On-line aptamer affinity solid-phase extraction capillary electrophoresis-mass spectrometry for the analysis of blood α-synuclein.

In this paper, an on-line aptamer affinity solid-phase extraction capillary electrophoresis-mass spectrometry method is described for the purification, preconcentration, separation, and characterization of α-synuclein (α-syn) in blood at the intact protein level. A single-stranded DNA aptamer is used to bind with high affinity and selectivity α-syn, which is a major component of Lewy bodies, the typical aggregated protein deposits found in Parkinson's disease (PD). Under the conditions optimized with recombinant α-syn, repeatability (2.1 and 5.4% percent relative standard deviation for migration times and peak areas, respectively) and microcartridge lifetime (around 20 analyses/microcartridge) were good, the method was linear between 0.5 and 10 µg•mL-1 and limit of detection was 0.2 µg•mL-1 (100 times lower than by CE-MS, 20 µg•mL-1). The method was subsequently applied to the analysis of endogenous α-syn from red blood cells lysate of healthy controls and PD patients.

propan-2-ol/water with 0.05% (v/v) of HFor and was delivered at a flow rate of 3.3 µL·min -1 by a KD Scientific 100 series infusion pump (Holliston, MA, USA).

Recombinant human α-syn expressed in Escherichia coli was purchased from
Analytik Jena (Jena, Germany). The solution provided by the manufacturer (5000 µg·mL -1 in phosphate buffered saline (PBS)) was aliquoted and stored in a freezer at -20ºC. Aliquots were thawed before use and working standard solutions were prepared by dilution in water. These solutions were stored in the fridge at 5ºC when not in use.
Human blood samples from patients were provided by the Basque Biobank/Biodonostia Node (www.biobancovasco.org). Samples were processed following standard operation procedures with appropriate approval of the Ethical and Scientific Committees. Three healthy donor blood samples and three PD patient blood samples (one at stage III and two at stage IV of the disease) were analyzed. All the samples corresponded to males and females aged between 60 and 80 years.

Pretreatments of red blood cells lysates
Red blood cells (RBCs) lysates were prepared from blood as described in the Supporting Information 34 .
The RBCs lysates were precipitated with ethanol-chloroform to deplete hemoglobin as described elsewhere 34 , with some changes: 350 µL of cold ethanol and 200 µL of cold chloroform were added to 250 µL of RBCs lysate. The mixture was shaken for 5 min at 4ºC and centrifuged at 3000 g for 10 min at 4ºC. The supernatant was collected and low Mr compounds were removed with 10,000 Mr cut-off (MWCO) cellulose acetate centrifugal filters (Amicon Ultra-0.5, Millipore).
Thermo-enrichment to deplete the non-thermostable proteins was performed on the RBCs lysates as following 31 : 350 µL of RBCs lysate were heated at 90ºC for 10 min in a thermoshaker. The mixture was centrifuged at 12000 g for 5 min at 4ºC and the supernatant (i.e. thermo-enriched (TE) RBCs lysate) was filtered using a 0.22 µm polyvinylidene difluoride centrifugal filter (Ultrafree-MC, Millipore, Bedford, MA, USA) at 12000 g for 5 min.

CE-MS
Fused silica capillaries were supplied by Polymicro Technologies (Phoenix, AZ, USA). All CE-MS experiments were performed in a 7100 CE coupled with an orthogonal G1603A sheath-flow interface to a 6220 oa-TOF LC/MS spectrometer (Agilent Technologies, Waldbronn, Germany). ChemStation and MassHunter softwares (Agilent Technologies) were used for the CE and TOF mass spectrometer control, data acquisition and processing. The TOF mass spectrometer was operated in ESI+ mode and the optimized parameters are presented in the Supporting Information.
Separations were performed at 25°C in a 72 cm long (LT) × 75 µm i.d. × 365 µm o.d. capillary. All capillary rinses were performed flushing at 930 mbar. For new capillaries or between workdays, the capillaries were flushed off-line with 1 M NaOH (15 or 5 min, respectively), water (15 or 10 min), and BGE (30 or 15 min) to avoid the unnecessary contamination of the MS system. Samples were hydrodynamically injected at 50 mbar for 10 s (54 nL, i.e. 1.7% of the capillary, estimated using the Hagen-Poiseuille equation 35 ), ,and a separation voltage of +25 kV (normal polarity, cathode in the outlet) was applied. The autosampler was kept at 10°C using an external water bath (Minichiller 300, Peter Huber Kältemaschinenbau AG, Offenburg, Germany). Between runs, the capillary was conditioned flushing with water (2 min) and BGE (2 min).

AA-SPE-CE-MS
AA-MBs were prepared following the manufacturer recommendations. small volume of eluent with 100 mM NH4OH (pH 11.2) was injected at 50 mbar for 20 s (100 nL 35 ). For a rapid and repeatable protein elution, the small plug of eluent was pushed with BGE at 50 mbar for 100 s, before applying the separation voltage (+25 kV) and a small pressure (25 mbar) to compensate for the microcartridge counter-pressure.
Between consecutive runs, to avoid carry-over, the capillary was flushed with water for

Quality parameters
The details regarding the limit of detection (LOD), limit of quantification (LOQ), repeatability of migration time and peak area, linearity, and microcartridge lifetime in CE-MS and AA-SPE-CE-MS are given in the Supporting Information.

LC-Orbitrap-MS/MS
The details about TE RBCs lysate bottom-up proteomics workflow including LC-Orbitrap-MS/MS analysis are given in the Supporting Information.

CE-MS
In general, the best results for the analysis of intact proteins by CE-MS in ESI+ are obtained using acidic volatile BGEs and sheath liquids because protein ionization is maximized and the best sensitivity is achieved. Different conditions were tested for the analysis of recombinant human α-syn (i.e. 50 mM HAc: 50 mM HFor (pH 2.3); 100 mM HAc (pH 2.9) or 10 mM NH4Ac (pH 7.0, 8.0 or 9.0) as BGEs combined with 60:40 (v/v) propan-2-ol/water with 0.05 or 0.25% (v/v) of HFor as sheath liquids). The best results for the analysis of α-syn were obtained with a BGE of 100 mM HAc (pH 2.9) and a sheath liquid of 60:40 (v/v) propan-2-ol/water with 0.05% (v/v) of HFor. As an example, Figure 1 shows the extracted ion electropherogram (EIE) (A), mass spectrum (B) and deconvoluted mass spectrum (C) for the CE-MS analysis of a 100 µg·mL -1 standard solution of recombinant human α-syn in the optimized conditions. The only detected proteoform was free α-syn because the recombinant human α-syn expressed in E. coli was not supposed to undergo post-translational modifications (PTMs) during bacterial synthesis (the minor peaks in the deconvoluted mass spectrum of Figure 1-C are mostly due to Na + and K+ adducts to the ion species of the mass spectrum of Figure   1-B). Table S-1 shows the theoretical average Mr of free α-syn and the relative error (Er) for the experimental deconvoluted average Mr. Mass accuracy was good (Er<20 ppm).
Under the optimized conditions, consecutive analyses of the α-syn standard were repeatable in terms of migration time and peak area (%RSD (n=3) were 0.9 and 5.6% at 100 µg·mL -1 ). The LOD was 20 µg·mL -1 , better than the 50 µg·mL -1 LOD by CE-MS using a BGE of 10 mM NH4Ac (pH 7.0) and a sheath liquid of 60:40 (v/v) propan-2ol/water with 0.25% (v/v) of HFor. These latter conditions are typically required to analyze proteins by IA-SPE-CE-MS 12,13 .

AA-SPE-CE-MS optimization
The DNA aptamer M5-15 was selected to prepare the aptamer affinity (AA) sorbent because it has been described to selectively bind with high affinity to α-syn monomer 32 .
In contrast to our previous studies by IA-SPE-CE-MS 12,13 , it was observed that in AA-SPE-CE-MS the AA sorbent was stable even using acidic BGEs. Sensitivity and repeatability were investigated with different combinations of BGE and sheath liquid (see below), using a basic volatile eluent of 100 mM NH4OH (pH 11.2) 12,13 . Using a sheath liquid of 60:40 (v/v) propan-2-ol/water with 0.05% (v/v) of HFor, a BGE of 100 mM HAc (pH 2.9) allowed obtaining similar peak areas to a BGE of 50 mM HAc: 50 mM HFor (pH 2.3). However, in the second case, a slight decrease of peak areas was detected after consecutive injections, probably because of AA sorbent deterioration due to the lower pH value. Using the basic eluent, the acidic BGEs provided higher sensitivity than the BGEs of 10 mM NH4Ac at pH 7.0, 8.0 or 9.0, even with a sheath liquid of 60:40 (v/v) propan-2-ol/water with an increased amount of HFor (i.e. 0.25% (v/v) of HFor). To confirm that no analyte was eluted under acidic conditions, BGEs of detected. Therefore, the BGE of 100 mM HAc (pH 2.9) was selected for the rest of the experiments.
The investigation of the volatile eluents was extended, testing different hydroorganic mixtures in the presence or absence of 100 mM NH4OH. First, acetonitrile:water mixtures at 40% and 80% (v/v) were tested to disrupt analyte-aptamer interaction 14,23 . However, these acetonitrile:water eluents were rapidly discarded because electropherograms and mass spectra were extremely poor, probably due to sorbent deterioration during the elution. Results with 60% (v/v) MeOH in the presence or absence of 100 mM NH4OH were neither satisfactory as shown in Figure 2A. The best sensitivity and repeatability of peak areas and migration times were obtained with the aqueous basic eluent of 100 mM NH4OH (pH 11.2). A higher concentration than 100 mM of NH4OH in the aqueous eluent was not tested to prevent aptamer denaturation and expand the sorbent lifetime. The volume of the eluent plug was investigated injecting the eluent at 50 mbar for 10, 20 and 40 s (50, 100 and 200 nL 35 ).
A higher amount of α-syn was detected injecting the eluent 20 s instead of 10 s.
However, for the 40 s eluent injection, the analyte peak broadened, and peak area decreased. Between consecutive runs of a 10 µg·mL -1 α-syn standard solution, a small amount of α-syn was detected as carry-over when only rinsing with water between injections. Therefore, to prevent carry-over, the capillary was rinsed with water, a small plug of eluent, and again water between injections.
With regard to the sample loading, the standard solutions were prepared in water because lower peak areas were observed when using PBS, probably due to the smaller binding efficiency in a salty environment. The sample loading time was studied introducing a 1 µg·mL -1 α-syn standard solution at 930 mbar from 3 to 15 min. As can be seen in Figure 2B, the maximum amount of α-syn was detected at around 5 min.
When loading for a longer time, the sample breakthrough volume was exceeded and the α-syn washed away was higher than the amount retained, causing a significant decrease of peak area. Therefore, to reduce the total analysis time and to obtain the highest recoveries, a sample loading time of 5 min was selected for the rest of the experiments.
Under the optimized conditions, consecutive analyses of the standard were repeatable in terms of migration time and peak area. At 1 µg·mL -1 , the %RSDs (n=3) were 2.1 and 5.4%, respectively, similar to the values in CE-MS. As can be seen in Figure 2C, the method was satisfactorily linear (R 2 >0.994) between 0.5 and 10 µg·mL -1 .
The LOQ was 0.5 µg·mL -1 and the LOD was 0.2 µg·mL -1 , which was an improvement of about 100 times compared to the CE-MS method. The lifetime of the microcartridges was around 20 analyses (at 1 µg·mL -1 ). As an example, Figure 1A shows the AA-SPE-CE-MS analysis of a 1 µg·mL -1 α-syn standard. Compared to CE-MS ( Figures 1B, 1C and Table S-1), the mass spectrum and mass accuracy for experimental deconvoluted average Mr of free α-syn was similar (data not shown).

Analysis of α-syn in blood samples
The AA-SPE-CE-MS method optimized with standards was applied to the analysis of blood. The concentration of α-syn has been shown to be higher in blood than in cerebrospinal fluid (CSF) and, furthermore, drawing blood is less invasive than lumbar puncture to obtain CSF from patients 29,30 . More than 99% of the α-syn in human blood resides in the red blood cells (RBCs) 37 . N-terminal acetylation of proteins is a common occurrence, especially for those proteins initiated in Met residue, and this PTM of α-syn is the most abundant in blood and also in brain cytosol 26 .  Table S-2). These results were consistent with the fact that ethanol-chloroform extraction has been also historically used to achieve quantitative removal of Hb in purification of carbonic anhydrases 39 . However, an alternative purification method was necessary for α-syn.
Taking into account that α-syn is a thermostable protein, and the solubility and the PTMs pattern of α-syn is not supposed to be altered upon heating 27,31,40  and to obtain separate EIEs ( Figure 4A). With regard to the migration time increase detected in TE RBCs lysates, it was probably due to the modification of the inner wall of the separation capillary induced during sample loading by the non-retained components of the complex sample matrix. This modification happened after the first analysis with a new microcartridge and was permanent, because repeatability was high (see the quality parameters below). Another proof of this permanent modification was that free α-syn was also detected at this increased migration time when a standard α-syn solution was analyzed after a TE RBCs lysate sample.
As can be also observed in Figures 3 and 4, N-acetylated α-syn slightly comigrated with ubiquitin, which was also retained by the AA sorbent. In addition to ubiquitin, very small amounts of apolipoprotein A-I were detected (Figure S-4). Several experiments were made to investigate non-specific adsorption on the AA sorbent. It is worth highlighting that compared to the typical aptamer-or antibody-based biosensors or bioassays, a great advantage of AA-SPE-CE-MS is that the electrophoretic separation and the selectivity of the MS detection prevent the possibility of a false positive or an erroneous quantification of the target protein in the presence of non-specific adsorption.
A microcartridge containing blank sorbent (i.e. activated and endcapped MBs, without aptamer) was tested by SPE-CE-MS. When loading a 1 µg·mL -1 α-syn standard solution a very small amount of α-syn was detected (6.4% compared to the AA sorbent). This suggested that non-specific adsorption of α-syn was very limited because it was not efficiently retained by the endcapped sorbent without aptamer, hence also confirming that the aptamer was required for the binding of α-syn. In contrast, when the TE RBCs

S-31
Supporting Information

Table of contents Preparation of RBCs lysates S-32
Optimized MS parameters S-32 Quality parameters S-32  S-32 was collected and the membranes were discarded.

Optimized MS parameters
The TOF-MS parameters were optimized analyzing by CE-MS a 100 µg·mL -1 αsyn standard solution: capillary voltage 4000 V, drying gas temperature 300 ºC, drying gas flow rate 4 L·min -1 , nebulizer gas 7 psig, fragmentor voltage 325 V, skimmer voltage 80 V, OCT 1 RF Vpp voltage 300 V. Data were collected in profile at 1 spectrum/s between 100 and 3200 m/z, with the mass range set to high resolution mode (4 GHz).

Quality parameters
All quality parameters were calculated from data obtained by measuring

S-35
Six digested TE RBCs lysate samples were analyzed with two replicates of each sample. Proteins identified with "low" or "medium" protein confidence values or detected in less than 9 of the 12 analyses were not reported. a) Relative error (Er) was calculated in ppm as: (Mr exp -Mr theo)/Mr theo × 10 6 (exp = experimental and theo = theoretical). Mr exp was obtained as an average of three replicates. 14,502.14 -6 X a) Relative error (Er) was calculated in ppm as: (Mr exp -Mr theo)/Mr theo × 10 6 (exp = experimental and theo = theoretical). Mr exp was obtained as an average of three replicates.

S-37
b) This Uniprot accession number corresponds to ubiquitin-40S ribosomal protein S27a and was assigned taking into account the results of the LC-Orbitrap-MS/MS analysis (see Table S-3).