Article Cite This: Anal. Chem. XXXX, XXX, XXX−XXX
pubs.acs.org/ac
Reductive Amination Combining Dimethylation for Quantitative Analysis of Early-Stage Glycated Proteins Qing-He Tong,†,∥ Tian-Yang Yan,§,∥ Tao Tao,† Lei Zhang,† Li-Qi Xie,*,† and Hao-Jie Lu*,†,‡,§ †
Shanghai Cancer Centre and Institutes of Biomedical Sciences, and ‡Key Laboratory of Glycoconjuates Research Ministry of Public Health, Fudan University, Shanghai 200032, People’s Republic of China § Department of Chemistry, Fudan University, Shanghai 200433, People’s Republic of China S Supporting Information *
ABSTRACT: Due to the critical role glycation plays in many serious pathological conditions, such as diabetes, it is of great significance to discover protein glycation at an early stage for precaution and prediction of the disease. Here, a method of reductive amination combining dimethylation (RAD) was developed for the quantification of early-stage glycated proteins. The quantitative analysis was first carried out by reducing the samples using NaBH3CN or NaBD3CN, resulting in a 1 Da mass shift and the stabilization of early-stage protein glycation. The two samples were then digested and isotopically dimethylated to achieve the mass shift of 4m + 3n (m represents the number of N-termini and Lys residues, and n represents the number of glycated sites) between light- and heavy-labeled glycated peptides for quantification. Consequently, the false positive result can be removed according to the different mass shifts of glycated peptides and non-glycated peptides. In quantification of glycated myoglobin, RAD showed good linearity (R2 > 0.99) and reproducibility (CVs ≤ 1.6%) in 2 orders of magnitude (1:10−10:1). RAD was then applied to quantify the endogenous glycated proteins in the serum of diabetic patients, revealing significant differences in the glycation level between the patients with complicated retinal detachment and those without. In conclusion, RAD is an effective method for quantifying endogenous glycated proteins.
G
Glycation is a highly disease-relevant modification, associating with diabetes, angiocardiopathy, and other chronic complications of age-related degenerative diseases.7,8 High blood glucose level is the principal feature of diabetes mellitus, which is one of the top three chronic incurable diseases.9 The glycated hemoglobin (HbA1c) has been widely incorporated into the management and the diagnosis of diabetic patients for decades.10 Poorly controlled hyperglycemia may lead to multiple serious complications, retinopathy, diabetic nephropathy, neuropathy, and other macrovascular diseases, which all require sophisticated medications.11 There are, however, no early symptoms or signs for diabetic retinopathy before focal blurring or vitreous or retinal detachment.12 To find more sensitive and informative protein biomarkers in precaution and prediction of these diseases,
lycated proteins are formed by nonenzymatic reaction between reducing sugars (such as glucose, fructose, ribose, or their derivatives) and amino groups located in the N-terminal or in lysine and arginine residues.1,2 In contrast to enzymemediated glycosylation, glycation is slow, reversible, and quite dynamic in nature. Take glucose, the major free sugar in the human body, for example; about 0.2% of proteins or 0.01% of amino groups are glycated a day at normal (5 mM) blood glucose concentration in vivo.3 Also, the rates of glycation are different for different proteins. For example, the glycation of albumin is about 9 times faster than that of hemoglobin.4 Accordingly, it is necessary to incorporate an enrichment step to enable efficient identification of these low-level post-translational modifications.5 Since the boronate ion has strong and reversible coordination with cis-diols, the m-aminophenylboronic acid material was used to selectively enrich glycated peptides or proteins. In addition, glycation is often thermally and chemically labile when removed from the physiological environment.6 Accordingly, the protein glycation needs to be enriched and stabilized for better identification. © XXXX American Chemical Society
Received: September 7, 2017 Accepted: February 21, 2018 Published: February 21, 2018 A
DOI: 10.1021/acs.analchem.7b03668 Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
for 1 h at room temperature. The reaction was terminated with 4 μL of 4% ammonia. RAD Labeling of Serum Samples. Sera of diabetic patients were obtained from patients during physical examination at Affiliated Hospital of Nantong University. All investigations described in this study were performed after informed consent was obtained and in accordance with an institutional review board protocol approved by the Human Research Ethic Committee at Affiliated Hospital of Nantong University. To apply RAD to analyze serum, sera of 15 diabetic patients without complicated retinal detachment were mixed isovolumetrically. An amount of 100 μL of the mixture was then reduced with 75 mM sodium borohydride (NaBH4) for 4 h at 37 °C. Similarly, 100 μL of equally mixed serum from 15 diabetic patients with complicated retinal detachment was reduced with 75 mM sodium borodeuteride (NaBD4). The heavy- or light-labeled serum was then ultrafiltered to remove reductant and glucose. After that, two sets of solution were reacted with 100 mM DTT for 1 h at 37 °C and 250 mM IAA for 45 min at room temperature in the dark. Digestion was carried out using trypsin at a final enzyme to protein ratio of 1:30 for 16 h at 37 °C. A C18 column was used to remove salt. Heavy- and light-labeled samples after digestion were dimethylated with CH2O and CD2O, respectively, and then enriched with boric acid for 16 h at 37 °C, 1100 rpm. Mass Spectrometry Analysis. RAD-labeled standard samples were analyzed by a matrix-assisted laser desorption ionization time-of-flight time-of-flight mass spectrometer (MALDI-TOF-TOF-MS) (5800 Proteomics Analyzer, Applied Biosystems, Framingham, MA, U.S.A.). Serum samples were separated on an Easy Nano LC system using a C18 column (75 μm i.d. × 50 cm, 2 μm i.d. 100 Å pore size, Thermo Fisher Scientific, U.S.A.) coupled online with an Orbitrap Fusion mass spectrometer (Thermo Fisher Scientific, U.S.A.). The gradient used for separation was from 2% to 35% acetonitrile in 130 min at a flow rate of 300 nL/min. The fraction A was 0.1% formic acid (FA) in water, and the fraction B was 0.1% FA in ACN. The mass spectrometer was run at positive mode (ESI+) and data-dependent mode (DDA, top speed). The mass range of a full scan was set as mass-to-charge ratio (m/z) from 350 to 1800 with a mass resolution of 120 000. The automatic gain control settings during fullMS and tandem mass spectrometry (MS/MS) were 500 000 and 10 000, respectively. The normalized collision energy value was set at 35%, and the previously fragmented peptides were excluded for 30 s. Data Analysis. All MS/MS spectra were searched using Mascot Daemon 2.3. The searching parameters for light labeling were set up as follows: protein database was human proteins; missed cleavages of trypsin K/R was up to 4; dimethylation of N-terminal (+28.031 Da) and carbamidomethylation of cysteine (+57.01 Da) were set up as fixed modifications; glycation and methylation of lysine (+178.084 Da), dimethylation of lysine (+28.031 Da), and oxidation of methionine (+15.995 Da) were set as variable modifications; a single-isotope peak was used to retrieve peptides; and because of the high resolution of the Orbitrap analyzer, the peptide mass tolerance was 5 ppm, and the fragment ion tolerance was 0.02 Da. All the tandem mass spectra underwent manual correction to ensure accurate m/z and matching of fragments. Charge numbers of the peptides were 2, 3, and 4. Consequently, quality control chose peptides with a false discovery rate (FDR) less than 1% or that contained two peptides with a confidence limit of 95%. Similarly, the searching parameters for heavy labeling were set up as the same as above except that dimethylation of N-terminal (+32.056 Da) and
comprehensive proteomic studies are required to identify those glycated proteins whose altered structure may contribute to pathology. Therefore, it is of great importance to develop a method for quantitative analysis of early-stage glycated proteins. Mass spectrometry is often the method of choice for detection and quantification of glycation adduct content of biological samples where glycation on all proteins in complex mixtures could be simultaneously assessed.13 Presently, the mass spectrometry based quantification of early-stage glucose glycated products is mainly relying on stable isotope labeling, including enzymatic 16O/18O labeling, in vitro metabolic isotope labeling, and chemical labeling.14 The enzymatic 16O/18O labeling strategy was successfully applied to the quantification of human serum albumin (HSA) glycation under different circumstances and the recognition of glycation sites.15,16 The Priegocapote group2,17 incubated samples with [13C6] glucose or [12C6] glucose, respectively, causing a 6 Da mass shift per glycation site. The method realized the comparison of glycation under different blood sugar levels. However, these methods could only be used to quantify in vitro glycation. To quantify in vivo glycated proteins of complicated samples, Liu et al. reduced the early-stage glycation samples using NaBH3CN or NaBD3CN, resulting in 1 Da mass shift between glycated samples, and meanwhile stabilized the glycation of proteins.18 However, it is difficult to distinguish the isotope peaks because of the minor 1 Da difference between the heavy and light labeling. Therefore, its application on complicated practical samples is also impeded. To address current problems in the quantification of early-stage glycated proteins, we described a novel quantitative strategy, reductive amination combining dimethylation (RAD), shown in Figure 1. The quantitative analysis was first carried out by reducing the samples using NaBH3CN or NaBD3CN, resulting in a 1 Da mass shift between samples and the stabilization of early-stage protein glycation. The two samples were then digested with trypsin and labeled with dimethylation agent (CH2O, CD2O), respectively. In this way, the isotope peaks could be easily recognized and the false positive result could be removed according to the different mass shift of glycated peptides and non-glycated peptides. Meanwhile, quantification at the protein and peptide level were both achieved. When applied to complicated samples, RAD succeeded in quantifying the endogenous glycated proteins in the serum of a diabetic patient, revealing significant level differences of glycation between the patients with and without complicated retinal detachment. These results indicated that RAD could be used for accurate quantitative analysis of earlystage glycated proteins.
■
MATERIALS AND METHODS All chemicals used in this study were purchased from Sigma (St. Louis, MO) unless specified otherwise. Deuterated formaldehyde (CD2O) was purchased from Isotech. Acetonitrile (ACN) was purchased from Merck. Glycation of Standard Sample. Myoglobin or standard peptides (4 mg/mL) were glycated in 1 M glucose and 10 mM PBS with or without reductant (300 mM sodium cyanoborohydride, NaBH3CN) for 24 h at 37 °C. The labeled myoglobin was ultrafiltered (MWCO = 3 kDa) to remove reductant and glucose. A solution of 10 mM TEAB was used to replace the solution. Dimethylation of Glycated Proteins. The labeled myoglobin was then digested with trypsin (myoglobin/trypsin = 50:1) at 37 °C for 16 h and denatured at 100 °C for 10 min. After that, 100 μg portions of digested peptides were heavy (light) labeled with 4 μL of 4% CD2O (CH2O) and 4 μL of 600 mM NaBH3CN B
DOI: 10.1021/acs.analchem.7b03668 Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
■
Figure 1. Workflow of RAD.
RESULTS AND DISCUSSION Establishment of RAD. RAD is a quantification method for protein glycation based on reductive amination and dimethylation (Figure 1). Briefly, two early-stage glycation products are reduced by either NaBH3CN or NaBD3CN and meanwhile
carbamidomethylation of cysteine (+57.01 Da) were set up as fixed modifications and glycation and methylation of lysine (+181.102 Da), dimethylation of lysine (+32.056 Da), and oxidation of methionine (+15.995 Da) were set as variable modifications. C
DOI: 10.1021/acs.analchem.7b03668 Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
Figure 2. Mass spectra of myoglobin peptide ALELFR detected instantly after glycation (a and b) and 24 h after the removal of glucose (c and d). The peptide was glycated with (a and c) or without reducing reagent (b and d). The intensity of the non-glycated form and that of the glycated form was compared.
the molecular weight by 162 Da and the quantity of glycated peptides was much less than that of the non-glycated ones, while the glycation level was increased when NaBH3CN was used. After the removing of glucose for 24 h, the glycated peptide almost vanished, while the reduced glycated peptides remain unchanged (Figure 2, parts c and d). Therefore, reductive amination should be used to stabilize protein glycation before digestion. To test the feasibility of RAD, a glycated peptide YIGIVKgQAGLER (MW 1508.60) was heavy-RAD-labeled to test the efficiency of reduction. The efficiency of the first step of reduction reached 95.04%, and the second step of reduction was nearly complete (Figure S-5). After that, glycated myoglobin was reduced, digested, and dimethylated. Taking the light-labeled myoglobin as an example, after dimethylation, the molecular weight of a non-glycated peptide increased by 28 Da due to the dimethylation of the N-terminal (Figure S-1). Meanwhile, the molecular weight increase of a peptide with two glycation sites was 56 Da, which contained 28 Da of N-terminal dimethylation and 14 Da of methylation on each glycation site. Then, heavyand light-labeled peptides were mixed and detected. As show in Figure S-2, the mass shift of light- and heavy-labeled glycated peptides turned out to be 10 Da [(4 × 1) + (3 × 2)] or 14 Da [(4 × 2) + (3 × 2)], indicating that RAD is a viable method for quantification of early-stage glycated proteins. Accuracy, Reproducibility, and Dynamic Range of RAD in Quantifying Standard Exogenous Glycated Proteins. First of all, glycated peptides were mixed at light-to-heavy mole ratios 10:1, 5:1, 2:1, 1:1, 1:2, 1:5, 1:10. Three replicates for technical reproducibility were processed. As shown in Figure 3,
stabilized. Every glycation site thus gets a 1 Da difference between heavy and light labeling. Then, the two samples are digested and respectively isotopically dimethylated (CH2O, CD2O). During the reaction, the free amino group takes on two methyl residues while the secondary amine of the glycated site (obtained from reduction) takes on one methyl residue. As a result, the final mass shift between light- and heavy-labeled glycated peptides expands to 4m + 3n (m represents the number of N-termini and Lys residues, and n represents the number of glycated sites). The 3 Da mass shift of the glycation site is composed of a 1 Da difference caused by differential reduction and a 2 Da difference caused by the isotopic methylation of the secondary amine. In addition, the light- and heavy-RAD-labeled non-glycated peptides with 4m mass shift can also be used for quantification. A formula R(L/H) = (S1/S1′)/(S2/S2′) was used to compare the differences between light- and heavy-labeled glycation peptides from a certain protein. S1 and S1′ represented the signal strength of light- or heavy-labeled glycated peptides. S2 and S2′ were the median of the signal strength of all light- and heavy-labeled non-glycated peptides that belong to the protein, respectively, which showed the difference between the amount of the protein in the two samples. As a result, quantification can be achieved on both modification sites and the protein level. Because the early-stage glycation products were in the midst of a reversible dynamic balance, when glucose was removed from the system, the shift of the chemical equilibrium resulted in the decreasing of the products. To study the effect, the peptides of trypsin-digested myoglobin were glycated with or without reducing reagent. As shown in Figure 2, single glycation increased D
DOI: 10.1021/acs.analchem.7b03668 Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
Figure 3. Mass spectra of light- and heavy-labeled glycated peptide NDIAAKgYKgELGFQG mixed with the ratio of (a) 1:1, (b) 2:1, (c) 5:1, and (d) 10:1.
reversible covalent ester reacting with 1,2- and 1,3-diols in the solution, which are structures contained in the early-stage glycation products,5 commercial boric acid material was used to enrich the glycated peptides. To optimize the condition, we considered the ratio of the peptides and the material. As shown in Figure S-3, before the enrichment, peaks were mainly non-glycated peptides. As the amount of boric acid increased, the signal strength of glycated peptides enlarged. When the ratio of the material and the peptides reached 200:1, the signal became saturated. And if we continued adding the material, instead of amplifying the signal, nonspecific adsorption increased. As a result, we chose 200:1 as the optimized ratio of the boric acid material and the peptides in the following enrichment process. Meanwhile, to evaluate the interferences from other glycosylated peptides during enrichment, trypsin-digested asialofetuin (ASF) and heavy-RAD-labeled glycated peptides were mixed at the ratio of 20:1 for boric acid enrichment according to the ratio of glycosylated and glycated proteins in serum.19,20 After that, the same amount of light-RAD-labeled and enriched glycated peptide was added and detected as shown in Figure S-4. Peak intensities of light- and heavy-labeled glycated peptide are almost the same, which indicated that there were no obvious interferences from glycosylated peptide during enrichment. Since we focus on the quantification of glycated proteins, unless both the HPLC retention time and mass-to-charge ratio of the glycated and glycosylated peptides are the same, there is no interference. Application of RAD in Quantifying Endogenous Glycated Proteins. Since there is a certain amount of glucose existing in the serum, reduction might result in exogenous glycation. In order to avoid exogenous glycation and optimize the conditions,
the measured and theoretical ratios of the glycated peptides were close and the coefficients of variation (CVs) were between 0.8% and 1.6%, showing good accuracy and stability. Meanwhile, the correlation coefficients, R2, of dual-logarithm plots of all the theoretical and measured ratios (Figure 4) were greater than 0.99,
Figure 4. Dual-logarithm plots of the theoretical and measured ratios.
showing good linearity. We can therefore reach a conclusion from the results above that, in quantification of glycated proteins, RAD showed good linearity (R2 > 0.99) and reproducibility (CVs ≤ 1.6%) in 2 orders of magnitude (1:10−10:1). Enrichment of Glycated Peptides. Because there are very few endogenous glycated peptides compared to the non-glycated ones, it is of great importance to selectively enrich the glycated peptides before the quantification. Since boric acid can form a E
DOI: 10.1021/acs.analchem.7b03668 Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
Figure 5. Mass spectra of RAD-labeled (a) glycated peptide VTKgCCTESLVNR and (b) non-glycated peptide LVTDLTK in the serum of diabetic patients.
samples of serum. Among them, 61 peptides had a single glycation site and seven peptides had two glycation sites. There were six glycation peptides only identified in diabetic patients without retinal detachment, and 20 glycation peptides only identified in diabetic patients with complicated retinal detachment. For the 42 glycation peptides identified in both samples, the glycation difference between these two kinds of serum was quantified, resulting in 39 peptides with R(L/H) below 1 and three peptides above, which indicate higher protein glycation level in retinal detachment serum. HSA is commonly considered as the indicator of diabetes and hyperglycemia. In our experiment, both glycated and unmodified HSA peptides were identified. For example, the heavy- and lightlabeled glycated peptide VTKgCCTESLVNR from HSA shows a mass shift of 7 Da (Figure 5a). Meanwhile, the heavy- and lightlabeled non-glycated peptide LVTDLTK has an 8 Da mass shift (Figure 5b). On the basis of these mass spectra, we first calculated the ratio of all HSA peptides in the two samples to quantify HSA at the protein level. The abundance of HSA is 1.14-fold lower in diabetic patients with complicated retinal detachment. We then determined the ratio of glycated HSA peptides and combined it to the ratio of HSA proteins. Among all the glycation sites detected, the HSA glycation of diabetic patients without retinal detachment was on average less than that of diabetic patients with complicated retinal detachment. Glycation of all the other proteins detected both in the serum of patients with and without retinal detachment was also examined and is listed in Table S-3. As the toxicity of glucose was generally considered the main reason for diabetic complication, these diseases usually occur several years after diabetes.21,22 Chronic hyperglycemia can increase the concentration of blood sugar by inhibiting pancreas cell activity, hence increasing the rate of early diabetes. This vicious cycle may eventually cause all the pancreatic β cells to stop insulin release.23,24 Therefore, glycation is often deemed to be associated with diabetes, renal failure, and other chronic complications of age-related degenerative diseases.7,8 It means that the HSA glycation level of diabetic patients with complication ought to be higher than that of those without, which corresponds to the results of RAD. In conclusion, RAD can be used to quantify endogenous glycated proteins in complicated, practical samples. In addition, pair peaks of heavy- and light-labeled glycated peptides that resulted with a mass shift of 4m + 3n enable us to eliminate false positive results. For instance, after a search in the database, the triply charged peptide with m/z of 594.35 resulted
we studied the exogenous glycation level generated by the reductive reaction as time grew. As shown in Table S-1, reductant concentration of 75 mM and reaction time of 4 h were the optimized conditions. With the same reductant concentration, the exogenous glycation level increased over the reaction time. Meanwhile, the glycation depended on the concentration of the reductant within the same reaction time. There were altogether 20 glycation hot spots in myoglobin, including 19 lysine residues and one N-terminal. We can indicate from Table S-1 that, with the reductant concentration of 300 mM, the number of exogenous glycation sites reached 19 after 108 h of reaction. However, glycation did not increase over time when there was too little reductant, mainly because most of the reductant had been deactivated in H2O after a long reaction time. Besides, to test the applicability of this optimized reductive condition to low-abundant glycated proteins, we designed the following experiment. Glycated peptide YIGIVKgQAGLER was diluted with water or 5 mM glucose, respectively, to different concentrations (1, 0.1, 0.01, 0.001 mg/mL). The two sets of samples were then treated with 75 mM NaBH3CN and NaBD3CN, respectively, for 4 h, and then labeled with dimethylation agent (CH2O, CD2O), respectively. Mass spectra of 1:1 mixed light- and heavy-RAD-labeled peptide with the same concentration are shown in Figure S-6. Reduced peptides with the concentration of 0.001 mg/mL can be well-detected as well. Meanwhile, there were no obvious differences of the peak intensities between those reduced in water or 5 mM glucose, indicating there is no exogenous glycation. To sum up, a reductant concentration of 75 mM and reaction time of 4 h were chosen as the optimized conditions, for endogenous Schiff base was well-reduced, while no exogenous glycation was introduced under these conditions. Application of RAD in Analyzing Serum of Diabetic Patients. RAD was applied to quantify endogenous glycation proteins in the serum of diabetic patients with complicated retinal detachment and those without. We first reduced the serum of diabetic patients with complicated retinal detachment with NaBD3CN and the serum of diabetic patients without retinal detachment with NaBH3CN. Samples were then desalted, digested, heavily or lightly dimethylated, mixed, enriched, and finally quantified. We manually checked all the MS/MS spectra identified as glycated peptide; only those including fragment ions containing glycation sites were picked out and exhibited in the Supporting Information. As shown in Table S-2, all together 75 endogenous glycation sites of 68 peptides belonging to 30 proteins were found in the two F
DOI: 10.1021/acs.analchem.7b03668 Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
■
in peptide VLGEKgETLLYENK of SYCP1HUMAN protein, which was a heavy-labeled glycated peptide. However, according to the original spectra, there was another heavy-labeled peak behind the peak of m/z = 594.35 and the mass shift was exactly 11 Da (Figure S-7). This indicated that the peptide with m/z of 594.35 ought to be a light-labeled peptide, thus removing the false positive results.
CONCLUSION On the basis of reductive amination and dimethylation, we successfully proposed a comprehensive and improved strategy targeting quantification of early-stage glycated proteins, which we termed RAD, and validated its effectiveness by applying it in the different serum samples. The experimental results indicated that this novel strategy had good linearity (R2 > 0.99) and reproducibility (CVs ≤ 1.6%) in 2 orders of magnitude (1:10−10:1) in the quantification of standard glycated myoglobin. Moreover, it managed to remove the false positive results based on the specific mass shift. When applied to complicated samples, RAD succeeded in quantifying the endogenous glycated proteins in the serum of diabetic patients at both the protein and peptide level, revealing significant level differences of glycation between the patients with complicated retinal detachment and those without. RAD, in light of its simplicity, accuracy, and reliability, promises to be applied and popularized in the quantitative proteomics field. ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.7b03668. Number of exogenous glycation sites of myoglobin generated under different reaction times and reductant concentrations, ratio of endogenous glycated peptides detected in the serum of diabetic patients without and with complicated retinal detachment, comparison of protein glycation in the serum of diabetic patients with or without complicated retinal detachment, and MALDI-TOF mass spectra (PDF) MS/MS spectra of peptides (PDF)
■
REFERENCES
(1) Arena, S.; Salzano, A. M.; Renzone, G.; D'Ambrosio, C.; Scaloni, A. Mass Spectrom. Rev. 2014, 33, 49−77. (2) Priegocapote, F.; Scherl, A.; Muller, M.; Waridel, P.; Lisacek, F.; Sanchez, J. Mol. Cell. Proteomics 2010, 9, 579−592. (3) Zhang, Q.; Ames, J. M.; Smith, R. D.; Baynes, J. W.; Metz, T. O. J. Proteome Res. 2009, 8, 754−769. (4) Garlick, R. L.; Mazer, J. S. J. Biol. Chem. 1983, 258, 6142−6146. (5) Zhang, Q.; Tang, N.; Brock, J. W. C.; Mottaz, H. M.; Ames, J. M.; Baynes, J. W.; Smith, R. D.; Metz, T. O. J. Proteome Res. 2007, 6, 2323− 2330. (6) Thornalley, P. J.; Rabbani, N. Biochim. Biophys. Acta, Gen. Subj. 2014, 1840, 818−829. (7) Hipkiss, A. R. Exp. Gerontol. 2006, 41, 464−473. (8) Brownlee, M. Nature 2001, 414, 813−820. (9) Wild, S. H.; Roglic, G.; Green, A.; Sicree, R.; King, H. Diabetes Care 2004, 27, 1047−1053. (10) Welsh, K. J.; Kirkman, M. S.; Sacks, D. B. Diabetes Care 2016, 39, 1299−1306. (11) Liu, L.; Tang, J.; Cheng, Y.; Agrawal, A.; Liao, W.; Choudhary, A. In Proceedings of the 22nd ACM International Conference on Information and Knowledge Management; ACM: New York, 2013; pp 279−288. (12) Klein, R.; Klein, B. E. K. JAMA, J. Am. Med. Assoc. 2016, 315, 1778−1779. (13) Rabbani, N.; Ashour, A.; Thornalley, P. J. Glycoconjugate J. 2016, 33, 553−568. (14) Zhang, M.; Xu, W.; Deng, Y. Diabetes 2013, 62, 3936−3942. (15) Barnaby, O. S.; Cerny, R. L.; Clarke, W.; Hage, D. S. Clin. Chim. Acta 2011, 412, 277−285. (16) Barnaby, O. S.; Cerny, R. L.; Clarke, W.; Hage, D. S. Clin. Chim. Acta 2011, 412, 1606−1615. (17) Ramirezboo, M.; Priegocapote, F.; Hainard, A.; Gluck, F.; Burkhard, P. R.; Sanchez, J. J. Proteomics 2012, 75, 4766−4782. (18) Liu, H.; Ponniah, G.; Neill, A.; Patel, R.; Andrien, B. J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 2014, 958, 90−95. (19) Zhang, Q.; Tang, N.; Brock, W. C. J.; Mottaz, M. H.; Ames, M. J.; Baynes, W. J.; Smith, D. R.; Metz, O. T. J. Proteome Res. 2007, 6, 2323− 2330. (20) Zhang, Y.; Zhang, C.; Jiang, H.; Yang, P.; Lu, H. Chem. Soc. Rev. 2015, 44, 8260−8287. (21) Reusch, J. E. B. J. Clin. Invest. 2003, 112, 986−988. (22) Brunner, Y.; Schvartz, D.; Priegocapote, F.; Coute, Y.; Sanchez, J. J. Proteomics 2009, 71, 576−591. (23) Leroith, D. Am. J. Med. 2002, 113, 3−11. (24) Dubois, M.; Vacher, P.; Roger, B. t.; Huyghe, D.; Vandewalle, B.; Kerrconte, J.; Pattou, F.; Moustaidmoussa, N.; Lang, J. Endocrinology 2007, 148, 1605−1614.
■
■
Article
AUTHOR INFORMATION
Corresponding Authors
*E-mail:
[email protected]. Fax: +86-21-54237961. Phone: +86-21-54237416. *E-mail:
[email protected]. Fax: +86-21-54237961. Phone: +86-21-54237618. ORCID
Hao-Jie Lu: 0000-0003-3477-7662 Author Contributions ∥
Q.-H.T. and T.-Y.Y contributed equally to this work.
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The work was supported by the National Key Research and Development Program (2016YFA0501303), NSF (Grants 21335002 and 31670835), the Ph.D. Programs Foundation of Ministry of Education of China (20130071110034), and Shanghai Projects (Eastern Scholar, 15JC1400700 and B109). G
DOI: 10.1021/acs.analchem.7b03668 Anal. Chem. XXXX, XXX, XXX−XXX
