Determination of Deamidation Artifacts Introduced by Sample

Jul 3, 2012 - In this method, b ions were used for the calculation of Asn deamidation that .... Lorenzi, Preeti Purwaha, Valeda Yong, David H. Hawke, ...
0 downloads 0 Views 1MB Size
Download PDF
Technical Note pubs.acs.org/ac

Determination of Deamidation Artifacts Introduced by Sample Preparation Using 18O-Labeling and Tandem Mass Spectrometry Analysis Yi Du, Fengqiang Wang, Kimberly May, Wei Xu, and Hongcheng Liu* Merck Research Laboratory, 1011 Morris Avenue, Union, New Jersey 07083, United States S Supporting Information *

ABSTRACT: The sites and levels of Asn deamidation in proteins are often determined by LC−MS analysis of peptides obtained from enzymatic digestion. However, deamidation that occurs during sample preparation steps results in overestimation of the original level of deamidation. The inherent deamidation and those introduced by sample preparation can be differentiated by preparing samples in 18O water. When using H218O, the formation of isoAsp and Asp by Asn deamidation during sample preparation results in a molecular weight increase of 3 Da due to the incorporation of the 18O atom to the side chains of isoAsp or Asp; in contrast, inherent deamidation only results in a molecular weight increase of 1 Da. In addition, up to two 18O atoms can also be incorporated into the peptide C-terminal carboxyl group during enzymatic digestion. Therefore, the 2 Da molecular weight difference at the deamidation sites can only be used to differentiate deamidation that occurs prior to or during sample preparation under conditions that a fixed number of 18O atoms are incorporated into the peptide C-terminal carboxyl groups. Otherwise, it is challenging to apply this procedure because of the resulting complicated isotopic distributions. Here, a new procedure of using 18 O-water for sample preparation coupled with tandem mass spectrometry (MS/MS) was established to calculate the deamidation artifacts. In this method, b ions were used for the calculation of Asn deamidation that occurred prior to or during sample preparation, which eliminated the complicated factor of various number of 18O-atoms to the peptide carboxyl groups. This procedure has the potential to be applied under the general peptide mapping conditions.

D

alkylation does not cause a significant level of Asn deamidation since these steps are usually performed within a short time period.22 Similarly, LC−MS analysis that is typically carried out at low pH and elevated temperature does not induce a significant amount of deamidation.6 In contrast, significant levels of isoAsp and Asp have been detected after 4 h enzymatic digestion and the level of deamidation artifact continues to increase with longer incubation.6,8,22,23 Two strategies have been reported in the literature to accurately determine the level of Asn deamidation in protein samples. The first one is to reduce the amount of Asn deamidation during sample preparations by digesting proteins at lower temperature and at lower pH.24 However, such conditions are not optimal for the most commonly used proteases such as trypsin and Lys-C. The second strategy is to differentiate Asn deamidation that occurs prior to or during sample preparation by preparing samples in 18O-water.8,25 The 18 O atom can be incorporated into peptides at the deamidation site and the newly generated peptide C-terminal carboxyl

eamidation of asparagine (Asn) is a common modification and a major degradation pathway of recombinant monoclonal antibodies (mAbs).1−15 Asn deamidation under mild pH conditions usually occurs through the β-elimination mechanism with the formation of succinimide, which can be readily hydrolyzed to form isoaspartate (isoAsp) and aspartate (Asp).16 Asn deamidation of recombinant mAbs not only causes antibody heterogeneity but may also significantly impact its safety and efficacy. Although no direct correlation between mAbs’ immunogenicity and the presence of isoAsp has been established,17 it has been demonstrated using other proteins that isoAsp can trigger an immunoresponse to self-proteins and thus may contribute to autoimmune diseases.18,19 Direct evidence from studies using recombinant monoclonal antibodies has demonstrated that replacement of the original Asn in the complementary determination regions (CDR) by Asp from Asn deamidation or isoAsp from both deamidation and isomerization resulted in a significant decrease of binding affinity and potency.1,2,14,20,21 Therefore, it is critical to characterize the occurrence of Asn deamidation in recombinant monoclonal antibodies. The major challenge of LC−MS analysis of Asn deamidation is the generation of artifacts during sample preparation, especially enzymatic digestion. Denaturation, reduction, and © 2012 American Chemical Society

Received: May 16, 2012 Accepted: July 3, 2012 Published: July 3, 2012 6355

dx.doi.org/10.1021/ac3013362 | Anal. Chem. 2012, 84, 6355−6360

Analytical Chemistry

Technical Note

Figure 1. A diagram demonstrating the 18O labeling procedure. Incorporation of one 18O into the peptides containing isoAsp and Asp as a result of Asn deamidation during sample preparation will differentiate these peptides containing isoAsp and Asp that were originally present in the sample. Incorporation of 18O-atom into the peptide carboxyl groups is homogeneous only under specific experimental conditions.

procedure, except using buffers and reagents all prepared in normal water. LC−MS and LC−MS/MS. An Agilent Infinity 1290 UHPLC (Santa Clara, CA) coupled with an Agilent 6538 UHD Q-TOF mass spectrometer was used to analyze the peptides. A volume of 20 μL of each sample was loaded onto a Proto C18 column (250 mm × 1 mm i.d., 5 μm particle size, Higgins Analytical Inc., Mountain View, CA) with 98% mobile phase A (0.02% TFA and 0.08% formic acid in water) and 2% mobile phase B (0.02% TFA and 0.08% formic acid in acetonitrile) at a flowrate of 50 μL/min. The peptides were separated and eluted using a gradient from 2% mobile phase B to 25.4% in 105 min, remaining at 25.4% B for 30 min and then increase to 98% in 10 min. The column was washed and equilibrated before the next injection. The column oven was heated at 60 °C. The MS was operated in positive ion mode with a full mass scan from m/z 200 to 2000. The MS/MS experiment was set up at the fragmentation voltage of 35 V targeted at the Asp and isoAsp containing peptide precursor ions.

groups. When the Asn deamidation intermediate is hydrolyzed, one 18O atom is incorporated into the side chain of Asp and isoAsp, resulting in peptide molecular weight increase of 3 Da; 1 Da from deamidation and 2 Da from incorporation of an 18Oatom to the side chains of Asp and isoAsp. In contrast, Asn deamidation that occurs prior to sample preparation results in peptides containing either Asp or isoAsp with a molecular weight increase of 1 Da. The molecular weight difference of 2 Da can be used to differentiate whether or not Asn deamidation was introduced by sample preparation. However, the fact that up to two 18O-atom can also be incorporated into the peptide C-terminal carboxyl group results in complicated mass spectra containing several overlapping isotopic distributions (Figure 2).26−29 Therefore, it is challenging to apply this procedure to determine the level of Asp and isoAsp introduced by sample preparation. To overcome this issue, a method of using trypsin digestion in 18O water coupled with MS/MS fragmentation was established. In the newly established method, b ions generated from fragmentation of Asp or isoAsp containing peptides, which were not interfered with by the variation of 18O atom incorporation into peptide C-terminal carboxyl groups, were used for the calculation of deamidation artifacts.

■

RESULTS AND DISCUSSION Principle of the Method. The procedure for sample preparations in 18O-water is outlined in Figure 1, which represents the ideal outcome. Asn deamidation in normal water results in a peptide molecular weight increase of 1 Da. However, deamidation that is introduced during sample preparation in 18O-water results in an additional 2 Da molecular weight increase because of the incorporation of an 18O-atom into the side chain carboxyl group of Asp and isoAsp. This molecular weight difference of 2 Da can be used to differentiate Asn deamidation that occurs prior to or during sample preparation. However, up to two 18O-atoms can also be introduced into the peptide C-terminal carboxyl groups during enzymatic digestion. The extent of the second 18O-atom incorporation is highly dependent on the experimental conditions. For example, higher trypsin to antibody ratios and longer digestion time resulted in more peptides and higher levels of the second 18O incorporation to the peptide Cterminal carboxyl groups. On the other hand, lower trypsin to antibody ratio and shorter digestion time resulted in fewer peptides and lower levels incorporation of the second 18O to the peptide terminal carboxyl groups. The presence of denaturing reagent can also inhibit the second 18O atom incorporation. If either none or complete incorporation of the second 18O-atom can be confirmed, the outlined procedure can be used to determine the level of deamidation during sample preparation as previously reported.8 However, in most cases, the incorporation of the second 18O-atom varies with

■

MATERIALS AND METHODS Materials. The recombinant monoclonal antibody was produced and purified at Merck (Union, NJ). Ammonium bicarbonate, dithiothreitol, iodoacetamide, urea, and 18Oenriched water (97.1% purity) were purchased from Sigma (St. Louis, MO). Acetonitrile, formic acid, and trifluoroacetic acid (TFA) were purchased from J.T. Baker (Phillipsburg, NJ). Trypsin was purchased from Worthington Biochemical Corporation (Lakewood, NJ). Trypsin Digestion. The antibody sample was first buffer exchanged into 100 mM ammonium bicarbonate prepared in 18 O-water using Zeba spin desalting columns (Thermo Scientific, Rockford, IL). To ensure a complete removal of the water in the sample, buffer exchange was performed twice. The stock solutions of 1 M DTT, 8 M urea, and 0.5 M iodoacetamide were also prepared in 18O-water. The buffer exchanged sample was denatured and reduced at 37 °C for 30 min using 10 mM DTT in the presence of 6 M urea in the ammonium bicarbonate buffer. It was further alkylated using 30 mM iodoacetamide at 37 °C for 30 min. The sample was then buffer exchanged into 100 mM ammonium bicarbonate prepared in 18O-water using NAP-5 columns (GE Healthcare, Piscataway, NJ). Trypsin was added to each sample to a final ratio of 1:10 (trypsin/antibody, w:w) and then incubated at 37 °C for 8 h. A control sample was digested following the same 6356

dx.doi.org/10.1021/ac3013362 | Anal. Chem. 2012, 84, 6355−6360

Analytical Chemistry

Technical Note

Figure 2. A diagram illustrating the complexity of the peptide mixtures generated under general experimental conditions. The various molecular weights and the expected elution positions corresponding to peaks in the EIC chromatogram in Figure 3B are also shown in the figure.

experimental conditions.26−29 Therefore, two overlapping isotopic distributions are expected to be associated with the Asn containing peptide and four overlapping isotopic distributions are expected to be associated with each of the isoAsp or Asp containing peptides (Figure 2), which makes it challenging to apply the previously established procedure to determine the level of Asn deamidation that occurs during sample preparation under various trypsin digestion conditions. In the current study, mAb sample was prepared following the same 18O labeling procedure as previously described; however, for mass spectra analysis, b ions from MS/MS fragmentation of Asp and isoAsp containing peptides that contain only Nterminal amino acids were used to calculate the level of Asn deamidation artifacts to prevent complicated overlapping isotopic distributions. Two examples were present in this study to demonstrate the utility of this proposed method. In the first example, the peptide with the amino acid sequence of GFYPSDLAVEWESNGQPENNYK was used. The focus of this study was on the deamidation of the first Asn residue since it is followed by a glycine residue and highly susceptible to deamidation. In the second example, a peptide containing an Asn residue that is followed by a serine residue was employed and the data is provided as Supporting Information. Full Scan Mass Spectra. Typical extracted ion chromatograms (EICs) of the peptide are shown in Figure 3. Similar EIC chromatograms were obtained from analysis of the peptide from digestion of the mAb in either normal water (Figure 3A) or 18O-water (Figure 3B). The peak identities have been determined previously8 and labeled accordingly in Figure 3 as peak 1 (isoAsp), peak 2 (Asn), and peak 3 (Asp). The full scan mass spectra of the three peaks from digestion of the antibody in normal water are shown in Figure 4. As expected, the molecular weight of peak 1 and peak 3 is 1 Da higher than the molecular weight of peak 2. Full scan mass spectra of the corresponding peaks from the antibody that was digested in 18 O-water are shown in Figure 5. Clearly, each spectrum contains overlapping isotopic distributions, as expected from Figure 2. The full scan mass spectrum from peak 2 contains two overlapping peak distributions from the incorporation of either one or two 18O atoms to the peptide C-terminal carboxyl groups. Peak 1 contains the peptide with isoAsp residue with four overlapping isotopic distributions. Peak 3 contains the peptide with Asp residue with four overlapping isotopic distributions. Therefore, it is almost impossible to apply the

Figure 3. EIC chromatograms of the peptide obtained from digestion in normal water (A) or 18O-water (B). The peak identities are labeled as shown. The peak between peaks 2 and 3 corresponds to the peptide with Asp from deamidation of the second Asn residue in the peptide.

previously established 18O labeling procedure to determine the level of Asn deamdiation that occurred prior to or during sample preparation. MS/MS Spectra and Calculation of Deamidation Artifacts. As discussed previously, the complicated full scan mass spectra were formed due to the variation of 18O incorporation into the peptide C-terminal carboxyl groups. This can be overcome by using b ions generated from MS/MS spectra because b ions do not contain the C-terminal carboxyl groups. The particular b ion was selected based on the following criteria. First, the b ion should include the residues of interest, either isoAsp or Asp. Second, the b ion molecular weight should be unique, meaning its m/z is different enough from other fragment ions. Third, the b ion should have a strong signal and good isotopic resolution. On the basis of these criteria, the b ion that contains amino acids up to Gln was selected (Figure 6), which has a calculated m/z of 1780.78. The experimental m/z of this b ion from peak 2 is shown in Figure 7B, which is in good agreement with the calculated m/z. The 6357

dx.doi.org/10.1021/ac3013362 | Anal. Chem. 2012, 84, 6355−6360

Analytical Chemistry

Technical Note

Figure 6. An MS/MS spectrum of the peptide containing the original Asn. The amino acid sequence and the major fragment ions are shown on top of the figure. The Y and B ion m/z are calculated based on the amino acid sequence. The observed Y ion m/z will be 2 or 4 Da higher than the calculated molecular weights because of the incorporation of either one or two 18O-atoms to the peptide C-terminal carboxyl groups. The observed B ion m/z will be the same as calculated for the amino acid sequence.

Figure 4. Full mass spectra of the doubly charged peptide containing either isoAsp (A), Asn (B), or Asp (C) from digestion of the antibody in normal water.

Figure 7. Enlarged MS/MS spectra of the B ion of peptide containing either isoAsp (A), Asn (B), and Asp (C). The isotopic distributions were used for the calculation.

overlapping isotopic distributions, one with a monoisotopic m/ z of 1781.78 and the other with a monoisotopic m/z of 1783.78 Da. The m/z of 1781.78 corresponds to the peptide with deamidation of the Asn residue that occurred prior to sample preparation. The m/z 1783.78 corresponds to the peptide with the deamidation of the Asn residue that occurred during sample preparation including incorporation of an 18O-atom to the newly formed carboxyl groups of isoAsp and Asp. Higher intensity of the peak with m/z of 1783.78 compared to that of m/z 1781.78 indicates a higher level of Asn deamidation that occurred during sample preparation. The peak ratios of m/z 1781.78 and 1783.78 can be used to calculate the level of Asn deamidation that occurred prior to or during sample preparation.

Figure 5. Full mass spectra of the doubly charged peptide containing either isoAsp (A), Asn (B), or Asp (C) from digestion of the antibody in 18O-water.

isotopic distributions of this b ion from peaks 1 and 3 are shown in Figure 7A (isoAsp) and 7C (Asp), respectively. Clearly the mass spectra in Figure 7A,C are the results of two 6358

dx.doi.org/10.1021/ac3013362 | Anal. Chem. 2012, 84, 6355−6360

Analytical Chemistry

■

CONCLUSIONS In the present study, a method to differentiate Asn deamidation prior to and during sample preparation was established. Sample preparation including trypsin digestion in buffers prepared in 18 O-water resulted in a molecular weight increase of 3 Da at the deamidation site if deamidation occurred during sample preparation, while deamidation that occurred prior to sample treatment results in a molecular weight increase of 1 Da. This molecular weight difference has been used previously to differentiate Asn deamidation that occurred prior to or during sample preparation under a specific condition. However, variation in the incorporation of 18O-atom into the peptide C-terminal carboxyl groups limits its application under general peptide mapping conditions. This limitation was overcome by using the b ion generated from fragmentation of the peptides containing either isoAsp or Asp for calculation. This newly improved method can thus be generally applied to determine the level of Asn deamidation that is present in the samples of interest without interference of sample preparation artifacts. Such a method in combination with the methods of identifying isoAsp will allow the use of LC−MS for an accurate quantitation of isoAsp and Asp in proteins.

The percentage of Asn deamidation that occurred prior to sample preparation was calculated by the following steps (Table 1). Step 1, the percentage of the total isoAsp and Asp at the Table 1. Calculation of Percentage of the Asn Deamidation in the Samplea step 1 step 2 step 3 a

calculation methods

peak 1 (isoD)

peak 3 (D)

EIC peak areas (Figure 3) ratios (Figure 7) step 1 × step 2

22.1 ± 1.8 20.7 ± 2.6 4.6 ± 0.5

5.2 ± 0.13 27.9 ± 2.7 1.5 ± 0.2

Technical Note

The data represent an average of triplicate experiments.

time of analysis was calculated using the relative extracted ion chromatogram peak areas as shown in Figure 3. The results showed that 22.1% of the peptide contains isoAsp and 5.2% of the peptide contains Asp. Step 2, the percentage of isoAsp and Asp that were generated prior to or during sample preparation was calculated using mass spectra in Figure 7. The peak of m/z of 1781.78 represents the level of deamidation that occurred prior to the sample preparations. The peak of m/z of 1783.78 represents the level of deamidation that occurred during sample preparations and the overlapping from the third peak of the m/ z 1781.78 isotopic peak series. The overlap was calculated by multiplying the peak intensity of m/z of 1781.78 (Figure 7A,C) by the ratio of the third peak over the first peak in Figure 7B. The peak intensity of m/z of 1781.78 divided by the sum of the peak intensities of m/z of 1781.78 and m/z of 1783.78 after subtraction of the overlap results in the percentage of the deamidation that occurred prior to the sample preparation. This percentage was further corrected by following the equation provided as Supporting Information because the purity of the 18O-water is only 97.1%. As summarized in Table 1, only 20.7% isoAsp and 27.9% Asp of the total isoAsp and Asp detected are present in the original samples. Step 3, the percentage of isoAsp and Asp that was present prior to the sample preparations was calculated by multiplying the relative EIC peak areas (step 1) by the percentage from step 2, which resulted in 4.6% for isoAsp and 1.5% for Asp. So, although, at the time of analysis, the samples contained 22.1% isoAsp and 5.2% Asp, the majority of the isoAsp and Asp were formed during sample preparations. To further demonstrate the utility of this method, a second peptide containing an Asn residue that is followed by a serine residue was also studied by following the same procedure. As shown in the Supporting Information, 63.3% isoAsp and 44.1% Asp that was detected at the time of analysis is present in the sample prior to sample treatment. A much lower level artifact was observed for this Asn residue compared to the one followed by a glycine residue, as expected. Attempt was also made to calculate the level of artifact of the second Asn residue in the first peptide. However, b ions with the same criteria cannot be identified. Therefore, methods such as chemical derivation to enhance the generation of b ions will be worthwhile to explore for future studies. In summary, the newly established method was able to determine the level of Asn deamidation that occurred prior to or during sample preparation using a particular b ion from LC− MS/MS fragmentation of isoAsp and Asp containing peptides. This method along with the recent development of several chemical and enzymatic methods to identify isoAsp residue30−32 will allow an accurate identification and quantitation of isoAsp and Asp in proteins.

■

ASSOCIATED CONTENT

* Supporting Information S

Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: 908-820-6162. Notes

The authors declare no competing financial interest.

■

REFERENCES

(1) Harris, R. J.; Kabakoff, B.; Macchi, F. D.; Shen, F. J.; Kwong, M.; Andya, J. D.; Shire, S. J.; Bjork, N.; Totpal, K.; Chen, A. B. J. Chromatogr., B: Biomed. Sci. Appl. 2001, 752, 233−245. (2) Huang, L.; Lu, J.; Wroblewski, V. J.; Beals, J. M.; Riggin, R. M. Anal. Chem. 2005, 77, 1432−1439. (3) Lyubarskaya, Y.; Houde, D.; Woodard, J.; Murphy, D.; Mhatre, R. Anal. Biochem. 2006, 348, 24−39. (4) Zheng, J. Y.; Janis, L. J. Int. J. Pharm. 2006, 308, 46−51. (5) Harris, R.; Shire, S. J.; Winter, C. Drug Dev. Res. 2004, 61, 137− 154. (6) Chelius, D.; Rehder, D. S.; Bondarenko, P. V. Anal. Chem. 2005, 77, 6004−6011. (7) Terashima, I.; Koga, A.; Nagai, H. Anal. Biochem. 2007, 368, 49− 60. (8) Gaza-Bulseco, G.; Li, B.; Bulseco, A.; Liu, H. C. Anal. Chem. 2008, 80, 9491−9498. (9) Sinha, S.; Zhang, L.; Duan, S.; Williams, T. D.; Vlasak, J.; Ionescu, R.; Topp, E. M. Protein Sci. 2009, 18, 1573−1584. (10) Wang, L.; Amphlett, G.; Lambert, J. M.; Blättler, W; Zhang, W. Pharm. Res. 2005, 22, 1338−1349. (11) Mukherjee, R.; Adhikary, L.; Khedkar, A.; Iyer, H. Rapid Commun. Mass Spectrom. 2010, 24, 879−884. (12) Liu, Y. D.; van Enk, J. Z.; Flynn, G. C. Biologicals. 2009, 37, 313−322. (13) Perkins, M.; Theiler, R.; Lunte, S.; Jeschke, M. Pharm. Res. 2000, 17, 1110−1117. (14) Vlasak, J.; Bussat, M. C.; Wang, S.; Wagner-Rousset, E.; Schaefer, M.; Klinguer-Hamour, C.; Kirchmeier, M.; Corvaia, N.; Ionescu, R.; Beck, A. Anal. Biochem. 2009, 392, 145−154. 6359

dx.doi.org/10.1021/ac3013362 | Anal. Chem. 2012, 84, 6355−6360

Analytical Chemistry

Technical Note

(15) Zhang, W.; Czupryn, M. J. J. Pharm. Biomed. Anal. 2003, 30, 1479−1490. (16) Geiger, T.; Clarke, S. J. Biol. Chem. 1987, 262, 785−794. (17) Niebecker, R.; Kloft, C. Curr. Drug Safety 2010, 5, 275−286. (18) Mamula, M. J.; Gee, R. J.; Elliott, J. I.; Sette, A.; Southwood, S.; Jones, P. J.; Blier, P. R. J. Biol. Chem. 1999, 274, 22321−22327. (19) Moss, C. X.; Matthews, S. P.; Lamont, D. J.; Watts, C. J. Biol. Chem. 2005, 280, 18498−18503. (20) Cacia, J.; Keck, R.; Presta, L. G.; Frenz, J. Biochemistry 1996, 35, 1897−1903. (21) Rehder, D. S.; Chelius, D.; McAuley, A.; Dillon, T. M.; Xiao, G.; Crouse-Zeineddini, J.; Vardanyan, L.; Perico, N.; Mukku, V.; Brems, D. N.; Matsumura, M.; Bondarenko, P. V. Biochemistry 2008, 47, 2518− 2530. (22) Ren, D.; Pipes, G. D.; Liu, D.; Shih, L. Y.; Nichols, A. C.; Treuheit, M. J.; Brems, D. N.; Bondarenko, P. V. Anal. Biochem. 2009, 392, 12−21. (23) Krokhin, O. V.; Antonovici, M.; Ens, W.; Wilkins, J. A.; Standing, K. G. Anal. Chem. 2006, 78, 6645−6650. (24) Stroop, S. D. Rapid Commun. Mass Spectrom. 2007, 21, 830− 836. (25) Li, X.; Cournoyer, J. J.; Lin, C.; O’Connor, P. B. J. Am. Soc. Mass Spectrom. 2008, 19, 855−864. (26) Schnolzer, M.; Jedrzejewski, P.; Lehmann, W. D. Electrophoresis 1996, 17, 945−953. (27) Hajkova, D.; Rao, K. C.; Miyagi, M. J. Proteome Res. 2006, 5, 1667−1673. (28) Miyagi, M.; Rao, K. C. Mass Spectrom. Rev. 2007, 26, 121−136. (29) Fenselau, C.; Yao, X. J. Proteome Res. 2009, 8, 2140−2143. (30) Alfaro, J. F.; Gillies, L. A.; Sun, H. G.; Dai, S.; Zang, T.; Klaene, J. J.; Kim, B. J.; Lowenson, J. D.; Clarke, S. G.; Karger, B. L.; Zhou, Z. S. Anal. Chem. 2008, 80, 3882−3889. (31) Ni, W.; Dai, S.; Karger, B. L.; Zhou, Z. S. Anal. Chem. 2010, 82, 7485−7491. (32) Liu, M.; Cheetham, J.; Cauchon, N.; Ostovic, J.; Ni, W.; Ren, D.; Zhou, Z. S. Anal. Chem. 2012, 84, 1056−1062.

6360

dx.doi.org/10.1021/ac3013362 | Anal. Chem. 2012, 84, 6355−6360