Assessment of Protein Carbonylation in Type 2 Diabetes Mellitus

ADAM-DIMA Elena-Ines1*, NICOLESCU Florica1, PURDEL Nicoleta Carmen1*, MARGINĂ Denisa2* 

1 Dept. of Toxicology, Faculty of Pharmacy, University of Medicine and Pharmacy of Bucharest, 6, Traian Vuia Street, 020956, Bucharest (ROMANIA)

2Dept. of Biochemistry, Faculty of Pharmacy, University of Medicine and Pharmacy of Bucharest, 6, Traian Vuia Street, 020956, Bucharest (ROMANIA)

Corresponding author: *Authors with equal contribution


Oxidative stress is characterized by massive ROS production, which eventually affects numerous tissues and organs, being associated with many types of diseases, including diabetes mellitus. One of the targets are the proteins, whom physiology is impaired as a consequence of their carbonylation. In this study, we correlated the protein carbonylation induced in vitro by different hydroxyl radical generating systems with in vivo carbonylation process in type 2 diabetes mellitus patients using two different analytical techniques. The results proved there are premises for establishing a carbonylation pattern in diabetics’ serum, identifying well separated protein fraction with low or medium molecular weight.

Keywords: oxidative stress, protein carbonyls, pattern of carbonylation, capillary electrophoresis


Protein carbonylation is a chemical process associated with oxidative stress. Protein carbonyls have been incriminated for causing a variety of metabolic pathologies including insulin resistance, neurodegeneration and aging. Diabetes is a complex disease, involving inflammatory and metabolic pathogenicity, linked to increased ROS levels. Moreover, well-known dermatologic manifestations, such as acanthosis nigricans or diabetic dermopathy are correlated with oxidative imbalance [1]. 

The standard method for studying protein carbonylation in various biological materials (such as plasma, cellular extracts, erythrocytes, or isolated proteins) involves the reaction of the carbonyl groups with 2,4-dinitrophenylhydrazine (DNPH) to obtain the 2,4-dinitrophenylhydrazones (DNP) [2]. This stable reaction product (yellow) has an absorption peak around 370 nm, thus it can be spectrophotometrically assessed [3,4,5]. A major disadvantage of the method is related to protein loss during washing steps, practiced to remove additional chromophores and traces of free DNPH [6,7].  In the same time DNP can be separated using capillary electrophoresis (CE) and structurally characterized, based on molecular weight.  

This study aimed to correlate the results of the protein carbonylation induced in vitro by different hydroxyl radical generating systems with in vivo carbonylation process in type 2 diabetes mellitus patients (T2DM) serum using two different techniques: capillary electrophoresis (CE) with DAD detector and spectrophotometric assay.



All used reagents were of analytically grade and all solutions were prepared with bidistilled water. Iron(II) sulfate, iron(III) chloride, cooper(II) sulfate, DNPH, ascorbic acid, potassium ascorbate, sodium tetraborate, dextran 70, hydrochloric acid, trichloroacetic acid (TCA) (99% GC), guanidine hydrochloride (>99 %), bovine serum albumin (BSA> 98%  66,000 Da), were bought from Sigma Aldrich. Ethyl acetate (ULC-MS) and ethanol were purchased from Scharlau Chemie SA, and human serum albumin (HSA) perfusion solution 20 g/L from Baxter. 

Biological samples

All T2DM patients’ samples came as laboratory waste, after other blood tests were performed, with no specific blood draw. Patients with hematological or cancer pathologies were excluded from the study. Fasting venous blood samples drawn in clot activator vacutainers were centrifuged for serum isolation and the serum was used.

In vitro carbonylation process 

The formation of protein carbonyls was investigated on BSA and HSA. Three different carbonylation systems were tested: 1 mM iron(III) chloride / 25 mM sodium ascorbate; 0.95 mM iron(II) sulfate / 1.5 mM ascorbic acid, and 0.05mM cooper(II) sulfate / 2 mM hydrogen peroxide, each one being applied to 1mg/ml or 5 mg/ml BSA solutions, using incubation at 37oC for 1 to 18 hours. 

Protein carbonyls were evaluated by derivatization, with DNPH 10 mM / HCl 2,5 mM at room temperature for 15 minutes, followed by precipitation with TCA 10% and then incubation for 30 minutes at -100C. The separation of hydrazone pellets was done according to Levine’s method [16], with two major technique optimizing changes: (1) centrifugation at 4000 g for 20 minutes at 40C and (2) two instead of three successive rinses with ice cold ethanol–ethyl acetate mixture (50:50 v/v) for the free DNPH removal, thus diminishing the protein loss. The obtained pellets were redissolved in 6.0 M guanidine hydrochloride solution at 37oC, for 9 hours.  The samples were analyzed at Cary 100 Bio (Varian Inc.) UV-Vis absorption spectrophotometer using 6 M guanidine hydrochloride as blank.

CE analysis 

The assessment of the carbonylation pattern in the HSA carbonylated samples and T2DM serum samples was performed on a G1600A (Agilent) capillary electrophoresis (CE) with diode-array detector (DAD) controlled by Agilent ChemStation ver. B.0 2.0x software. CE method used was developed and validated in our laboratory [8]. Electrophoretic separation was performed using an aqueous electrolyte system (20 mM borate buffer) with pH 9.0. A 15% dextran 70 in 20 mM borate buffer was used for column coating. The filtration through dextran helps the individualization of peaks representing species with close molecular mass. Injection was performed in the hydrodynamic mode at 0.5 psi for 10s, and the applied voltage was 25 kV. The detection was performed at three wavelengths, two regarding PC assay, 370 nm and 365 nm, and 214 nm to confirm the protein structure.

Results and discussions 

In vitro assays focused on the comparative assessment of protein carbonyls generated by the three oxidizing systems using BSA or HSA and the dynamics of carbonylation process for the three oxidizing systems, while CE analysis investigated the carbonyl protein molecular weights.

The carbonyl generation potency of the three oxidizing systems mentioned above was investigated on different concentrations of BSA (1 mg/ml and 5 mg/ml). The protein carbonylation capacity varied decreasingly: FeCl3/sodium ascorbate > ascorbic acid/FeSO4 > CuSO4/H2O2. Therefore, the most efficient carbonyl-generating degrading system is considered FeCl3/sodium ascorbate, with a carbonyl level significantly higher compared with the physiological one (over 30 nmol/mg protein vs. 1 nmol/mg protein) [9] (Fig. 1).


Fig. 1. Carbonyl level generated by the three oxidizing systems in different concentrations of BSA

Similar spectra were observed in the spectrophotometric analysis, when equal concentrations of BSA and HSA (1 up to 5 mg/ml) were degraded using each of the three oxidizing systems mentioned above. The greatest amount of protein carbonyls resulted in the samples that underwent oxidation with FeCl3/sodium ascorbate, for both tested proteins. The results are in line with our previous data.

A direct dependency between the protein carbonyl levels generated by FeCl3/sodium ascorbate system and the incubation time during the first 9 hours of exposure was observed, while for the ascorbic acid/FeSO4 system, the carbonylation process slows down after the first 5 hours. Independently of the oxidation system used, prolonging the incubation up to 18 hours decreased the protein carbonyl concentrations because of their instability. 

CE analysis of T2DM serum samples revealed the presence of 11 DNPs that migrated within 35 minutes, which translates into small or medium carbonyl protein molecular weights (Fig. 2.). 5 of the 11 carbonyls are also present in the in vitro FeCl3/sodium ascorbate degraded HSA samples (Table I).  


Fig. 2.  Electropherograms of DNPs, belonging to T2DM patients (detection at 365 nm)


Table I. DNPs obtained by CE (peaks detected at 365 nm; corresponding peaks are in bold)

Peak nr. Time (min) of migration (expressed as range) Peak nr. Time (min) of migration (expressed as range)
HSA carbonylated in vitro Serum samples from T2DM patients
1 3,846 -3,946 1 4,34 – 4,52
2 4,129 – 4,16 2 7,11 – 7,43
3 4,36 – 4,45 3 7,75 -7,97
4 5,58 – 5,64 4 9,50 – 9,66
5 5,76 – 5,78 5 10,26 – 10,59
6 5,95 – 6,04 6 12,30 – 12,42
7 7,74 – 7,80 7 13,05 – 13,27
8 8,41 – 8,49 8 14,24 – 14,56
9 8,67 – 8,94 9 15,40 – 15,40
10 9,883 – 9,93 10 29,59 – 30,19
11 10,27 – 10,31 11 31,69 – 32,59
12 11,02 – 11,34
13 11,5 – 11,68
14 12,4 – 12,75
15 17,73 – 18,54
16 21,18 – 21,51
17 28,42 – 28,88
18 30,32 – 30,68
19 31,50 – 31,74

In Fig. 3 it can be noticed that the protein carbonyls, detected as DNPs, appear to the end of the elution time. The peaks are well individualized, although they have close molecular weights. This proves the efficient separation allowed by ionization and filtration through dextran 70. The peptide nature was confirmed by the signal at 214 nm.


Fig. 3. 3D plot of electropherogram of a T2DM patient serum sample 


By using two complementary techniques, this study shows clearly that in vitro data were partially reproducible in vivo. The most potent in vitro oxidizing system was FeCl3/sodium ascorbate, leading to the highest level of protein carbonyls. CE analysis of FeCl3/sodium ascorbate degraded HSA samples revealed the presence of nineteen DNPs, with good velocity. Five of these are also present in T2DM serum samples, this suggesting that in vitro generating systems can be useful to establish a carbonylation pattern in metabolic pathologies. Further studies should investigate whether carbonylation and protein lysis take place simultaneously or not and if dermatologic manifestations are directly correlated with this process.


This paper was co-financed from the Human Capital Operational Program 2014-2020, project number POCU / 380/6/13/125245 no. 36482 / 23.05.2019 “Excellence in interdisciplinary PhD and post-PhD research, career alternatives through entrepreneurial initiative (EXCIA)”, coordinator The Bucharest University of Economic Studies”


  1. Hecker, M., & Wagner, A. (2017). Role of protein carbonylation in diabetes. Journal of Inherited Metabolic Disease, 41(1), 29-38.
  2. Levine, R., Garland, D., Oliver, C., Amici, A., Climent, I., & Lenz, A. et al. (1990). Determination of carbonyl content in oxidatively modified proteins. Oxygen Radicals In Biological Systems Part B: Oxygen Radicals And Antioxidants, 464-478. 
  3. Himmelfarb, J., McMonagle, E., & McMenamin, E. (2000). Plasma protein thiol oxidation and carbonyl formation in chronic renal failure. Kidney International, 58(6), 2571-2578. 
  4. Winterbourn, C., & Buss, I. (1999). Protein carbonyl measurement by enzyme-linked immunosorbent assay. Methods In Enzymology, 106-111. 
  5. Aksenov, M., Aksenova, M., Butterfield, D., Geddes, J., & Markesbery, W. (2001). Protein oxidation in the brain in Alzheimer’s disease. Neuroscience, 103(2), 373-383. 
  6. Fagan, J., Sleczka, B., & Sohar, I. (1999). Quantitation of oxidative damage to tissue proteins. The International Journal of Biochemistry & Cell Biology, 31(7), 751-757. 
  7. Purdel, N.C., Dima, I., Margina, D., Gradinaru, D. and Ilie, M., (2014). Current methods used in the protein carbonyl assay. Annual Research & Review in Biology, 4(12), pp.2015-2026. 
  8. Purdel, C., Dima, I., Margină, D., Grădinaru, D., Ilie, M. (2015). Investigation of the carbonylation process of protein induced “in vitro” by different hydroxyl radical generating systems. Revista de Chimie, 66(3), pp.320-333.
  9. Dalle-Donne, I., Rossi, R., Giustarini, D., Milzani, A. and Colombo, R. (2003). Protein carbonyl groups as biomarkers of oxidative stress. Clinica Chimica Acta, 329(1-2), pp.23-38