3D model of phospholipase domain
The model of Parvovirus B19 phospholipase domain has been built with a help of SwissModel server based on a template of phospholipase A2. Such model represents typical architecture of phospholipase A2 [22] that features three alpha helices and a long relatively hydrophobic loop between the first and the second alpha helices (Fig. 1a). That loop has been shown to dig into the membrane letting the second ester bond of phospolipids go straight into the active center of the enzyme [4].
To check stability of phospholipase domain we used PentUnFOLD server [11]. According to the results of PentUnFOLD 1D version that considers just amino acid sequence, the only stable element of secondary structure for the whole domain is the second alpha helix. However, 2D version of the same server that considers the information on secondary structure, described the first alpha helix as a stable one together with the second alpha helix (see supplementary Material, Table S1). Interestingly, PentUnFOLD 3D considered only the first alpha helix as a fragment that cannot turn into disordered state. As one can see in Fig. 1, the first and the second helices form numerous contacts with each other.
The PentaFOLD 3.0 algorithm [13] predicted that both helices are stable enough, unlike the third metastable one (Fig. 1b). To decrease hydrophobicity of the peptide we reduced the length of the long loop between the first and the second alpha helices. To decrease the level of noise in spectral data we also deleted unstable third alpha helix. As a result, we concentrated our attention on interactions between the first and the second alpha helices of phospholipase A2 domain. According to our hypothesis, the second alpha helix is responsible of the formation of the first alpha helix due to numerous interactions, and the shift from random coil to alpha helix may be a mechanism of phospholipase domain activation in the acidic environment. So, we used the synthetic peptide named SK30 in our experiments, as well as the conjugate of that peptide with BSA through additional N-terminal cysteine residue and SMCC linker.
Circular dichroism of SK30 peptide and CSK31 peptide conjugate with BSA
The shape of CD spectra for SK30 peptide solution in 0.01M phosphate buffer is almost the same both at pH = 7.4 and pH = 5.0. Moreover, that shape of spectra recorded by us is quite similar to the known one but for the complete phospholipase A2 domain [6]. In our experiment with just two fragments of that domain included in SK30 peptide the structure was represented by alpha helix (~ 12%) and beta sheet (~ 20%) according to the analysis with BeStSel server [15]. In the previous experiment with the complete domain, percentage of structured regions (~ 10% for alpha helices and ~ 25% for beta sheet) was also quite low [6]. So, the model (Fig. 1) does not correspond to the observed shape of the domain and the peptide. Indeed, the minimum at 202 nm on those spectra (Fig. 2) is rather indicating high percent of random coil than the presence of expected alpha helices [23]. According to our hypothesis, beta structure should be intermolecular in oligomers of both the domain and the peptide. No shifts in secondary structure content caused by the change in pH or the heating were detected by us for the SK30 peptide.
In contrast, spectra of the CSK31 peptide conjugate with BSA are different at pH = 7.4 (Fig. 3a) and pH = 5.0 (Fig. 3b). Even through at both pH values the minimum at 208–222 nm indicates the alpha-helical conformation of the polypeptide [23], at pH = 5.0 (Fig. 3b) that minimum is flatter (does not contain such a sharp minimum at 208 nm), while the maximum at 190 nm is about two times higher than that at pH = 7.4 (Fig. 3a). These features of those spectra have been interpreted by the BeStSel server [15] as follows: at pH = 7.4 the percent of residues in alpha helix is about 30%, but at pH = 5.0 the percent of residues in alpha helix is about 40%. Such random coil to alpha helix shift is caused by the change in conformation of the peptide, and not the carrier protein, since BSA itself has previously been shown to lose its alpha-helical conformation with the acidification of the medium in the CD study: from 75% at pH = 7.0 to 50% at pH = 4.0 [24].
Taken together, conjugation of the SK30 peptide through N-terminal cysteine and SMCC linker prevented formation of oligomers and the most of intermolecular beta sheet. That is why conjugation to a carrier can help to reveal structural shifts characteristic to full length protein.
Infra-red spectroscopy and membrane filtration of SK30 peptide and CSK31 peptide conjugate with BSA
Infra-red spectra of SK30 and CSK31 conjugate with BSA have been recorded with multiple horizontal attenuated total reflectance accessory. Because of this, we are able to check the secondary structure of partially dehydrated polypeptides. As one can see in Fig. 4, at both pH = 7.4 and pH = 5.0 SK30 peptide spectra show two peaks of the amide I band: the first one corresponds to intermolecular beta sheet (1625 cm− 1) [25], the second one corresponds to alpha helix (1656 cm− 1) [25]. For the CSK31 conjugate with BSA one can observe just a single peak at 1656 cm− 1 corresponding to alpha helix at both pH values, as well as for BSA itself. Only residual beta structural peak at 1625 cm− 1 can be observed in those spectra, and it is especially prominent at pH = 5.0 (Fig. 4). Filtration of SK30 solution through the centrifuge filter with 10 kDa molecular mass cut off has shown that together with bigger oligomers connected by intermolecular beta sheet there are dimers that contain no beta structure and show just a single peak at 1658 cm− 1. Interestingly, solution that passed through the filter with 30 kDa molecular mass cut off demonstrated both alpha-helical and beta-structural peaks (Fig. 4).
Gel electrophoresis of SK30 peptide
Blue native gel electrophoresis results confirmed the presence of the major fraction of rather large oligomers of SK30 peptide in a solution along with a minor fraction of smaller oligomers. Results were almost identical at both pH = 7.4 (Fig. 5) and pH = 5.0 in all replica with different electrophoresis chambers and staining methods. Namely, the major band of SK30 is situated between the second and the third bands of cytochrome c, while the minor band of SK30 is under the first band of cytochrome c that was used as a molecular weight marker. Since molecular weight of cytochrome c monomer is 12 kDa, the major oligomer of SK30 peptide has a mass between 24 and 36 kDa. Taking into account molecular mass of SK30 monomer (3.42 kDa), one can suggest that abovementioned synthetic peptide prefers to dissolve in water solution in form of octamers (3.42×8 = 27.36 kDa). Such octamers indeed should pass freely through the filter with 40 kDa molecular mass cut off, but should retain on a filter with 10 kDa molecular mass cut off.
Immunological confirmation of CSK31 conjugate reliability
We immunized two rabbits by SK30 peptide, and two other rabbits by CSK31 conjugate with BSA according to the standard protocol for polyclonal antibodies production [26]. Antibodies against SK30 were collected using AminoLink column for affine chromatography with immobilized SK30 peptide. Antibodies against CSK31 conjugate with BSA were collected using SulfoLink column with immobilized conjugate. Quantitatively, humoral immune response to the conjugate was stronger than that against the peptide (Fig. 6). Polyclonal antibodies against CSK31 conjugate with BSA have been checked for their ability to bind immobilized SK30 peptide and recombinant VP1u domain of the VP1 capsid protein of Parvovirus B19 with a help of affine chromatography. Some part of those antibodies were able to form complexes with SK30 peptide and with the full length recombinant VP1u that were not destroyed even at low pH of 2.5 of the IgG elution buffer. Existence of such stable complexes was confirmed by the reverse experiments as well, when collected filtered antibodies against CSK31 conjugate with BSA had been immobilized, while solution of VP1u was applied on those columns. In those affine chromatography experiments (Fig. 7) antigen-antibody complexes were broken after the heating of the column during 10 minutes at the temperature of 60°C only. Antibodies against CSK31 have been eluted from the column with immobilized SK30 peptide mostly after the second turn of the heating, and the shape of the fluorescence spectrum was characteristic for native IgG (with maximum of emission at 343 nm). Same antibodies have been eluted from the column with immobilized recombinant VP1u mostly after the first turn of the heating, but the maximum on a spectrum was at 308 nm, that is the evidence of the denaturation of IgG and the break of the energy transfer from Tyr to Trp residues (Fig. 7a). Recombinant VP1u protein has been eluted from the column with immobilized antibodies against CSK31 mostly in the third eluate, and there was maximum at 308 nm on that spectrum (Fig. 7a). However, the differential spectrum between the third and the second eluates from the same experiment has the shape characteristic to the native (not denatured) recombinant VP1u with the maximum at 343 nm (Fig. 7b). It means that some part of VP1u has been denatured, while another part has not.
The data described in this section confirm that some epitopes of SK30 peptide have been reproduced in the CSK31 conjugate with BSA, as well as that some epitopes of full length VP1u are the same as those of SK30 and CSK31 conjugate with BSA. It is likely that those common epitopes correspond to the second (stable) alpha helix of phospholipase domain, while the rest of the sequence of SK30 peptide might be involved in intermolecular beta sheet that is hidden inside oligomers.