 [m, 10 H, (CH₂OCH₂)₂CH₂OCH₂-

CH₂NH₂], 7.19–7.49 [m, 15 H, (C₆H₅)₃COCH₂] ppm. ¹³C NMR

(75 MHz, CDCl₃): δ = 42.2, 63.8, 70.8, 71.09, 71.11, 71.14, 71.2,

11-(Triphenylmethyloxy)-3,6,9-trioxaundecan-1-ol (4): Triphenyl-

methyl chloride (27.9 g, 100 mmol) in dry dichloromethane

(500 mL) was added dropwise to tetraethylene glycol (214 g, 1.10

mol) and triethylamine (20.2 g, 27.8 mL, 200 mmol) in dry dichlo-

romethane (1500 mL). This solution was stirred under argon at

room temperature for 24 h. The reaction was then concentrated in

vacuo to give a mixture of product and starting material as a yellow

oil. This oil was then redissolved in dichloromethane (1000 mL),

partitioned with satd. aq. NaHCO₃(500 mL), water (3ϫ400 mL)

and finally satd. aq. NaCl (500 mL). The organic layer was sepa-

rated, dried with anhydrous sodium sulfate and concentrated in

vacuo to give 4 as a yellow oil (42.2 g, 95%) which was used with-

out further purification. ¹H NMR (300 MHz, CD₃CN): δ = 2.75

(br. s, 1 H, CH₂OH), 3.16 [t, J = 5.1 Hz, 2 H, (C₆H₅)₃COCH₂],

3.47–3.63 [m, 14 H, (CH₂OCH₂)₃CH₂OH], 7.25–7.49 [m, 15 H,

(C₆H₅)₃COCH₂] ppm. ¹³C NMR (75 MHz, CD₃CN): δ = 60.8,

63.2, 69.87, 69.91, 70.02, 70.04, 70.2, 72.1, 86.1, 126.8, 127.5, 128.3,

73.7, 87.0, 127.3, 128.2, 129.2, 144.6 ppm. IR (oil) ν = 3380 (N–

˜

H), 3060 (aryl-H), 2870 (C–H), 1600 (C=C Ar) cm^–1. MS (positive

ion FAB): m/z (%) = 436 (14) [M + H]⁺, 243 (99) [(C₆H₅)₃C]⁺.

HRMS (positive ion FAB): calcd. for [M + H]⁺436.2487; found

436.2483.

1-(Fluoren-9-ylmethyloxycarbonylamino)-11-(triphenylmethyloxy)-

3,6,9-trioxaundecane (6): Fluoren-9-ylmethyl chloroformate (9.03 g,

35 mmol) in dry dichloromethane (40 mL) was added dropwise to

1-amino-11-(triphenylmethyloxy)-3,6,9-trioxaundecane (5) (8.70 g,

20 mmol) and N-methylmorpholine (4.4 mL, 4.05 g, 40 mmol) in

dry dichloromethane (100 mL) under argon at 0 °C. Once addition

was complete, the reaction was stirred at room temperature for

24 h. Dichloromethane (100 mL) was added and the resulting solu-

tion partitioned with citric acid solution (pH 6, 150 mL). The or-

ganic phase was then partitioned with satd. aq. NaCl (200 mL),

dried with anhydrous sodium sulfate and concentrated in vacuo.

The product was isolated by normal-phase silica gel chromatog-

raphy, eluting with chloroform only, to yield 6 as a viscous yellow

144.0 ppm. IR (oil): ν = 3440 (O–H), 3057 (aryl-H), 2872 (C–H),

˜

1597 (C=C) cm^–1.

1

1-Amino-11-(triphenylmethyloxy)-3,6,9-trioxaundecane (5): Meth-

anesulfonyl chloride (7.74 mL, 11.45 g, 100 mmol) in dry dichloro-

methane (50 mL) was added dropwise to 11-(triphenylmethyloxy)-

3,6,9-trioxaundecan-1-ol (4) (21.82 g, 50 mmol) and triethylamine

(15.33 mL, 11.13 g, 110 mmol) in dry dichloromethane (150 mL)

under argon at 0 °C. Once addition was complete, the reaction was

oil (12.93 g, 95%). H NMR (300 MHz, CDCl₃): δ = 3.25 [t, J =

5.2 Hz, 2 H, (C₆H₅)₃COCH₂], 3.36 (q, J = 4.5 Hz, 2 H,

CH₂NHFmoc), 3.55–3.68 [m, 12 H, (CH₂OCH₂)₃CH₂NHFmoc],

4.21 (t, J = 6.7 Hz, 1 H, NHCOOCH₂CH), 4.40 (d, J = 6.7 Hz, 2

H, NHCOOCH₂CH), 5.40 (br. s, 1 H, NH), 7.22–7.49 (m, 15 H,

aryl Fmoc and trityl), 7.59 (d, J = 7.40 Hz, 2 H, aryl Fmoc), 7.77

stirred for 3 h at room temperature. Dichloromethane (200 mL) (d, J = 7.40 Hz, 2 H, aryl Fmoc) ppm. ¹³C NMR (75 MHz,

was then added and the organic solution was partitioned with satd. CDCl₃): δ = 41.4, 47.7, 63.8, 67.0, 70.5, 70.8, 71.07, 71.12, 71.2,

2906

www.eurjoc.org

Eur. J. Org. Chem. 2008, 2900–2914

Non-Viral Gene Delivery

87.0, 120.3, 125.5, 127.4, 128.1, 127.4, 128.2, 129.2, 141.7, 144.5,

bend) cm^–1. MS (positive ES): m/z (%) = 452 (99) [M + Na]⁺.

HRMS (positive ESI): calcd. for [M + Na]⁺452.16797; found

452.16899.

144.6, 156.9 ppm. IR (oil) ν = 3350 (N–H), 3060 (aryl-H), 2860

˜

(C–H), 1700 (C=O), 1570–1460 (N–H bend and C=C) cm^–1. MS

(positive ESI): m/z (%): 697 (19) [M + K]⁺, 681 (99) [M +

Na]⁺. HRMS (positive ESI): calcd. for [M + Na]⁺680.29826; found

680.29881.

8-(Triphenylmethyloxy)-3,6-dioxaoctan-1-ol (8):^[54]Triphenylmethyl

chloride (6.07 g, 30 mmol) in dry dichloromethane (150 mL) was

added dropwise to triethylene glycol (90.1 g, 600 mmol) and trieth-

ylamine (6.07 g, 8.36 mL, 60 mmol) in dry dichloromethane

(500 mL). This solution was stirred under nitrogen at room tem-

perature for 24 h. The reaction was then concentrated in vacuo to

give a mixture of product and starting material as a yellow oil. This

oil was redissolved in dichloromethane (400 mL), partitioned with

satd. aq. NaHCO₃(200 mL), water (4ϫ250 mL) and finally satd.

aq. NaCl (250 mL). The organic layer was separated, dried with

anhydrous sodium sulfate and concentrated in vacuo to give a yel-

low oil. This oil was purified by normal-phase silica gel chromatog-

raphy, eluting with a gradient from chloroform (100%) to chloro-

form/methanol (98:2), to yield 8 as a pale-yellow oil (12.0 g, 100%)

spectroscopically identical to the literature.^{[54] 1}H NMR (300 MHz,

CDCl₃): δ = 2.35 (br. s, 1 H, CH₂OH), 3.25 [t, J = 5.1 Hz, 2 H,

(C₆H₅)₃COCH₂], 3.58–3.71 [m, 10 H, (CH₂OCH₂)₂CH₂OH], 7.18–

7.47 [m, 15 H, (C₆H₅)₃COCH₂] ppm. ¹³C NMR (75 MHz, CDCl₃):

δ = 62.2, 63.7, 71.0, 71.1, 71.2, 73.0, 87.1, 127.4, 128.2, 129.1,

11-(Fluoren-9-ylmethyloxycarbonylamino)-3,6,9-trioxaundecan-1-ol

(7): 1-(Fluoren-9-ylmethyloxycarbonylamino)-11-triphenylmeth-

yloxy-3,6,9-trioxaundecane (6) (12.93 g, 19 mmol) was dissolved in

a solution of dichloromethane, trifluoroacetic acid and triethylsi-

lane (100 mL, 8:1:1). The reaction was stirred at room temperature

under argon for 1 h. Water (100 mL) was then added to the reac-

tion and the resulting mixture was adjusted to pH 7 with satd. aq.

NaHCO₃whilst being vigorously stirred. The two layers were sepa-

rated and the aqueous solution was partitioned with another vol-

ume of dichloromethane (100 mL). The organic fractions were

combined, dried with anhydrous sodium sulfate and concentrated

in vacuo. The product was isolated by normal-phase silica gel

chromatography, using a gradient from chloroform (100%) to ethyl

1

acetate (100%), to yield 7 as a yellow oil (7.56 g, 94%). H NMR

(300 MHz, CDCl₃): δ = 3.40 (br. s, 2 H, CH₂NHFmoc), 3.57–3.74

[m, 14 H, HOCH₂(CH₂OCH₂)₃], 4.21 (t, J = 6.6 Hz, 1 H,

NHCOOCH₂CH), 4.42 (d, J = 6.6 Hz, 2 H, NHCOOCH₂CH),

4.76, 6.06 (br. s, 2 H, NH and OH), 7.30 (t, J = 7.4 Hz, 2 H), 7.38

(t, J = 7.4 Hz, 2 H), 7.61 (d, J = 7.4 Hz, 2 H, aryl Fmoc), 7.76 (d,

J = 7.4 Hz, 2 H, aryl Fmoc) ppm. ¹³C NMR (75 MHz, CDCl₃): δ

= 41.4, 47.8, 61.7, 67.1, 70.1, 70.4, 71.5, 71.3, 72.5, 120.2, 125.4,

144.5 ppm. IR (oil): ν = 3440 (O–H), 3057 (aryl-H), 2872 (C–H),

˜

1597 (C=C) cm^–1. MS (positive ion FAB): m/z (%) = 415 (25) [M

+ Na]⁺. HRMS (positive ion FAB): calcd. for [M + Na]⁺415.1885;

found 415.1876.

8-(Triphenylmethyloxy)-3,6-dioxaoctyl Methanesulfonate (9):^[55]

Methanesulfonyl chloride (5.80 mL, 8.59 g, 75 mmol) in dry dichlo-

romethane (40 mL) was slowly added to 8-(triphenylmethyloxy)-

3,6-dioxaoctan-1-ol (8) (12.0 g, 30 mmol) and triethylamine (9.1 g,

12.5 mL, 90 mmol) in dry dichloromethane (100 mL) under nitro-

gen at 0 °C. Once addition was complete, the reaction was left to

warm to room temperature. After 2.5 h, excess satd. aq. NaHCO₃

was added and the reaction was stirred for 10 min. Additional

dichloromethane (200 mL) was added, the aqueous and organic

layers were separated and the organic solution was partitioned with

satd. aq. NaCl (100 mL). The aqueous solution was back-extracted

with dichloromethane (150 mL), the organic solutions were com-

bined, dried with anhydrous sodium sulfate and concentrated in

vacuo to give 9 as an orange oil. The oil was then azeotroped sev-

eral times with dry toluene. This oil became a red solid (14.1 g,

99.8%) after several hours in high vacuo. The product, identical by

NMR to the literature,^[55]was used without further purification.

¹H NMR (300 MHz, CDCl₃): δ = 2.94 (s, 3 H, CH₃SO₃CH₂), 3.25

[t, J = 5.1 Hz, 2 H, CH₂OC(C₆H₅)₃], 3.64–3.76 [m, 6 H,

CH₃SO₃CH₂CH₂O(CH₂)₂OCH₂], 3.78 (m, 2 H, CH₃SO₃CH_2-

CH₂O), 4.35 (m, 2 H, CH₃SO₃CH₂CH₂O), 7.20–7.48 [m, 15 H,

OC(C₆H₅)₃] ppm. ¹³C NMR (75 MHz, CDCl₃): δ = 38.0, 63.8,

69.5, 69.7, 71.15, 71.2, 87.0, 127.4, 128.2, 129.1, 144.5 ppm.

127.4, 127.9, 141.6, 144.3, 157.6 ppm. IR (oil) ν = 3324 (N–H and

˜

O–H), 3067, 3041, 3020 (aryl-H), 2882 (C–H), 1672 (C=O), 1535

(N–H) cm^–1. MS (positive ion FAB): m/z (%): 454 (92) [M + K]⁺,

438 (99) [M + Na]⁺, 179 (34) [C₁₄H₁₁]⁺. HRMS (positive ion FAB):

calcd. for [M + Na]⁺438.1893; found 438.1886.

11-(Fluoren-9-ylmethyloxycarbonylamino)-3,6,9-trioxaundecanoic

Acid (Fmoc-Haa4, 1):^[29]Chromium trioxide (3.00 g, 30 mmol) in

sulfuric acid (1.5 , 60.0 mL, 90 mmol) was very slowly added to

vigorously stirring 11-(fluoren-9-ylmethyloxycarbonylamino)-3,6,9-

trioxaundecan-1-ol (7) (4.15 g, 10 mmol) in acetone (100 mL) at

0 °C. Once addition was complete the reaction was left to warm to

room temperature and to stir for 24 h. Water (300 mL) was then

added to the reaction along with reversed-phase silica gel (100 g).

The reaction mixture was concentrated in vacuo until the volume

had decreased by approximately one quarter and more water

(100 mL) was added. The reaction mixture was once again concen-

trated in vacuo until no acetone could be detected in the mixture.

The mixture was then filtered to recover the reversed-phase silica

gel, which was washed with copious amounts of water until the

silica gel was almost colourless. The reversed-phase silica gel was

then washed with acetonitrile (4 ϫ 200 mL) and chloroform

(3ϫ150 mL). The organic fractions were combined and concen-

trated in vacuo to give a pale-green oil. The product was isolated

by normal-phase silica gel chromatography, eluting with a gradient

from chloroform (100%) to chloroform/methanol (95:5) and then

to chloroform/methanol/acetic acid (80:5:5), to yield 1 as a yellow

8-(Phenylmethyloxy)-3,6-dioxaoctan-1-ol (10):^[56,57]Sodium hydride

(1.20 g, 60% w/w in mineral oil, 30 mmol) was added cautiously

to triethylene glycol (30.0 g, 200 mmol) and tetrabutylammonium

bromide (0.5 g, cat.) in dry tetrahydrofuran (150 mL). After stirring

oil (2.93 g, 68%) spectroscopically identical to the literature.^{[29] 1}

H

NMR (300 MHz, CDCl₃): δ = 3.39 (m, 2 H, CH₂NHFmoc), 3.54– for 20 min, benzyl chloride (2.53 g, 20 mmol) was added and the

3.71 [m, 10 H, OCH₂(CH₂OCH₂)₂CH₂NHFmoc], 4.12 (s, 2 H, reaction was heated at reflux under nitrogen for 24 h. The reaction

HO₂CCH₂), 4.21 (t, J = 6.7 Hz, 1 H, NHCOOCH₂CH), 4.39 (d, J was then cooled to room temperature and concentrated in vacuo

= 6.7 Hz, 2 H, NHCOOCH₂CH), 5.70 (br. s, 1 H, NH), 7.30 (t, J

= 7.4 Hz, 2 H), 7.38 (t, J = 7.4 Hz, 2 H), 7.59 (d, J = 7.4 Hz, 2 H),

to give an oil. This oil was dissolved in dichloromethane (300 mL)

and partitioned with satd. aq. NaCl (2 ϫ200 mL). The organic

7.74 (d, J = 7.4 Hz, 2 H) ppm. ¹³C NMR (75 MHz, CDCl₃): δ = phase was dried with anhydrous sodium sulfate and concentrated

41.3, 47.7, 67.1, 69.1, 70.4, 70.6, 70.8, 71.5, 120.3, 125.5, 127.5,

in vacuo to give a yellow oil. This was purified by normal-phase

128.1, 141.7, 144.4, 157.3, 173.1 ppm. IR (oil) ν = 3336 (COO–H), silica gel chromatography, eluting with a gradient from chloroform

˜

3060 (aryl-H), 2880 (C–H), 1715, 1700 (C=O), 1611 (N–H

(100%) to chloroform/methanol (90:10), to yield 10 as a pale-yel-

2907

Eur. J. Org. Chem. 2008, 2900–2914

www.eurjoc.org

A. B. Tabor et al.

FULL PAPER

low oil (4.94 g, 99 %) spectroscopically identical to the litera-

(4.00 g, 60% w/w in mineral oil, 100 mmol) was added to 17-(phen-

ture.^{[56,57] 1}H NMR (300 MHz, CDCl₃): δ = 2.74 (br. t, 1 H, ylmethyloxy)-3,6,9,12,15-pentaoxaheptadecan-1-ol (12) (16.38 g,

CH₂OH), 3.60–3.70 [m, 12 H, OCH₂(CH₂OCH₂)₂CH₂OH], 4.57 (s,

43.9 mmol) in dry dimethylformamide (250 mL). The reaction was

2 H, C₆H₅CH₂O), 7.25–7.35 (m, 5 H, C₆H₅CH₂O) ppm. ¹³C NMR left to stir for 4 d at room temperature after which time it was

(75 MHz, CDCl₃): δ = 62.1, 69.8, 70.8, 71.0, 71.1, 72.9, 73.7, 128.0,

concentrated in vacuo. The reaction mixture was redissolved in di-

ethyl ether (400 mL) and cooled in an ice bath to 0 °C. Water was

then slowly added until hydrogen was no longer evolved. The or-

128.1, 128.7, 138.1 ppm. IR (oil): ν = 3449 (O–H), 3030 (aryl-H),

˜

2868 (C–H) cm^–1. MS (positive ion FAB): m/z (%) = 263 (4) [M +

Na]⁺, 241 (99) [M + H]⁺, 91 (100) [C₇H₇]⁺. HRMS (positive ion ganic phase was then partitioned twice with aq. satd. NaCl with a

FAB): calcd. for [M + H]⁺241.1440; found 241.1442.

little NaHCO₃(2ϫ300 mL), dried with anhydrous sodium sulfate

and concentrated in vacuo to give a viscous dark-brown oil. The

product was isolated by normal-phase silica gel chromatography,

eluting with ethyl acetate only to yield 14 as a yellow oil (24.2 g,

75.0%). ¹H NMR (300 MHz, CDCl₃): δ = 3.24 [t, J = 5.2 Hz, 2 H,

CH₂OC(C₆H₅)₃], 3.57–3.69 [m, 34 H, C₆H₅CH₂OCH₂(CH₂-

OCH₂)₈], 4.56 (s, 2 H, C₆H₅CH₂OCH₂), 7.21–7.48 [m, 20 H,

C₆H₅CH₂OCH₂(CH₂OCH₂)₈CH₂OC(C₆H₅)₃] ppm. ¹³C NMR

(75 MHz, CDCl₃): δ = 63.4, 69.5, 70.6, 70.7, 70.8, 73.2, 86.6, 126.9,

127.6, 127.7, 128.4, 128.7, 138.4, 144.2 (two signals superim-

1-(Triphenylmethyloxy)-17-(phenylmethyloxy)-3,6,9,12,15-penta-

oxaheptadecane (11): Sodium hydride (2.4 g, 60% w/w in mineral

oil, 60 mmol) was added to 8-(phenylmethyloxy)-3,6-dioxaoctan-1-

ol (10) (4.94 g, 20 mmol) and 8-(triphenylmethyloxy)-3,6-dioxaoc-

tyl methanesulfonate (9) (9.88 g, 21 mmol) in dry dimethylform-

amide (50 mL) under argon. This solution was stirred for 5 d at

room temperature. The reaction was then concentrated in vacuo

and then redissolved in diethyl ether (200 mL) and water (200 mL).

The organic phase was then partitioned with aq. satd. NaCl with

a little NaHCO₃(200 mL), dried with anhydrous sodium sulfate

and concentrated in vacuo to give a viscous dark-brown oil. This

oil was purified by normal-phase silica gel chromatography, eluting

with a gradient from chloroform (100%) to chloroform/methanol

(98:2), to yield 11 as an orange oil (9.11 g, 74.1 %). ¹H NMR

(300 MHz, CDCl₃): δ = 3.25 [t, J = 5.2 Hz, 2 H, CH₂OC(C₆H₅)₃],

3.64–3.69 [m, 22 H, OCH₂(CH₂OCH₂)₅], 4.57 (s, 2 H,

C₆H₅CH₂OCH₂),7.20–7.52[m, 20H,C₆H₅CH₂OCH₂(CH₂OCH₂)₅-

CH₂OC(C₆H₅)₃] ppm. ¹³C NMR (75 MHz, CDCl₃): δ = 63.8, 69.9,

71.03, 71.07, 71.10, 71.2, 73.7, 87.0, 127.3, 128.0, 128.1, 128.2,

posed) ppm. IR (oil): ν = 3058, 3031 (aryl-H), 2868 (C–H), 1962

˜

(C=C), 1596 (C=C stretch) cm^–1. MS (positive ion FAB): m/z (%)

= 769 (75) [M + Na]⁺, 243 (39) [(C₆H₅)₃C]⁺. HRMS (positive ion

FAB): calcd. for [M + Na]⁺769.39222; found 769.39298.

26-(Phenylmethyloxy)-3,6,9,12,15,18,21,24-octaoxahexaeicosan-1-ol

(15):^[58]A solution of trifluoroacetic acid (20 mL) and triethylsilane

(20 mL) in dichloromethane (160 mL) was added to 1-(triphenyl-

methyloxy)-26-(phenylmethyloxy)-3,6,9,12,15,18,21,24-octaoxa-

hexaeicosane (14) (24.0 g, 32.1 mmol). The reaction was left to stir

at room temperature for 1 h, after which satd. aq. Na₂CO₃was

slowly added to the reaction with vigorous stirring until the reac-

tion reached pH 12. The two layers were partitioned and the aque-

ous fraction was back-extracted several times with dichlorometh-

ane (9ϫ100 mL). The organic fractions were combined, dried with

anhydrous sodium sulfate and concentrated in vacuo to give a yel-

low solid/oil. This oil was redissolved in water (200 mL) and the

aqueous solution was partitioned twice with hexane (2ϫ200 mL).

The aqueous solution was then concentrated in vacuo to give a

yellow oil. The resulting opaque oil was dried by azeotropic distil-

lation with anhydrous acetonitrile until 15 was obtained as a

colourless oil (16.2 g, 99.9%), spectroscopically identical to the lit-

erature data.^{[58] 1}H NMR (300 MHz, CDCl₃): δ = 3.57–3.73 [m, 36

H, C₆H₅CH₂OCH₂(CH₂OCH₂)₈CH₂OH], 4.55 (s, 2 H,

C₆H₅CH₂O), 7.21–7.33 (m, 5 H, C₆H₅CH₂O) ppm. ¹³C NMR

(75 MHz, CDCl₃): δ = 61.6, 69.5, 70.2, 70.4, 70.5, 70.55, 70.62,

72.4, 73.2, 127.6, 127.7, 128.3, 138.3 (two signals superim-

128.7, 129.2, 138.8, 144.6 ppm. IR (oil): ν = 3030 (aryl-H), 2869

˜

(C–H), 1600, 1570 (C=C) cm^–1. MS (positive ion FAB): m/z (%) =

637 (0.6) [M + Na]⁺, 613 (1) [C₃₈H₄₅O₇]⁺, 243 (100) [C₁₉H₁₅]⁺.

HRMS (positive ion FAB, NOBA matrix): calcd. for [M + Na]⁺

637.3141; found 637.3148.

17-(Phenylmethyloxy)-3,6,9,12,15-pentaoxaheptadecan-1-ol

(12):^[57,58]A solution made up of dichloromethane, triethylsilane

and trifluoroacetic acid (100 mL, 80:10:10) was added to 17-(phen-

ylmethyloxy)-1-(triphenylmethyloxy)-3,6,9,12,15-pentaoxaheptade-

cane (11) (9.11 g, 14.82 mmol). The reaction was stirred at room

temperature for 1 h. Satd. aq. Na₂CO₃was then slowly added to

the reaction with vigorous stirring until the reaction reached pH 12.

The two layers were partitioned and the aqueous fraction was back-

extracted several times with dichloromethane (6ϫ100 mL). The or-

ganic fractions were combined, dried with anhydrous sodium sul-

fate and concentrated in vacuo to give a yellow solid/oil. This oil

was redissolved in water (100 mL) and the aqueous solution was

partitioned twice with hexane (2ϫ100 mL). The aqueous solution

was then concentrated in vacuo to give a yellow oil. The resulting

oil was dried by azeotropic distillation with anhydrous acetonitrile

until 12 was recovered (5.51g, 99%) as an oil, spectroscopically

identical to the data given in the literature,^[57,58]which required no

posed) ppm. IR (oil): ν = 3333 (O–H), 3030 (aryl-H), 2869 (C–H),

˜

1455 (C=C) cm^–1. MS (positive ESI): m/z (%) = 543 (3) [M + K]⁺,

527 (99) [M + Na]⁺. HRMS (positive ESI): calcd. for [M + Na]⁺

527.2826; found 527.2850.

1-Azido-26-(phenylmethyloxy)-3,6,9,12,15,18,21,24-octaoxa-

hexaeicosane (16): Methanesulfonyl chloride (7.43 mL, 11.0 g,

96 mmol) in dry dichloromethane (75 mL) was slowly added to 26-

(phenylmethyloxy)-3,6,9,12,15,18,21,24-octaoxahexaeicosan-1-ol

(15) (16.2 g, 32.1 mmol) and triethylamine (11.3 g, 15.60 mL,

112 mmol) in dry dichloromethane (200 mL) under nitrogen at

0 °C. Once addition was complete, the reaction was left to warm

to room temperature. After 4 h, excess satd. aq. NaHCO₃was

added and the reaction was stirred for 10 min. Additional dichloro-

methane (300 mL) was added, the aqueous and organic layers were

separated and the organic solution was partitioned with satd. aq.

NaCl (250 mL). The aqueous solution was back-extracted with

dichloromethane (2ϫ250 mL), the organic solutions were com-

1

further purification. H NMR (300 MHz, CD₂Cl₂): δ = 3.62–3.79

[m, 22 H, (CH₂OCH₂)₅CH₂OH], 4.60 (s, 2 H, C₆H₅CH₂O), 7.29–

7.38 (m, 5 H, C₆H₅CH₂O) ppm. ¹³C NMR (75 MHz, CD₂Cl₂): δ

= 60.9, 69.1, 69.79, 69.82, 69.9, 70.10, 70.14, 70.2, 70.5, 72.1, 73.9,

127.6, 127.7, 128.3, 138.1 (two signals superimposed) ppm. IR (oil):

ν = 3474 (O–H), 3030 (aryl-H), 2868 (C–H) cm^–1. MS (positive ion

˜

FAB): m/z (%) = 395 (5) [M + Na]⁺, 373 (16) [C₁₉H₃₃O₇]⁺. HRMS

(positive ion FAB, NOBA matrix): calcd. for [M + Na]⁺395.2046;

found 395.2051.

1-(Triphenylmethyloxy)-26-(phenylmethyloxy)-3,6,9,12,15,18,21,24-

octaoxahexaeicosane (14):^[43b]8-(Triphenylmethyloxy)-3,6-dioxaoc- bined, dried with anhydrous sodium sulfate and concentrated in

tyl methanesulfonate (9) (28.23 g, 60 mmol) and sodium hydride

vacuo to give an orange oil. The oil was then azeotroped several

2908

www.eurjoc.org

Eur. J. Org. Chem. 2008, 2900–2914

Non-Viral Gene Delivery

times with dry toluene. This oil became a red solid after several

hours under high vacuo. This red solid was redissolved in dry di-

solution, cooled to 0 °C. The reaction was then warmed to room

temperature and left to stir for 6 h. After 6 h the solution was acidi-

methylformamide (200 mL) and sodium azide (10.4 g, 160 mmol) fied to pH 6 by addition of solid citric acid and then concentrated

was added. The reaction was stirred for 72 h at room temperature

under argon, after which the reaction was concentrated in vacuo.

Xylene (250 mL) was then added and the reaction was once again

concentrated in vacuo. The reaction mixture was triturated with

diethyl ether (750 mL) and filtered, the filtrate being concentrated

in vacuo to give an orange oil. The product was isolated by normal-

phase silica gel chromatography, eluting with a gradient from ethyl

acetate (100%) to ethyl acetate/methanol (95:5), to yield 16 as a

yellow oil (14.4 g, 85%). ¹H NMR (300 MHz, CDCl₃): δ = 3.36 (t,

in vacuo to remove dioxane. The remaining aqueous solution was

partitioned three times with chloroform (3ϫ250 mL). The organic

fractions were combined, dried with anhydrous sodium sulfate and

concentrated in vacuo to give a yellow oil. This oil was purified by

normal-phase silica gel chromatography, eluting with ethyl acetate/

1

methanol (4:1), to yield 18 as a yellow oil (6.10 g, 88%). H NMR

(300 MHz, CDCl₃): δ = 2.71 (br. s, 1 H, OH), 3.39 (m, 2 H,

CH₂NHFmoc), 3.55–3.73 [m, 34 H, HOCH₂(CH₂OCH₂)₈], 4.22 (t,

J = 6.8 Hz, 1 H, NHCOOCH₂CHC₁₂H₈), 4.41 (d, J = 6.8 Hz, 2

J = 5.1 Hz, 2 H, CH₂N₃), 3.58–3.70 [m, 34 H, CH₂(CH₂OCH₂)₈- H, CH₂NHCOCH₂), 5.50 (br. s, 1 H, NH), 7.31 (t, J = 7.4 Hz, 2

CH₂N₃], 4.55 (s, 2 H, C₆H₅CH₂OCH₂), 7.21–7.33 (m, 5 H, H, aryl Fmoc), 7.39 (t, J = 7.4 Hz, 2 H, aryl Fmoc), 7.61 (d, J =

C₆H₅CH₂O) ppm. ¹³C NMR (75 MHz, CDCl₃): δ = 50.7, 69.5, 7.4 Hz, 2 H, aryl Fmoc), 7.76 (d, J = 7.4 Hz, 2 H, aryl Fmoc) ppm.

70.0, 70.6, 70.7, 73.2, 127.5, 127.7, 128.3, 138.3 (signals superim-

¹³C NMR (75 MHz, CDCl₃): δ = 41.0, 47.3, 61.7, 66.6, 70.1, 70.4,

posed) ppm. IR (oil): ν = 3020 (aryl-H), 2866 (C–H), 2106 (N ),

70.6, 72.6, 119.9, 125.1, 127.0, 127.7, 141.3, 144.0, 156.6 ppm. IR

˜

3

1454 (C=C) cm^–1. MS (positive ES): m/z (%) = 568 (2) [M + K]⁺,

552 (99) [M + Na]⁺, 524 (3) [C₂₅H₄₃O₉N + Na]⁺. HRMS (positive

ion FAB): calcd. for [M + Na]⁺552.2897; found 552.2906.

(oil): ν = 3325 (O–H and N–H), 3066, 3041, 3020 (aryl-H), 2883

˜

(C–H), 1672 (C=O stretch and N–H bend), 1452 (C=C) cm^–1. MS

(positive ion FAB): m/z (%) = 658 (8) [M + Na]⁺, 636 (32) [M +

H]⁺, 414 (21) [C₁₈H₄₀O₉N]⁺, 179 (68) [C₁₄H₁₁]⁺. HRMS (positive

ion FAB): calcd. for [M + H]⁺636.3384; found 636.3380.

1-Amino-26-(phenylmethyloxy)-3,6,9,12,15,18,21,24-octaoxa-

hexaeicosane (17): Triphenylphosphane (2.89 g, 11 mmol) was

added to 1-azido-26-(phenylmethyloxy)-3,6,9,12,15,18,21,24-oc-

taoxahexaeicosane (16) (5.29 g, 10 mmol) in tetrahydrofuran

(50 mL). The reaction was stirred for 2 h at room temperature and

water (0.8 mL, 44 mmol) was then added. The reaction was then

stirred for 48 h at room temperature. The reaction mixture was then

concentrated in vacuo, redissolved in toluene (100 mL) and then

concentrated in vacuo once more. The product was isolated from

the resulting semi-solid by normal-phase silica gel chromatography,

eluting with a gradient from chloroform (100%) to chloroform/

methanol/triethylamine (90:5:5), to afford 17 as a pale-yellow oil

26-(Fluoren-9-ylmethyloxycarbonylamino)-3,6,9,12,15,18,21,24-

octaoxahexaeicosanoic Acid (Fmoc-Haa9, 3): Chromium trioxide

(2.88 g, 28.8 mmol) in sulfuric acid (1.5 , 57.6 mL, 86.4 mmol)

was very slowly added to vigorously stirring 26-(fluoren-9-ylmeth-

yloxycarbonylamino)-3,6,9,12,15,18,21,24-octaoxahexaeicosan-1-

ol (18) (6.10 g, 9.60 mmol) in acetone (125 mL) at 0 °C. Once ad-

dition was complete the reaction was left to warm to room tem-

perature and stirred for 24 h. Water (400 mL) was then added to the

reaction along with reversed-phase silica gel (120 g). The reaction

mixture was concentrated in vacuo until the volume had decreased

(4.86 g, 97 %). ¹H NMR (300 MHz, DMSO): δ = 2.67 (t, J = by approximately one quarter and more water (100 mL) was added.

5.8 Hz, 2 H, CH₂NH₂), 3.38 (t, J = 5.8 Hz, 2 H, CH₂CH₂NH₂),

3.49–3.61 [m, 32 H, C₆H₅CH₂O(CH₂CH₂O)₈], 4.51 (s, 2 H,

C₆H₅CH₂O), 7.27–7.39 (m, 5 H, C₆H₅CH₂O) ppm. ¹³C NMR

(75 MHz, DMSO): δ = 42.2, 70.1, 70.5, 70.7, 72.9, 73.8, 128.2,

The reaction mixture was once again concentrated in vacuo until

no acetone could be detected in the mixture. The mixture was then

filtered to recover the reversed-phase silica gel which was washed

with copious amounts of water until the silica gel was almost

colourless. The reversed-phase silica gel was then washed with ace-

128.4, 129.1, 139.4 (signals superimposed) ppm. IR (oil): ν = 3269

˜

(N–H), 3050 (aryl-H), 2869 (C–H), 1601 (N–H bend), 1452 tonitrile (4ϫ200 mL) and chloroform (3ϫ150 mL). The organic

(C=C) cm^–1. MS (positive ES): m/z (%) = 526 (8) [M + Na]⁺, 504

fractions were combined and concentrated in vacuo to give a pale-

(99) [M + H]⁺, 524 (13) [C₁₈H₄₀O₉N]⁺. HRMS (positive ESI): green oil. The product was isolated by normal-phase silica gel

calcd. for [M + H]⁺504.31671; found 504.31645.

chromatography, eluting on a gradient from chloroform to chloro-

form/methanol (95:5) and finally to chloroform/methanol/acetic

26-(Fluoren-9-ylmethyloxycarbonylamino)-3,6,9,12,15,18,21,24-

octaoxahexaeicosan-1-ol (18): A solution of 1-amino-26-(phenyl-

methyloxy)-3,6,9,12,15,18,21,24-octaoxahexaeicosane (17) (5.51 g,

10.9 mmol) in dry diethyl ether (50 mL) under argon was cooled to

–78 °C. Ammonia was condensed into this solution until the vol-

ume of the solution had almost doubled (ca. 100 mL). Sodium pel-

lets were added to the solution under argon at –78 °C until a dark-

blue colour persisted. The reaction was warmed from –78 to –30 °C

and then cooled back to –78 °C. Methanol was slowly added to the

reaction at –78 °C under argon until the solution was no longer

blue in colour. The reaction was then warmed to room temperature

and the majority of the ammonia was allowed to evaporate. The

reaction was concentrated in vacuo and then kept under high vac-

uum for 24 h to ensure that as little ammonia as possible was pres-

ent in the remaining residue. Water (100 mL) was then added to

the remaining residue and the pH was adjusted to 4 (concd. hydro-

chloric acid). The pH of the solution was then adjusted to 7 by

addition of solid NaHCO₃. Once the solution had been neutralised,

NaHCO₃(1.38 g, 16.41 mmol) was again added, followed by water

(50 mL) and dioxane (100 mL). 9-Fluorenylmethyl chloroformate

(4.25 g, 16.41 mmol) in dioxane (50 mL) was slowly added to this

1

acid (80:15:5), to yield 3 as a yellow oil (4.38 g, 70.0%). H NMR

(300 MHz, CDCl₃): δ = 3.38 (m, 2 H, CH₂NHCOCH₂C₁₃H₉),

3.54–3.73 [m, 30 H, HO₂CCH₂OCH₂(CH₂OCH₂)₇], 4.14 [s, 2 H,

H O ₂C C H ₂( O C H ₂C H ₂) ₇] , 4 . 2 1 ( t , J = 6 . 9 H z , 1 H ,

N H C O O C H ₂C H C _{1 2}H ₈) , 4 . 3 9 ( d , J = 6 . 9 H z , 2 H ,

CH₂NHCOCH₂C₁₃H₉), 5.54 (br. s, 1 H, NH), 7.31 (t, J = 7.4 Hz,

2 H, aryl Fmoc), 7.38 (t, J = 7.4 Hz, 2 H, aryl Fmoc), 7.59 (d, J

= 7.4 Hz, 2 H, aryl Fmoc), 7.74 (d, J = 7.4 Hz, 2 H, aryl

Fmoc) ppm. ¹³C NMR (75 MHz, CDCl₃): δ = 40.9, 47.3, 66.6,

70.1, 70.3, 70.5, 70.6, 71.1, 119.9, 125.1, 127.1, 127.7, 141.3, 144.0,

156.7, 171.6 ppm. IR (oil): ν = 3336 (COO–H), 3060 (aryl-H),

˜

2881, 1714 (C=O), 1700 (C=O), 1600 (C=C and N–H bend) cm^–1.

MS (positive ES): m/z (%) = 672 (2) [M + Na]⁺, 650 (99) [M +

H]⁺, 428 (4) [C₁₈H₃₈O₁₀N]⁺. HRMS (positive ESI): calcd. for [M

+ Na]⁺672.29905; found 672.299822.

1-Azido-20-(phenylmethyloxy)-3,6,9,12,15,18-hexaoxaeicosane (20):

Methanesulfonyl chloride (15.48 mL, 22.91 g, 200 mmol) in dry

dichloromethane (100 mL) was added dropwise to 11-(tri-

phenylmethyloxy)-3,6,9-trioxaundecan-1-ol (4) (43.65 g, 100 mmol)

and triethylamine (30.66 mL, 22.26 g, 220 mmol) in dry dichloro-

Eur. J. Org. Chem. 2008, 2900–2914

www.eurjoc.org

2909

A. B. Tabor et al.

FULL PAPER

methane (250 mL) under argon at 0 °C. Once addition was com-

added. The reaction was then stirred at room temperature for 48 h,

plete, the reaction was stirred for 5 h at room temperature. A fur- after which the reaction was concentrated in vacuo. Water (500 mL)

ther volume of dichloromethane (250 mL) was then added to the

reaction and the organic solution was partitioned with satd. aq.

NaHCO₃(2ϫ300 mL). The organic phase was then partitioned

with satd. aq. NaCl (400 mL), dried with anhydrous sodium sulfate

and concentrated in vacuo to give a brown oil. This oil was dried

by azeotropic distillation several times with toluene and then

placed under high vacuum until the oil became an orange solid. 8-

(Phenylmethyloxy)-3,6-dioxaoctan-1-ol (10) (28.34 g, 100 mmol) in

dry dimethylformamide (200 mL) was added to the orange solid

under argon followed by sodium hydride (12.0 g, 300 mmol, 60%

w/w in mineral oil). The reaction was stirred for 5 d at room tem-

perature. The reaction was then concentrated in vacuo and then

redissolved in diethyl ether (1000 mL) and water (500 mL). The

organic phase was then partitioned with aq. satd. NaCl with a little

NaHCO₃(500 mL), dried with anhydrous sodium sulfate and con-

centrated in vacuo to give 19 as a brown oil. A solution made up

of dichloromethane, triethylsilane and trifluoroacetic acid (300 mL,

80:10:10) was added to this brown oil. The reaction was left to stir

at room temperature for 2 h. After 2 h, satd. aq. Na₂CO₃was