Rabbit PrP Is Partially Resistant to in vitro Aggregation Induced by Different Biological Cofactors

Prion diseases have been described in humans and other mammals, including sheep, goats, cattle, and deer. Since mice, hamsters, and cats are susceptible to prion infection, they are often used to study the mechanisms of prion infection and conversion. Mammals, such as horses and dogs, however, do not naturally contract the disease and are resistant to infection, while others, like rabbits, have exhibited low susceptibility. Infection involves the conversion of the cellular prion protein (PrPC) to the scrapie form (PrPSc), and several cofactors have already been identified as important adjuvants in this process, such as glycosaminoglycans (GAGs), lipids, and nucleic acids. The molecular mechanisms that determine transmissibility between species remain unclear, as well as the barriers to transmission. In this study, we examine the interaction of recombinant rabbit PrPC (RaPrP) with different biological cofactors such as GAGs (heparin and dermatan sulfate), phosphatidic acid, and DNA oligonucleotides (A1 and D67) to evaluate the importance of these cofactors in modulating the aggregation of rabbit PrP and explain the animal’s different degrees of resistance to infection. We used spectroscopic and chromatographic approaches to evaluate the interaction with cofactors and their effect on RaPrP aggregation, which we compared with murine PrP (MuPrP). Our data show that all cofactors induce RaPrP aggregation and exhibit pH dependence. However, RaPrP aggregated to a lesser extent than MuPrP in the presence of any of the cofactors tested. The binding affinity with cofactors does not correlate with these low levels of aggregation, suggesting that the latter are related to the stability of PrP at acidic pH. The absence of the N-terminus affected the interaction with cofactors, influencing the efficiency of aggregation. These findings demonstrate that the interaction with polyanionic cofactors is related to rabbit PrP being less susceptible to aggregation in vitro and that the N-terminal domain is important to the efficiency of conversion, increasing the interaction with cofactors. The decreased effect of cofactors in rabbit PrP likely explains its lower propensity to prion conversion.


INTRODUCTION
The cellular prion protein (PrP C ) is a constitutive protein that is mostly found attached to the extracellular membrane. PrP C occurs naturally in the cells of all mammals, primarily in the nervous system (Prusiner, 1991), and is a conserved protein that exhibits a high sequence and structural identity. The protein has an N-terminal unstructured and flexible region formed by residues 23 to 121. It has an octapeptide region known as an octarepeat, which is comprised of a sequence of eight amino acids (PHGGGWGQ) that are repeated five times. The globular C-terminal region is composed of residues 122 to 231, which have three α-helices (H1 to H3), with H2 and H3 connected by a disulfide bridge and one small antiparallel β-sheet (Hornemann et al., 1997). Minor differences have been observed between the globular domain conformation in different species related to specific amino acid substitutions, including surface charge potential and dynamics (Pastore and Zagari, 2007;Wen et al., 2010a;Srivastava and Lapidus, 2017).
The prion protein has been physiologically related to several functions in the central nervous system, including neuronal protection, neuroplasticity, and cell signaling (Chen et al., 2003;Lopes et al., 2005;Linden, 2017;Wulf et al., 2017). However, it is most prominently associated with the development of pathologies known as transmissible spongiform encephalopathies (TSEs) or prion diseases. PrP C can undergo changes in its native conformation that turn it into its pathogenic isoform, PrP scrapie (PrP Sc ), a structure rich in β-sheets (Prusiner, 1982). PrP Sc can arise because of errors in the protein folding process, post-translational conformational changes, and conversion by direct contact between PrP Sc and PrP C in an autocatalytic process (Castilla and Requena, 2015).
Naturally occurring TSEs have been described in several mammal species. These include scrapie in sheep and goats, bovine spongiform encephalopathy (BSE) in cattle, chronic wasting disease (CWD) in deer, and Creutzfeldt-Jakob disease (CJD) in humans. Scrapie has been experimentally transmitted to rats and cats, although with varying degrees of difficulty (Gibbs and Gajdusek, 1973). The molecular mechanisms that determine transmissibility between species and barriers to transmission remain poorly understood . For example, some species do not develop TSEs at all, such as rabbits, dogs, and horses (Sigurdson and Miller, 2003;Bian et al., 2017). These mammals' PrP C have been reported as being resistant to conversion by PrP Sc samples from other species (Vorberg et al., 2003;Polymenidou et al., 2008). However, prions have been generated in vitro from rabbit material using protein misfolding cyclic amplification (PMCA), and are capable of infecting leporids (Chianini et al., 2012) and transgenic mice expressing rabbit PrP (Vidal et al., 2015). Species previously considered completely resistant to prions have therefore been shown to be slightly susceptible to TSE infections (Castilla et al., 2005;Saá et al., 2006;Chianini et al., 2013;Vidal et al., 2015).
Although human and rabbit PrPs have very similar primary sequences (88%) (Yan et al., 2014), some amino acid residues in the rabbit PrP C sequence appear to contribute significantly to its inability to convert into PrP Sc and are thus closely associated with rabbits' low susceptibility to TSEs (Loftus and Rogers, 1997;Vorberg et al., 2003;Wen et al., 2010b;Yuan et al., 2013;Eraña et al., 2017). The primary sequence alignment between rabbit and mouse PrP has nine different amino acid residues in the flexible N-terminus and 14 in the structured C-terminus (Figure 1). In rabbits, PrP has a larger positively charged surface than it does in mice (Wen et al., 2010a;Silva et al., 2011).
Although the conversion of PrP C is caused by PrP Sc , several cofactors have been identified as important adjuvants in this process. These include glycosaminoglycans (GAGs), lipids, and nucleic acids (Deleault et al., 2007;Silva and Cordeiro, 2016;Vieira and Silva, 2016).
Several studies have linked GAGs to prion biology. Dermatan sulfate is one of the most commonly found components in amyloid deposits (Hirschfield and Hawkins, 2003). Membrane heparan sulfate plays a crucial role in the development of prion disease as a receptor for PrP Sc (Horonchik et al., 2005;Vieira and Silva, 2016). However, while these molecules have been observed to stimulate conversion in some instances, at other times, they inhibit the conversion, thus showing a paradoxical effect. Low molecular weight heparin interacts with murine PrP, leading to a transient oligomerization/aggregation process . The octapeptide region has also been shown to be important for interaction at neutral pH Kim et al., 2020), while a second binding site in the C-terminal region of the protein under acidic conditions has also been suggested . In addition, low molecular weight heparin acts as an inhibitor of PrP fibrillation, indicating its therapeutic potential (Vieira et al., , 2014. Lipids also play a role in the conversion of PrP C , and some phospholipids have been described as important for conversion and aggregation (Supattapone, 2014). Lipids such as phosphatidylserine (PS), phosphatidic acid (PA), phosphatidylethanolamine (PE), and phosphatidylinositol (PI) lead to the aggregation of PrP C (Tsiroulnikov et al., 2009), converting it to a protease-resistant form (Wang et al., 2007). Palmitoyloleoyl-phosphatidylglycerol (POPG) vesicles have been shown to induce PrP structural changes, leading to intermediates that, with the addition of RNA, form aggregates with infectious characteristics (Miller et al., 2013).
Besides lipids, GAGs, and proteins (Linden et al., 2008;Satoh et al., 2009), PrP also binds nucleic acids, both DNA and RNA. These interactions have been the subject of research over recent decades. Nucleic acids have been proposed to act as catalysts in the conversion of PrP C to a PrP Sc -like form (Cordeiro et al., 2001;Cordeiro and Silva, 2005). More recently, DNA-induced prionlike conversion in a physiological process was also observed, with specific DNA aptamers modulating PrP liquid-liquid phase separation (Matos et al., 2020). The binding of nucleic acids can thus influence the unfolding of proteins in both disease-related and physiological situations.
Some regions of murine PrP have already been observed to bind to nucleic acids, such as the lysine-rich regions 23-27 and 101-110 and the C-terminal globular domain (Lima et al., 2006;Cavaliere et al., 2013). DNA can bind to PrP in vitro, modulating protein aggregation; this interplay changes the native conformation of PrP and increases the presence of FIGURE 1 | Rabbit and murine PrP structural alignment and comparison. (A) RaPrP 23−231 sequence (UNIPROT Q95211) was aligned with MuPrP 23−231 sequence (UNIPROT P04925) using the T-Coffee server. The figure was produced with the ESPrip server, using PDB 2FJ3 for RaPrP 23−231 and PDB 1XYX for MuPrP 23−231 secondary structure depiction. Aligned amino acid residues are in red boxes to distinguish them from low similarity residues. The secondary structures (alpha-helix and beta-strands) are shown on the top of the alignment for RaPrP 23−231 and the bottom for MuPrP 23−231 . (B) RaPrP 23−231 3D structure is shown in red, and non-homologous residues are colored gray. ARG 155, HIS186, and ASP201 (rabbit sequence numbering) are depicted in blue, showing hydrogen bonds and saline bridges between these residues. (C) MuPrP 23−231 3D structure in red with ARG 156, HIS187, and ASP 202 (mouse sequence numbering) are depicted in blue. No hydrogen bond is observed between ARG 156 and HIS187, and ARG 156-ASP 202 shows one salt bridge. ChimeraX was used for structure visualization and H-bond and salt bridge identification.
β-sheets (Cordeiro et al., 2001;Nandi et al., 2002). A complex formed by the interaction with a GC-rich 21 bp DNA (D67) has been shown to be toxic to a murine neuroblastoma cell line (Macedo et al., 2012). Recently, PrP-DNA interactions were investigated in a more targeted manner, selecting short single-stranded NA sequences that bind with high affinity and specificity to the extended C-terminal domain of the prion protein (PrP 90−231 ). A selected 25-mer apatmer (A1) was able to modulate the phase transition of PrP -from LLPS to aggregation (Matos et al., 2020).
Since several cofactors are important in modulating the conversion and aggregation of PrP, the interaction of these molecules with different PrP sequences may help explain differences in susceptibility between species. In this study, we compare the interactions of rabbit and mouse PrP with different biological cofactors such as GAGs (heparin and dermatan sulfate), phosphatidic acid (PA), and DNA oligonucleotides (A1 and D67), demonstrating that the interaction with these cofactors is related to a lower susceptibility of RaPrP to aggregation in vitro.

Preparation of PA Large Unilamellar Vesicles
Phosphatidic acid (L-α-phosphatidic acid, monosodium salt of chicken egg, cod. 840101) was purchased from Avanti Polar Lipids, Inc. (Alabaster, AL, United States). Powdered PA was solubilized in chloroform (code 102445, Merck, Darmstadt, Germany). For the formation of the lipid film, chloroform was evaporated using nitrogen gas. Subsequently, the film was resuspended in PBS buffer. To form large unilamellar vesicles (LUVs), the sample was extruded (code 610000) against polycarbonate membranes with a pore diameter of 0.1 µm (code 610005), all from Avanti Polar Lipids, Inc. (Alabaster, AL, United States).

DNA Oligonucleotides
Two DNA sequences previously studied by the group were chosen, one single-stranded (A1) and one double-stranded (D67). A1 contains 25 nucleotides (5 -CCGCGTACAATCGAGCTCGGGTGTC-3 ) and D67 has 21 nucleotides (5 -AAAGGACGCGCGCGCGCGTTA-3 ). The oligonucleotides were synthesized and purified by Integrated DNA Technologies (Coralville, IA, United States). The DNA material was annealed by heating at 95 • C for 5 min and subsequently slow cooling overnight before the experiments. The DNA concentration was determined by absorbance at 260 nm using the corresponding molar extinction coefficient of each oligonucleotide.

Prion Protein Expression and Purification
The expression in E. coli and purification by high-affinity chromatography of recombinant full length (PrP 23−231 ) and truncated (PrP 90−231 ) murine (Mu) and rabbit (Ra) proteins were performed as described in Vieira and Silva (2019).

Spectroscopic Measurements
Protein intrinsic fluorescence and light scattering spectra were acquired using an FP-8200 spectrofluorometer (Jasco Corp., Tokyo, Japan) or a PC1 spectrofluorometer (ISS, Champaign, IL, United States) in an "L" geometry (at 90 • relative to the excitation light). All samples were prepared in one of two solutions at 25 • C: 50 mM Tris (pH 7.4) and 100 mM NaCl, or 20 mM sodium acetate buffer (pH 5.5), and 100 mM NaCl. For all PrP constructs, the concentration used was 2 µM for interactions with GAG and PA. To assess the interaction with the DNA sequences, a concentration of 5 µM was used for PrP 90−231 constructs and 2 µM for PrP 23−231 due to signal intensities. For light scattering, the samples were illuminated at 450 nm, and data were acquired from 430 to 470 nm (DNA analysis) or else at 320 nm with data acquisition from 300 to 340 nm (GAG and PA analysis). Intrinsic fluorescence measurements were performed by exciting the sample at 290 nm and collecting the fluorescence emissions between 300 and 420 nm (DNA analysis) or exciting the sample at 280 nm and collecting the fluorescence emissions between 300 and 420 nm. The Stern-Volmer constant (K SV ) was obtained from the linear regression equation of fluorescence data (F 0 /F) from quencher cofactors and can be interpreted as the association constant of the complex (K a ) according to the following equation (Jameson, 2014): Where F 0 is the free protein initial fluorescence and F is the fluorescence signal at each quencher concentration, represented by [Q]. PrP interaction with PA data were analyzed as the center of spectral mass and fitted using GraphPad Prism with a one-site binding non-linear regression: (2) Where CM is the variation of the center of fluorescence spectral mass, Bmax is the maximum number of binding sites, [PA] is the concentration of PA, and K d is the equilibrium dissociation constant. K d was converted to its inverse, K a .
PrP:GAG fluorescence was measured after protein disaggregation. Samples were prepared at each specified PrP:GAG concentration in Eppendorf Protein LoBind Tubes at 25 • C, and measurements were taken after 4 h.

Heparin Affinity Chromatography
A HiTrap Heparin column was used (code 17040701, GE Healthcare, Little Chalfont, United Kingdom). The PrP sample was applied to the column, followed by a washing step with 20 mL of buffer (20 mM sodium acetate solution at pH 5.5, or 50 mM Tris solution at pH 7.4). Protein elution was carried out with 70 mL of 20 mM sodium acetate with 1 M NaCl at pH 5.5, or 50 mM Tris solution with 1 M NaCl at pH 7.4. The flow rate used in all steps was 1 mL/min. The data were normalized by the maximum absorption at 280 nm of each tested condition, and negative absorbance values were considered equal to zero. Retention factor values were calculated to allow comparison of the relative affinities for the immobilized cofactor. It is calculated by the formula (Gjerde et al., 2009): Where t r is the retention time of the test substance, and t m is the column dead time.

Aggregation and Interaction of Murine and Rabbit PrP With GAGs
In the cell, GAGs (especially heparan sulfate) and PrP molecules can both be found in the plasma membrane and endocytic compartments (Godsave et al., 2015) and thus are able to experience pH values close to neutral and acid . We therefore evaluated the PrP:low molecular weight heparin interaction at these two pH values. We found an interaction with a binding site in the N-terminal domain of murine PrP (MuPrP) at pH 7.4 and an interaction with a second site located at the C-terminal domain of MuPrP at pH 5.5 . Low molecular weight heparin was also shown to induce transient aggregation in MuPrP .
We used light scattering measures to evaluate and compare the effect of GAGs on the aggregation of MuPrP 23−231 and RaPrP 23−231 . The results showed that increasing heparin concentrations at pH 7.4 and 5.5 were accompanied by an increase in light scattering for MuPrP 23−231 and RaPrP 23−231 , suggesting the formation of aggregates (Figures 2A,B, respectively). However, the effect was 200% greater in MuPrP 23−231 (Figure 2A) than in RaPrP 23−231 ( Figure 2B). Meanwhile, dermatan sulfate triggered aggregation only at pH 5.5 in MuPrP 23−231 and RaPrP 23−231 (Figures 2C,D). The aggregation was 60% greater for MuPrP, further demonstrating the contrasting effect between the two PrPs at an acidic pH. Figures 1A,B), consistent with observations for low molecular weight heparin with MuPrP .

Aggregation was transient in all conditions (Supplementary
We also compared the interaction of GAGs with PrP 90−231 to explore the importance of the unstructured N-terminal region for the observed effect (Figure 3). A robust aggregation was only observed at pH 5.5 for heparin ( Figure 3A), though the effect was 500% lesser than for PrP 23−231 , corroborating data obtained for low molecular weight heparin that showed only one binding site at the C-terminal domain at acidic pH . The same effect was observed for dermatan sulfate (Figure 3C), suggesting that heparin and dermatan sulfate share the same binding regions but with different affinities. RaPrP 90−231 showed a negligible aggregation for both heparin ( Figure 3B) and dermatan sulfate ( Figure 3D).
To evaluate binding and affinities, we performed affinity chromatography with a heparin column (Figure 4). The result showed that all PrP 23−231 types were able to interact with the resin, with a higher concentration of NaCl required to displace this interaction at pH 5.5. A retention factor (k) difference between pH 5.5 and 7.4 was of 12.3 for MuPrP and 13 for RaPrP, reflecting a greater affinity at pH 5.5. We observed that k difference between MuPrP 23−231 and RaPrP 23−231 was of 1.4 at pH 7.4 and 0.7 at pH 5.5, suggesting a low difference in the affinities using this methodology (Figure 4). To further analyze affinity differences, we also performed protein intrinsic fluorescence measurements (Figure 5) after protein disaggregation. The results showed that binding led to a decrease in fluorescence intensity with increasing concentrations of heparin (Figures 5A,B) and dermatan sulfate (Figures 5C,D) at pH 7.4 and 5.5. Binding therefore led to fluorescence quenching, probably due to the approximation of glucosamine and galactosamine moieties (by heparin and dermatan sulfate, respectively) to tryptophan residues. Table 1 shows the results of the Stern-Volmer linear regression to compare affinity. All conditions were analyzed except for the interaction of PrP 23−231 (murine and rabbit) with heparin at pH 5.5 since disaggregation was not complete under heparin treatment (Supplementary Figure 1C), and remaining aggregates can interfere with intensity measurements, promoting an inner filter effect (Jameson, 2014).
We did not observe any significant differences in terms of affinity when comparing the fluorescence data obtained for the interactions of MuPrP 23−231 and RaPrP 23−231 with heparin ( Figure 5A) at pH 7.4. Meanwhile, we observed an eightfold affinity increase for MuPrP 90−231 at pH 5.5 ( Figure 5B and Table 1). Dermatan sulfate exhibited a low affinity for all PrP constructs (Figures 5C,D) except for PrP 23−231 at pH 5.5 that it was twice as high for MuPrP than for RaPrP ( Table 1). All GAGs had a greater affinity for PrP 23−231 than PrP 90−231 .

Evaluation of PA Interaction With Murine and Rabbit PrP
Lipids have also been identified in several studies as adjuvants of PrP conversion (Shyng et al., 1993;Nishina et al., 2004). To detect any differences in the interactions between RaPrP and a cofactor with different chemical and structural characteristics from GAGs, we also performed intrinsic fluorescence measurements to investigate the interaction with PA (Supplementary Figure 2). PA leads to an increase in fluorescence emission in all constructs at pH 7.4 and 5.5 (Supplementary Figure 2). This increase was accompanied by a blue shift of the emission spectra, indicating a shift to a more non-polar environment in the chemical environment of tryptophan. This suggests an interaction with the hydrophobic region of the LUVs and/or a reorganization of the structure where these amino acids are located (Supplementary Figure 2).
Variation of the center of mass ( CM) from intrinsic fluorescence data was obtained and fitted with a one-site binding non-linear curve, obtaining K a values ( Table 1) for all PrP constructs. Although CM was greater at pH 5.5, the affinity was higher at pH 7.4 (Figures 6A-D and Table 1), suggesting that the structural changes observed as a result of the interaction do not depend directly on affinity. CM and K a were statistically different between pH values except concerning the interaction between PA and RaPrP 90−231 ( Figure 6C and Table 1). Affinity differences were only observed between MuPrP and RaPrP at pH 5.5, where RaPrP was 1.5-fold greater than MuPrP (Table 1). Affinity was also 1.5-fold greater in the presence of the complete N-terminal domain at pH 7.4 ( Table 1). In PrP 90−231 , affinity is twice greater for RaPrP at pH 5.5 (Table 1).
Light scattering measurements of MuPrP 23−231 and RaPrP 23−231 showed that PA was able to induce protein aggregation and that its effect was also pH-dependent (Figures 7A,B), with greater aggregation at pH 5.5. The aggregation of MuPrP (Figures 7A,C) was higher than RaPrP  Table 1). This confirms that, in this case, structural changes and aggregation are not directly related to affinity.

Interaction Between PrP C-Terminal Extended and DNA Oligonucleotides
DNA can bind with high affinity to the prion protein in vitro and modulate its aggregation (Cordeiro et al., 2001;Macedo et al., 2012). It has also been shown that the DNA structure leads to distinct interactions with MuPrP 90−231 (Matos et al., 2020). In this study, we analyzed two DNA sequences previously described as high-affinity cofactors of MuPrP and evaluated whether they have different binding profiles in their interactions with MuPrP and RaPrP. We did this by monitoring intrinsic protein fluorescence and light scattering on a titration curve.
We began by evaluating the interaction of PrP 90−231 and two DNA sequences previously studied by the group, A1 and D67, at two different pH values. There are evident differences when analyzing the light scattering of different combinations of the two proteins (MuPrP 90−231 and RaPrP 90−231 ) with both oligonucleotides (Figure 8). These changes are also linked to the pH, with acidic environments leading to fourto twenty-fold increase in light scattering for MuPrP 90−231 and RaPrP 90−231 when interacting with the DNA sequences. Aggregation was greater with A1 ( Figures 8A,B) than with D67 (Figures 8C,D). A 20-fold increase was observed for the MuPrP 90−231 :A1 sample ( Figure 8A) relative to the initial light scattering, while the increase in RaPrP 90−231 was only sixfold (Figure 8B). The effect of D67 was less than A1 (Figures 8C,D). The same profile was observed at pH 7.4, though with a lower relative difference than at the lower pH.
The intrinsic fluorescence quenching upon binding with oligonucleotides revealed that MuPrP 90−231 had a higher affinity than RaPrP 90−231 for both A1 (1.2 and 1.4 times at pH 7.4 and 5.5, respectively) and D67 (1.8 and 1.9 times at pH 7.4 and 5.5, respectively) ( Figure 9 and Table 1). However, in contrast to the light scattering results, the differences between the samples containing A1 were not significant. The association constants show that the MuPrP:D67 interaction has the highest affinity of any treatment ( Table 1).
To prove that any quenching effect observed was related to the addition of cofactors, the PrP 90−231 fluorescence was evaluated with the addition of only reaction medium (Supplementary Figure 3). The fluorescence was not suppressed in a manner consistent with the observations for cofactors, GAGs or DNA.

Interaction Between Full-Length PrP and DNA Oligonucleotides
In addition to the nucleic acid binding sites also present in PrP 90−231 , PrP 23−231 has an additional predicted binding site in the lysine-rich region (residues 23-27) located at the extreme N-terminus (Yin et al., 2008;Cavaliere et al., 2013). We therefore sought to assess whether this region would change the interactions with the two DNA sequences.
Depending on the aptamer used, we observed a 2to 9-fold increase in the relative light scattering in the lowest stoichiometric ratio (Figure 10). In contrast to the results for PrP 90−231 , the light scattering of RaPrP 23−231 with both oligonucleotides at pH 5.5 had a slightly higher scattering than the MuPrP 23−231 samples (Figure 10). The profile for the two DNA sequences was similar to the observations for PrP 90−231 : A1 induced more aggregation than D67. Another difference between PrP 23−231 and PrP 90−231 was that the neutral pH treatments showed a higher relative light scattering for MuPrP (Figures 10A,C).
Similar to the findings for the PrP 90−231 C-terminal extended protein, the interaction between DNA and the full-length PrP caused fluorescence suppression in every treatment analyzed (Figure 11). As expected, since PrP 23−231 has more nucleic acid binding sites, the association constants were higher than those found for PrP 90−231 (Table 1). Surprisingly, the interaction that showed the greatest association constant was between RaPrP and D67 at pH 7.4 (Table 1), which had the lowest light scattering increase ( Figure 10D).

DISCUSSION
In this study, we compared the interaction of recombinant murine and rabbit PrP with different cofactors that have been identified as adjuvant molecules for conversion and aggregation of PrP, namely heparin, dermatan sulfate, PA, and DNA oligonucleotides (Tsiroulnikov et al., 2009;Silva et al., 2010). We show that RaPrP is less susceptible to aggregation by any of the cofactors tested, suggesting that the limited effect of these molecules may be associated with greater resistance to prion protein conversion observed in rabbits (Fernandez-Funez et al., 2011). Recombinant rabbit PrP c exhibits different electrostatic potential distribution than MuPrP, which has a larger positively charged surface (Wen et al., 2010a;Silva et al., 2011;Matos et al., 2020). This may affect interactions with negatively charged cofactors and, therefore, likely vary with pH. We found that all tested cofactors induce both MuPrP and RaPrP aggregation, and this effect does not correlate with differences on biding affinity, suggesting that the positive charge area found on RaPrP, covering residues 125-135, 150-160, and 180-190 (Wen et al., 2010a), is not related to RaPrP resistance to cofactor-induced aggregation.
The N-terminus of PrP does not seem to interfere with conversion to PrP Sc -like species, since truncated PrP can be converted, even if with lower efficiency (Lawson et al., 2001). However, this region is markedly different in rabbits (Figure 1; Myers et al., 2020), and is the main site for interaction with FIGURE 6 | PA-induced structural change is enhanced at acid pH but to a lesser extent for rabbit PrP. Effect of increasing the concentration of PA on the center of mass values obtained from the fluorescence spectra of MuPrP 23−231 (A), RaPrP 23−231 (B), MuPrP 90−231 (C), and RaPrP 90−231 (D) at pH 5.5 (red) and 7.4 (blue). All proteins were analyzed at 2 µM. The experiments were performed in 50 mM Tris buffer containing 100 mM NaCl at pH 7.4 or in 20 mM sodium acetate buffer containing 100 mM NaCl at pH 5.5. All experiments were performed at 25 • C. The error bars represent the standard deviation of at least two independent experiments. many of the studied cofactors . We observed the same effect for PrP 23−231 and PrP 90−231 for both MuPrP and RaPrP, although the overall interaction with cofactors was pronounced in the case of the full-length protein, suggesting that the N-terminal domain is determinant for interaction with cofactors resulting in an increased conversion efficiency, but not related to cofactor-induced aggregation resistance. Horse PrP has four saline bridges (GLU196 -ARG156-HIS187, ARG156-ASP202, and GLU211-HIS177). RaPrP has a strong ASP177-ARG163 saline bridge, which keeps the β2-α2 loop attached (Zhang, 2011;Zhang and Zhang, 2014). ASP201-ARG155, which connects helices H3 and H1, makes two saline bridges in RaPrP, while MuPrP has only one (Figure 1). All these saline bridges contribute to the increased stability in RaPrP, especially at neutral pH, since losing these salt bridges at low pH reduces thermostability (Zhang, 2011;Zhang and Zhang, 2014). This effect of pH was consistent across the cofactors studied. However, even at low pH, aggregation of RaPrP was lower than that of the murine protein, suggesting that other interactions may be important.
Rabbit PrP SER173 forms a double hydrogen bond, forming a helix-capping motif, which decreases the tendency of RaPrP to populate a β-structured state at low pH . This may also contribute to RaPrP being less prone to cofactor-mediated aggregation, as observed here. The interaction site of low molecular weight heparin in the globular region of PrP is close to HIS186 . Rabbit PrP shows a less dynamic hydrogen bond between HIS185 and ARG155 than murine PrP at neutral pH (Figure 1; Zhang, 2010Zhang, , 2011. This HIS residue shows reduced pKa (∼5) in different species, and its protonation affects PrP stability. The HIS186ARG mutation, which introduces a positively charged residue, is correlated with familial CJD and destabilizes the murine protein (Singh and Udgaonkar, 2016). This more stable hydrogen bond may hinder the aggregation induced by low molecular weight heparin and other GAGs, although it does not prevent interaction due to residue protonation. The same must be important for the other studied cofactors. The introduction of a positive charge (due to protonation or residue mutation) should favor the interaction with negatively charged cofactors, such as those tested in this study, but the effect on RaPrP structural conversion and aggregation should be limited.
We observed some differences in the effects of the various cofactors tested. The aggregation profile of PrP in the presence of A1 and D67 was distinct from that in the presence of GAGs or PA. This may be due to differences in the regions of interaction, dissimilar charge and polarity, and/or even in the folding of these molecules. Intriguingly, the overall effect was identical.
We also evaluated the effect of a 21-mer double-stranded and of a 25-mer single-stranded DNA oligonucleotide (D67 and A1, respectively), on PrP aggregation. A1 induced higher aggregation than D67, possibly due to differences in binding affinity and/or DNA conformation (Macedo et al., 2012;M. Passos et al., 2021). Due to secondary structure, length, and sequence, nucleic acids may have different effects on PrP, in the extent of aggregation and toxicity of the aggregated species, as previously shown (Gomes et al., 2008;Macedo et al., 2012).
Interaction with PA induces PrP aggregation, but in a different manner than observed with GAGs and nucleic acids. The changes induced by PA led to a blue shift and increased fluorescence, suggesting more pronounced structural changes. PA interacted with full-length and truncated PrP, with a fivefold higher affinity at pH 7.4, contrasting with observations with phosphatidylserine, another anionic lipid, which actively interacted with human PrP 20−231 at pH 5.0 but showed no interaction with human PrP 90−231 (Morillas et al., 1999). The negative head group in PA may interact with positively charged amino acids. The N-terminal domain of PrP has two positively charged clusters, one between residues 23 and 30 and the other between residues 101 and 110, both of which are candidate sites for the interaction with PA, possibly explaining the differences observed with the two PrP constructs (fulllength and truncated). Interestingly, dipolar phospholipids have also been shown to interact with and induce PrP aggregation (Kazlauskaite et al., 2003;Tsiroulnikov et al., 2009), indicating that the effect may not be necessarily related to the charge of the polar head group.
The present study is the first report of the direct interaction in vitro of dermatan sulfate with PrP. Dermatan sulfate is a GAGs found in the central nervous system, important for brain physiology, but is also involved in diseases that involve deficiency in the degradation of GAGs and consequent accumulation in endocytic vesicles -such as ataxia, intellectual disabilities, spasticity and other neurological symptoms are found in patients with mucopolysaccharidoses (Rauch and Kappler, 2006;Schwartz and Domowicz, 2018). Dermatan sulfate enhances the efficiency of PrP conversion by PMCA (Yokoyama et al., 2011), but there is no information available on its interaction with PrP and on the importance of this interaction for prion replication in vivo.
Our data also show that the interaction of PrP with dermatan sulfate leads to a similar effect to that observed with heparin, although with lower affinity and reduced aggregation. The interaction was not observed at pH 7.4, suggesting that structural differences between these GAGs are important at a neutral pH. Low molecular weight heparin binds to the octarepeat region of PrP at pH 7.4, and to helices H2 and H3 at pH 5.5 . The fact that interactions with dermatan sulfate only took place at pH 5.5, with a higher affinity for PrP 23−231 , suggests that N-and C-terminal domains must be available to allow interaction. Interestingly, the differences observed for dermatan sulfate may be related to a relevant role in prion pathology.

CONCLUSION
In summary, our findings demonstrate that most cofactors induce milder effects on RaPrP. Aggregation of RaPrP is weaker than that of MuPrP, specifically at acidic pH, suggesting that the mechanism involved in resistance to cofactors may be linked to RaPrP conformational/structural stability at low pH, and less to the physicochemical characteristics of the cofactors. The decreased effect of cofactors contributes to a better understanding of PrP conversion mechanisms and susceptibility among different mammalian species.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

AUTHOR CONTRIBUTIONS
JA and YP: data collection and analysis and manuscript writing. JB: data collection. JS: data analysis and study design. YC: data analysis, study design, and manuscript writing. TV: data collection and analysis, study design, and manuscript writing. All authors read and approved the final manuscript.