Modified Protein Misfolding Cyclic Amplification Overcomes Real-Time Quaking-Induced Conversion Assay Inhibitors in Deer Saliva To Detect Chronic Wasting Disease Prions

Saliva samples.

White-tailed deer were sourced from the University of Georgia Warnell School of Forestry and Natural Resources (a region where CWD is not endemic; see Acknowledgments) and were housed in an indoor CWD research facility at Colorado State University. We maintained all deer in accordance with protocols approved by the Colorado State University Institutional Animal Care and Use Committee. Saliva was obtained from anesthetized deer by inserting a 3-ml syringe into the cheek pouch closest to the ground and aspirating. Samples were aliquoted and frozen at −80°C until testing. Deer were inoculated as follows: deer 121, 1.0 g of CWD⁺ mule deer brain intracerebrally; deer 139, 0.25 g of CWD⁺ transgenic 1536 mouse brain intravenously (32); deer 142, 0.25 g of CWD⁺ transgenic 1536 mouse brain intracerebrally (32); deer 372, 250 ml of whole blood from CWD⁺ deer intravenously (10, 33); deer 4116, 300 ml of whole blood from CWD⁺ deer intravenously (10, 33); deer 777, 783, and 786, 1.0 g of CWD⁺ white-tailed deer brain per os (52); and deer 813, 815, 816, and 817, 0.1 g of CWD⁺ white-tailed deer brain via aerosolization (34).

Western blotting and Coomassie stain.

We pooled replicate RT-QuIC wells and combined 20 μl of RT-QuIC product or saliva with 20 μl of 2× Laemmli sample buffer (LSB)–β-mercaptoethanol (β-ME) (Novagen) and boiled the samples for 5 min. We loaded 30 μl of each sample to a 12-well, 12% bis-Tris, precast polyacrylamide gel (Novagen) and electrophoresed it at 150 V for 1 h. For Coomassie-stained gels, we added Coomassie stain (0.25% [wt/vol] Coomassie brilliant blue R250, 7.5% [vol/vol] glacial acetic acid, 50% [vol/vol] methanol) to the gel and rocked it at room temperature for 2 h, then removed the stain, rinsed the gel, and destained (7.5% [vol/vol] glacial acetic acid, 50% [vol/vol] methanol) it for 2 h at room temperature. For Western blots, we transferred the proteins to polyvinylidene fluoride (PVDF) membranes for 1 h at 80 V on ice. We blocked the membranes in 5% nonfat dry milk and then exposed the membranes to anti-PrP antibody Bar224 (Cayman Chemicals) overnight at 4°C, followed by secondary antibody (goat anti-mouse IgG_2a-horseradish peroxidase [HRP] [SeraCare]) for 1 h at room temperature. We developed the Western blots with Pierce ECL substrate (Thermo Scientific). For both Western blots and Coomassie-stained gels, we collected images with an ImageQuant LS (GE).

Inhibitory saliva pools.

We made a large pool of CWD⁻ saliva to be used for the experiments in this study. The pool was comprised of 1 ml of baseline saliva samples (preinoculation) from each of the following deer: 1201, 1204, 1205, 1211, and 1218.

Protease treatment.

We treated saliva samples with proteinase K (PK; final concentration of 4 μg/ml) at 45°C for 1 h at 900 rpm. We added 10 mM phenylmethylsulfonyl fluoride (PMSF) to each sample and boiled the samples for 5 min to inactivate the PK. Next, we spiked 1 μl of 0.01% brain homogenate into 9 μl of treated saliva.

Protease inhibitor treatment.

To test whether the proteases present in saliva were destroying the recombinant PrP (rPrP) upon which the RT-QuIC assay relies, we preincubated saliva with rPrP, with or without protease inhibitors. We substituted one-third of the water in the RT-QuIC mastermix (1 mM EDTA, 0.1 mg/ml of rPrP, 320 mM NaCl, 10 μM thioflavin T [ThT], H₂O) with saliva (for a final concentration of 20% [vol/vol]). We added protease inhibitor cocktail (EDTA-free protease inhibitor cocktail; Thermo Scientific; 78437) and incubated the mixture for 1 h, 2 h, 6 h, or overnight at 42°C with 1 min of shaking (500 rpm) alternating with 1 min of rest. We used these samples as the substrate for RT-QuIC or for electrophoresis and then Western blotting or Coomassie staining.

Mucin treatment.

We made fresh mucin daily (25 mg/ml of bovine submaxillary mucin [Sigma] in 10 mM HEPES buffer, pH 7.6) and diluted it in phosphate-buffered saline (PBS) or saliva electrolytes for final concentrations of 0.5 mg/ml, 1.25 mg/ml, 2.5 mg/ml, 5.0 mg/ml, 6.25 mg/ml, and 12.5 mg/ml. The use of 10 mM HEPES as a diluent for bovine submaxillary mucin was published previously (35).

Saliva electrolyte treatment.

We created a solution with the electrolyte concentrations of ruminant saliva based on a 1948 report by McDougall (36). The solution composition is as follows: 117 mM NaHCO₃, 26 mM Na₂HPO₄·12H₂O, 8 mM NaCl, 8 mM KCl, 0.2 mM CaCl₂ anhydrous, 0.3 mM MgCl₂ anhydrous. We dissolved these salts in deionized water and filtered the solution. We added 2% SDS for a final concentration of 0.1% SDS and diluted brain homogenate into the SDS-electrolyte solution.

Saliva screen.

To determine the frequency with which negative saliva samples contain inhibitors of RT-QuIC, we devised a screening protocol. First, we diluted positive brain to a final concentration of 0.01% or 0.001% in saliva and 0.01% SDS and tested for inhibition in triplicate in RT-QuIC. To test the dilutional response of the inhibitor, we added 0.001% positive brain in 30% saliva or 60% saliva, added SDS for a final concentration of 0.01%, and added the sample to triplicate wells in RT-QuIC.

RT-QuIC.

We purified Syrian hamster rPrP (codons 90 to 231) as previously described (37). We expressed the cDNA in E. coli BL21-Star cells overnight under the following inducing conditions: LB medium, kanamycin and chloramphenicol, 20× NPS [0.5 M (NH₄)₂SO₄, 1 M KH₂PO₄, 1 M Na₂HPO₄], 50× 5052 [0.5% glycerol, 0.05% glucose, 0.2% lactose], and 1 mM MgSO₄. We harvested inclusion bodies with lysozyme (0.25 mg/ml), DNase (1 μg/ml), and MgCl₂ (5 mM) in 1× Bugbuster (Novagen) by following the manufacturer's protocol (Novagen). We solubilized inclusion body pellets in 8 M guanidine hydrochloride (GdnHCl) overnight and then bound the solubilized protein to Ni-agarose resin, rotating at room temperature for 45 min (GE Healthcare Life Sciences). We refolded the rPrP with a gradient from 6 M GdnHCl in 100 mM Na₂HPO₄ and 10 mM Tris to the same buffer with no GdnHCl on a GE fast-performance liquid chromatograph (FPLC; AktaPure; GE Healthcare Life Sciences). We eluted with a gradient from 0.0 M to 0.5 M imidazole and dialyzed the rPrP in two changes of 20 mM NaH₂PO₄ overnight and then it stored at 4°C.

We added 0.1 mg/ml of rPrP to 10 μM ThT, 320 mM NaCl, 1 mM EDTA, and 1× PBS in 1 well of a 96-well plate (Nunc black, optical-bottom, 96-well plates; Thermo Scientific). Sample preparation was variable, depending on the treatment, and is described in the previous sections, but we always added 2 μl of the sample to each well. We either prepared the sample in 0.1% SDS or added 2 μl of 0.1% SDS to the plate separately. For controls, we included brain (positive and negative) diluted into PBS and into inhibitory saliva on every plate. We used a BMG Fluostar Omega microplate reader to shake the plates for 1 min (double orbital shaking at 700 rpm), followed by 1 min of rest, for 62.5 h at 42°C. The fluorescence was recorded every 15 min with a 450-nm excitation wavelength, a 480-nm emission wavelength, and a gain of 1,700.

To test PMCA products by RT-QuIC, we added 0.09 mg/ml of rPrP to 10 μM ThT, 320 mM NaCl, 1 mM EDTA, and 1× PBS in 1 well of a 96-well plate. We diluted PMCA products 1,000-fold in 0.1% SDS–1× PBS and then added 2 μl of PMCA product to each well.

The RT-QuIC data are displayed as lag phases. To calculate the lag phase, we established a threshold fluorescence (5 standard deviations above the average baseline fluorescence on the plate) and determined the time that the fluorescence signal crosses the threshold (lag phase). For samples that did not cross the threshold during the 62.5-h experiment, we assigned a lag phase of 70 h.

Sonication.

We added saliva (undiluted or diluted) to 0.2-ml tubes and sonicated the tubes in a horn sonicator (QSonica LLC) in a 37°C incubator. We used a program consisting of 29.5 min of rest, followed by 30 s of sonication with an amplitude of 30. For an experiment lasting 72 h, the samples were typically exposed to ∼1,000,000 J.

PMCA.

Our PMCA reaction was adapted from previous publications (9, 22, 38). We used a substrate of 10% (wt/vol) brain homogenate from transgenic mice that overexpress elk PrP^C (Tg5037 mice [39]), homogenized in 1× PBS, 150 mM NaCl, and 1% Triton X-100. Brain homogenate was stored at −80°C until use. We seeded 90 μl of brain homogenate substrate with 10 μl of saliva in 0.2-ml Eppendorf tubes containing three 1.59-mm-diameter and two 2.38-mm-diameter polytetrafluoroethylene beads. In the first round, samples were subjected to cycles of 29.5 min of rest followed by 30 s of pulse sonication for 72 h in a Qsonica microplate horn sonicator housed within a 37°C incubator and received energy levels of approximately 1,000,000 J. All subsequent rounds used 50 μl of 10% brain homogenate as the substrate and 20 μl of PMCA product from the previous round. Rounds 2 to 5 were also subjected to cycles of 29.5 min of rest and 30 s of pulse sonication for 24 h, receiving energy levels of approximately 300,000 J. PMCA reactions were continued for 5 rounds. A temperature probe measuring the water temperature of the sonicator indicated that the operating temperature was between 48 and 50°C during the experiment. PMCA samples were evaluated for PrP^Sc amplification by Western blotting or RT-QuIC.

Western blot detection of PMCA products.

First, 10-μl volumes of PMCA products (10% brain homogenate) were digested with 1 μl of solution containing 50 μg of proteinase K (Invitrogen) and incubated at 45°C for 1 h with shaking. Following PK digestion, samples were mixed with a reducing agent (10× LDS–4× sample buffer; Invitrogen) at a final concentration of 1× and boiled for 10 min. A total of 15 μl of sample was run on a 26-well 12% bis-Tris gel (Bio-Rad) at 125 V for 1.5 h. Proteins were transferred to a PVDF membrane at 80 V for 1 h on ice. The PVDF membrane was blocked in 5% nonfat dry milk for 1 h at room temperature and then incubated with anti-PrP antibody Bar224 (Cayman Chemicals) diluted in Tris-buffered saline–0.01% Tween (TBS-T) to 0.2 μg/ml overnight at 4°C, followed by goat anti-mouse secondary antibody conjugated to HRP (goat anti-mouse IgG_2a-HRP [SeraCare]) diluted in TBS-T to 0.4 μg/ml for 1 h at room temperature. We developed the Western blots with Pierce ECL substrate (Thermo Scientific) and captured images with an ImageQuant LS (GE).

Saliva samples often require dilution to seed RT-QuIC.

We previously demonstrated that prions in saliva are infectious and have used RT-QuIC to quantify prions in saliva and other excreta (5, 6, 10, 13, 19, 20, 38). In RT-QuIC experiments, some saliva samples appeared negative until they were diluted at least 10-fold (examples in Fig. 1A); this suggested that saliva contained an inhibitor of RT-QuIC which must be diluted to allow prion seeding.

First, we investigated how commonly saliva inhibited RT-QuIC. We performed a series of spiking experiments, in which CWD⁺ brain was diluted into 39 CWD⁻ saliva samples, and examined the ability of the mixture to seed RT-QuIC. Roughly 25% of the saliva samples completely abolished seeding when 200 ng of CWD⁺ brain homogenate was added. That percentage increased to nearly 50% when only 20 ng of CWD⁺ brain homogenate was used as the seed (Fig. 1B). Next, we added 200 ng of brain to serial dilutions of saliva in 1× PBS. More saliva samples completely inhibited RT-QuIC at higher saliva concentrations and less inhibition was observed at lower saliva concentrations, confirming a dilutional response (Fig. 1C). Even in samples that displayed some seeding activity, it was common for the lag phase (an indication of seeding efficiency) to be prolonged in the presence of increasing saliva concentrations (Fig. 1D). These data confirmed the presence of an RT-QuIC reaction inhibitor in saliva.

Saliva inhibits seeding activity in RT-QuIC.

We created a pool of inhibitory saliva from CWD-naive deer (n = 5) to ensure that we had a consistent sample of saliva for subsequent experiments. First, we confirmed that the CWD⁻ saliva pool inhibited RT-QuIC by repeating the CWD⁺ brain spiking experiments described above. We diluted CWD⁺ brain in 100% saliva plus 0.1% SDS or in PBS plus 0.1% SDS, a required component of the RT-QuIC assay. Seeding for all but the most concentrated dilution of brain homogenate was completely blocked by dilution in saliva but not PBS. The dilution containing the highest concentration of brain had an increased lag phase compared to that of the PBS-SDS control, confirming that our CWD⁻ saliva pool inhibited RT-QuIC even in the presence of high concentrations of brain prions (Fig. 2A).

Next, we repeated the saliva dilutional series by adding 200 ng of CWD⁺ brain to serial dilutions of saliva in 1× PBS. Again, there was a dilutional response to saliva inhibition; seeding activity was completely inhibited in the presence of 100% saliva and 60% saliva, delayed at 30% saliva, and uninhibited at 0% saliva (Fig. 2B). We also tested dilutions of CWD⁻ brain in PBS–0.1% SDS or saliva–0.1% SDS to ensure that our saliva pool did not cause spontaneous misfolding of the rPrP substrate (Fig. 2C). We did not observe any spontaneous misfolding under these conditions. Finally, to ensure that saliva did not inhibit RT-QuIC by rendering SDS ineffective, we examined the impact of increased SDS concentrations in the reaction (Fig. 2D). Increased SDS concentrations failed to rescue the inhibition by saliva.

Saliva contains proteases that degrade rPrP but do not affect the RT-QuIC reaction.

We hypothesized that the RT-QuIC inhibitor in saliva may damage one of the essential components of the RT-QuIC assay, especially rPrP or ThT. To test whether saliva impeded the RT-QuIC assay by degrading rPrP, we incubated saliva with rPrP for 1 h and assessed degradation by Western blotting. The addition of large concentrations of saliva (20 to 33% total volume, 10 to 15 times the volume normally used in an RT-QuIC experiment) resulted in degradation of the rPrP substrate, which confirmed that salivary proteases are active under typical experimental conditions (Fig. 3A). Not surprisingly, when we used the rPrP samples preincubated with saliva as the RT-QuIC substrate with a CWD⁺ brain seed, we observed no amyloid formation (Fig. 3B, QuIC products D to F).

To evaluate the effect of proteases in normal RT-QuIC conditions, we diluted CWD⁺ brain homogenate in the inhibitory saliva pool and added 2 μl to the RT-QuIC assay with fresh rPrP substrate (this constituted a final saliva concentration in the RT-QuIC reaction of approximately 2%). In this case, we observed the previously described inhibition of amyloid formation without substantial degradation of the rPrP (Fig. 3B, QuIC products A and B). Excess rPrP remained at the end of the experiment. These results supported the tenet that RT-QuIC inhibition was not due to degradation of rPrP; the concentration of salivary proteases was too low in the face of excess rPrP. To confirm this hypothesis, we treated saliva and rPrP with a protease inhibitor cocktail, sampled the mixture over a 16-h incubation, and then used the incubated samples as the substrate for the RT-QuIC assay. The addition of protease inhibitors reduced the minor degradation of rPrP by saliva proteases, but seeding was not rescued (Fig. 3C). We concluded that saliva did not inhibit RT-QuIC by degrading the rPrP substrate.

To assess whether saliva interfered with the function of ThT, which is required for detection of amyloid in the RT-QuIC assay, we compared the baseline fluorescence and the maximum fluorescence in positive wells with or without saliva (Table 1). Neither the baseline nor the maximum fluorescence values differed significantly among wells with and without saliva, suggesting that saliva had no effect on ThT function (P = 0.8229 and 0.119, respectively; two-sided, unpaired t test).

Proteinase K removes the RT-QuIC inhibitor in CWD⁻ saliva but only occasionally rescues seeding in CWD⁺ saliva samples.

Based on the high protein content of ruminant saliva, we suspected that the inhibitor may be a protein, so we treated our CWD⁻ saliva pool with several concentrations of PK (26). This produced changes in the total protein content in saliva samples treated with 1.0 to 5.0 μg/ml of PK (Fig. 4A). We tested the ability of PK-treated saliva to inhibit RT-QuIC (after PMSF treatment to stop the PK activity). Upon treatment with ≥1.0 μg/ml of PK, the inhibitory saliva pool was no longer inhibitory (Fig. 4B). We also PK treated PBS and stopped the reaction with PMSF to (i) ensure that we were removing all PK activity and (ii) ensure that PMSF did not alter the RT-QuIC reaction. PMSF eliminated all detectable PK activity and had no discernible effect on the RT-QuIC reaction. Based on the effects of PK on total protein composition (Fig. 4A), we examined the effect of 4 μg/ml of PK on the inhibitory saliva pool. PK treatment, but not sham treatment, rescued seeding (Fig. 4C).

We next examined the effect of 4 μg/ml of PK on individual potentially inhibited saliva samples. PK rescued seeding in two of the three saliva samples from CWD⁺ deer that are normally inhibited in RT-QuIC (Fig. 4D). For the sample (121) in which seeding was not rescued by PK, we hypothesized that if the inhibitor remained unaffected by PK, dilution would result in positive seeding (as it did when sample 121 was not PK treated). However, dilution did not enable seeding, suggesting that the prions in sample 121 may have been less PK resistant than “typical” prions (Fig. 4D). Together, these data support the notion that the RT-QuIC inhibitor is a protein and imply that prions in some saliva samples may be relatively protease sensitive.

Electrolytes in saliva are compatible with RT-QuIC.

The major components of saliva besides proteins are electrolytes. We created a solution (117 mM NaHCO₃, 26 mM Na₂HPO₄·12H₂O, 8 mM NaCl, 8 mM KCl, 0.2 mM CaCl₂ anhydrous, 0.3 mM MgCl₂ anhydrous) to mimic the electrolyte milieu of ruminant saliva and diluted brain homogenate into the solution instead of saliva (36). This electrolyte solution did not inhibit RT-QuIC (Fig. 5A).

Mucin inhibits RT-QuIC.

The major protein component of saliva is the mucin family, a complex group of glycoproteins that provide the viscosity of saliva (40, 41). We compared several reported physiological concentrations of bovine submaxillary mucin with PBS as a diluent for CWD⁺ brain homogenate (28, 29, 40, 41). We observed a dose-dependent inhibition of RT-QuIC seeding activity with increasing mucin concentrations but not in the buffer-only controls (Fig. 5B and C). Interestingly, mucin inhibited the RT-QuIC assay at lower concentrations when combined with the saliva electrolyte solution described above (Fig. 5C). This is likely due to the effect of electrolytes on mucin structure; i.e., with increased ionic strength, mucin becomes more compact and may be a better “trap” for prions (42).

We tested the ability of PK to rescue the inhibition caused by mucin. Treatment with 4 μg/ml of PK removed the inhibition caused by 5 mg/ml of mucin, but the sham control was still inhibited (Fig. 5D). At low (but still physiologically relevant) mucin concentrations (≤2.5 mg/ml), the sham control (45°C for 1 h, followed by PMSF and boiling) was sufficient to eliminate the inhibition from mucin (Fig. 5D). Together, these data indicate that mucin inhibits the RT-QuIC assay and can be rescued by PK treatment.

Sonication removes RT-QuIC inhibitor in some CWD⁺ saliva samples but reduces detection specificity.

Because we suspected that mucin may be the inhibitor in the RT-QuIC assay, and that it may function by trapping prions, we sonicated the CWD⁻ saliva pool for 72 h to fragment mucin, then spiked in CWD⁺ brain, and added it to the RT-QuIC assay. Sonication of the CWD⁻ saliva removed inhibition (Fig. 6A, open circles). Encouraged by the results of sonication of the CWD⁻ saliva pool, we sonicated inhibited CWD⁺ saliva samples. We hypothesized that sonication would sufficiently damage the mucin matrix to liberate the prions in CWD⁺ saliva and enable seeding in RT-QuIC. For two of three samples tested, sonication for 72 h did enable seeding (Fig. 6A). In the third sample (142, not the same sample that resisted PK treatment), sonication did not relieve inhibition, and the sample still required dilution for seeding activity to be detected. Importantly, we also found that CWD⁻ saliva sonicated in parallel as assay negative controls was much more prone to induce spontaneous RT-QuIC seeding (Fig. 6B), which discouraged us from broader adoption of sonication alone to remove the inhibitor.

Modified PMCA protocol enables sensitive detection of CWD prions in saliva.

The effect of sonication on RT-QuIC inhibitors prompted us to test our inhibited saliva samples with PMCA, which uses sonication to break apart amyloid fibrils and form new reaction seeds (43). We used CWD⁺ saliva, which normally is inhibited in RT-QuIC, to seed transgenic mouse brain homogenate substrate in PMCA. In studies which detected BSE prions in saliva with PMCA, Murayama et al. suggested that a higher temperature and the subsequent increased energy output improved sensitivity (24, 25). We performed our PMCA reaction at 50°C with an average energy of 1,000,000 J in the first round over 3 days and 300,000 J in all subsequent 24-h rounds. PK-resistant prion-amplified material was detectable by WB at round 5 for all RT-QuIC-inhibited samples (Fig. 7A). We hypothesized that amplification would be consistently detectable before round 5 if we analyzed PMCA products by RT-QuIC. Indeed, we observed significant seeding activity by RT-QuIC in all saliva-seeded PMCA products starting at round 3, with lag phases ranging from 5 to 40 h. By comparison, Western blotting of the same round 3 PMCA products showed signal in only one of the three inhibited saliva samples. We detected seeding activity in the remaining saliva-seeded PMCA products at round 4, with lag phases less than 15 h (Fig. 7B). We did not observe any nonspecific RT-QuIC amplification in negative-control PMCA products (seeded with five individual CWD⁻ saliva samples). Therefore, saliva samples with inhibited seeding in RT-QuIC could successfully seed PMCA, and RT-QuIC analysis of PMCA products provided detection of saliva-seeded prion amplification earlier than did Western blotting.

Because the PMCA reaction involves dilution of saliva to seed the reaction and through subsequent rounds, we made the same dilution of saliva in normal brain homogenate that was used in round 1, but skipped the sonication and incubation, and used the dilution to seed RT-QuIC. None of the brain-diluted saliva samples displayed substantial seeding activity, indicating that the sonication and incubation components of PMCA are essential for the eventual detection of prions in saliva by RT-QuIC and that the effect is not due to dilution alone (Fig. 7C). Finally, we performed the full PMCA reaction with brain homogenate from a transgenic PrP^C knockout (KO) mouse (44, 45). The presence of brain (without PrP^C), sonication, time, and heat did not result in saliva seeding of RT-QuIC, which suggested that the combination of sonication, PrP^C substrate, and the PMCA reaction conditions was necessary to overcome the inhibition (Fig. 7D).

Prions in inhibited saliva amplify in modified PMCA.

After demonstrating the efficacy of PMCA with samples we have characterized and know contain both prion seeding activity and an RT-QuIC inhibitor, we analyzed samples whose status was less clear. We identified 11 saliva samples with the following characteristics: (i) samples that produced ≤25% positive replicates in RT-QuIC (which would not be considered positive) and (ii) RT-QuIC negative saliva samples from deer in which both earlier and later collected samples were positive (Fig. 8A). We tested whether dilution of the ambiguous saliva samples alone would enable detection by RT-QuIC, and neither a 1:10 nor a 1:100 dilution resulted in consistent detection of prions in the saliva samples (Fig. 8B and C). We found that for 9 of the 11 ambiguous samples, seeding activity was detected by PMCA (Fig. 8D). We did not observe any prion-seeded amplification in the CWD-negative control saliva (Fig. 8D). Therefore, PMCA increased prion detection sensitivity over that of RT-QuIC alone.