ToF-SIMS and Principal Component Analysis Investigation of ... - Doi.org

Subscriber access provided by CORNELL UNIVERSITY LIBRARY

Article

ToF-SIMS and Principal Component Analysis Investigation of Denatured, Surface-Adsorbed Antibodies Nicholas G. Welch, Robert M.T. Madiona, Judith A. Scoble, Benjamin W. Muir, and Paul J. Pigram Langmuir, Just Accepted Manuscript • DOI: 10.1021/acs.langmuir.6b02754 • Publication Date (Web): 07 Oct 2016 Downloaded from http://pubs.acs.org on October 12, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Langmuir is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 15

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

ToF-SIMS and Principal Component Analysis Investigation of Denatured, Surface-Adsorbed Antibodies Nicholas G. Welch1,2, Robert M. Madiona1,2, Judith A. Scoble2, Benjamin W. Muir2, Paul J. Pigram1* 1 Centre for Materials and Surface Science and Department of Chemistry and Physics, School of Molecular Sciences, La Trobe University, VIC, 3086, Australia 2 CSIRO Manufacturing, VIC, 3168, Australia

Keywords: Denaturation; ToF-SIMS; principal component analysis; antibody; plasma polymer.

ABSTRACT: Antibody denaturation at solid-liquid interfaces plays an important role in the sensitivity of protein assays such as enzyme-linked immunosorbent assays (ELISAs). Surface immobilized antibodies must maintain their native state, with their antigen binding (Fab) region intact, to capture antigens from biological samples and permit disease detection. In this work, two identical sample sets were prepared with whole antibody IgG, F(ab′)2 and Fc fragments, immobilized to either a silicon wafer, or a diethylene glycol dimethyl ether plasma polymer surface. Analysis was conducted on one sample set at Day 0, and the second sample set after 14 days in vacuum, with time-of-flight secondary-ion mass spectrometry (ToF-SIMS) for molecular species representative of denaturation. A 1003 mass fragment peak list was compiled from ToF-SIMS data and compared to a 35 amino acid mass fragment peak list using principal component analysis. Several ToF-SIMS secondary-ions, pertaining to disulfide and thiol species, were identified in the 14 day (presumably denatured) samples. A substrate and primary-ion independent marker for denaturation (ageing) was then produced using a ratio of mass peak intensities according to;

. + . + . + . + . + . ∶

. + . + . + . + . The ratio successfully identifies denaturation on both the silicon and plasma polymer substrates and for spectra generated with Mn+, Bi+, and Bi3+ primary ions. We believe this ratio could be employed to as a marker of denaturation of antibodies on a plethora of substrates. INTRODUCTION Understanding protein denaturation and conformation at surfaces is of paramount importance in a multitude of biological contexts ranging from biomedical implantation to maximizing the efficiency of protein assay detection sensitivity1. As a protein begins to interact with a surface, it may undergo 2 conformational changes that negatively affect the overall tertiary structure . In the case of an enzymelinked immunosorbent assay (ELISA), the immobilized protein may be an antibody specific to a 3 disease biomarker or antigen . As the antibody binds to the surface the antigen capture site Fab may be compromised, reducing the capability of that antibody to bind specifically and effectively to the antigen. Each antibody compromised by conformational changes represents lost antigen detection 4 capacity and a loss of overall assay performance . Despite being robust proteins, antibodies may be denatured by interaction with the substrate or from unfavourable changes in the local environment2, 5. For example, proteins can be readily denatured by pH, temperature, and dehydration6. Time-of-flight secondary-ion mass spectrometry (ToF-SIMS) provides a means to study the outermost 1 - 5 nm of a sample surface, providing rich structural, chemical and molecular information7 however it is a high vacuum technique. Surface adsorbed proteins and associated interfaces have been studied in-depth with ToF-SIMS to yield information regarding protein structure, conformation and orientation8, 9, 10, 11, 12, 13. Surface adsorbed thermally denatured proteins have been studied 14, 15 1 quantitatively and methods for their removal investigated by chemical means . ToF-SIMS mass 1, 6, 9 fragments indicative of protein denaturation have been found , however secondary-ions produced from fragmentation of a surface are complex in that they may arise from multiple components of the 16 system . As a result, multivariate analysis techniques such as principal component analysis (PCA) are frequently employed to reduce the dimensionality of a data set and to elucidate information on

ACS Paragon Plus Environment

Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

composition, form and function from large and complex data sets17. ToF-SIMS and PCA are frequently used in combination for the investigation of biological systems. With proteins it is typical to perform PCA using a limited number of ToF-SIMS mass peaks particularly related to amino acid 6, 10, 11, 18, 19 fragments . This is a sound method in many cases as it limits the input from other components of the samples, such as substrate, and focuses on the analysis of protein-specific secondary-ions. However, one of the strengths of ToF-SIMS is molecular specificity where intricately related samples can be differentiated by considering a large range of secondary-ions, fully accessing the utility of the technique20. In this work we have prepared proteins including whole antibody immunoglobulin (IgG), and the F(ab′)2 and Fc portions of the antibody, and adsorbed these to both a silicon wafer and a complex polymer thin film produced from plasma deposition of diethylene glycol dimethyl ether (DG). We have interrogated the protein surfaces with ToF-SIMS before and after vacuum ageing and employed PCA in a strategic manner to investigate the relationship between samples analysed at Day 0 and those at Day 14 in vacuum. We make the assumption that the Day 0 samples will be more native-like than the Day 14 samples understanding that the adsorption and vacuum drying may directly affect the state of the proteins in both samples. A direct comparison is made between using a large mass peak list, and 12, 21, 22 a list containing only literature-derived amino acid specific mass fragments. From this we were able to produce a substrate and primary-ion independent marker for denatured antibodies and associated proteolytic fragments, providing a universal means for determining antibody denaturation at the surface of a plethora of substrates.

EXPERIMENTAL SECTION Substrate Preparation. Silicon (Si) wafers were cut into 10 mm x 10 mm squares and sonicated for 30 minutes in a solution of 2 % RBS™ detergent (Sigma), 2 % ethanol and Milli-Q water (18.2 MΩ.cm) and dried with nitrogen. Half of the wafers were coated with plasma polymer thin films using a custom built plasma reactor. The plasma vessel cylinder (Pyrex glass) was housed between two poly(vinyl chloride) (PVC) plates grooved to support two o-rings. The top positive electrode was circular and mounted through the top PVC plate. A large negative circular electrode was used that was mounted through the bottom PVC plate and held the silicon wafer samples. Electrodes were made from copper and were separated by 100 mm. The diethylene glycol dimethyl ether monomer (Sigma) flow was controlled with a manual gauge. The plasma parameters for plasma polymer deposition were a radio frequency of 125 kHz, load power of 50 W, initial monomer pressure of 0.2 mbar, and a treatment time of 2 x 120 s. Immediately after deposition, DG coated wafers were placed in 500 µL Milli-Q water. Preparation and purification of Fc and F(ab′)2 fragments from whole IgG. To prepare the Fc fragment, 0.24 mg papain (Sigma, activated with dithiothreitol (Sigma) immediately prior to use) was added to 11.8 mg of anti-epidermal growth factor receptor (anti-EGFR23) IgG antibody and the digestion allowed to proceed for 16 h at 37 °C before being stopped with the addition of iodoacetamide. The reaction mixture was applied to a 1 mL Protein A FF column (GE Healthcare). The bound fraction containing the Fc fragment was eluted with 0.1 M citrate, 0.15 M NaCl, pH 3.0 and again subjected to gel filtration (Superdex S200 10-30, GE Healthcare) in Tris-buffered saline (TBS). To prepare the F(ab′)2 fragment, the anti-EGFR was digested with pepsin (50 µg per mg antibody) in 0.1 M citrate pH 3.6 for 45 min at 37 °C before the F(ab′)2 fragment was isolated by gel filtration purification as above. Protein Adsorption. Antibody and antibody fragments were prepared in separate solutions of TBS at 7 nM. Si and DG samples were incubated in 400 µL of protein solution at room temperature for 1 hour. Samples were further rinsed with TBS and 0.05 % Tween 20 (to remove unbound protein) and Milli-Q water to remove excess protein and salt. The first replicate of each sample type was mounted immediately, awaiting analysis at Day 0. The second replicate was stored under vacuum for 14 days with the pressure in the storage chamber maintained at 1 x 10-7 mbar. Time-of-Flight Secondary Ion Mass Spectrometry (ToF-SIMS). ToF-SIMS data were acquired using a TOF-SIMS 5 instrument (ION-TOF GmbH, Münster, Germany) equipped with a


Page 2 of 15

Page 3 of 15

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Bi/Mn liquid-metal ion source operating at 30 kV in pulsed, bunched mode. All spectra were collected in positive polarity from a 100 µm × 100 µm region under static conditions (total ion dose < + 12 –2 + + 1 × 10 ions cm ) using three primary-ions: Mn , Bi , and Bi3 . Spectra were acquired from three locations for each primary ion. Charge compensation was achieved by flooding the sample surface + with low-energy electrons, resulting in a mass resolution (m/∆m) for the C2H3 (27.023 m/z) peak of + + + greater than 6887 for Bi , 6323 for Bi3 , and 7045 for Mn . Data were acquired for 150 µs between primary-ion pulses, providing an accessible mass range of up to 2000 m/z. The mass scale was + + + + + + + + + calibrated using peaks assigned to the C , CH , CH2 , CH3 , C2H3 , C3H3 , C4H3 , C5H3 , C6H5 and + -9 C7H7 ions. The pressure in the analysis chamber during data acquisition was 3 x 10 mbar. Peak lists used for multivariate analysis were generated using SurfaceLab 6 software (ION-TOF GmbH, Münster, Germany; version 6.5). Principal Component Analysis (PCA). A list of 1010 mass spectral peaks with maximum 2 intensity greater than 10 was generated over the m/z range 1 - 650. Mass peaks were selected on an individual basis by manually removing a linear background. In the case of significant overlap, the whole area encompassing both masses was selected. Peak areas were normalised by total ion intensity per spectrum to take account of differences in beam current with each primary-ion source. Seven peaks (Si+, K+, SiH+, SiOH+, Na+, Al+, Cs+) were removed from the peak list to minimize data skewing due to strong intensities. The resulting matrix consisted of 108 sample spectra (3 spectra per primary-ion per sample) with 1003 mass spectral peaks ("all peaks", denoted in this work as "AP"). A subset of this peak list containing 35 amino acid specific mass fragments derived from literature12, 21, 22 was prepared (denoted in this work as "AA"). This amino acid peak list is shown in Table S1. Peak areas were normalized within each set of mass spectral peaks and mean-centred before each data analysis to mitigate systematic differences between samples. No cross-validation methods were applied and 6 principal components were selected. PCA was conducted using PLS_Toolbox (version 8.1) (Eigenvector Research, Manson, WA) utilizing MATLAB R2015b (version 8.6) (The MathWorks Inc., USA). Throughout this work, principal components (PC1, PC2, etc.) are listed with the percentage of variance they encompass in brackets; for example PC1 (89.5 %). RESULTS AND DISCUSSION PCA was employed to investigate ToF-SIMS spectra of sample sets with increasing levels of complexity to identify markers for denaturation of an antibody and its proteolytic fragments according to Scheme 1. 1.

Simple peak ratio marker for denaturation

• PCA of Day 0 and Day 14 IgG samples immobilized to Si, analysed with Bi3+. • Ratio of the most positive loading fragment and the most negative loading fragment as a marker for denaturation using the AP and AA approaches.

2.

Broad application of peak ratio marker

• Application of the above ratio marker to the Day 0 and Day 14 samples, IgG, F(ab')2 and Fc antibody fragments immobilized to Si and DG substrates, with analysis using Bi3+, Bi+ and Mn+ primary-ions.

3.

PCA of all samples

• PCA of the entire data set using the AP and AA approaches. • Assessment of usable denaturation markers based on principal components.

4.

PCA of individual sample types

• "PC1 Approach"; PCA conducted on each sample type individually for Day 0 and Day 14 sample separation using AP and AA aproaches. All other parameters (primary-ion, substrate, and protein) remained constant.

5.

Identification of denaturation marker

• Ratio of summed secondary-ion intensities used to produce a marker for denaturation.

Scheme 1. Systematic analysis for producing a substrate and primary-ion independent marker for denaturation.


Langmuir

Identifying Peak Ratio Markers for Denaturation. ToF-SIMS (Bi3+ primary-ion acquisition) and PCA utilizing the amino acid (AA) peak list were employed to investigate IgG immobilized to Si samples. Principal components are comprised of contributions from all mass fragments used in analysis. The loadings plot details the amount of contribution from each mass fragment to the principal component. The scores plot details how well each sample is described by that principal component. In Figure 1A, principal component 1 (PC1), that encompasses 89.5 % of the variance in the data shows good discrimination between Day 0 and Day 14 samples. This indicates that PC1 is an appropriate principal component to use to describe differences between these samples. The loadings plot for PC1 (Figure 1B) showed the ToF-SIMS mass fragment at 61.01 m/z, corresponding to methionine, loaded strongly with the Day 14 samples. Conversely, the mass fragment at 91.05 m/z, corresponding to both phenylalanine and tyrosine, loaded strongly with the Day 0 samples. The ratio of these two peak intensities was developed as a marker for denaturation and good discrimination was shown for IgG samples (Figure 1C). We also employed this peak ratio marker with F(ab′)2 and Fc samples immobilized on Si and found reasonable agreement. This marker was less effective for Fc samples. This may be due to the slightly lower content (6.6%) of phenylalanine and tyrosine in the Fc as compared to the F(ab′)2 (7.7%). In any case, this ratio was not a suitable marker for all antibody fragments. A)

B) 1 PC1 Loadings (89.5 %)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

0.8 0.6 0.4 0.2 0 -0.2 -0.4 -0.6

Fragment Mass (m/z)

Figure 1. Producing a marker for denaturation from PCA utilizing a 35 amino acid peak list. A) The scores plot of PC1, encompassing 89.5 % of the variation, discriminates between three points collected on Day 0 (blue) and three points collected on Day 14 (red) samples of IgG immobilized on Si substrate with Bi3+ acquisition. B) The corresponding PC1 loadings plot where negative loadings are those more prevalent in Day 0 samples and positive loadings are those more prevalent in Day 14 samples. The most positively loading mass fragment peak is 61.01 m/z and the most negatively loading mass fragment peak is 91.05 m/z. C) Peak ratios of the ion intensities [61.01/91.05 m/z] applied to the IgG sample and also F(ab′)2 and Fc, Day 0 and Day 14 samples. Columns are average of three points per sample. Error bars ± standard deviation. Amino acid specific list available in Table S1.


Page 4 of 15

Page 5 of 15

In a similar mode, the entire 1003 mass fragment (AP) peak list was employed with PCA to investigate IgG samples. PC1 showed good discrimination of the Day 0 and Day 14 samples (Figure 2A). It is notable that the variance captured by PC1 increased to 95.8 %, indicating that the two sample sets were better resolved with the more comprehensive peak list. The loadings plot for PC1 shows the large and complex set of contributions from the 1003 mass fragments (Figure 2B). As before, we selected the most positively loading mass fragment (66.96 m/z) and the most negatively loading mass fragment (39.02 m/z). The ratio of these peak intensities gave good discrimination for the IgG, and also for the F(ab′)2 and Fc fragments alone (Figure 2C). The 66.96 m/z mass fragment is attributed to CSNa+. The increase in this mass fragment population is associated with the exposure of the internal disulfide-bonded cysteines accompanying the denaturation of the antibody, or its + fragments. The 39.02 m/z mass fragment is attributed to C3H3 , which is associated with all protein. A)

B) 0.8 0.6 0.4 0.2 0 -0.2 -0.4

1.01 23.98 29.99 39.96 43.06 47.01 52.94 56.01 58.99 62.93 66.95 69.06 71.98 74.10 79.02 82.94 85.94 88.05 91.96 95.06 98.92 102.94 105.95 108.97 112.03 116.06 120.08 123.08 127.10 134.01 139.02 145.91 150.88 155.08 160.90 166.07 172.85 178.91 188.81 197.90 202.87 212.05 220.81 229.06 243.12 255.24 265.09 276.05 288.07 301.07 312.80 337.07 357.66 387.08 411.07 488.66

PC1 Loadings (95.8 %)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Fragment Mass (m/z)

Figure 2. Producing a marker for denaturation from PCA utilizing a 1003 mass fragment peak list. A) The scores plot of PC1, encompassing 95.8 % of the variation, discriminates between three points collected on Day 0 (blue) + and Day 14 (red) samples of IgG immobilized on Si substrate with Bi3 acquisition. B) The corresponding PC1 loadings plot where negative loadings are more prevalent in Day 0 samples and positive loadings are more prevalent in Day 14 samples. The most positively loading mass fragment peak is 66.96 m/z and the most negatively loading mass fragment peak is 39.02 m/z. C) Peak ratios of the ion intensities [66.96/39.02 m/z] applied to the IgG sample and also F(ab′)2 and Fc, Day 0 and Day 14 samples. Columns are average of three points per sample. Error bars ± standard deviation.

Broad Application of Peak Ratio Markers. We further sought to identify and confirm a peak ratio that would serve as a marker for denaturation of antibodies and associated proteolytic fragments, across a range of parameters., We tested the peak ratio [61.01 m/z / 91.05 m/z] (derived from the amino acid peak list), and the peak ratio [66.96 m/z / 39.02 m/z] (derived from the 1003 mass fragment peak list) against a larger sample set including the proteins immobilized to the Si substrate,


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

and to the DG hydrocarbon polymer substrate, and with ToF-SIMS data collected using Bi3+, Bi+ and + Mn primary-ions. Considering the amino acid peak ratio and the Si substrate (Figure 3), we see that good discrimination is achieved between the Day 0 and Day 14 samples with the Bi3+ spectra. However, + + discrimination is less convincing with Bi spectra and poor with Mn spectra. This is due to the lighter primary-ions producing significantly different protein fragmentation patterns24. When this peak ratio was applied to protein immobilized on DG substrate, poor discrimination between Day 0 and Day 14 samples was achieved in all cases (Figure 4). In this instance, the peak ratio often favoured the Day 0 samples over the Day 14 samples, making it a poor marker for denaturation. The analysis is + + complicated by the hydrocarbon overlaps from the substrate such as C3H2Na and C5H .

Figure 3. Peak ratios of the ion intensities for the most positively loading (61.01 m/z) and the most negatively loading (91.05 m/z) mass fragments, applied to the IgG, F(ab′)2 and Fc, Day 0 and Day 14 samples, immobilized + + + on Si substrate with A) Bi3 , B) Bi and C) Mn acquisition. Columns are average of three points per sample. Error bars ± standard deviation.

Figure 4. Peak ratios of the ion intensities for the most positively loading (61.01 m/z) and the most negatively loading (91.05 m/z) mass fragments, applied to the IgG, F(ab′)2 and Fc, Day 0 and Day 14 samples, immobilized + + + on DG substrate with A) Bi3 , B) Bi and C) Mn acquisition. Columns are average of three points per sample. Error bars ± standard deviation.

We tested the approach again using the 1003 mass fragment peak list ratio [66.96 m/z / 39.02 m/z] for the Si substrate (Figure 5), and for the DG substrate (Figure 6). For the Si substrate, generally the peak ratio worked well for the Bi3+ acquisition, reasonably for the Bi+ acquisition, and + + poorly for the Mn acquisition. In the case of Mn , the poor results were attributed to contributions from Fc, likely due to poor ion intensity. For the DG substrate, poor results were observed in all cases.


Page 6 of 15

Page 7 of 15

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

The ratio calculated for the Bi3+ spectra, similar to the amino acid peak ratio, produced larger values for the Day 0 samples. Further, the lighter primary-ions were inconsistent, with large errors. We conclude then that neither the amino acid peak ratio, nor the 1003 mass fragment peak ratio was sufficient to characterise or indicate denaturation over this broad range of parameters.

Figure 5. Peak ratios of the ion intensities for the most positively loading (66.96 m/z) and the most negatively loading (39.01 m/z) mass fragments, applied to the IgG, F(ab′)2 and Fc, Day 0 and Day 14 samples, immobilized + + + on Si substrate with A) Bi3 , B) Bi and C) Mn acquisition. Columns are average of three points per sample. Error bars ± standard deviation.

Figure 6. Peak ratios of the ion intensities for the most positively loading (66.96 m/z) and the most negatively loading (39.01 m/z) mass fragments, applied to the IgG, F(ab′)2 and Fc, Day 0 and Day 14 samples, immobilized + + + on DG substrate with A) Bi3 , B) Bi and C) Mn acquisition. Columns are average of three points per sample. Error bars ± standard deviation.

Principal Component Analysis of All Samples. PCA was conducted on the entire data set, that included the Si and DG substrates, IgG, F(ab′)2 and Fc proteins, and Bi3+, Bi+, and Mn+ primaryions, using the 35 amino acid peak list. PC1 (78.2 %) was plotted against PC2 (17.5 %) to give good discrimination of the two substrates (Figure 7A). PC3 (2.4 %) was plotted against PC4 (0.7%), revealing the complexity of the data set (Figure 7B). The majority of the Bi3+ samples loaded positively with PC4, but discrimination between Day0 and Day 14 samples was not clear. + PCA was conducted on a reduced data set comprising only samples using Bi3 acquisition. PC1 (86.2 %) was plotted against PC2 (9.9 %) revealing good discrimination of the two substrates (Figure 8A). PC3 (3.4 %) was plotted against PC4 (0.2 %) to show good discrimination of the different protein samples (Figure 8B). The proteins separated both in type, and also Day 0 from Day 14, though not in a consistent manner (Figure 8B). Similar results were observed using the 1003 mass + fragment list when the entire data set was considered (Figure S1) and when just Bi3 samples was considered (Figure S2). Given the complexity of the sample set, it was clear this method would be


Langmuir

insufficient to resolve the Day 0 and Day 14 samples from one another. Subsequently, we considered each individual sample independently.

A)

B) 0.2 0.15 0.1

Scores on PC 4 (0.7%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 15

0.05 0 -0.05 -0.1 -0.15 -0.2 -0.4

-0.3

-0.2

-0.1 0 0.1 0.2 Scores on PC 1 (78.2%)

0.3

0.4

Figure 7. Principal component analysis of the whole sample set utilizing all primary-ions and the 35 amino acid specific peak list. A) PC1 plotted against PC2 showing discrimination of Si substrate (light) from DG substrate (dark). B) PC3 plotted against PC4 with primary-ions; Bi3+ (), Bi+ () and Mn+ (). Day 0 samples shown in blue and Day 14 samples shown in red. Protein fragments are not defined as they were not discriminated here.

A)

B) 0.03

0.15

0.02

0.1

0.01 0.05

0 0

-0.01 -0.05

-0.02 -0.1

-0.15 -0.3

-0.03

-0.2

-0.1

0

0.1

Scores on PC 1 (86.2%)

0.2

0.3

-0.04 -0.08

-0.06

-0.04

-0.02

0

0.02

0.04

0.06

0.08

Scores on PC 3 (3.4%) +

Figure 8. Principal component analysis of the whole sample set utilizing the Bi3 primary-ion and the 35 amino acid specific peak list. A) PC1 plotted against PC2 showing discrimination of Si substrate (light) from DG substrate (dark). B) PC3 plotted against PC4 with protein fragments; IgG (), F(ab’)2 () and Fc (). Day 0 samples shown in blue and Day 14 samples shown in red.

Principal Component Analysis of Individual Sample Types. We used PCA to investigate differences between Day 0 and Day 14 sample types when all other parameters (substrate, protein, primary-ion, and peak list) remained constant. Using the 1003 mass fragment peak list (AP) and the amino acid peak list (AA) approaches independently we compared the triplicate of Day 0 to the triplicate of Day 14 with the assumption that the first principal component, PC1, would reveal fragments particularly pertaining to denaturation (we denote this the “PC1 Approach”). Figure 9 shows the scores plots associated with each of the individual PC1s. A large variance captured by PC1 indicates that strong differentiation between Day 0 and Day 14 could be made. Across the sample set, + PC1 using the AP approach, captured a minimum variance of 51.4 % (DG, Mn , F(ab′)2) and a maximum of 97.9 % (DG, Bi3+, F(ab′)2). Similarly, using the AA approach, the minimum variance was + + 44.3 % (Si, Mn , F(ab′)2) and the maximum 98.4 % (Si, Bi3 , F(ab′)2). In some cases, the differences in the triplicate meant that some samples that were in fact analysed at Day 0 would load on the scores plot with samples from Day 14, and vice-versa. These sample inconsistencies did not trend with the variance captured by the PC1s. Significantly, the number of samples overall that loaded inconsistently


Page 9 of 15

35 Amino Acid Fragments (AA) IgG F(ab′)2 Fc

Bi+ Mn+

Bi+

Bi3+

Mn+

Si

Bi3+

1003 Mass Fragments (AP) IgG F(ab′)2 Fc

DG

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

Figure 9. "PC1 Approach". Principal component analysis showing PC1 for individual groups of Day 0 compared to Day 14 using the all peaks approach (AP) or the amino acid only peaks approach (AA). 18 individual PC1s were produced for each mass fragment list. Blue markers represent triplicate samples analysed at Day 0, and red markers represent triplicate samples analysed at Day 14. Table 1. Mass fragments loading with Day 14 for all 18 PCs (Group 1) and for 17 of 18 PCs (Group 2) produced using the PC1 approach. Assignments to the mass fragments and associated deviation from exact mass for the all peak approach are shown. ToF-SIMS Potential Exact Mass of Deviation from Measured Measured Mass Assignments Assignment Mass (ppm) + Group 1 61.9534 KNa 61.9529 -7.69 (18 out of 18 PCs) 62.9846 Na2OH+ 62.9838 -12.92 + 122.9547 C3H2SNK 122.9545 -1.87 NaS2O2H4+ 122.9550 2.19 + Group 2 78.9555 KNaOH 78.9557 1.98 + (17 out of 18 PCs) S2NH 78.9545 -13.22 + 84.9609 NaSNO 84.9598 -12.38 + Na3O 84.9637 33.52 120.9461 C2ONa2Cl+ 120.9428 -27.32 + C2HS2O2 120.9412 -40.55 NaS2O2H2+ 120.9394 -55.43


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Table 2. Mass fragments loading with Day 0 for all 18 PCs (Group 1) and for 17 of 18 PCs (Group 2) produced using the PC1 approach. Assignments to the mass fragments and associated deviation from exact mass for the all peak approach are shown. ToF-SIMS Potential Exact Mass of Deviation from Measured Measured Mass Assignments Assignment Mass (ppm) Group 1 (18 out of 18 PCs) + Group 2 30.9979 CF 30.9979 -0.12 + (17 out of 18 PCs) SiH3 30.9999 64.39 + 42.9991 AlNH2 42.9997 13.44 SiCH3+ 42.9999 18.10 + 73.0660 C4H9O 73.0648 -16.46 C2H7N3+ 73.0634 -35.62 + 147.0780 Si2C4H15N2 147.0774 3.90 + C8H12ONa 147.0786 -4.26 + C6H10N3Na 147.0772 5.26 + C5H11N2O3 147.0770 6.62 C8H9N3+ 147.0796 -11.06 + C6H13ONa2 147.0762 -12.06 SiCH15N2O4+ 147.0801 -14.46 + C3H9N5O2 147.0756 -16.14 + C4H11N3Na2 147.0748 -21.58 207.0376 C4H14O4Na2Cl+ 207.0376 -0.05 + C5H14O3KNa2 207.0375 0.53 Na2SN5O3H11+ 207.0378 0.92 + C5H3N8O2 207.0379 1.40 + C6H9NO7 207.0379 -1.40

was 11 for the AA approach and only 5 for the AP approach. This indicates that the AP approach was much better at differentiating between two sample types. This is not unexpected as utilizing 1003 mass peaks should give much better sensitivity to small changes in protein structure in comparison with the use of the 35 mass fragment (AA) peak list. We also compared the average variance captured by PC1 across like factors (Figure S3). We found that the using the AA approach, the variance captured by PC1 decreased with the use of lighter primary-ions, implying that it was more difficult to differentiate between Day 0 and Day 14. This was less apparent for the AP approach, which showed good differentiation across all primary-ion sources. Next, using the loadings from each of the 18 PC1s associated with the AP and AA approaches independently, we assessed the individual fragments loading with the Day 14 samples. When we investigate the loadings for the AA approach, no one fragment consistently loaded with denatured protein in all 18 PC1s. The most prominent fragment, occurring in 14 of 18 loadings, was + 43.0554 m/z. This mass can be attributed to C3H7 , an identifier for amino acids leucine and isoleucine. Increasing the list size to 13 of 18 loadings, another fragment, 72.0445 m/z, was identified. This mass could be attributed to C3H6NO+ associated with amino acids alanine and threonine. Conversely, we assessed the PCA for fragments loading with Day 0 samples and found that in 16 of 18 loadings, 73.0660 m/z, attributed to C2H7N3+, and amino acid arginine, was present. The tabulated data with potential overlaps of other mass fragments in available in Table S2 and S3. 9 It has been reported that variation in amino acid secondary-ions arising from cysteine residues can be used as a marker for denaturation as they are internal in a native antibody conformation. This population is particularly useful when taken as a ratio to an external residue fragment, such as arginine. This approach is effective, however, may be compromised when the substrate produces secondary-ions that overlap with amino acid peaks. This is the case with 44.97 + + m/z, where the CHS marker for cysteine overlaps with the SiOH silicon substrate peak. A similar + case arises with 76.02 m/z where the C2H6SN marker for cysteine is very close to the C2H4O3+ hydrocarbon fragment ion. In the current study, mass fragment peaks, produced with the PC1 approach, identified leucine, isoleucine, alanine and threonine as loading with denaturation; the first three of which have hydrophobic side chains. These are prevalent in the internal structure of the


Page 10 of 15

Page 11 of 15

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Langmuir

antibody that may be exposed during denaturation, increasing their populations. As a new marker for denaturation, we produced a ratio of the summed ion intensity of the two fragments loading with Day 14 and the fragment loading with Day 0 from the 18 PC1 analyses using the AA list (Tables S2 and S3) as described above;

:

!" . + . $%& '() + %&'*+,) - /012,425 + 61!,789 : # # # . 69; !" %.'(+&)

This new marker for denaturation was applied to all samples as shown in Figure 10 but was a poor predictor for samples immobilized on Si substrate. Particularly, the Mn+ sample had poor intensities for one or more mass fragments, in one or more of the triplicate samples, resulting in incomplete results. In the case of the DG substrate, this new peak ratio marker was a very good indicator for denatured samples, regardless of primary-ion. As there are some overlaps with hydrocarbon mass fragments, we believe that the increase in secondary ion-yields on the DG substrate provided an extra accuracy to separate the minor differences in amino acid secondary ionyields. Overall, this marker for denaturation served well for the DG substrate, but not for the Si substrate. Given that there were no mass fragments loading with Day 14 samples in all 18 PC1s, this was not unexpected.

Figure 10. Peak ratios of the summation of ion intensities, derived from the amino acid list, loading with Day 14 samples (43.06 and 72.04 m/z) to those loading with the Day 0 samples (73.07 m/z), applied to the IgG, F(ab′)2 + + + and Fc, Day 0 and Day 14 samples, immobilized on A-C) Si or D-F) DG substrate with Bi3 , Bi and Mn acquisition. Columns are average of three points per sample. Error bars ± standard deviation.


Langmuir

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

In the same manner, using the AP approach, we identified three fragments loaded with Day + 14 samples in all 18 PC1s (Table 1). The fragment masses were 61.9534 m/z attributed to KNa , + + 62.9846 m/z attributed to Na2OH , and 122.9547 m/z attributed to perhaps both NaS2O2H4 and C3H2SNK+. The two former fragments are simply related to salts which may indicate a difference in local salt stabilization from the denatured protein. The latter fragment ion, however, is sulfurcontaining, and likely to be related to sulfur/disulfide species in the protein. An antibody, and its fragments, contain internal disulfide-bonded cystines that, after denaturation may become exposed, and subsequently surface available to ToF-SIMS analysis, increasing their population in the Day 14 samples. Increasing the list size by considering the denaturing loadings for 17 of 18 PCs, we find that another three fragments are present. The fragment masses were 78.9555 m/z which can be attributed + + + + to KNaOH or S2NH , the mass 84.9609 m/z which can be attributed to Na3O or NaSNO , and finally the mass at 120.9461 m/z which may be attributed to C2ONa2Cl+, NaS2O2H2+ or C2HS2O2+. This shortlist also contains salt peaks and sulfur/disulfide containing species which further reinforces their importance for identifying denatured protein. Conversely, five fragments loading in 17 of the 18 PC1s of the Day 0 samples were identified; 30.9979, 42.9991, 73.0660, 147.0780 and 207.0376 m/z (Table 2). Amine-containing fragments, including the 73.0660 m/z fragment associated with the external arginine, were among the list. As a new marker for denaturation, we produced a ratio of the summed ion intensity of the fragments loading with Day 14 and the fragments loading with Day 0;

ToF-SIMS and Principal Component Analysis Investigation of ... - Doi.org

Recommend Documents