New highly efficient, chirally clean methods of segment coupling and stepwise solid phase in situ peptide synthesis

Since loss of configuration at the carboxylic acid residue during peptide coupling remains one of the most challenging problems in peptide synthesis whether one is concerned with stepwise or convergent synthesis, a main focus of the present work was to find an approach to overcome this difficulty. A first attempt involved an examination of the Fm group as a backbone protectant for two reasons. First, since the Fm group 43 precludes oxazolone formation it was expected that Fm amino acids could be substituted for Fmoc acids in stepwise solid phase syntheses while eliminating any chance for loss of configuration. Second, if an Fm group were introduced at the C-terminal position of a peptide segment it was expected that such segments would undergo coupling with the N-terminal amino group of a second segment without loss of chirality. Initial attention was paid to this aspect of the convergent approach to peptide synthesis. An appropriate segment was synthesized and some tests were conducted which showed loss of configuration to be minimal. Unfortunately however it was found that the synthesis of the appropriate Fm-backbone protected peptides proved to be difficult because of the large size of the Fm residue. This problem led to an examination of methods to overcome these slow coupling reactions. One of the methods examined involved the synthesis of the indazole analog 194 of HOAt 84 since it was expected that removal of one of the triazole N-atoms should allow the corresponding HATU-like uronium/guanidinium coupling species to be more efficient in directing the coupling process. A method for the synthesis of 7-aza-N-hydroxyindazole 208 was developed and the structure of the new compound established by normal methods which included a definitive X-ray structural analysis. Unfortunately it proved to be impossible to convert 208 into a stand-alone analog of HATU 99 and when 208 was used as an additive results were not encouraging. Therefore we abandoned the possible use of 208 and spent time on the development of some theoretically interesting pyrazine analogs of HOAt. Unfortunately these materials proved difficult to synthesize and this line of attack was dropped. At the time of this development we were engaged in trying to develop optimized synthetic routes to O-HAPyU 241 a reagent more effective than either N-HAPyU 242 or N-HATU 48 for coupling reactions, but which was only available in very low yield via standard methods. One of the methods examined involved study of the very fast reaction between the potassium salt of HOAt and the chlorouronium salt BTCFH 248. In this reaction the initially formed species is O-HAPyU 241 but this compound rapidly undergoes base-catalysed isomerization to the less reactive N-HAPyU 242. By working quickly it was hoped to be able to isolate the more reactive O-isomer. While carrying out these studies it occurred to us that this might be a general technique for optimizing other ordinary coupling reactions which involve the use of O-HATU 99b and O-HBTU 99a. All such O-substituted reagents are sensitive to reaction conditions and are quickly isomerized to the corresponding N-isomers. Thus the superior properties of the O-derivatives are lost. By generating these O-forms in situ we hoped for better results. In fact we were pleasantly surprised to find that precisely such results were observed. Indeed when tested on our standard models for examining the loss of chirality in segment coupling the new methods were remarkably more efficient than the standard techniques. For example in one case leading to a tripeptide 212 via [2+1] coupling the older technique always gave 5-10% of the LDL-isomer whereas our newer method can achieve this coupling with absolutely no loss of configuration. Several models were examined, even some extremely difficult ones and in all cases significant improvements were observed. These methods were also applied to solid phase syntheses of longer peptides and here again the results were superior. Thus we believe the new method represents a remarkable, even astonishing, improvement over previously described methods.