New clustering methods for population comparison on paternal lineages.

The goal of this study is to show two new clustering and visualising techniques developed to find the most typical clusters of 18-dimensional Y chromosomal haplogroup frequency distributions of 90 Western Eurasian populations. The first technique called "self-organizing cloud (SOC)" is a vector-based self-learning method derived from the Self Organising Map and non-metric Multidimensional Scaling algorithms. The second technique is a new probabilistic method called the "maximal relation probability" (MRP) algorithm, based on a probability function having its local maximal values just in the condensation centres of the input data. This function is calculated immediately from the distance matrix of the data and can be interpreted as the probability that a given element of the database has a real genetic relation with at least one of the remaining elements. We tested these two new methods by comparing their results to both each other and the k-medoids algorithm. By means of these new algorithms, we determined 10 clusters of populations based on the similarity of haplogroup composition. The results obtained represented a genetically, geographically and historically well-interpretable picture of 10 genetic clusters of populations mirroring the early spread of populations from the Fertile Crescent to the Caucasus, Central Asia, Arabia and Southeast Europe. The results show that a parallel clustering of populations using SOC and MRP methods can be an efficient tool for studying the demographic history of populations sharing common genetic footprints.

Facile solid-phase synthesis of C-terminal peptide aldehydes and hydroxamates from a common Backbone Amide-Linked (BAL) intermediate.

C-Terminal peptide aldehydes and hydroxamates comprise two separate classes of effective inhibitors of a number of serine, aspartate, cysteine, and metalloproteases. Presented here is a method for preparation of both classes of peptide derivatives from the same resin-bound Weinreb amide precursor. Thus, 5-[(2 or 4)-formyl-3,5-dimethoxyphenoxy]butyramido-polyethylene glycol-polystyrene (BAL-PEG-PS) was treated with methoxylamine hydrochloride in the presence of sodium cyanoborohydride to provide a resin-bound methoxylamine, which was efficiently acylated by different Fmoc-amino acids upon bromo-tris-pyrrolidone-phosphonium hexafluorophosphate (PyBrOP) activation. Solid-phase chain elongation gave backbone amide-linked (BAL) peptide Weinreb amides, which were cleaved either by trifluoroacetic acid (TFA) in the presence of scavengers to provide the corresponding peptide hydroxamates, or by lithium aluminum hydride in tetrahydrofuran (THF) to provide the corresponding C-terminal peptide aldehydes. With several model sequences, peptide hydroxamates were obtained in crude yields of 68-83% and initial purities of at least 85%, whereas peptide aldehydes were obtained in crude yields of 16-53% and initial purities in the range of 30-40%. Under the LiAlH4 cleavage conditions used, those model peptides containing t-Bu-protected aspartate residues underwent partial side chain reduction to the corresponding homoserine-containing peptides. Similar results were obtained when working with high-load aminomethyl-polystyrene, suggesting that this chemistry will be generally applicable to a range of supporting materials.

Evaluation of geranylazide and farnesylazide diphosphate for incorporation of prenylazides into a CAAX box-containing peptide using protein farnesyltransferase.

Protein farnesyltransferase (PFTase) catalyzes the attachment of a geranylazide (C10) or farnesylazide (C15) moiety from the corresponding prenyldiphosphates to a model peptide substrate, N-dansyl-Gly-Cys-Val-Ile-Ala-OH. The rates of incorporation for these two substrate analogs are comparable and approximately twofold lower than that using the natural substrate farnesyl diphosphate (FPP). Reaction of N-dansyl-Gly-Cys(S-farnesylazide)-Val-Ile-Ala-OH with 2-diphenylphosphanylbenzoic acid methyl ester then gives a stable alkoxy-imidate linked product. This result suggests future generations whereby azide groups introduced using this enzymatic approach are functionalized using a broad range of azide-reactive reagents. Thus, chemistry has been developed that could be used to achieve highly specific peptide and protein modification. The farnesylazide analog may be useful in certain biological studies, whereas the geranylazide group may be more useful for general protein modification and immobilization.

Preparation of protected peptidyl thioester intermediates for native chemical ligation by Nalpha-9-fluorenylmethoxycarbonyl (Fmoc) chemistry: considerations of side-chain and backbone anchoring strategies, and compatible protection for N-terminal cysteine.

Native chemical ligation has proven to be a powerful method for the synthesis of small proteins and the semisynthesis of larger ones. The essential synthetic intermediates, which are C-terminal peptide thioesters, cannot survive the repetitive piperidine deprotection steps of N(alpha)-9-fluorenylmethoxycarbonyl (Fmoc) chemistry. Therefore, peptide scientists who prefer to not use N(alpha)-t-butyloxycarbonyl (Boc) chemistry need to adopt more esoteric strategies and tactics in order to integrate ligation approaches with Fmoc chemistry. In the present work, side-chain and backbone anchoring strategies have been used to prepare the required suitably (partially) protected and/or activated peptide intermediates spanning the length of bovine pancreatic trypsin inhibitor (BPTI). Three separate strategies for managing the critical N-terminal cysteine residue have been developed: (i) incorporation of N(alpha)-9-fluorenylmethoxycarbonyl-S-(N-methyl-N-phenylcarbamoyl)sulfenylcysteine [Fmoc-Cys(Snm)-OH], allowing creation of an otherwise fully protected resin-bound intermediate with N-terminal free Cys; (ii) incorporation of N(alpha)-9-fluorenylmethoxycarbonyl-S-triphenylmethylcysteine [Fmoc-Cys(Trt)-OH], generating a stable Fmoc-Cys(H)-peptide upon acidolytic cleavage; and (iii) incorporation of N(alpha)-t-butyloxycarbonyl-S-fluorenylmethylcysteine [Boc-Cys(Fm)-OH], generating a stable H-Cys(Fm)-peptide upon cleavage. In separate stages of these strategies, thioesters are established at the C-termini by selective deprotection and coupling steps carried out while peptides remain bound to the supports. Pilot native chemical ligations were pursued directly on-resin, as well as in solution after cleavage/purification.

Facile preparation of disulfide-bridged peptides using the polymer-supported oxidant CLEAR-OX.

Formation of disulfide bonds in synthetic peptides is one of the more challenging transformations to achieve in peptide chemistry, in view of the possible formation of oligomeric by-products and other side reactions, as well as occasional solubility problems in aqueous oxidizing media. It was shown previously that 5,5'-dithiobis(2-nitrobenzoic acid) (DTNB identical with Ellman's reagent), when attached to polyethylene glycol-polystyrene (PEG-PS), controlled-pore glass (CPG), or modified Sephadex supports, was an effective oxidizing agent that promoted disulfide formation under mild conditions. More recently, this work was extended to Cross-Linked Ethoxylate Acrylate Resin (CLEAR) supports, because of their compatibility with both organic and aqueous solvent mixtures. The resultant new tool, termed CLEAR-OX, was used to conveniently produce several model cyclic disulfides with improved purities and yields, when compared with solution oxidations. A particularly striking example was the gram-scale oxidation of a urotensin II antagonist peptide containing a hindered penicillamine unit.

Practical protocols for stepwise solid-phase synthesis of cysteine-containing peptides.

This study details a series of conditions that may be applied to ensure 'safe' incorporation of cysteine with minimal racemization during automated or manual solid-phase peptide synthesis. Earlier studies from our laboratories [Han et al. (1997) J. Org. Chem. 62, 4307-4312] showed that several common coupling methods, including those exploiting in situ activating agents such as N-[(dimethylamino)-1H-1,2,3-triazolo[4,5-b]pyridin-1-ylmethylene]-N-methylmethanaminium hexafluorophosphate N-oxide (HATU), N-[1H-benzotriazol-1-yl)-(dimethylamino)methylene]-N-methylmethanaminium hexafluorophosphate N-oxide (HBTU), and (benzotriazol-1-yl-N-oxy)tris(dimethylamino)phosphonium hexafluorophosphate (BOP) [all in the presence of N-methylmorpholine (NMM) or N,N-diisopropylethylamine (DIEA) as a tertiary amine base], give rise to unacceptable levels (i.e. 5-33%) of cysteine racemization. As demonstrated on the tripeptide model H-Gly-Cys-Phe-NH(2), and on the nonapeptide dihydrooxytocin, the following methods are recommended: O-pentafluorophenyl (O-Pfp) ester in DMF; O-Pfp ester/1-hydroxybenzotriazole (HOBt) in DMF; N,N'-diisopropylcarbodiimide (DIPCDI)/HOBt in DMF; HBTU/HOBt/2,4,6-trimethylpyridine (TMP) in DMF (preactivation time 3.5-7.0 min in all of these cases); and HBTU/HOBt/TMP in CH(2)Cl(2)/DMF (1:1) with no preactivation. In fact, several of the aforementioned methods are now used routinely in our laboratory during the automated synthesis of analogs of the 58-residue protein bovine pancreatic trypsin inhibitor (BPTI). In addition, several highly hindered bases such as 2,6-dimethylpyridine (lutidine), 2,3,5,6-tetramethylpyridine (TEMP), octahydroacridine (OHA), and 2,6-di-tert-butyl-4-(dimethylamino)pyridine (DB[DMAP]) may be used in place of the usual DIEA or NMM to minimize cysteine racemization even with the in situ coupling protocols.

NMR-detected order in core residues of denatured bovine pancreatic trypsin inhibitor.

The NMR characteristics of [14-38]Abu, a synthetic variant of BPTI that is partially folded in aqueous buffer near neutral pH, support a model of early folding events which begin with stabilization of the nativelike, slow exchange core [Barbar, E., Hare, M., Daragan, V., Barany, G., and Woodward, C. (1998) Biochemistry 37, 7822-7833 (1)]. In partially folded [14-38]Abu, urea denaturation profiles for representative amide protons show that global unfolding is non-two-state and that core residues require a higher concentration of urea to unfold. Dynamic properties of pH-denatured [14-38]Abu and fully reduced and unfolded BPTI analogue were determined from heteronuclear NMR relaxation measurements at similar solution conditions. Differences at various sites in the polypeptide chain were evaluated from spectral density functions determined from T1, T2, and steady-state heteronuclear NOE data. Although denatured [14-38]Abu contains no persistent secondary structure, its most ordered residues are those that, in native BPTI, fold into the slow exchange core. The fully reduced analogue is significantly more mobile and shows less heterogeneous dynamics, but at 1 degree C, restricted motion is observed for residues in the central segments of the polypeptide chain. These observations indicate that there is a developing core or cores even in highly unfolded species. Apparently the effect of 14-38 disulfide on unfolded

Toward new designed proteins derived from bovine pancreatic trypsin inhibitor (BPTI): covalent cross-linking of two 'core modules' by oxime-forming ligation.

A 25-residue disulfide-cross-linked peptide, termed 'oxidized core module' (OxCM), that includes essentially all of the secondary structural elements of bovine pancreatic trypsin inhibitor (BPTI) most refractory to hydrogen exchange, was shown previously to favor nativelike beta-sheet structure [Carulla, N., Woodward, C., and Barany, G. (2000) Synthesis and Characterization of a beta-Hairpin Peptide That Represents a 'Core Module' of Bovine Pancreatic Trypsin Inhibitor (BPTI). Biochemistry 39, 7927-7937]. The present work prepares to explore the hypothesis that the energies of nativelike conformations, relative to other possible conformations, could be decreased further by covalent linkage of two OxCMs. Optimized syntheses of six approximately 50-residue OxCM dimers are reported herein, featuring appropriate monomer modifications followed by oxime-forming ligation chemistry to create covalent cross-links at various positions and with differing lengths. Several side reactions were recognized through this work, and modified procedures to lessen or mitigate their occurrence were developed. Particularly noteworthy, guanidine or urea denaturants that were included as peptide-solubilizing components for some reaction mixtures were proven to form adducts with glyoxylyl moieties, thus affecting rates and outcomes. All six synthetic OxCM dimers were characterized by 1D (1)H NMR; three of them showed considerable chemical shift dispersion suggestive of self-association and mutual stabilization between the monomer units.

Role of cysteine residues in structural stability and function of a transmembrane helix bundle.

To study the structural and functional roles of the cysteine residues at positions 36, 41, and 46 in the transmembrane domain of phospholamban (PLB), we have used Fmoc (N-(9-fluorenyl)methoxycarbonyl) solid-phase peptide synthesis to prepare alpha-amino-n-butyric acid (Abu)-PLB, the analogue in which all three cysteine residues are replaced by Abu. Whereas previous studies have shown that replacement of the three Cys residues by Ala (producing Ala-PLB) greatly destabilizes the pentameric structure, we hypothesized that replacement of Cys with Abu, which is isosteric to Cys, might preserve the pentameric stability. Therefore, we compared the oligomeric structure (from SDS-polyacrylamide gel electrophoresis) and function (inhibition of the Ca-ATPase in reconstituted membranes) of Abu-PLB with those of synthetic wild-type PLB and Ala-PLB. Molecular modeling provides structural and energetic insight into the different oligomeric stabilities of these molecules. We conclude that 1) the Cys residues of PLB are not necessary for pentamer formation or inhibitory function; 2) the steric properties of cysteine residues in the PLB transmembrane domain contribute substantially to pentameric stability, whereas the polar or chemical properties of the sulfhydryl group play only a minor role; 3) the functional potency of these PLB variants does not correlate with oligomeric stability; and 4) acetylation of the N-terminal methionine has neither a functional nor a structural effect in full-length PLB.

Experimental approaches to protein folding based on the concept of a slow hydrogen exchange core.

In a review of protein hydrogen exchange, we concluded that the slow exchange core is the folding core. By this we mean that the elements of secondary structure carrying the slowest exchanging backbone amides will tend to be the elements of secondary structure to fold first, that partially folded proteins will tend to be most organized in the core, and that peptides made to mimic the slow exchange core will tend to show nativelike structure. These generalizations have led us to ask several experimental questions that will be examined here: (1) In partially folded and unfolded proteins, how do the dynamics and structure of core regions differ from noncore regions? (2) Can we make protein 'core modules' as peptides corresponding to the slow exchange core? Can core modules be covalently linked to make a native state in which one conformation is significantly more stable than all other accessible conformations? (3) In a mutant perturbed outside the core, what are the effects on hydrogen exchange and folding?

Chemical syntheses and biological activities of lactam analogues of alpha-conotoxin SI.

Bicyclization represents an effective method for the introduction of conformational constraints into small, biologically important peptides. Several strategies have been developed for the preparation of bicyclic lactam analogues of alpha-conotoxin SI, a 13-residue peptide neurotoxin found in cone snail venom. Four analogues of the natural regioisomer of alpha-conotoxin SI were designed and synthesized, each with one of the two paired cysteines of the parent peptide being replaced by a side-chain lactam bridged glutamic acid/lysine pair. Solid-phase lactamization was studied to determine rates of formation of the two possible loops and to document the extent of dimerization and higher oligomerization. Radioligand binding assays were carried out on all synthesized peptides, including the naturally occurring two-disulfide form, in order to determine their affinities for nicotinic acetylcholine receptors (nAChRs). Replacement of the Cys(2)-Cys(7) loop of alpha-conotoxin SI with a lactam bridge resulted in complete loss of activity, whereas replacement of the Cys(3)-Cys(13) disulfide loop resulted in a approximately 60-fold reduction in affinity for one orientation and a approximately 70-fold increase in affinity for the other. The two active lactam analogues retain the selectivity exhibited by the naturally occurring peptide for the alpha/delta subunit of nAChRs, as judged by competition experiments with the curariform antagonist metocurine.

Synthetic null-cysteine phospholamban analogue and the corresponding transmembrane domain inhibit the Ca-ATPase.

Chemical synthesis, functional reconstitution, and electron paramagnetic resonance (EPR) have been used to analyze the structure and function of phospholamban (PLB), a 52-residue integral membrane protein that regulates the calcium pump (Ca-ATPase) in cardiac sarcoplasmic reticulum (SR). PLB exists in equilibrium between monomeric and pentameric forms, as observed by SDS-PAGE, EPR, and fluorescence. It has been proposed that inhibition of the pump is due primarily to the monomeric form, with both pentameric stability and inhibition dependent primarily on the transmembrane (TM) domain. To test these hypotheses, we have studied the physical and functional properties of a synthetic null-cysteine PLB analogue that is entirely monomeric on SDS-PAGE, and compared it with the synthetic null-cysteine TM domain (residues 26-52). The TM domain was found to be primarily oligomeric on SDS-PAGE, and boundary lipid spin label analysis in lipid bilayers verified that the isolated TM domain is more oligomeric than the full-length parent molecule. These results indicate that the stability of the PLB pentamer is due primarily to attractive interactions between hydrophobic TM domains, overcoming the repulsive electrostatic interactions between the cationic cytoplasmic domains (residues 1-25). When reconstituted into liposomes containing the Ca-ATPase, the null-cysteine TM domain had the same inhibitory function as that of the full-length parent molecule. We conclude that the TM domain of PLB is sufficient for inhibitory function, the oligomeric stability of PLB does not determine its inhibitory activity, and the three Cys residues in the TM domain are not required for inhibitory function.

Carbopeptides: carbohydrates as potential templates for de novo design of protein models.

De novo design of proteins has evolved into a powerful approach for studying the factors governing protein folding and stability. Among the families of structures frequently studied is the 'four-helix bundle' in which four alpha-helical peptide strands, linked by loops, form a hydrophobic core. Assembly of protein models on a template has been suggested as a way to reduce the entropy of folding. Here we describe the potential use of a carbohydrate as such a template. The monosaccharide D-galactose was per-O-acylated with (Nbeta-Fmoc-betaAla)2O to give a penta-substituted derivative, which was converted to the corresponding glycosyl bromide and used for the glycosylation of 4-hydroxymethylbenzoic acid pentafluorophenyl ester (HMBA-OPfp). The beta-glycosidic carbohydrate template (Nbeta-Fmoc-3Ala)4-beta-D-Galp-(1-O)-MBA-OPfp thus obtained was coupled to a PAL-PEG-PS resin and simultaneously extended at the four arms to yield, after cleavage from the solid support, a carbopeptide with four identical peptide strands. Extension of this concept to, for example, synthesis of novel multiple antigenic peptides (MAPs) and synthesis of carbohydrate clusters can be easily envisioned. The ability to efficiently synthesize such structures sets the stage for further studies to test whether the carbohydrate templates do indeed nucleate folding.

Synthesis and characterization of a beta-hairpin peptide that represents a 'core module' of bovine pancreatic trypsin inhibitor (BPTI).

A new strategy for the design and construction of peptide fragments that can achieve defined, nativelike secondary structure is presented. The strategy is based upon the hypothesis that 'core elements' of a protein, synthesized in a single polypeptide molecule, will favor nativelike structure, and that by incorporating a cross-link, nativelike core structure will dominate the ensemble as the more extended conformations are excluded. 'Core elements' are the elements of packed secondary structure that contain the slowest exchanging backbone amide protons in the native protein. The 'core elements' in bovine pancreatic trypsin inhibitor (BPTI) are the two long strands of antiparallel beta-sheet (residues 18-24 and 29-35) and the small beta-bridge (residues 43-44). To test the design strategy, we synthesized an 'oxidized core module', which contains the antiparallel strands connected by a modified reverse turn (A27 replaced by D), a natural disulfide cross-link at the open end of the hairpin, and N- and C-termini blocking groups. A peptide with identical sequence but lacking the disulfide cross-link at the open end was used as the 'reduced core module' control. The conformational behavior of both peptides was examined using (1)H NMR spectroscopy. Chemical shift dispersion, chemical shift deviation from random coil values, sequential and long-range NOEs, and H/D amide exchange rates were compared for the two peptides. We conclude that the ensemble of oxidized and reduced core module conformations samples both nativelike 4:4 and non-native 3:5 beta-hairpin structure, and that the oxidized module samples nativelike structure for a greater fraction of the time than the reduced module.

Solution structure of alpha-conotoxin SI.

The nuclear magnetic resonance solution structure of alpha-conotoxin SI has been determined at pH 4.2. The 36 lowest energy structures show that alpha-conotoxin SI exists in a single major solution conformation and is stabilized by six hydrogen bonds. Comparisons are made between the SI solution structure and the solution and crystal structures of alpha-conotoxin GI. Surprisingly, a high degree of similarity between the backbone conformations of the GI crystal and the SI solution structures is seen in the region of lowest sequence homology, namely residues Gly-8 to Ser-12. This similarity is more surprising when considering that in SI a proline replaces the Arg-9 found in GI. The correspondence in conformation in this region provides the definitive evidence that it is the loss of the arginine basic charge at residue 9 which determines the differences in toxicity between GI and SI, rather than any changes in conformation induced by the cyclic proline residue.

Solid-phase syntheses of heterocycles containing the 2-aminothiophenol moiety.

Efficient and general procedures have been developed for the solid-phase preparation of substituted benzothiazoles (1), 3, 4-dihydro-1,4-benzothiazines (2), 3,4-dihydro-1,4-benzothiazine-1, 1-dioxides (3), 3,4-dihydro-3-oxo-1,4-benzothiazines (4), and 3, 4-dihydro-3-oxo-1,4-benzothiazine-1,1-dioxides (5). All five classes of compounds were prepared from a common intermediate, resin-bound 2-amino-4-carboxythiophenol, in a minimal number of steps. This intermediate was generated by (i) coupling 4-fluoro-3-nitrobenzoic acid onto Wang resin, or onto an amino acid bound to the resin, (ii) substitution of the aryl fluoride with a protected thiol, (iii) reduction of the nitro group, and (iv) removal of sulfur protection. Reaction with the appropriate substrates and reagents to effect cyclization gave the substituted core structures, which were modified further to introduce additional point(s) of diversity. Following cleavages from the solid support, the compounds were obtained in high initial purities and good isolated yields after purification.

Chemical synthesis and receptor binding of catfish somatostatin: a disulfide-bridged beta-D-Galp-(1-->3)-alpha-D-GalpNAc O-glycopeptide.

The glycopeptide hormone catfish somatostatin (somatostatin-22) has the amino acid sequence H-Asp-Asn-Thr-Val-Thr-Ser-Lys-Pro-Leu-Asn-Cys-Met-Asn-Tyr-Phe-Trp-Lys-Se r-Arg-Thr-Ala-Cys-OH; it includes a cyclic disulfide connecting the two Cys residues, and the major naturally occurring glycoform contains D-GalNAc and D-Gal O-glycosidically linked to Thr5. The linear sequence was assembled smoothly starting with an Fmoc-Cys(Trt)-PAC-PEG-PS support, using stepwise Fmoc solid-phase chemistry. In addition to the nonglycosylated peptide, two glycosylated forms of somatostatin-22 were accessed by incorporating as building blocks, respectively, Nalpha-Fmoc-Thr(Ac3-alpha-D-GalNAc)-OH and Nalpha-Fmoc-Thr(Ac4-beta-D-Gal-(1-->3)-Ac2-alpha-D-GalNAc)-O H. Acidolytic deprotection/cleavage of these peptidyl-resins with trifluoroacetic acid/scavenger cocktails gave the corresponding acetyl-protected glycopeptides with free sulfhydryl functions. Deacetylation, by methanolysis in the presence of catalytic sodium methoxide, was followed by mild oxidation at pH 7, mediated by Nalpha-dithiasuccinoyl (Dts)-glycine, to provide the desired monomeric cyclic disulfides. The purified peptides were tested for binding affinities to a panel of cloned human somatostatin receptor subtypes; in several cases, presence of the disaccharide moiety resulted in 2-fold tighter binding.

Controlled syntheses of natural and disulfide-mispaired regioisomers of alpha-conotoxin SI.

Methods are reported for the unambiguous syntheses of all three possible disulfide regioisomers with the sequence of alpha-conotoxin SI, a tridecapeptide amide from marine cone snail venom that binds selectively to the muscle subtype of nicotinic acetylcholine receptors. The naturally occurring peptide has two 'interlocking' disulfide bridges connecting Cys2-Cys7 and Cys3-Cys13 (2/7&3/13), while in the two mispaired isomers the disulfide bridges connect Cys2-Cys13 and Cys3-Cys7 (2/13 & 3/7, 'nested') and Cys2-Cys3 and Cys7-Cys13 (2/3 & 7/13, 'discrete'), respectively. Alignment of disulfide bridges was controlled at the level of orthogonal protection schemes for the linear precursors, assembled by Fmoc solid-phase peptide synthesis on acidolyzable tris(alkoxy)benzylamide (PAL) supports. Side-chain protection of cysteine was provided by suitable pairwise combination of the S-9H-xanthen-9-yl (Xan) and S-acetamidomethyl (Acm) protecting groups. The first disulfide bridge was formed from the corresponding bis(thiol) precursor obtained by selective deprotection of S-Xan, and the second disulfide bridge was formed by orthogonal co-oxidation of S-Acm groups on the remaining two Cys residues. It was possible to achieve the desired alignments with either order of loop formation (smaller loop before larger, or vice versa). The highest overall yields were obtained when both disulfides were formed in solution, while experiments where either the first or both bridges were formed while the peptide was on the solid support revealed lower overall yields and poorer selectivities towards the desired isomers.