Introduction

Remarkable and progressive improvements in recent decades have been witnessed in the Nobel-Prize-winning solid phase peptide synthesis (SPPS) methodology (Merrifield 1963; Barany and Merrifield 1980; Fields and Noble 1990; Raibaut et al. 2015). For example, many types of studies have investigated not only the influence of the solid polymeric structure but also the peptide sequence and loading in the solvation properties of peptide resins (Sarin et al. 1980; Garcia-Martin et al. 2006; Deng et al. 2010; Nakaie et al. 2011). Experimental approaches have focused on variations in different aspects of the standard synthesis protocol such as temperature (Varanda and Miranda 1997), acylating conditions (Carpino et al. 1994; Kienhofer 2001) and the application of microwave irradiation (Kappe and Dallinger 2006). These approaches comprise part of the set of alternative experimental approaches used to improve the SPPS method. In this context, many efforts have also focused on using nuclear magnetic resonance (NMR) (Deber et al. 1989; Warrass et al. 2000), infrared (IR) (Yan 1998), fluorescence (Vaino et al. 2000) and CD (Pillai and Mutter 1981) spectroscopy.

In a conceptual departure from the great majority of existing approaches, here we examine peptide–resin solvation as a unique and complex physico-chemical example of a solute–solvent interaction process. By synthesizing different peptide resins as models and selecting solvent systems that encompassed nearly the entire polarity scale, we suggest some rules that control the swelling of peptide–resin beads (Cilli et al. 1996; Malavolta and Nakaie 2004). This different physico-chemical approach allowed us to propose a more practical and dimensionless polarity scale (Malavolta et al. 2002, 2008) that takes into account the electrophilic and nucleophilic values of solvent molecules (Gutmann 1978). Despite past efforts, the occurrence of sluggish amino acylation during peptide chain growth remains a significant limitation that is, in most cases, related to sequences with a strong propensity for aggregation (Hancock et al. 1973; Kent 1988; Malavolta et al. 2013).

In order to derive general criteria to overcome this serious and persistent problem in SPPS, we evaluated long transmembrane receptor fragments in terms of synthesis difficulty. In addition, we focused on the structural/conformational features of segments of some G-protein-coupled receptors (GPCR) to examine additional correlations with their complex mechanisms on the cellular level.

Consistent with earlier reports (Oliveira et al. 1997, 2002) that investigated some aspects of the chemical assembly of a second transmembrane helix (TM) fragment of the rat B2 receptor (McEachern et al. 1991), we focused on the peptide sequence (66–97)-TM32: VAEIYLGNLAGADLILASGLPFWAITIANNFD. The TM-32 fragment noted above was deliberately synthesized under very challenging conditions (i.e., using a very heavily homemade amino group-substituted 2.6 mmol/g MBHAR (methylbenzhydrylamine-resin) batch. The purpose of this alternative synthesis option was to verify the effect of the presence of a large quantify of peptide chains growing within resin beads with a peptide content in the final stage of synthesis (position #32) of roughly 90% (weight/weight).

For the present comparative investigation, we selected the same (66–97) fragment but from the MAS (TM32: FTVYITHLSIADISLLFSIFILSIDYALDYEL) (Young et al. 1988; Bader et al. 2014) and the rat AT1 (Peach 1977; Oliveira et al. 2007) receptors. In the latter peptide, an Ala residue was inserted at position #70 for use in a parallel project currently in progress (Ala70-TM32: LLNALALADLSFLLTLPLWAVYTAMEYRWPFG).

Boc chemistry was selected for the comparative synthesis of the (66–97) B2, MAS and AT1 receptors fragments. We used continuous coupling reaction monitoring and removed the peptide–resin portions at each of the eight-residue-long steps, thereby generating TM-8, -16, -24 and -32 peptide resins for each receptor. Microscopic swelling measurements of separated peptide resin beads in different solvents and analytical HPLC and LC/ESI-MS analyses of each cleaved and purified peptide from all of the peptide resins were also sequentially determined. We also examined the correlation between hydrophobicity and retention time (RT) in HPLC for each cleaved aggregating peptide fragment. In addition, to better estimate the solvent-dependent degree of peptide chain motion internally to the resin beads during the synthesis, all of the fractionated peptide resins of each receptor were labeled with the Toac (2, 2, 6, 6-tetramethylpiperidine-1-oxyl-4-amino-4-carboxylic acid) amino acid-type spin probe. This probe was previously introduced in peptide chemistry (Nakaie et al. 1981, 1983; Marchetto et al. 1993; Toniolo et al. 1995; Schreier et al. 2012). The conjugated microscopic swelling measurement–Toac-EPR spectroscopy of peptide resins used to evaluate their solvation properties has been applied with success for interpreting solvent-dependent peptide synthesis yields (Cilli et al. 1997, 1999; Marchetto et al. 2005). A clear correlation between the mobility of peptide chains or the average chain–chain distance values and the rate of coupling reactions has been demonstrated; this correlation persists regardless of the solvent system (Nakaie et al. 2006, 2011). Finally, to better understand the complex GPCR’s structure–function relationship, preliminary conformational data of purified TM32 transmembrane fragments of the three receptors were examined in the light of CD experiments performed in aqueous solutions but containing varying amounts of the strong electron acceptor TFE.

Materials and methods

Materials

All of the reagents and solvents used in the solid phase peptide synthesis were analytical grade and were used from freshly opened containers without further purification. Dimethylformamide (DMF) was distilled over P2O5, and ninhydrin was distilled under reduced pressure before use. The 0.55 mmol/g MBHAR batch was acquired from Bachem but the heavily substituted 2.6 mmol/g MBHAR batch was synthesized in our laboratory following guidelines laid out in previous reports (Marchetto et al. 1992; Nakaie et al. 2011). Protected amino acids were purchased from Bachem with the following side chain protections: Asp and Glu (OcHex), Ser and Thr (Bzl), Trp (For), Tyr (2-BrZ), His and Arg (Tos).

Methods

Peptide synthesis

The peptides were synthesized manually according to the standard Boc protocol. After coupling the C-terminal amino acid to the resin, the successive α-amino group deprotection and neutralization steps were performed in 30% TFA/DCM (30 min) and 10% DIEA/DCM (10 min), respectively. The amino acids were coupled using DIC/HOBt in DCM/DMF (1:1), and if necessary, TBTU in the presence of HOBt or HOAt and DIEA using 20% DMSO/NMP as a solvent system. After a 2-h coupling period, the qualitative ninhydrin test was performed to estimate the completeness of the reaction. To check the purity of the synthesized peptide sequence attached to the resin, cleavage reactions with small aliquots of resin were carried out with the anhydrous HF procedure.

Analytical RP-HPLC

Analyses were performed in a system consisting of two model 510 HPLC pumps (Waters, Milford, MA, USA), an automated gradient controller, a Rheodyne manual injector, a 486 detector and a 746 data module. Unless otherwise stated, the peptides were analyzed on a 4.6 × 150-mm column with a 300-Å pore size and a 5-μm particle size (C18; Vydac, Hesperia, CA, USA) using the following solvent systems: A (H2O containing 0.1% TFA) and B (60% or 90% MeCN in H2O containing 0.1% TFA). A linear gradient of 10-90% B was applied at a flow rate of 1.5 mL/min over 30 min, and the detection of peaks was carried out at 220 nm.

Preparative RP-HPLC

The peptides were purified using solvent A (H2O containing 0.1% TFA) or solvent B (90% MeCN in H2O containing 0.1% TFA). A linear gradient was applied that was dependent upon the RT determined in the HPLC analysis of the peptide using the same solvent systems. The flow rate was 10 mL/min, and the detection of peaks was carried out at 220 nm.

Swelling measurements of resin beads (dry and swollen)

Before the use for the peptide synthesis or microscopically measuring the bead sizes, most of the resin batches were sized by sifting them through metal sieves. In short, 150–200 dry and swollen beads of each resin were allowed to solvate overnight and were spread over a microscope slide and measured directly using an Olympus model SZ11 microscope coupled with Image-Pro Plus version 3.0.01.00 software. The resins were measured with their amino groups in deprotonated form and were obtained by sequential washes in TEA/DCM/DMF, DCM/DMF (1:1, v/v) and DCM. The peptide resins were dried in a vacuum using an Abderhalden-type apparatus with reflux in MeOH.

Amino acid analysis

Peptide composition was controlled by an amino acid analysis that was performed on a Biochrom 20 Plus amino acid analyzer (Pharmacia LKB Biochrom Ltd., Cambridge, England) equipped with an analytical cation-exchange column. The peptides were hydrolyzed with 6 M HCl in sealed tubes in a nitrogen atmosphere at 110 °C for 72 h. The samples were concentrated in high vacuum, suspended in 0.2 M sodium citrate buffer, adjusted to pH 2.2 and automatically injected into the analyzer.

Mass spectrometry

The LC/ESI-MS experiments were performed on a system consisting of a Waters Alliance model 2690 separation module and a model 996 photodiode array detector (Waters, Eschborn, Germany) controlled with a Compaq AP200 workstation coupled to a Micromass model ZMD mass detector (Micromass, Altrincham, Cheshire, England). The samples were automatically injected on a Waters narrow bore Nova-Pak column C18 (2.1 × 150 mm, 60 Å pore size, 3.5 µm particle size). The elution was carried out with solvents A (0.1% TFA/H2O) and B (60% or 90% acetonitrile/0.1% TFA/H2O) at a flow rate of 0.4 mL/min using a linear gradient ranging from 5 to 95% B over 30 min.

EPR study

Spectra were obtained at 9.5 GHz in a Bruker ER 200 spectrometer at room temperature (22 ± 2 °C) using flat quartz cells from Wilmad Glass Co. (Buena, NJ, USA). The magnetic field was modulated with amplitudes less than one-fifth of the line widths, and the microwave power was set to 5 mW to avoid saturation effects. The concentration of peptides was 10−4 M in 0.02 M phosphate buffer at pH = 7. Roughly, 5% of the total amine groups present in each peptide resin was labeled with the Toac spin probe before the EPR experiments, as already detailed in the literature (Cilli et al. 1997, 1999; Marchetto et al. 2005).

CD study

Spectra were obtained on a Jasco J-810 spectropolarimeter that was continually flushed with ultra-pure nitrogen at room temperature. The peptide concentration was 10−4 M in 0.02 M phosphate buffer with a pH of 7.0 with the addition of TFE. Equivalent results were found in triplicate for the EPR and CD experiments.

Results and discussion

Synthesis and hydrophobicity data of transmembrane peptide fragments

Figure 1 comparatively highlights the progression of the synthesis difficulties of the three receptors’ fragments by quantifying the number of times that the coupling reaction was necessary at each position of the peptide sequence. The need for the re-coupling process began in residues 9 and 10 for the AT1 and MAS, respectively, and in residue 7 for B2 receptor. As discussed earlier, the result displayed for the latter peptide (Fig. 1c) is due to the fact that this B2 fragment was synthesized under more difficult conditions (using the 2.6 mmol/g MBHAR batch). This point accordingly explains the faster onset of the need for re-coupling reactions (at residue 7) during the synthesis of this peptide sequence. In any case, these results are in agreement with those of a previous study (Kent 1988). That author found that the increase in difficulty for acylation reaction typically began at the 7–10 regions onwards. However, Kent (1988) noted that the possibility of occurrence in other regions of the sequence could not be ruled out. Figure 1 also shows several other regions in the three peptide sequences where re-coupling steps were necessary until the end of synthesis. We note that it is also possible to depict an early need for the double recoupling process in the B2 segment at positions 12–13 (Fig. 1c). This finding is in close agreement with our previous results (Oliveira et al. 2002), which indicated the need for substituting DMF for 20% DMSO/NMP during the acylation steps of the synthesis of this transmembrane sequence.

Fig. 1
figure 1

Monitoring of the coupling reactions during the synthesis of TM-32 fragments: AT1 (a); MAS (b) and B2 (c)

Full size image

Table 1 lists the sequences of the examined transmembrane peptides that are composed of a set of segments taken from eight residues up to a total of 32 (henceforth denoted TMs-8, -16, -24 and -32). This table also lists the estimated values of peptide synthesis yield based on the peak area of the desired peptide on analytical HPLC and LC/ESI-MS chromatograms as well as their hydrophobicity indexes calculated according to a previous report (Meek and Rossetti 1981) and the corresponding RTs from analytical HPLC.

Table 1 Amino acid sequences, estimated synthesis yields (%), hydrophobicity indexes and retention times of TM-32’s fragments of AT1, MAS and B2 receptors
Full size table

Variations in the estimated synthesis yields of each TM’s fragments as a function of the peptide chain elongation are listed in Table 1. A progressive decrease in the synthesis yields of all three transmembrane segments was observed with minimum values of roughly 25% (AT1 and MAS) and 18% (B2). We note that this latter value is probably due to the fact that the B2 segment was carried out in very high peptide content values. These data reinforce the persistent occurrence of a set of side reactions that have not yet been overcome in the complex SPFS methodology. We deliberately opted to synthesize highly differentiated peptide sequences (i.e., long, hydrophobic and aggregating) that were well known as typical examples of challenging peptide structures to chemically assemble.

Finally, the analytical HPLC data in Table 1 reveal a direct relationship between hydrophobicity indexes and RTs in the analytical HPLC column of all transmembrane peptides that we examined. These interesting findings accordingly reveal a pronounced similarity among the three TM segments in terms of the synthesis and structural characteristics found for the GPCR-type AT1, MAS and B2 receptors.

Peptide–resin solvation studies

Next, we compared the (66–97) second transmembrane fragments of the AT1, MAS and B2 receptors in terms of the solvation degree of peptide chains spread within the resin beads in DCM, DMF and DMSO.

Microscopic measurements of peptide resins

By applying previously described methods (Sarin et al. 1980; Tam and Lu 1995) with minor alterations (Cilli et al. 1996; Marchetto et al. 2005), we succeeded in more properly monitoring the level of solvation of peptide–resin beads. This process is strongly dependent not only on the solvent system but also on the type of starting resin and the amount and sequence of peptide chains coupled to the polymeric support. This alternative approach allowed us to compare the total volume of the peptide–resin bead occupied by solvent molecules (as a percentage) and also to estimate important structural parameters such as the average distance between peptide chains (Marchetto et al. 2005) and its influence on the yield of the coupling reaction (Nakaie et al. 2006, 2011).

The degrees of swelling of the peptide–resin beads bearing 8-, 16-, 24- and 32-mer TM fragments in DCM, DMF and DMSO are comparatively shown in Fig. 2. In general, due to the progressive increase in the aggregation process occurring within the resin matrix as a function of peptide length, a decreasing in the swelling of beads in DCM was observed. Conversely, improved solvation was verified in the more polar and strong electron-donating DMSO, which is known to disrupt resin-coupled inter-chain associations (Malavolta and Nakaie 2004). This observation is also in close agreement with the dissolution rule of intractable and aggregated peptide segments, even those that are free in solution, which was proposed earlier by our group based on the electrophilic (AN) and nucleophilic (DN) strengths of each solvent system (Malavolta et al. 2006). Larger differences between AN and DN values correspond to higher degrees of dissolution of insoluble peptides. Hexafluoroisopropanol (HFIP) and DMSO are typical representatives of stronger electrophilic and nucleophilic solvents of these types of solvents; they have AN and DN values of 88.0 and 29.8, respectively.

Fig. 2
figure 2

Swelling values (%) of TM-32 fragments of AT1 (a); MAS (b) and B2 (c) in DCM, DMF and DMSO

Full size image

Finally, we need to point out that we used the DCM/DMF mixture (1:1) for the coupling step rather than the widely applied DMF. The choice of the former solvent system was due to the apolar characteristics of both the starting resin and the synthesized hydrophobic peptide sequences. Furthermore, we already observed in our earlier study (Oliveira et al. 2002) the occurrence of a very strong resin bead shrinking process at about position 12 of the B2-TM second transmembrane fragment when we used the electron-donating DMF rather than DCM. The use of a particular solvent is therefore very dependent on many factors, including the polarity of the starting resin, the side chain-protected or not amino acid residues and also the degree of peptide aggregation propensity. In this study, we selected DCM/DMF. However, 20% DMSO/NMP is also a valuable option for the assembly of these peptide segments.

Estimating peptide chain mobility using EPR spectroscopy

To complement the microscopic measurements of the peptide–resin beads, we also carried out a comparative investigation of the dynamics of peptide chains spread within beads using DCM, DMF and DMSO. We examined the EPR spectra of Toac spin-labeled TM fragments. Figure 3 shows the EPR spectra of AT1’s TM-16 and TM-32 peptide resins in these three solvents. This spectroscopic approach is valuable because higher peptide chain motion within resin beads corresponds to faster coupling reactions (Nakaie et al. 2006, 2011). In agreement with comparative swelling data of these peptide resins shown in Fig. 2, narrower EPR spectral peaks, which are indicative of greater mobility where the Toac spin labels are coupled to the resin, were mainly observed in DMSO.

Fig. 3
figure 3

EPR spectra of TM-16 (a) and TM-32 (b) peptide resins of the AT1 receptor obtained in DCM, DMF and DMSO

Full size image

The solvent-dependent effect on the degree of peptide chain mobility can also be seen in Fig. 4. The measured spectral central peak line-width (W0) parameter of EPR is known to be sensitive to the motion degree of resin-bound peptide chains (Oliveira et al. 2002; Nakaie et al. 2006; Malavolta et al. 2013). In this case, lower mobility of the site where Toac is attached to the resin corresponds to higher values of W0. The variation in the values of this EPR spectral parameter for all TM segments revealed that slower peptide chain motion is observed in DCM, particularly in the case of longer peptide segments.

Fig. 4
figure 4

EPR’s W0 values of TM-32 fragments of AT1 (a); MAS (b) and B2 (c) in DCM, DMF and DMSO

Full size image

Therefore, by conjugating the peptide–resin solvation data from microscopic measurements of resin beads and EPR experiments of spin-labeled peptide resins, we stress the similarity among the peptide fragments of the three receptors. As a general rule, one can conclude that DCM should be replaced by other more polar solvents during the critical amino acid acylation step when long, hydrophobic and aggregating sequences are to be synthesized.

CD studies of TM-32 purified fragments

Figure 5 shows the CD spectra of the TM-32 fragments of these three receptors as a function of increasing amounts of TFE. This solvent is known to be a secondary structure-inducing organic solvent (Buck 1998). The CD experiments revealed that TM-32 AT1 (Fig. 5a) exhibited no evidence of a classical helical conformation even in the presence of high TFE concentrations. Negative bands appeared at roughly 202 and 218 nm in the CD spectra of this transmembrane peptide regardless of the TFE concentration. These features are not indicative of a classical α-helix conformation; they appear to be due to the presence of a mixture of different helicoidal conformers (Perczel and Turns 1996). In this case, also the known contribution of aromatic side chains (2 Phe, 2 Tyr and 2 Trp) present in greater amount in the TM-32 AT1 sequence comparatively to other TM-32 partners may not be also completely disregarded.

Fig. 5
figure 5

CD spectra of TM-32 fragments of the AT1 (a), MAS (b) and B2 (c) receptors in different TFE concentration

Full size image

Conversely, this conformational characteristic found for TM-32 AT1 was no longer observed for the MAS and B2 TM fragments (Fig. 5b, c, respectively). Typical α-helix spectra (an upward band at 190 nm and two negative bands near 207 and 222 nm) were observed when the percentage of this strong electron acceptor alcohol exceeded roughly 30% for the MAS and B2 receptors’ fragments. These initial findings suggest the need for other complementary spectroscopic methods to more appropriately evaluate the conformational characteristics of TM32 fragments and also their shorter TM-8, -16 and -24 sequences.

Conclusions

The investigated second TM of AT1, MAS and B2 GPCR-type receptors (66–97 regions) revealed a set of relevant findings using multiple approaches. Interesting similarities were detected, for instance, associated with the level of synthesis difficulty, the direct correlation between peptide segment hydrophobicity versus the RT from analytical HPLC and also when data from microscopic swelling of the peptide–resin study was correlated with EPR’s peptide chain motion approach. The CD experiments revealed different structural features of the TM32 AT1 fragment compared with the TM32 MAS or B2 partners. Typical α-helicoidally structures were only detected for TM32 MAS or B2 when they were in aqueous solutions containing more than roughly 30% (v/v) TFE. In the case of the TM32 AT1 peptide, a mixture of different and complex helical conformers appears to contribute to its structural properties. Consistent with earlier studies that examined conformational evaluation of different extra- and intracellular domains of the AT1 (Pertinhez et al. 1997; Franzoni et al. 1997; Salinas et al. 2002) and B2 (Lopes et al. 2013) receptors, this study intended to initiate a more extended comparative conformational investigation of different fragments of some GPCRs. Taken together, the present findings demonstrate interesting aspects of this type of AT1, MAS and B2 fragments. They support the use of complementary investigations to reveal relevant aspects of the mechanism of action of these macromolecules on the cellular level.