Inr. J . Peptide Prorein Res. 37, 1991, 252-256

Effect of tertiary amine on the carbodiimide-mediated peptide synthesis MICHAEL BEYERMANN, PETER HENKLEIN', ANNEROSE KLOSE, REINHARD SOHR' and MICHAEL BIENERT

Institute of Drug Research, Academy of Sciences of the G.D.R.. *Institute of Pharmacology and Toxicology, Humboldt-University of Berlin, Berlin, G.D.R.

Received 19 March, accepted for publication 11 August 1990

The effect of tertiary amine (DIEA) on reaction rate and product purity of a carbodiimide/HOBt-mediated peptide synthesis was studied. It was found that very rapid activation can be achieved using carbodiimidel HOBt in non-polar solvents, such as DCM. Although the HOBt is poorly soluble in DCM, the activation proceeds within 2 min, probably forming the HOBt-ester. By such a preactivation followed by a coupling in the presence of DIEA the rate of coupling is comparable with other rapid methods using BOP or TBTU, and no racemization was found in a model coupling ( < 0.1%). For comparison, syntheses of neurotensin by means of different coupling reagents (BOP, TBTU, OPfp-esters) and the DIEA-catalyzed coupling after carbodiimide/HOBt-activation under comparable conditions have shown that these procedures are of the same value in view of coupling efficiency and product purity. Key words: carbodiimide: HOBt; increased efficiency; solid phase peptide synthesis; tertiary amine

amine will make possible rapid peptide couplings. As reported by Fournier et al. (2) the BOP-mediated coupling of N-alkoxycarbonyl protected amino acid with an excess of DIEA (2.3equiv.) proceeds without racemization ( < 0.2%). Nevertheless, there are some indications that the presence of DIEA may cause racemization especially for sensitive amino acids (8) or for segment condensation (9). Interestingly, additiqn of a tertiary amine, such as DIEA, during peptide syntheses may also function to disrupt H-bonds which are stabilizing conformations and consequently increase the amino group accessibility (10). With regard to the used excess of base it was found that the higher the base excess the higher the coupling rate (1 1). Using carbodiimide/HOBt, Otteson ef al. (12) added DIEA at the end of coupling and found an Abbreviations follow the recommendations of the IUPAC-IUB improved coupling efficiency in the BOC synthesis but Commission on Biochemical Nomenclature (Biochem. J. 126, 773no effect in the Fmoc synthesis. Because of that we 780 (1972)). Additional abbreviations: DCC, dicyclohexylcar- were interested to study the DIEA-catalysis for bodiimide; DIPCDI, diisopropylcarbodiimide; HOBt, hydroxy- carbodiimide/HOBt-mediated peptide bond formbenzotriazole; BOP, benzotriazol-l-yl-oxy-tris-(dimethylamino)- ation.

Many kinds of coupling techniques have been developed for the formation of amide bonds for the solid phase peptide synthesis ( I ) . At present the application of BOP and HBTU or their analogs seems to be preferred (2-4). An alternative is given by the remarkable catalysis of an active ester coupling by salts of N-hydroxy compounds (5-7). Recently, we reported that the catalysis of pentafluorophenyl ester or Fmoc amino acid chloride by a mixture of HOBt and DIEA results in a coupling rate comparable with BOP and HBTU (5). Assumedly all of these coupling methods follow the same mechanism via formation of HOBtester and coupling in the presence of DIEA (4). If this is true, independent of the mode of preparation, preformed HOBt-ester in combination with a tertiary

phosphonium hexafluorophosphate; TBTU, benzotriazol-l-ylI ,1,3,3-tetramethyluroniumtetrafluoroborate; HOPfp, pentafluorophenol; DIEA, diisopropylethylamine; DMF, dimethylformamide; DMA, dimethylacetamide; DCM, methylene chloride; MBHA, p-methylbenzhydrylamine-resin(polystyrene/I ?hdivinylbenzene).

252

RESULTS AND DISCUSSION

To study various coupling procedures, we carried out the reaction of Fmoc-Val (0.05 M) to resin-bound

Carbodiimide-mediated peptide synthesis amine (0.025 M) for 3 min. After several washes of the resin to remove the excess of reagents, the coupling yield was determined by treatment of the resin with 20% piperidine/DMF and reading the UV absorption at 301 nm. In first experiments we examined the direct influence of solvent, HOBt, and DIEA on the DIPCDImediated coupling (see Table 1). The highest rate of coupling was observed using DIPCDI in neat DCM without any additive (Ia). The addition of HOBt in DCM led to a diminished coupling rate, probably caused by a competition of amide bond and HOBt-ester formation (Ib). The higher the HOBt-excess the lower the coupling yield (not shown here) in neat DCM within 3min. The addition of DIEA in neat DCM (Ic) had a disadvantageous effect on the coupling rate in the presence of HOBt. In DMA and DCM/DMA (1/1) the rate of a carbodiimide-mediated coupling is in general lower (Id-Ii) than in neat DCM, as reported by other authors (3, 13). Contrary to the analogous reaction in DCM, the carbodiimide-mediated peptide formation in DMA (lh)can be accelerated by addition of HOBt. According to Bates et af. (14), the activation in DMF may depend on an equilibrium (R-COOH + DIPCDI 0-acylisourea), which can be shifted to the right side by catching the 0-acylisourea with HOBt forming the HOBt-ester. Independent of the solvent, the direct addition of DIEA had a negative effect on coupling rates (Ic, If, Ii). A surprising result, a substantial increase of the coupling rate, was observed by separating the activation using DCC/HOBt in DCM from the coupling step in the presence of DIEA (LId). A change from DIPCDI to DCC makes no difference in view of activation, as reported before (3). An advantage of using DIPCDI is the possibility of following the activation progress. In DCM the HOBt is practically TABLE 1 Effect of additives on a DIPCDI (0.Immol)-mediated coupling of Fmoc-VaCOH (0.1 mmol) 10 MEHA-resin (0.0Smmol amine); reaction for 3 min Additives (0.1 mmol) HOBt DIEA

DCM

Solvents (v = 2 mL) DCM/DMA

Yield

("/.I DMA

(1/1) a

-

b

+

-

c

+

d

-

+

e

+

f

+

g

-

h i

+ +

-

+ -

+

+ + +

+ + +

insoluble, but in the presence of Fmoc-Val after addition of DIPCDI within 1-2min a clear solution is obtained indicating the formation of the corresponding HOBt-ester, which rapidly reacts under DIEAcatalysis. Following the same activation procedure the coupling without DIEA proceeds much slower (IIc). An activation by DCC only, probably during formation of the symmetrical anhydride (15), and an addition of HOBt/DIEA (IIe) or DIEA (1%) to the coupling step did not result in comparable coupling rates to (IId). The comparison of (IIe) and (IIf) demonstrates that DIEA should be avoided during the activation step, which should explain the results in Table 1 (Ic, f, i). Since the presence of a polar solvent, such as DMF, is desirable in peptide synthesis to decrease inter- and intrachain hydrogen bonding, we investigated the solvent effect on the preactivation (Table 3) with the following results: (i) a nonpolar solvent, such as DCM, is favorable for the preactivation (IIIc), (ii) the presence of a high proportion of a polar solvent causes a dramatic decrease in the coupling yield (IID, c) and (iii) even a prolongated preactivation does not compensate for the disadvantages of using DMF (IIId). Thus, any DMF should be added to the coupling step only. Next, we compared the efficiency of the DCC/HOBt//DIEA-method (see IIIc) with couplings TABLE 2 Effect of HOBt and DIEA (O.1mmoi) on the DCC (0.lmmol)mediated preactivation ( P ) for 2min of Fmoc- Val (0.1 mmol) in DCM (1.0mL) and on the coupling ( C ) for 3min to MBHA-resin (0.05mmol amine) in DCM (2.OmL) HOBt (P)

HOBt (C)

DlEA (P)

a b

-

-

-

C

+

d e f

+

-

-

-

-

-

+ +

Yield (%)

-

57 44 47

+ + +

-

+

-

83 53 39

TABLE 3 Effect of solvent on DCCIHOBt (0.1mmoll-preactivation ( P ) pf Fmoc- Val-OH (0.1 mmol) and coupling to MEHA-resin (0.OSmmol amine) in DCMIDMF ( I l l ) , coupling for 3min in the presence of DlEA (0.1mmol)

44 25 22 I

+ +

+

-

DlEA (C)

10 4

a b

< I 5 < I

C

d e

Solvent (P)

Preactivation Time (min)

DMF DCM/DMF (l/l) DCM DMF DCM/DMF (1/1)

3 3 3 15 15

Yield (YO)

13 32 87 32 51

253

M. Beyermann et al. TABLE 4 Comparison of different procedures for the coupling of Fmoc- Val-X (0.1mmol) lo MBHA-resin (0.05 mmol amine), reaction for 3 min in 2.0mL DCMIDMF (111)

X

a b c

d

Reagent (0.1 mmol)

DCC* BOP TBTU -

OH OH OH OPfp

Additive HOBt (mmol)

DIEA (mmol)

Yield

0. I 0.1

0.I 0.25 0.25 0.1

84 83 19 81

(Yo)

‘According to I l k .

using BOP, TBTU, and OPfp-ester in the same model reaction (see Table 4). Our results demonstrate that in view of the rate of couplings the DCC/HOBt//DIEA-procedure is of the same value as widely applied methods for solid phase peptide synthesis, such as BOP, TBTU, OPfp-ester (Wa-d). The acceleration of a carbodiimide/HOBt-mediated peptide bond formation by tertiary amine may be accompanied by racemization (16). To study the effect of DIEA on racemization, Fmoc-Phe-OH was coupled by DIPCDI/HOBt//DIEA to resin-bound valine. After deblocking by piperidine/DMF, the dipeptide amide (L-Phe-Val-NH,) was detached from the Dod-resin (1 7) and analyzed by HPLC. The same coupling was carried out using BOP and TBTU, respectively. Independent of the reagent, no significant racemization was observed (D-Phe-ValNH, < 0.1%). To compare the effectiveness of DIPCDI/HOBt// DIEA with OPfp-ester, BOP, and TBTU in actual peptide synthesis, neurotensin (Pyr-Leu-Tyr-GluAsn-Lys-Pro-Arg-Arg-Pro-Tyr-Ile-Leu-NH,) was prepared. The syntheses, using the Fmoc-protection m ul

v!

N

strategy, were performed with brief couplings (1 5 min) in DCM/DMF using 2equiv. of the carboxylic components. For OPfp-esters the synthesis was performed automatically on a MilliGen 9050 PepSynthesizer using a standard program (couplings for 30min in DMF, 4 equiv.). All other syntheses were carried out manually. If necessary, a recoupling was carried out with 1 equiv. for 15 min, although the ninhydrin-test (18) indicated a nearly complete first coupling in all these cases. Recouplings were necessary using BOP (for Ile, Tyr, Pyr), TBTU (for Ile, Pyr), and DIPCDI/ HOBt//DIEA (for Asn, Pyr), respectively. The HPLCchromatograms of the crude products after detachment from resin (see Fig. 1) demonstrate that all syntheses were closely similar with regard to the purity of product obtained. Our results indicate that the DIEA-catalyzed peptide coupling via HOBt-ester using BOP, TBTU, or DCClDIPCDI is of the same value in view of the coupling efficiency. The procedure using carbodiimide is limited by the solubility of amino acid derivatives in DCM, but most of them were dissolved as a consequence of the preactivation step. The formation of HOBt-esters in the absence of base using carbodiimide may prove to be an advantage compared to BOP (or TBTU), if the sensible step in view of racemization consists in the formation of the acylphosphonium (uronium) salt in the presence of DIEA/HOBt (9). Further studies are in progress. MATERIALS A N D METHODS

Reagents and solvents Fmoc-protected amino acids were purchased from Milligen and Novabiochem. The Dod-resin (0.5 mequiv./g), the MBHA-resin (0.5 mequiv./g), and the Wang-resin (0.63 mequiv./g) were obtained from Bachem Inc. and Novabiochem, respectively. DMA,

0 N

rs

N

L

T I M E (minutes)

254

FIGURE 1 HPLC-chromatogramsfrom syntheses of neurotensin by BOP (a), DIPCDI/HOBt//DIEA (b), OPfp-esters (c), TBTU (d).

Carbodiimide-mediated peptide synthesis HOBt, BOP, DIPCDI, and DCC were purchased from Fluka. Tfa and DIEA were obtained from Merck-Schuchardt. DCM was distilled from anhydrous Na,CO, and kept over 4A molecular sieves. DMF (VEB Laborchemie Apolda) was kept prior to use over 4A molecular sieves for at least 2 weeks. TBTU was prepared according to Knorr et al. (4, 19). Analytical methods UV measurements were carried out on a Specord M40 (VEB Carl Zeiss, Jena). Analytical HPLC analyses were performed on Nucleosil 300, C 18, 5 p column 25 x 0.4cm) with a Shimadzu LC-8A instrument monitored at 220 nm. Conditions were: (a) characterization of crude products (neurotensin); CH,OH/ 0.025 M NH,Ac, pH 6.0 (40/60), flow 1.O mL/min, (b) racemization studies; buffer A = 0.2% Tfa/H20, buffer B = 0.16% Tfa - 83% CH,CN/H,O, linear gradient 5 to 25% B/40min, flow l.OmL/min, tR (L-Phe-Val-NH,): 10.2min, tR (D-Phe-Val-NH,): 31.9 min. Peptide samples were hydrolyzed in 6 N HCl for 24 h (20), and the amino acid analyses were performed after derivatization with dansylchloride by HPLC according to Knecht & Chang (21). The procedures used for experiments are listed in Tables 1-4. For all experiments, an aliquot of the same sample of deprotonated MBHA-resin was used. The deprotonated resin was prepared by washes of 5 g of the MBHA-resin with DCM, triethylaminelDCM, DMF, DCM and by drying under vacuum for several days, until no further loss of weight was observed.

resin was washed with DMF and DCM. For determination of the coupling yield see Ia-i.

IIIa-e: Fmoc-Val (0.1 mmol), DCC (0.1 mmol), and HOBt (0.1 mmol) were dissolved in 1.0mL DMF (a, d), DCM (c), or DMF/DCM (b, e), respectively, under intensive stirring. After preactivation, the mixture was transferred into a reaction vessel containing 100mg deprotonated MBHA-resin swollen in 1.0mL DCM (a, d), DMF (c), or DMF/DCM (b, e), respectively, and DIEA (0. I mmol). After 3 min, the reaction mixture was rapidly removed, and the resin was washed with DMF and DCM. For determination of the coupling yield see Ia-i. IVa-f Fmoc-Val (HOBt) and the appropriate coupling reagent (in each case 0.1 mmol) were dissolved in 1.O mL DCM within 2 min (preactivation), and the mixture was transferred into a reaction vessel containing 100mg of the deprotonated MBHA-resin swollen in I.OmL DMF and DIEA (0.1 or 0.25 mmol). After 3min, the resin was rapidly washed with DMF and DCM. For determination of the coupling yield see Ia-i. Racemization study Fmoc-Val was coupled to Dod-resin by DIPCDII HOBt without DIEA. To Fmoc-Phe (39mg) and HOBt (15mg) in 0.35mL DCM 16pL DIPCDI were added. After preactivation for 2 min, the mixture was transferred into a reaction vessel containing 100mg of H-Val-Dod-resin (0.5 mmol/g) swollen in 0.35 mL DMF with DIEA (17pL). After 15min, the reaction mixture was removed, and the resin was washed with DMF (3 x). The deprotection was performed by 20% piperidine/DMF (2 x 5min) and the resin was washed with DMF (3 x ) and DCM (3 x ). The dipeptide amide was detached from the resin by treatment with Tfa/5% H,O (2.0mL) for 1 h at ambient temperature. The solution was separated, evaporated tb dryness, and the residue was dissolved in 0.2% Tfa/ H 2 0 for HPLC studies. For analogous reactions, instead of HOBt and DIPCDI, BOP or TBTU (0.1 mmol) were used with 0.2mmol DIEA for the coupling of Fmoc-Phe to the H-Val-Dod-resin.

Ia-i: 0.1 mmol Fmoc-Val and of the additives were dissolved in 2.0mL of the appropriate solvent on 100mg of deprotonated MBHA-resin. After addition of 0.1 mmol DIPCDI, the reaction was carried out under intensive stirring for 3min. When neat DCM was used, the insoluble HOBt went into solution immediately after addition of DIPCDI. The reaction was stopped by rapid removal of the reaction mixture and intensive washes with DMF (3 x ) and DCM (3 x). To determine the coupling yield, the air-dried resin was treated with 3.6 mL of 20% piperidine/DMF for 15min. After sedimentation of the resin, 0.1 mL of Syntheses of neurotensin the supernatant was diluted (1/100) with 20% Nu-Fmoc-protectionwas used for all amino acids and piperidine/DMF for UV measurement. By UV deter- side-chain protecting groups were: Arg(Mtr); Tyr mination of the dibenzofulvene-piperidine adduct (tBu); Lys(B0C); Glu(0tBu). Using OPfp-esters, the (22), we calculated the yield of coupling. synthesis was performed automatically on 1.5 g FmocIIa-f Fmoc-Val (0.1 mrnol) was dissolved in 1.OmL Leu-PepSyn-KA-resin (Milligen, 0.078 mequiv./g). DCM containing the appropriate additives Following a standard protocol, the coupling con(0.1 mmol). DCC (0.1 mmol) was added and after ditions were: 4 equiv. Fmoc-amino acid OPfp-ester stirring for 2 min, the mixture was transferred into a and couplings for 30 min (with an exception for p-Glu: reaction vessel containing 100mg of deprotonated 1 h). All other syntheses were carried out manually by MBHA-resin swollen in 1 .O mL DCM and the appro- the following protocol: DMF (3 x ), 20% piperidine/ priate additives (0.1 mmol). After reaction for 3 min, DMF (2 x 5min), DMF (3 x), coupling, DMF the reaction mixture was rapidly removed, and the (3 x ). Fmoc-Leu-Wang-resin was prepared by cou255

M. Beyermann et al. pling of Fmoc-Leu (3 equiv.) with BOP (3 equiv.) and DJEA (6 equiv.) to Wang-resin (0.63 mequiv./g) in DMF for 1h (0.35 mequiv./g). For a synthesis 0.5 g of that resin were used. Two equivalents (0.35 mmol) of Fmoc amino acid were coupled in 2.0 mL DCM/DMF (l/l) for 15min. With BOP or TBTU 2equiv. of coupling reagent and 4equiv. DIEA were used. For the DIPCDI/HOBt//DIEA-mediated synthesis, to Fmoc-protected amino acid (2 equiv.) and HOBt (2equiv.) in DCM (1 mL) DTPCDI (2equiv.) was added. After preactivation for 2 min, the solution was transferred into a reaction vessel containing the resin in 1 mL DMF with DIEA (2equiv.). Arginine was incorporated by Fmoc-Arg(Mtr)-OPfp (2 equiv.) in the presence of HOBt (2 equiv.) and DIEA (2 equiv.). After the incorporation of p-Glu, the peptide-resins were washed with DMF, DCM, acetic acid, CH,OH, DCM. The detachment of the peptide from resin was achieved by mixture K (23); 8.25 mL TfalO.5g phenol/ 0.5 mL thioanisole/0.5 mL H,0/0.25 mL 1,2ethanedithiole, at ambient temperature for 24 h. After separation of the resin, the solution was dropped into ether (150mL). The precipitated product was collected, washed with ether and dried. As a standard for HPLC studies, the product of the BOP-mediated synthesis was characterized by amino acid analysis: Asx 1.1 (1); Glx 1.9 (2); Pro 1.8 (2); Arg 2.0 (2); Ile 1.O (1); Leu 2.0 (2); Lys 1.O (I); Tyr 1.8 (2). ACKNOWLEDGMENT The authors thank Prof. L.A. Carpino and his staffat the University of Massachusetts for performing the automated synthesis of neurotensin bv the OPfD-ester method.

REFERENCES 1. Barany, G., Kneib-Cordonier, N. & Mullen, D.G. (1987) Inr. J. Peptide Protein Res. 30,705-739 2. Fournier, A., Wang, C.-T. &Felix, A.M. (1988) Int. J. Peptide Protein Res. 31, 86-87 3. Hudson, D. (1988) J. Org. Chem. 53, 617-624 4. Knorr, R., Trzeciak, A., Bannwarth, W. & Gillessen, D. (1989) Proceedings of the 20th European Peptide Symposium (Jung, G. & Bayer, E., eds.), pp. 37-39, W. de Gruyter, Berlin 5. Beyermann, M., Granitza, D., Klose, A., Heinrich, G.,

256

6. 7. 8. 9.

10. 11.

12.

13.

14.

15.

16. 17.

18. 19. 20. 21. 22.

23.

Bienert, M., Niedrich, H., Chao, H.-G. & Carpino, L.A. (1990) Proceedings of the 11th American Peptide Symposium (Rivier, J.E. & Marshall, G.R., 4s.). pp. 925-927, ESCOM, Leiden Horiki, K. & Murakami, A. (1989) Heterocycles 28,615-622 Bodanszky, M. & Bednarek, M.A. (1989) 1.Prorein Chem. 8, 461-469 Forest, M. & Fournier, A. (1990) Int. J. Peptide Protein Res. 35,89-94 Steinauer, R., Chen, F.M.F. & Benoiton, N.L. (1989) Int. J. Peptide Protein Res. 34, 295-298 Tam, J. (1987) I n f . 1.Pepride Protein Res. 29, 421431 Gausepohl, H., Kraft, M. & Frank, R. (1989) Proceedings of the 20th European Peptide Symposium (Jung, G. & Bayer, E., eds.), pp. 241-243, W. de Gruyter, Berlin Otteson, K.M., Harrison, J.L., Ligutom, A. & Ashcroft, P. (1989) ABI-presentation at the 11th American Peptide Symposium, La Jolla, CA Barany, G. & Merrifield, R.B. (1980) in The Pepiides (Gross. E. & Meienhofer,J., eds.), Vol. 2, p. 130, Academic Press, New York Bates, H.S., Jones, J.H., Ramage, W.I. &Witty, M.J. (1981) Proceedings of the 16th European Peptide Symposium (Brunfeldt, K., ed.), pp. 185-190, Scriptor, Copenhagen Benoiton, N.L. & Chen. F.M.F. (1981) Proceedings of the 7th American Peptide Symposium (Gross, E. & Rich, D.H., 4 s . ) . pp. 105-109, Pierce Chemical Company, Rockford, IL Kolodziejczyk, A.M. & Slebioda. M. (1986) Inr. J. Peptide Protein Res. 28, 444449 Breipohl, G., Knolle, J. & Stiiber. W. (1987) TefrahedronLeft. 28, 5651-5654 Kaiser, E., Colescott, R.L.. Bossinger. C.D. & Cook, P.I. (1970) Anal. Biochem. 34, 595-598 Knorr, R., personal communication Bidlingmeyer, B.A. (1984) J. Chrornatogr. 336, 93-104 Knecht, R. & Chang, J.-Y. (1986) Anal. Chem. 58,2375-2379 Meienhofer, J., Waki, M., Heimer, E.P., Lambros, T.J., Makofske, R.C. & Chang, C.-D. (1979) Inf. J. Pepride Protein Rex 13, 3 5 4 2 Fields, C.D. & Fields, G.B. (1990) see Beyermann et al. (5). pp. 928-930

Address:

Dr. Michael Beyermann Institute of Drug Research Academy of Sciences of the GDR Alfred-Kowalke-Strak 4 Berlin - I136 G.D.R.