The synthesis of polyamide-oligonucleotide conjugate molecules

Nucleic Acids Research, Vol. 18, No. 3 © 1990 Oxford University Press 493

The synthesis of polyamide-oligonucleotide conjugatemolecules

Jim Haralambidis*, Lucy Duncan, Karin Angus and Geoffrey W.TregearHoward Florey Institute of Experimental Physiology and Medicine, University of Melbourne, Parkville,Victoria 3052, Australia

Received October 24, 1989; Revised and Accepted December 18, 1989

ABSTRACT

We have developed methods for the synthesis ofpeptide-oligodeoxyribonucleotide conjugate moleculesin particular, and polyamide-oligonucleotide conjugatesin general. Synthesis is carried out by a solid-phaseprocedure and involves the assembly of a polyamideon the solid support, conversion of the terminal aminogroup to a protected primary aliphatic hydroxy groupby reaction with a,w-hydroxycarboxylic acidderivatives, and finally oligonucleotide synthesis usingphosphoramidite chemistry. The conjugate moleculescan be used as DNA probes, with the polyamidecomponent carrying one or more non-radioactivemarkers. These conjugates also have the potential tobe used as anti-sense inhibitors of gene expression,with the peptide segment acting as a targeting moiety.

INTRODUCTION

Synthetic oligonucleotides are widely used as probes for specificnucleic acids, by selectively hybridizing to a complementarytarget sequence1-2. The usual method of detecting hybridizationis by the attachment of a radioactive label (eg. 32P or 35S) to theoligonucleotide. Non-radioactive labels offer a number ofsignificant advantages including greater safety and stability ofthe probe. In cases where tissue sections are probed they alsooffer the ability to directly detect hybridization under themicroscope.

A major problem with attempts to develop non-radioactiveprobes has been the lack of suitable synthetic procedures for theattachment of the appropriate labels. Although a number ofprocedures have been reported recently, they usually involve theaddition of a single primary aliphatic amino3"8, orsulphydryl9"11 group at either the 3' or, more commonly, the5'-end of the oligonucleotide or an internal primary aliphaticamino group12"13. This reactive moiety is then used for thesubsequent attachment of a variety of labels such as fluorophores,biotin, chemiluminescent groups or enzymes. A limitation of thisapproach is that the probes often do not have the requiredsensitivity because only a single label can be attached to eacholigonucleotide molecule. Higher sensitivity is required when

probing nucleic acids rather than proteins, due to the much lowernumber of target molecules. We14, and others15"17, havereported procedures for the synthesis of oligonucleotidescontaining multiple amino groups attached to the nucleotideheterocyclic base through a linker arm. However, we have foundthat even this procedure is limiting, since the number ofnucleotides in the probe will dictate the maximum possiblenumber of labels that can be attached. Furthermore thehybridization efficiency of the probe can be affected when a highdegree of derivatization is used14. In this report we describe ageneral method for the preparation of oligonucleotides containinga multi-functional polyamide moiety at the 3'-end. This moietyis attached to the oligonucleotide by a stable link and its structurecan be designed and manipulated to accommodate any numberof labels at any distance from the oligonucleotide. Preliminarycommunications of this work have been published18"19.

RESULTS

There are two major points to consider in the synthesis ofpolyamide-oligonucleotide conjugate molecules. Firstly, whetherthe oligonucleotide or the polyamide should be synthesized first.We decided to synthesize the polyamide first, since we plannedto utilize the general procedures of solid phase peptide synthesis.These methods employ harsher conditions than those ofoligonucleotide synthesis, and could cause degradation and/orpremature deprotection of a preformed oligonucleotide. Secondly,the linkage between the two moieties had to be stable to theconditions of synthesis, deprotection, purification and storage.We decided to use a phosphodiester link, similar to the internallinks in the oligonucleotide. In order to achieve this, we had tomodify the amino terminus of the peptide so that it terminatedin a primary aliphatic hydroxy group rather than an amino group.

Design of the linkage between the peptide and theoligonucleotideA number of compounds were prepared that could be used toform the linkage between the peptide and the oligonucleotide.These are essentially derivatives of a,w-hydroxycarboxylic acids,and were synthesized as outlined in Scheme I. These compounds

* To whom correspondence should be addressed

Downloaded from https://academic.oup.com/nar/article-abstract/18/3/493/1114025by gueston 12 February 2018

494 Nucleic Acids Research

NaO2C(CH2)3OHRCI -NH 2

HO(CH2)6NH2

:C(CH2)3OR

1a R = DMTr1b R=Px

HO(CH 2 ) 6 NHC(CH2 ) 2 CO2

| DMTrCI0

RO(CH2 ) 6 NHC(CH 2 ) 2 CO 2

RO(CH2)6NHC(CH2)2C;

SCHEME I

have the hydroxyl group protected as either a di-(p-meth-oxyphenyl)phenymethyl (dimethoxytrityl) or 9-phenylx-anthen-9-yl (pixyl) ether and the carboxy terminus activated asthe p-nitrophenyl active ester. They can react with the aminoterminus of the peptide, giving rise to a protected primaryaliphatic hydroxy group as the new terminus. The hydroxyprotecting groups were chosen so that they are similar to thoseused in conventional DNA synthesis. Reaction of the deprotectedprimary hydroxy group with the first nucleoside phosphoramiditewill give rise (after oxidation) to a phosphate triester initially,which after deprotection will be converted to a phosphodiester.In this way, the link between the peptide and oligonucleotidemoieties contains only functionalities mat are normally found asinternal links in peptides and oligonucleotides, i.e. amides andphosphodiesters.

Of the three compounds prepared as linkers, lb is preferredsince it is a crystalline compound that can be made analyticallypure, unlike l a and 2 that are oils. It does have the disadvantagehowever that at low resin loadings the colorimetric measurementof the pixyl group after acid release is unreliable, and soquantitative ninhydrin tests are carried out to determine the extentof amino reaction.

We have also found that these compounds are excellent forthe initial derivatization of the Aminopropyl Controlled PoreGlass (AP-CPG) resin, giving a readily quantifiable loading ofprotected hydroxy group on to which (after deprotection) thepolyamide is attached. Unreacted amino groups can be cappedby acetylation after addition of the linker.

We also tested the use of 7-butyrolactone in place of the abovementioned compounds. However, we found that although the useof this reagent in the initial derivatization of the CPG did notgive rise to any problems, its use as the linker between the peptideand the oligonucleotide gave rise to byproducts, probably becauseof the long time and elevated temperatures (7 days at 60°C)necessary for reaction.

Synthesis of peptide-oligonucleotide conjugatesWe chose to synthesize a conjugate containing the peptide (Ala-

a.b

(V/^-NHC(CH2)3OH

c,d,e

O O-NHC(CH2)3OCCHNH2

CH3

f.g.h

N 0 2

NO2

o o fo 1NH<i(CH2)3O&CHNH CCHNH H

CH3 [ R J n

i.b

O O fo l oNHC(CH2)3OCCHNH CCHNH C

I CH3 [ R J n\ DNA SYNTHESIS

SCHEME n

Reagents : (a) 1 or 2, DMAP; (b) CI2CHCO2H/CH2C12; (c) (Boc-Ala)2O,DMAP; (d) TFA/H2O; (e) Et3N/CH2Cl2; (f) Fmoc-Xaa-OPfp, HOBt; (g)piperidine/DMF; (h) repeat steps f and g n times; (i) 1 or 2, HOBt.

Lys-)5Ala to illustrate the synthesis of such compounds. Weused the Fmoc (fluoren-9-ylmethoxycarbonyl) peptide synthesismethodology20 to prepare the peptide segment. The peptide wasdesigned to include multiple lysine residues that couldsubsequently be used as attachment sites for non-radioactive labelsand alanine residues that served as spacers between the lysines.

We derivatized Aminopropyl CPG by reaction with la (SchemeII). The amount of hydroxy group introduced could be quantitatedby colorimetric assay of the dimethoxytrityl cation released uponacid treatment (trityl test). Following deprotection of the hydroxylgroup with dichloroacetic acid (DCA), the first amino acid (Ala)was introduced as the ferf-butoxycarbonyl (Boc) symmetricalanhydride, with dimethylaminopyridine (DMAP) as catalyst. TheBoc amino protecting group was then removed by trifluoroaceticacid (TFA) and the amines neutralized with triethylamine indichloromethane. Standard Fmoc peptide synthesis methodologywas then utilized, using the pentafluorophenyl (Pfp) active estersof N-a-Fmoc protected amino acids (lysine and alanine) as themonomers. Following removal of the last N-a-Fmoc aminoprotecting group, the solid support was reacted with 2. Thissubstrate was then used for oligonucleotide synthesis in anApplied Biosystems Inc. automated DNA synthesizer, usingmethyl N,N-diisopropyl nucleoside phosphoramidites21'22. Thefirst phosphoramidite was linked to the deprotected terminalaliphatic hydroxyl group. This reaction was quantitative asassayed by the trityl test. Following the synthesis of the 30meroligonucleotide d(GGGCTTCACAACATCTGTGATGT-CAGCAGG) (KPIB, complementary to a region of mousekallikrein mRNA that is common to all the mouse kallikreins23),the methyl protecting groups on the phosphotriesters wereremoved using thiophenoxide ion, and the acid sensitiveprotecting groups (e-Boc on Lys residues and 5'-DMTr) removedby treatment with 90% TFA/10% ethanedithiol14. Normal


FmocNHS

DCC ROH

3 R = Pfp

4 R = NHS

DCC HOPfp

SCHEME

Nucleic Acids Research 495

A B C D E F G H

workup with aqueous ammonia14-24 gave the peptide-oligonucleotide conjugate. The ammonia treatment removes thestandard protecting groups on adenine, guanine and cytosineresidues, and also cleaves the peptide-ester bond to the CPG resin,releasing the oligonucleotide-polyamide conjugate probably asa mixture of the C-terminal amide and carboxylate. Analysis ofthe crude product by polyacrylamide gel electrophoresis (PAGE)gave the product as the major band (Figure 1, lane B), runningslower than the underivatized 30mer (Figure 1, lane A). It waspurified by preparative PAGE. Amino acid analysis of the productgave the expected ratio of 6 Ala:5 Lys, with 1 mole of (Ala-Lys-)5Ala per mole of KPIB (Table I). This indicates that thepeptide bonds are stable to oligonucleotide synthesis anddeprotection. The product was resistant to snake venomphosphodiesterase (blocked 3'-end) and it was only partiallydigested ( « 1 0 nucleotides from 5'-end) by spleenphosphodiesterase. This enzyme is known to be very sensitiveto secondary structure25, and the interaction of the positivelycharged lysine residues on the peptide with the phosphodiesterbackbone may be inhibiting further digestion of the DNA.Appropriate conditions were found however for the completedigestion of the product with Pi nuclease, which gave the sameHPLC profile of nucleoside and nucleotides as digestion of theunderivatized oligonucleotide. A number of other peptideoligonucleotide conjugates were prepared using these procedures,mainly containing various numbers of lysine residues with avariable number of spacer amino acids.

Synthesis ofconjugates

non-peptide polyamide oligonucleotide

It was realised that the use of non-peptidic amino acids wouldintroduce greater flexibility in the design of the moleculararchitecture of the polyamide. Even though lysine residues werefound to be quite appropriate for the attachment of labels (seefollowing paper), the natural a-amino acids provided too smalla distance between peptide bonds to act as efficient spacersbetween the lysine residues. A spacing of appropriate dimensionswould require multiple residues of a-amino acids, which couldgive rise to synthetic problems as the size of the peptide grewtoo large. However, any a,w-aminocarboxylic acid could act asa spacer. The readily available 6-aminohexanoic acid (eAhx) was

Figure 1. Autoradiograph of a 20% polyacrylamide gel on which variousconjugates were electrophoresed. Lane A, KPIB; lane B, KPIB-LL-(Ala-Lys-)5

Ala; lane C, KPffi-LL-(cAhx-)4Lys-Ala, synthesized by the Fmoc method; laneD, KPIB-LL-(eAhx-)4Lys-Ala, Boc method; lane E, KPIB-SL-(eAhx-Lys-)|oAla;lane F, KPffi-SL-[(cAhx-)2Lys-]10Ala; lane G, KPB-SL-Ala-[Lys-(eAhx-)4]9-Lys-Ala; lane H, KPIB. LL stands for the linker derived from 2 and SL thatfrom 1.

Table I. Amino acid analysis results of oligonucleotide-polyamide conjugates

KPffi-LL-(Ala-Lys-)5AlaKPIB-LL-(e Ahx-)4Ly s- Ala'KPIB-LL-(cAhx-)4Lys-Ala2

KPIB-SL-Ala-ILys-teAhx-^ljLys-Ala

Ala

6.11.30.92.4

Lys

4.91.21.110.4

€Ahx

-3.54.035.2

nPeptide

nKPIB1.10.91.20.8

1 By Fmoc method2 By Boc method

chosen as the standard unit, but any other similar aminocarboxylicacid could be used. Initially, the N-Fmoc pentafluorophenyl activeester derivatives 3 and 6 were synthesized and used successfullyusing standard Fmoc peptide synthesis methodology. The dimer6 could be prepared in good yield from the monomer 4 and6-aminohexanoic acid (Scheme IE). It was found that the useof the N-hydroxysuccinimide (NHS) active ester 4 in this reactiongave a much higher yield of the acid 5 (100%) than the Pfpderivative 3 (50%). The acid 5 was also used directly in solidphase synthesis by utilizing the BOP (benzotriazol-1-yl-oxy-tris-(dimethylamino)phosphonium hexafluoroborate) method-ology26-27. This proved to be a very efficient way of introducingthis spacer avoiding the need to prepare the active ester.

To illustrate the synthesis of an oligonucleotide containing asingle primary aliphatic amino group well removed from theoligonucleotide segment, the polyamide (eAhx-)4Lys(Boc)-Alawas synthesized using the methodology described in the previoussection, and utilizing two couplings of the active ester 6 to providethe spacing to the oligonucleotide. KPIB was then synthesized,after addition of the linker 2. It was deprotected as before andon PAGE analysis gave the product as the major band, runningslower than KPIB (Figure 1, lane C, and Table I). This was found



to be a convenient way of introducing labels while the polyarruaeis still on the solid support (see following paper), but we soughtan even more straightforward way to prepare the samecompound. For this we chose to use solely Boc chemistry forthe synthesis of the polyamide. We used a-Boc-e-Fmoc-Lys-OPfp(or the corresponding carboxylic acid with BOP methodology)as the lysine derivative and Boc-(eAhx-)2OPfp (7) as the spacer.Boc-(eAhx-)2OPfp could be prepared from the correspondingmonomer Boc-eAhx-OPfp (8) by reaction with 6-aminohexanoicacid followed by acidification, extraction, concentration andrecondensation of the product Boc(-eAhx)2 and the releasedpentafluorophenol with DCC. The protected polyamide-oligonucleotide conjugate was deprotected in exactly the samemanner as a normal oligonucleotide, to give the same productas before (Figure 1, lane D and Table I). In this case the Fmocprotecting group on the lysine residue is cleaved during theammonia deprotection step.

A series of conjugates containing long polyamide moieties withten lysine residues in each were also prepared by the Fmocmethod, using a manual peptide synthesizer for polyamidesynthesis. In this series the spacing between the lysine residueswas varied from four to one aminohexanoyl residues. PAGEanalyses of these compounds, after purification, are shown inFigure 1, lanes E - G .

DNA synthesis with methyl versus cyanoethyl protectednucleoside phosphoramiditesThe use of the O-methyl protected phosphoramidites in thesynthesis of these conjugate molecules proceeds with few apparentside reactions, giving the desired conjugate as the major speciesin the crude product. However, we found that when the O-cyanoethyl phosphoramidites were used the yield of product waslow, sometimes zero, with a large number of smaller moleculesreplacing the main product. This was the case irrespective of theamino acid sequence of the peptide and oligonucleotidecomponents. After a number of attempts to find the reason forthis difference, it was found that the key variable was the lengthof the capping time during oligonucleotide synthesis. Theprotocols recommended by Applied Biosystems for use on theirDNA Synthesizer with the methyl protected phosphoramiditeshave a 120 sec capping wait step with the DMAP/AcjO solution,whereas those for use with the cyanoethyl protectedphosphoramidites have a 5 sec wait step. When we changed thislatter wait step to 120 sec, we obtained the same result as whenusing the methyl protected phosphoramidites (Figure 2). We thentried a variety of intermediate times so as to minimize anyremoval of the cyanoethyl groups by DMAP and finally decidedto use a 60 sec wait step. It appears that extended contact withDMAP during the capping step is required for good results. Apossible explanation for this observation is that thephosphoramidites react with the carbonyl oxygen of the peptidebonds, and DMAP is required to reverse this reaction. Suchreaction is known to take place with the amides on the protectedheterocyclic bases of the nucleotides28-29. Therefore, in all workin which cyanoethyl protected nucleoside phosphoramidites areused, a 60 or 120 sec capping wait step was employed.

A B C O E F G H

Figure 2. Autoradiograph of a 20% polyacrylamide gel on which the crude productmixtures from synthesis with variable capping times were 5'-end labelled andelectrophoresed. Lanes A and H have normal KPtB oligonucleotide. Lanes B - Gare KPIB-LL-(Ala-Lys-)5Ala conjugates except for lane F which is KPIB-SL-(eAhx-)4(Lys-)2Ala. In lane B the synthesis was carried out using O-methylprotected phosphoramidites. The products in lanes C - G were synthesized usingO-cyanoethyl protected phosphoramidites with the following capping times: C,5 sec; D, 15 sec; E, 30 sec; F, 1 min; G, 2 min. The syntheses in lanes B - Eand G were carried out on the same batch of peptide resin solid support.

a wide variety of stably linked polyamide-oligonucleotideconjugates. Both Boc and Fmoc solid phase peptide synthesismethodologies have been used for the synthesis of the polyamide,which usually contains one or more lysine residues, suitable forthe attachment of non-radioactive labels. The use of non-peptidicamino acids has added a much greater flexibility to the designof the polyamide component, and gives control over both theinter-lysine and thus inter-label spacing and the distance to theoligonucleotide. Once the polyamide has been synthesized on thesolid support and the linker added, aliquots of this can be usedfor the synthesis of any oligonucleotide without having toderivatize each oligonucleotide separately. In its simplest formthis means that oligonucleotides containing a single primaryaliphatic amino function as part of a polyamide moiety attachedat the 3'-end at an appreciable distance from the oligonucleotidecan be synthesized on an Automated DNA Synthesizer withouthaving to perform any additional steps. These will be very usefulfor the attachment of larger molecules to the probes, for example,enzymes. Finally, peptide-oligonucleotide conjugates can also beenvisaged where the peptide is a specific sequence that will impartcertain desirable properties to the oligonucleotide, for example,to facilitate its passage through cell membranes, in cases wherea physiological or pharmacological effect of the oligonucleotideis desired.

DISCUSSION

The use of a linking synthon that essentially converts the N-terminus of a polyamide from a primary aliphatic amino groupto a primary aliphatic hydroxy group has enabled us to synthesize

EXPERIMENTALGeneral'H and 13C NMR spectra were recorded on a Bruker AM300at 300 and 75 MHz respectively with tetramethylsilane as internal



reference. Dimethylformamide (DMF) was distilled underreduced pressure and used within 1-2 days. Pyridine wasdistilled from potassium hydroxide and stored over 5A molecularsieves. Microanalyses were determined by Amdel, Melbourne.Melting points were determined in open ended capillaries on anElectrothermal Melting Point Apparatus. Amino acid analyseswere carried out on a Beckman System 6300 Amino AcidAnalyser. Snake venom phosphodiesterase, bovine spleenphosphodiesterase and P| nuclease were from Pharmacia.Infrared spectra were recorded on a Perkin-Elmer 983G infraredspectrophotometer. HPLC was carried out on an Altex system,using a Vydac C )8 5/x 25 cmx4.6 mm column. Buffers usedwere A, 0.1 M triethylammonium acetate, pH 7.0 and B, 0.1M triethylammonium acetate containing 45% CH3CN, at pH7.0. Flash chromatography30 was carried out using MerckKieselgel # 9385.

p-Nitrophenyl 4-[di-(p-methoxyphenyl)phenylmethyloxy]-butyrate (la)Sodium 4-hydroxybutyrate (1.26 g, 10 mmol) and4,4'-dimethoxytrityl chloride (DMTrCl) (3.39 g, 10 mmol) werestirred in 30 mL of pyridine for 16 h, p-nitrophenol (1.39 g,10 mmol) and dicyclohexylcarbodiimide (DCC) (2.06 g,lOmmol) were added and stirred for a further 2 days. The reactionmixture was filtered, the solution was concentrated and flashchromatographed on 70 g of silica gel with 25%EtOAc/petroleum ether to give a light yellow oil (5.0 g, 95%).•H NMR (CDC13) 8 2.04 (m, 2H, H3), 2.7 (t, J=7.2 Hz, 2H,H2), 3.2 (t, J=5.9 Hz, 2H, H4), 3.77 (s, 6H, OCH3), 6.8-7.5(m, 15H, ArH), 8.2 (d, J=9.2 Hz, 2H, PhNO2 m-H). 13CNMR (CDCI3) 8 25.2 (C3), 31.6 (C2), 55.2 (OCH3), 62.0(C4), 86.0 (CAr3), 113.1 126.8, 127.1, 127.8, 127.9, 128.1,129.1, 130.0, 136.2, 145.0, 158.4 (DMTr), 122.4, 125.1, 145.2,155.4, (PhNO2), 171.1 (CO2). The product contained someEtOAc solvent impurity.

p-Nitrophenyl 4-[(9-phenylxanthen-9-yl)oxy]butyrate (lb)This was prepared using the same method used to prepared laby substituting pixyl chloride for dimethoxytrityl chloride, to givelb in 80% yield, mp 130-130.5 °C (EtOAc). 'H NMR(CDC13) 8 1.98 (m, 2H, H3), 2.7 (t, J=7.3 Hz, 2H, H2), 3.0(t, J=5.8 Hz, 2H, H4), 7.0-7.5 (m, 15H, ArH), 8.2 (d, J=7.1Hz, 2H, PhNO2 m-H). I3C NMR (CDC13) 8 25.0 (C3), 31.5(C2), 61.8 (C4), 75.4 (CPh3), 116.3, 123.2, 123.5 126.4,126.6, 127.9, 129.1, 129.4, 148.9, 151.3 (PxC), 122.4, 125.1,145.2, 155.4 (PhNO2 C), 171.0 (CO2). Anal. Calcd for C29 H23

NO6: C, 72.3; H, 4.8; N, 2.9. Found: C, 72.0; H, 4.4; N, 3.2.

p-Nitrophenyl 3-[6-(di-(p-methoxyphenyl)phenylmethyloxy)-hexylcarbamoyljpropanoate (2)A solution of succinic anhydride (1.0 g, 10 mmol) and6-aminohexanol (1.17 g, 10 mmol) in pyridine (10 mL) wasstirred for 4 d. DMTrCl (3.39 g, 10 mmol) was then added, itwas stirred for a further 4 h, andp-nitrophenol (1.39 g, 10 mmol)and DCC (2.06 g, 10 mmol) added and was stirred for a further2 d. The reaction mixture was filtered, the solution concentratedand flash chromatographed on 100 g of silica gel with 50%EtOAc/petroleum ether to give a light yellow oil (4.09 g, 64%).'H NMR (CDC13) 8 1.2-1.7 (m, 8H, CH2), 2.5 (t, J=6.5 Hz,2H, CH2), 2.93 (t, J=6.5 Hz, 2H, CH2), 3.01 (t, J=6.4 Hz,2H, CH2), 3.2 (t, J=6.4 Hz, 2H, CH2), 3.76 (s, 6H, OCH3),6.8-7.5 (m, 15H, ArH), 8.2 (d, J=9.2 Hz, 2H, PhNO2 m-H).

I3C NMR (CDC13) 8 25.83, 26.69, 29.53, 29.87, 30.55 (CH2),39.68 (C7/2NHCO), 55.14 (OCH3), 63.16 (DMTrOC#2), 85.60(CAr3), 112.91, 126.53, 127.64, 127.70, 127.79, 128.10,129.07, 129.95, 136.60, 145.32, 158.26, (DMTr C), 122.4,125.1, 145.31, 155.8 (PhNO2 C), 170.49, 170.75 (C=O). Thecompound contained some EtOAc solvent impurity that couldnot be easily removed.

Pentafluorophenyl N-(fluoren-9-ylmethoxycarbonyI)-6-amino-hexanoate (Fmoc-€Ahx-OPfp,(3)6-Aminohexanoic acid (2.62 g, 20 mmol) and Na2CO3 (5.30 g,50 mmol) were dissolved in 60 mL of H2O, 25 mL of dioxanadded, followed by N-(fluoren-9-ylmethoxycarbonyl-oxy)succinimide (Fmoc-NHS) (6.75 g, 20 mmol). It was stirredvigorously for 16 h. The cloudy reaction mixture was then pouredinto 1.2 L of H2O to give a clear solution. This was extractedwith EtOAc (2 x 300 mL) and the rapidly stirring aqueous layeracidified to pH 3 using = 10 mL of cone. HC1, to give avoluminous precipitate. This was kept at 4°C for 16 h, and wasthen filtered to give 6.05 g (86%) of Fmoc-eAhx.

To a solution of Fmoc-fAhx (1.77 g, 5 mmol) andpentafluorophenol (1.01 g, 5.5 mmol) in 8 mL of DMF wasadded a solution of DCC (1.03 g, 5 mmol) in 2 mL of DMF.This was stirred for 16 h, filtered, and the filtrate evaporatedin vacuo to dryness giving the crude ester. It was recrystallizedfrom 95% EtOH/1% AcOH (« 10 mL) to give 2.41 g (93%)of white needles, mp 128-129°C. "H NMR (CDC13) 81.4-1.8 (m, 6H, CH2), 2.7 (t, J = 7.2 Hz, 2H, CH2CO2), 3.2(m, 2H, NHC#2), 4.2 (t, J = 6.7 Hz, 1H, Fmoc CH), 4.4 (d,J = 6.8 Hz, 2H, Fmoc CH2), 4.8 (bs, 1H, Fmoc NH),7.3-7.8 (m, 8H, Fmoc ArH). 13C NMR (CDC13) 5 24.36(C4), 25.93 (C3), 29.62 (C5), 33.18 (C2), 40.73 (C6), 47.31(Fmoc CH), 66.54 (Fmoc CH2), 119.98, 125.00, 127.02,127.67 (Fmoc Aromatic CH), 141.34, 143.99 (Fmoc AromaticC), 156.46 (Fmoc C=O), 169.36 (ester CO2). IR (KBr) 1687(Fmoc C = O), 1782 (ester C=O) cm- ' . Anal. Calcd forC27HnNO4F5: C, 63.8; H, 2.2; N, 2.8. Found: C, 63.9; H,1.8; N, 3.2.

N-[N-(Fluoren-9-ylmethoxycarbonyl)-6-aminohexanoyl]-6-aminohexanoic acid (Fmoc(-eAhx)2,(5)To a solution of Fmoc-eAhx (prepared as above, 1.77 g, 5 mmol)and N-hydroxysuccinimide (0.575 g, 5 mmol) in 8 mL of DMFwas added a solution of DCC (1.03 g, 5 mmol) in 2 mL of DMF.It was allowed to stir for 16 h, filtered, and the filtrate evaporatedin vacuo to a syrup. This was recrystallized from isopropanol(=10 mL) to give 1.94 g (86%) of 4.

To a solution of 4 (0.912 g, 2 mmol) in 10 mL dioxan wasadded dropwise a solution of 6-aminohexanoic acid (0.524 g,4 mmol) and Na2CO3 (0.424 g, 4 mmol) in 10 mL of H2O. Theresulting suspension was stirred vigorously for 48 h, and wasthen poured into 100 mL of H2O to give a clear solution. ThepH of this rapidly stirring solution was reduced to 3 by thedropwise addition of = 10 mL of 1 M KHSO4. Voluminousprecipitate formed, it was kept at 4°C for 24 h, and then filteredto give a quantitative yield of the acid. This crude product wasrecrystallized from EtOAc to give 0.679 g (73%) of a whitepowder, mp 116.5-117.5°C. 'H NMR (CD3OD) 5 1.3-1.7(m, 12H, internal CH2), 2.1 (t, J = 7.4 Hz, 2H, CW.CONH),2.3 (t, J = 7.3 Hz, 2H, CH2CO2H), 3.0-3.2 (m, 4H,FmocNHC//2 and CH2CONHC7/2), 4.2 (t, J=6.8 Hz, 1H,Fmoc CH), 4.3 (d, J=6.8 Hz, 2H, Fmoc CH2), 7.2-7.8 (m,8H, Fmoc ArH). IR (KBr) 1632 (amide C=O), 1696 (Fmoc



C=O), 1718 (acid C=O) cm' 1 . Anal. Calcd for C21ftMN2O5:C, 69.5; H, 7.4; N, 6.0. Found: C, 69.2; H, 7.3; N, 5.8.

Pentafluorophenyl N-[N-(fluoren-9-ylmethoxycarbonyl)-6-aminohexanoylj-aminohexanoate (Fmoc-(eAhx-)2OPfp, 6)To a solution of 5 (233 mg, 0.5 mmole) and pentafluorophenol(101 mg, 0.55 mmole) in 1 mL of DMF was added DCC (103mg, 0.5 mmole). This was allowed to stir for 2 d, was thenfiltered, the filtrate evaporated in vacuo to a cream solid whichwas recrystallized from 95% EtOH/1 % AcOH (1 mL) to give180 mg (57%) of pure 6, mp 126-127°C. 'H NMR (CDC13)6 1.3-1.8 (m, 12H, internal CH2), 2.2 (t, J = 7.4 Hz, 2H,C#2CONH) 2.7 (t, J-7.3 Hz, 2H, CH2CO2), 3.1-3.3 (m, 4H,FmocNHC//2 and CH2CONHCH2), 4.2 (t, J=6.9 Hz, 1H,Fmoc CH), 4.4 (t, J = 6.9 Hz, 2H, Fmoc CH2), 4.8 (bs, 1H,Fmoc NH), 5.5 (bs, 1H, CONH), 7.2-7.8 (m, 8H, FmocArH). IR (KBr) 1639 (amide C = O), 1689 (Fmoc C=O), 1785(ester C = O) cm- ' . Anal. Calcd for C33H33F5N205: C, 62.7;H, 5.3; N, 4.4. Found: C, 62.5; H, 5.1; N, 4.7.

Pentafluorophenyl N-te/t-butoxycarbonyl-6-arninohexanoate(Boc-eAhx-OPfp, 7).To a solution of N-Boc-6-aminohexanoic acid (4.78 g, 20.8mmol) and pentafluorophenol (3.68 g, 20 mmol) in 50 mL ofEtOAc was added DCC (4.12 g, 20 mmol). This was allowedto stir for 16 h, was filtered and the filtrate evaporated in vacuoto a syrup. On standing, this crystallized to give 7.46 g (94%)of the ester. Recrystallization from isopropanol/1 % acetic acidgave 6.26 g of white needles, mp 81-83°C. 'H NMR(CDC13) 8 1.4-1.6 (m, 15H, Boc CH3 and internal CH2), 1.8(m, 2H, CH2CH2CO2), 2.67 (5, J = 7.3 Hz, 2H, CH2CO2),3.1 (m, 2H, NHC//2), 4.5 (b, 1H, NH). 13C NMR (CDC13)5 24.38 (C4), 26.00 (C3), 28.39 (Boc CH3), 29.70 (C5), 33.20(C2), 40.28 (C6), 79 (Boc C), 156 (Boc C=O), 169 (Cl). Anal.Calcd for C17H20NO4F5: C, 51.4; H, 5.1; N, 3.5. Found: C,51.5; H, 5.0; N, 3.4.

Pentafluorophenyl N-(N-terf-butoxycarbonyl-6-amino-hexanoyl)-6-aminohexanoate (Boc-(eAhx-)2OPfp, 8)To a solution of 6-aminohexanoic acid (1.31 g, 10 mmol) in 5mL of 1 M NaOH (5 mmol) and 5 mL H2O was added asolution of 7 (1.99 g, 5 mmol) in 10 mL of dioxan. The resultingfine suspension was stirred vigorously for 3 d, by which timeit was clear. It was added to 200 mL of H2O, and the pHdecreased to 3.5 by the dropwise addition of = 10 mL of 1 MKHSO4, the resulting solution extracted with EtOAc (3x100mL), dried (Na2SO4), and concentrated in vacuo to 3 mL.Another 10 mL of EtOAc was then added, followed by DCC(1.03 g, 5 mmol). It was stirred for 16 h, was filtered, the filtrateevaporated to dryness in vacuo, and the product was recrystalliz-ed from EtOH/H20 containing 1% AcOH (= 10 mL), to give1.75 g (67%) of white needles, mp 88-89°C. 'H NMR(CDCI3) 8 1.3-1.9 (m, 2H, CH2 and Boc CH3), 2.2 (t, J =7.5 Hz, 2H, C//2CONH), 2.7 (t, J = 7.3 Hz, 2H, CH2CO2),3.1 (m, 2H, Boc NHC#2), 3.3 (m, 2H, CONHC7/2), 4.6 (bs,1H, Boc NH), 5.6 (bs, 1H, CONH). 13C NMR (CDC13) 824.35, 25.31, 26.10, 26.41, 29.31, 29.83, 33.17, 36.62, 39.14,40.35 (CH2), 28.44 (Boc CH3), 79.12 (Boc central C), 156.05(BocC=O), 169.39 (ester C=O), 172.84 (amide C=O). Anal.Calcd for C23H3IN2O5F5:C, 54.1; H, 6.1; N, 5.5. Found: C,53.9; H, 5.9, N, 5.8.

Solid phase synthesisSolid phase synthesis using the Fmoc methodology was carriedout on a Cambridge Research Biochemicals (CRB) manualpeptide synthesizer. All other reactions were performed in areaction cell previously described14. DNA synthesis was carriedout on an Applied Biosystems 380A DNA Synthesizer.

Derivatization of the CPGTo Aminopropyl Controlled Pore Glass (AP-CPG, Fluka, poresize 500 A, 0.5 g, 20 /imole of amino groups) was added eitherla, lb or 2 (250 /tmol) and dimethylaminopyridine (DMAP)(30.5 mg, 250 /imol) in 2 mL of DMF. This was either shakenfor 3 h or left standing for 16 h. The CPG was then washed(DMF, CH2Cl2) and dried. The degree of functionalization wasquantitated by spectrophotometric assay of the dimethoxytritylor pixyl (X = 445 run, £=4770 M~') cation released on acidtreatment of a small amount of CPG. Residual amino groups(=10-20%) were then acetylated by treating the CPG withacetic anhydride (0.5 ml, 2.5 mmol) and DMAP (50 mg, 0.4mmol) in pyridine (2 mL) for 15 min.

When 7-butyrolactone was used to derivatize the CPG, CPG(0.5 g) and 7-butyrolactone (3 mL) were placed in an oven at60°C for 7 days. The extent of reaction was monitored by thedisappearance of the primary amino groups on quantitativeninhydrin analysis31.

Coupling of the first amino acidThe CPG obtained in the previous step was treated with 3%dichloroacetic acid in CH2C12 (2x5 min) and washed (CH2C12).It was then reacted with a solution of N-Boc-alanine symmetricalanhydride and DMAP (0.2 M in each) in DMF for 20 h. Afterwashing, the extent of reaction was determined by quantitativeninhydrin assay on a small amount of deprotected CPG. Anyresidual hydroxyl groups were then acetylated as before.

Peptide synthesisThe Boc group was removed from the first amino acid bytreatment with 90% trifluoroacetic acid (TFA)/H2O (30 min),followed by washing (CH2C12), neutralization (20%triethylamine/CH2Cl2), washing (CH2C12) and drying. Furtherpeptide synthesis was then carried out on the manual CRB peptidesynthesizer, using standard Fmoc chemistry20, by utilizing afourfold molar excess of Fmoc-amino acid pentafluorophenylester and 1-hydroxybenzotriazole (HOBt) in DMF.

Addition of linker synthon to peptideFollowing deprotection of the terminal amino group, the CPGwas reacted with la, lb or 2 (0.2 mmol) and1-hydroxybenzotriazole (0.2 mmol) in DMF (0.5 mL) for 16 h.Residual amino groups were acetylated and the CPG used forDNA synthesis. Reaction with la or lb gives rise to the shortlinker (SL) and with 2 the long linker (LL).

When 7-butyrolactone was used, the CPG (100 mg) wasreacted with 2 mL of 7-butyrolactone at 60°C for 7 days.

Polyamide synthesisSynthesis of polyamides containing a,oj-aminocarboxylic acidresidues was carried out as for normal peptide synthesis but usingfour equivalents of N-Fmoc amino acid active esters 3 or 6 atthe appropriate stage. Synthesis was also carried out using theN-Fmoc amino acid 5 utilizing the BOP (benzotriazol-1-yl-oxy-tris-(dimethylamino)phosphonium hexafluoroborate)



procedure2627. Briefly, a fourfold molar excess of 5, BOPreagent, N-methylmorpholine and HOBt was used, in DMF.Alternatively, in cases were Boc methodology was used, Bocprotected amino acid active esters 7 and 8 were used, utilizinga fourfold molar excess of the active amino acid and HOBt inDMF for 0.5 h.

Deprotection of conjugatesFollowing synthesis of the desired oligonucleotide on an AppliedBiosystems 380A DNA synthesizer14 using either methyl orcyanoethyl protected nucleoside phosphoramidites, the conjugatewas deprotected as required. If the lysine residues contained aside chain Fmoc group, then no extra steps were necessary, asthe N-Fmoc group is removed during the aqueous ammonia basedeprotection step. If the conjugate contained Boc groups on thelysine side chains then the syntheses were carried out leavingthe oligonucleotide fully protected, the solid support transferredto the manual reaction cell and treated with 90% TFA/10%ethanedithiol (1 mL) for 5 min, with shaking. It was then rinsed(CH2C12) and neutralized by repeated washes with 20%Et3N/CH2Cl2, followed by rinsing (CH2C12). Normaloligonucleotide cleavage and deprotection with aqueous ammoniathen followed14-24, to give the fully deprotected conjugate. Incases where methyl protected phosphoramidite nucleosides wereused, a thiophenoxide deprotection step14-32 was used prior tothe amino acid side chain deprotection.

Purification of conjugatesPurification was carried out by polyacrylamide gel electrophoresis(PAGE) using 10% gels2. Purification of the single lysinecontaining conjugate prepared by Boc chemistry was convenientlycarried out by reverse phase HPLC on a Vydac C18 column,since the 5'-DMTr group of this conjugate is intact. The DMTr-containing conjugate was initially purified using the followingconditions: isocratic at 33.3% B for 20 min, and then a gradientto 66.6% B over 30 min. DMTr-KPm-SL-(eAhx-)4Lys-Alaelutes at 44.0 min. Detritylation of the eluate (equal volume ofacetic acid, 15 min) and rechromatographing on a gradient of0 to 66.6% B over 30 min gave the pure product, eluting at 26.0min.

Analysis of conjugatesConjugates were routinely analysed by a number of methods.A sample was 5'-end labelled2 and subjected to PAGE, on 20%gels, to determine homogeneity and apparent size. Amino acidanalysis was used to determine the amounts and ratios of aminoacids in the conjugate. Digestion with nuclease enzymes andHPLC of the digest was used to determine thenucleoside/nucleotide composition. Digestions with Pj nuclease(conjugate containing 1 jig of DNA in 20 /tl of H2O, 2 /d of 0.5M NaOAc, pH 6.0 and 5 ng of nuclease P, (5 id, in 0.05 MNaOAc, pH 6.0, 50% glycerol, stored at -20°C)) at 37°C for30 min gave complete digestion. HPLC was carried out usinga gradient of 0 to 33.3% Buffer B over 30 min. Digestion withspleen phosphodiesterase (conjugate containing 1 fig of DNA in20 id H2O, 2.5 id of 30 mM KH2PO4, pH 6.0, 2.5 id of 3 mMEDTA, 0.025 units of bovine spleen phosphodiesterase (2 id,in 1.5 mM KH2PO4, pH 6.0, 50% glycerol, stored at -20°Q)at 37°C for 16 h generally gave only partial digestion, from the5'-end. Snake venom phosphodiesterase digestion was carriedout with 1 /tg of enzyme in 0.05 M K2HPO4, pH 10.0 at 45°Cfor 4 h. The conjugates were resistant to these conditions (PAGE)with only a small amount of digestion taking place.

ACKNOWLEDGEMENTS

We thank Scott Pownall for the radioactive labelling of conjugatesand Dr. John Wade and Professor John Swan for valuablediscussions. This work was supported by the National Health andMedical Research Council of Australia.

REFERENCES

1. Itakura, K., Rossi, J.J., Wallace, R.B. (1984) Ann. Rev. Biochem., S3,323-356.

2. Penschow, J.P., Haralambidis, J., Aldred, P., Tregear, G.W., Coghlan,J.P. (1986) Methods in Enzymology, 124, 534-548.

3. Agrawal, S., Christodoulou, C , Gait, M.J. (1986) Nucleic Acids Res. 14,6227-6245.

4. ChoUet, A., Kawashima, E.H. (1985) Nucleic Acids Res. 13, 1529-1541.5. Wachter, L., Jablonski, J-A., Ramachandran, K.L. (1986) Nucleic Acids

Res. 14, 7985-7994.6. Sproat, B.S., Beijer, B., Rider, P. (1987) Nucleic Acids Res. 15,6181 -6196.7. Li, P.; Medon, P.P.; Skingle, D.C.; Lanser, J.A.; Symons, R.H. (1987)

Nucleic Acids Res. 15, 5275-5286.8. Smith, L.M.; Fung, S.; Hunkapiller, M.W.; Hunkapiller, T.J.; Hood, L.F.

(1985) Nucleic Acids Res. 13, 2439-2502.9. Sproat, B.S.; Beijer, B.; Rider, P.; Neumer, P. (1987) Nucleic Acids Res.

15, 4837-4848.10. Zuckerman, R.; Corey D.; Schultz, P. (1987) Nucleic Acids Res. 15,

5305-5321.11. Connolly, B.A. (1985) Nucleic Acids Res. 13, 4485-4502.12. Jablonski, E.; Moomaw, E.W.; Tullis, R.H.; Ruth, J.L. (1986) Nucleic Acids

Res. 14, 6115-6128.13. Urdea, M.S.; Warner, B.D.; Running, J.A.; Stempien M., Clyne, J.; Horn,

T. (1988) Nucleic Acids Res. 16, 4937-4956.14. Haralambidis, J.; Chai, M.; Tregear, G.W. (1987) Nucleic Acids Res. 15,

4857-4876.15. Ruth, J.L.; Morgan, C , Pasko, A., (1985) DNA 4, 93.16. Ruth, J.L. (1984) DNA, 3, 123.17. Draper, D.E. (1984) Nucleic Acids Res. 12, 989-1002.18. Haralambidis, J.; Duncan, L.; Tregear, G.W. (1987) Tetrahedron Lett., 28,

5199-5202.19. Haralambidis, J.; Duncan, L.; Angus, K.; Chai, M.: Pownall, S.; Tregear,

G. (1988) Nucleic Acids Res. Symposium Series, No.20, 115-116.20. Atherton, E.; Sheppard, R.C. (1985) J. Chem. Soc., Chem. Commun.,

165-166.21. Beaucage, S.L.; Caruthers, M.H. (1981) Tetrahedron Lett. 22, 1859-1862.22. McBride, L.J.; Caruthers, M.H. (1983) Tetrahedron Lett. 24, 245-248.23. Drinkwater, C.C.; Evans, B.A., Richards, R.I. (1988) Trends in Biochemical

Sciences, 13, 169-172.24. Atkinson, T.; Smith, M. In: Oligonucleotide Synthesis: A Practical Approach;

Gait, M.J., Ed., IRL Press, Oxford, (1984) pp 35 -81 .25. Bernardi, A.; Bernardi, G. In: The Enzymes, 3rd edition, Volume 4; Boyer,

P.P. Ed., Academic Press, New York, (1971) pp 359-336.26. Castro, B.; Dormoy, J.R.; Evin, G.; Selve, C. (1975) Tetrahedron Lett.

1219-1222.27. Hudson, D. (1988) J. Org. Chem. 53, 617-624.28. Caruthers, M.H.; McBride, L.J.; Bracco, L.P.; Dubendorff, J.W. (1985)

Nucleosides and Nucleotides, 4, 95-105.29. Eadie, J.S.; Davidson, D.S. (1987) Nucleic Acids Res. 15, 8333-8349.30. Still, W.C.; Kahn, M.; Mitra, A. (1978) J. Org. Chem. 43, 2923-2925.31. Sarin, V.; Kent, S.; Tarn, J.; Merrifield, R. (1981) Anal. Biochem. 117,

147-157.32. Matteucci, M.D.; Caruthers, M.H. (1981) J. Am. Chem. Soc., 103,

3185-3191.



Documents

The synthesis of polyamide-oligonucleotide conjugate molecules