Metallic and magnetic nanoparticles provide versatile sensing platforms for biological and biomedical applications such as the detection and quantification of large biomolecules serving as diagnostic markers in human medicine. A new sandwich immuno-assay for a surface-enhanced Raman spectrometry (SERS) based determination in combination with the ID approach of tau-protein has been developed within this project.

The developed assay is used to separate the target analyte from the matrix with the help of magnetic iron oxide (Fe_xO_y) nanoparticles as well as to quantify the tau protein by utilisation of a sensitive SERS active marker coupled to gold nanoparticles. The linkage of both nanoparticles to the protein will be ensured by immunoreaction with specific antibodies (Figure 3). Due to the high Raman cross section of the marker 5,5'-disulfanediylbis(2-nitrobenzoic acid) (DTNB), this method is suitable for very small sample volumes and low analyte concentrations typically found in CSF samples for tau-protein while highest accuracy is achieved through the ID approach (Zakel et al. 2013). Therefore, an isotopically enriched form of the marker was used as perfect internal standard (spike). Due to the higher molecular weight of the spike, a specific Raman shift can be observed and used for quantification. For this quantification method a set of calibration mixtures is necessary containing different ratios of the assay in its natural and its isotopic enriched form. Figure 4 shows the Raman spectra of those measurements.

To control the characteristics and to ensure a batch to batch comparability, both gold and Fe_xO_y nanoparticles were produced and functionalised inhouse. The production of the gold nanoparticles includes a nucleation step and a growth step according to Tainuchi et al. (Tainuchi et al. 2007). The size and the Ostwald ripening of the particles were monitored over time to ensure repeatable Raman intensities. The Fe_xO_y nanoparticles were produced using a top-down process. Both the build-up of the gold and the Fe_xO_y nanoparticle side were carefully characterised using techniques like UV/Vis spectrometry, X-ray diffraction, transmission electron microscopy, dynamic light scattering, thermogravimetric analysis, single-angle X-ray scattering and zeta potential measurements.

Measurements of tau protein in artificial CSF and mouse brain tissue show the possibility to reach the low concentrations which comply with the natural biological range in humans.

Another accurate quantification method for peptides and proteins is via their intrinsic S- and/or P-containing amino acids by ICP-MS/MS after separation of the analyte from the matrix. In the literature several tryptic digestion methods are described, sometimes only partially diverting from each other. Considering that S and P were used to determine the concentration, methods which included the use of S- and/or P-containing chemicals are to be avoided where possible. Additionally, methods which require desalting the tryptic digest (due to the use of e.g. urea) were avoided due to the inherent risk of losing the analyte. The most promising methods were adapted from Percy et al. (Percy et al. 2013) and Proc et al. (Proc et al. 2010). They allowed reproducible digestion of standard proteins, like casein and albumin, and their quantification via S- and/or P-containing peptides. The transfer of this method to biological matrices, such as cytosolic brain proteins, HeLa-cells or other cell types, showed that the tryptic digest of the proteins is reproducible. Using ICP-MS/MS for analysis, it is obvious that different tissues contain a range of non-proteogenic P-containing compounds, which need to be removed before the peptide analysis. These compounds were identified as deoxyribonucleic acid (DNA) and/or ribonucleic acid (RNA) oligomers. It was possible to remove them successfully by a benzonase digestion before the tryptic digest. A further challenge when working with biological tissues is the high concentration of low molecular S- and P-containing compounds, which can overshadow especially hydrophilic peptides eluting early from the separation column. A desalting of the proteins after benzonase treatment, denaturation and reductive alkylation of the cysteine residues with a customary desalting column was found to be the best way to overcome this issue without the loss of proteins and/or peptides.

For a successful quantification of the proteins using ICP-MS/MS, it is essential to achieve a baseline separation of the peptides in the chromatographic separation. Biological matrices typically contain hundreds if not thousands of proteins, each single one capable of generating more than one S and/or P-containing peptide. Therefore, a purification of the sample is unavoidable. Several possibilities are available for this purpose. At the protein level the use of specific antibodies or simply separating heat stable from instable proteins by heat precipitation is possible. At the peptide level, after the tryptic digest, purification can be achieved, for example, by using a cation exchange column with high pH-reversed phase separation and subsequently collecting the fractions of interest. Another possibility is the use of specific affinity materials to enrich peptides with specific functional groups such as phosphate groups.

Affinity materials for the enrichment of P-containing proteins and/or peptides are widely used in “standard” proteomics research. Due to the fact that protein phosphorylation has a huge impact on cellular processes, there is a high interest in P-containing proteins in medical and biochemical research. Enrichment of phosphorylated proteins and/or peptides is important for the identification and quantification of these proteomics, since phosphorylated forms are generally present at much lower concentration than the non-phosphorylated ones. Furthermore, they have a lower ionisation efficiency in ICP and a worse detection limit in molecular MS.

In contrast to molecular MS, the detection of peptides with ICP-MS does not have such severe problems of compound specific signal intensity, but it requires the presence of either S or P in the peptide which could be detected. Nevertheless, to improve the separation efficiency a purification of the analyte from the matrix by specifically enriching phosphorylated peptides is necessary, assuming these are the target compounds. The most common methods for phosphorylated peptide enrichment are the use of titanium dioxide nanoparticles and Fe immobilised affinity chromatography (Fe-IMAC). In this study, both were evaluated regarding the reproducibility of peptide recovery and additionally a third method, using Mullite instead of titanium dioxide nanoparticles, was tested. The reproducibility of all three methods was investigated using a mix of casein and ALB and was found to be not as good as expected from the literature. Measurements showed relative standard deviations of 5 to 50 % depending on the peptide. The best reproducibility and recovery of only one phosphate containing peptides was found after enrichment with titanium dioxide nanoparticles followed by Fe-IMAC. The recovery of peptides containing multiple phosphate groups is lower than that of mono-phosphorylated peptides for all three methods. Applying either of the methods to the standard tryptic digest of cellular origin showed that DNA and/or RNA oligomers present in these extracts are enriching extremely well on all materials and, therefore, must be removed in advanced.

Another possibility to quantify tau-protein was attempted by biochemical means using commercially available tau-protein standards as calibrants. But this approach has been proven futile because the concentrations of tau-protein standards were not accurately indicated by the manufacturers. The comparison of different standards showed strong deviations in measured signals when same concentration, according to manufacturer’s information, were used. Therefore, a method for reliable, accurate and traceable protein quantification of pure proteins using ICP-IDMS was developed. Using this method, the amount of S is accurately quantified and can be used to calculate the amount of protein in solution, as S is present in the two amino acids cysteine and methionine and the amino acid sequence of the target proteins is known. By measuring well characterised S standards and samples in one measurement sequence, the results are traceable to the standard and are comparable between laboratories. To correct for S containing impurities in the protein solution, an offline separation method was developed to quantify and correct for free S in the sample.