Accurate Determination of Human Serum Transferrin Isoforms

However, CDT is not a single molecular entity but refers to a group of sialic acid-deficient transferrin .... Analytical Science and Technology 2014 2...
0 downloads 0 Views 248KB Size
Anal. Chem. 2006, 78, 8218-8226

Accurate Determination of Human Serum Transferrin Isoforms: Exploring Metal-Specific Isotope Dilution Analysis as a Quantitative Proteomic Tool M. Estela del Castillo Busto, Maria Montes-Bayo´n, and Alfredo Sanz-Medel*

Department of Physical and Analytical Chemistry, Faculty of Chemistry University of Oviedo, C/Julia´ n Claverı´a 8, 33006 Oviedo, Spain Carbohydrate-deficient transferrin (CDT) measurements are considered a reliable marker for chronic alcohol consumption, and its use is becoming extensive in forensic medicine. However, CDT is not a single molecular entity but refers to a group of sialic acid-deficient transferrin isoforms from mono- to trisialotransferrin. Thus, the development of methods to analyze accurately and precisely individual transferrin isoforms in biological fluids such as serum is of increasing importance. The present work illustrates the use of ICPMS isotope dilution analysis for the quantification of transferrin isoforms once saturated with iron and separated by anion exchange chromatography (Mono Q 5/50) using a mobile phase consisting of a gradient of ammonium acetate (0-250 mM) in 25 mM Tris-acetic acid (pH 6.5). Species-specific and species-unspecific spikes have been explored. In the first part of the study, the use of postcolumn addition of a solution of 200 ng mL-1 isotopically enriched iron (57Fe, 95%) in 25 mM sodium citrate/citric acid (pH 4) permitted the quantification of individual sialoforms of transferrin (from S2 to S5) in human serum samples of healthy individuals as well as alcoholic patients. Second, the species-specific spike method was performed by synthesizing an isotopically enriched standard of saturated transferrin (saturated with 57Fe). The characterization of the spike was performed by postcolumn reverse isotope dilution analysis (this is, by postcolumn addition of a solution of 200 ng mL-1 natural iron in sodium citrate/ citric acid of pH 4). Also, the stability of the transferrin spike was tested during one week with negligible species transformation. Finally, the enriched transferrin was used to quantify the individual isoforms in the same serum samples obtaining results comparative to those of postcolumn isotope dilution and to those previously published in the literature, demonstrating the suitability of both strategies for quantitative transferrin isoform determination in real samples. The detection and identification of protein variants as well as abnormally increased concentrations of modified proteins are of vital importance for clinical diagnosis. Some significant examples * To whom correspondence should be addressed. E-mail: [email protected].

8218 Analytical Chemistry, Vol. 78, No. 24, December 15, 2006

can be found in the glycation of hemoglobin and its use as an index in the evaluation and management of patients with diabetes or the relation of Cu/Zn superoxide dismutase variants with amyotrophic lateral sclerosis cases.1 Similarly, changes in the different glycoforms of transferrin have been used as clinical markers for different diseases including chronic alcohol consumption and congenital disorders of glycosylation (I and II).2 Normal transferrin has two fully occupied N-glycosylation sites and is a relatively homogeneous glycoprotein with more than 75% of its oligosaccharides being sialylated structures with two antennas.3 However, other minor forms of transferrin containing none, one, or two sialic acids (asialo, monosialo, or disialo), commonly called carbohydrate-deficient transferrin (CDT), are the only isoforms directly correlated with alcohol intake.4,5 A daily alcohol intake exceeding 60-80 g for more than 1 week results in a significant increase of CDT serum concentration which, with abstinence from alcohol, decreases with a half-life of 2 weeks.6 The main advantage of using CDT for alcohol consumption monitoring compared with earlier laboratory tests such as the measurement of γ-glutamyltransferase is its higher specificity for alcohol exposure. A current problem with CDT quantification is the lack of standardization. The most common analytical approach to analyze CDTs makes use of commercial kits based on the separation of these transferrin isoforms in anion exchange minicolums followed by radioimmunoassay, turbidimetric immunoassay (TIA), or nephelometric immunoassay for the final quantification.7 These methods give the total CDT concentration (the sum of asialo-, monosialo-, disialotransferrin and, some of them, part of trisialotranferrin) as a percentage of the total encountered transferrin in the samples. Indeed, the total CDT concentration obtained this way can be used to distinguish between teetotalers and alcoholism patients, but additional information about sialoform distribution (related, for instance, with other systemic diseases) is lost. When (1) Shimizu, A.; Nakanishi, T.; Kishikawa, M.; Miyazaki, A. J. Chromatogr., B 2002, 776, 15-13. (2) Helander, A.; Bergstrom, J.; Freeze, H. H. Clin. Chem. 2004, 50, 954-958. (3) Dell, A.; Morris, H. R. Science 2001, 291, 2351-2356. (4) Helander, A.; Eriksson, G.; Stibler, H.; Jeppsson, J.-O. Clin. Chem. 2001, 47, 1225-1233. (5) Anttila, P.; Ja¨rvi, K.; Latvala, J.; Blake, J. E.; Niemela¨, O. Alcohol Alcohol. 2003, 38, 415-420. (6) Martello, S.; Trettene, M.; Cittadini, F.; Bortolotti De Giorgio, F.; Chiarotti, M.; Tagliaron, F. Forensic Sci. Int. 2004, 141, 153-157. (7) Turpeinen, U.; Mathuen, T.; Alfthan, H.; Laitinen, K.; Salaspuro, M.; Stenman, U.-H. Clin. Chem. 2001, 47, 1782-1787. 10.1021/ac060956d CCC: $33.50

© 2006 American Chemical Society Published on Web 11/04/2006

the individual concentration of each of the sialoforms is required or the validation of the commercial kits has to be performed, the use of analytical separations by chromatography,8,9 electrophoresis (isoelectric focusing, IEF, or capillary zone electrophoresis, CZE)10 of the sialoforms is mandatory. IEF is highly selective and sensitive, and it is commonly used for separation of CDT isoforms and to obtain their relative concentrations.11 However, for absolute quantification of CDTs, IEF or chromatographic methods make use of an enzymatically prepared standard of transferrin with altered content of the naturally occurring sialoforms (by treatment with neuraminidase). In all cases, the isolation of CDTs (using minicolumns, analytical chromatographic columns, or electrophoresis) is based on their different isoelectric point (IP). However, the differences in the IP of the naturally occurring sialoforms are not only due to the different sialic acid content. Transferrin is the main iron transporting protein in serum, and its two binding sites for this metal are partially occupied (30%) under nonpathological conditions.12 Each Fe ion bound to transferrin leads to a pH decrease of 0.2 unit in the IP, and therefore, the charge (and so the separation of CDTs) can be affected by the different possible iron content of the different sialoforms. To avoid these variations, iron saturation of the protein has to be performed previous to any kind of charge-dependent separation of CDTs. The essential iron saturation of the different forms to carry out their separation facilitates the use of an iron-specific detector such as an inductively coupled plasma mass spectrometer (ICPMS) to detect Fe-bound transferrin sialoforms.13 An additional inherent advantage of using ICPMS detection is the possibility of conducting quantitative elemental and isotope analysis. The need for quantitative information in proteomics is enormous nowadays, while common mass spectrometry-based techniques (ESI-MS or MALDI-MS) are very limited to determine protein abundances.14 In this regard, isotope dilution analysis (IDA) has provided an extraordinary tool in elemental speciation to perform quantitative analysis, in particular for endogenous trace elements associated with proteins;15,16 for this purpose, an isotopically enriched spike of the element to be determined (in a chemical form that can be different from that of the analyte) is added postcolumn, allowing the quantification of endogenous trace elements in biomolecules of unknown identity.17 When the stoichiometry of the metalloprotein complex is known and the corresponding isotopically enriched species are available, it is possible to conduct the so-called species(8) Renner, F.; Kanitz, R.-D. Clin. Chem. 1997, 43, 485-490. (9) de la Calle Guntin ˜as, M. B.; Bordin, G.; Rodrı´guez, A. R. Anal. Bioanal. Chem. 2004, 378, 383-387. (10) Sanz-Nebot, V.; Gonza´lez, P.; Toro, I.; Ribes, A.; Barbosa, J. J. Chromatogr., B 2003, 798, 1-7. (11) Arndt, T.; Hackler, R.; Kleine, T. O.; Gressner, A. M. Clin. Chem. 1998, 44, 27-34. (12) Van Campenhout, A.; Van Campenhout, C.; Lagrou, A. R.; Manuel-y-Keenoy, B. Clin. Chem. 2004, 50, 1640-1649. (13) Del Castillo Busto, M. E.; Montes-Bayo´n, M.; Meija, J.; Blanco-Gonza´lez, E.; Sanz-Medel, A. Anal. Chem. 2005, 77, 5615-5621. (14) Bornstrup, M. Exp. Rev. Prot. 2004, 1, 503-512. (15) Schaumlo ¨ffel, D.; Prange, A.; Marx, G.; Heumann, K. G.; Bra¨tter, P. Anal. Bioanal. Chem. 2002, 372, 155. (16) Hinojosa Reyes, L.; Marchante-Gayon, J. M.; Garcia Alonso, J. I.; Sanz-Medel, A. J. Anal. At. Spectrom. 2003, 18, 1210. (17) Diaz Huerta, V.; Hinojosa Reyes, L.; Marchante-Gayo´n, J. M.; Fernandez Sanchez, M. L.; Sanz-Medel, A. J. Anal. At. Spectrom. 2003, 18, 1243.

specific spike (species-specific IDA).18 In this case, not only the concentration of the elements associated with the protein can be obtained but also the whole protein concentration (quantitative proteomics). Although both IDA methods are extremely precise and accurate, the species-specific IDA method is superior to the postcolumn addition of spike because any physical or chemical loss during the whole analytical procedure is corrected for the final measurement, once a total mix-up of the analyte and the spike is achieved. Some examples have been published by using speciesspecific IDA for quantification of small biomolecules,19 but very few cases have been documented for metalloprotein quantification, since the synthesis and application of such isotopically enriched biocompounds is still as a challenging task.20 The above-mentioned lack of standardization of CDT analysis has been a matter of debate in international forums of discussion concluding with the urgent need for the development of a highly sensitive HPLC method for CDT determinations that could serve as standard method for present CDT test-clinical analyzer validation.21 Thus, taking into account the excellent characteristics of IDA methods as standard methods of analysis,22 both possibilities, species-unspecific and species-specific IDA, are studied as possible standard methods for CDT analysis. In the first part of this work, the species-unspecific IDA is performed by postcolumn addition of isotopically enriched 57Fe, which permits us to quantify transferrin isoforms (from S2 to S6) with excellent precision. In the second part, the synthesis of the isotopically enriched (57Fesaturated) human serum transferrin is accomplished for the first time, aiming at the production of a metalloprotein standard containing an enriched metal isotope (to be used as spike material for species-specific isotope dilution analysis) to allow accurate quantification of transferrin isoforms by HPLC-ICP(ORS)-MS. Both, species-specific and species-unspecific are tested to analyze real human serum for transferrin isoforms accurate determination. EXPERIMENTAL SECTION Materials and Methods. Human serum transferrin (Tf) was purchased from Sigma-Aldrich (St. Louis, MO). Mobile phases for HPLC containing ( A) 25 mM trisaminomethane (Merck, Darmstadt, Germany)/acetic acid (Merck), pH 6.5, and (B) A + 250 mM ammonium acetate (Merck) were prepared by dilution of the solid salts with the 18 MΩ‚cm distilled deionized water (Millipore, Bedford, MA). Proteomics grade R(2f3, 6, 8, 9)neuraminidase from Arthrobacter ureafaciens was purchased from Sigma-Aldrich. For the species-unspecific IDA, an isotopically enriched iron solution (200 ng g-1) with relative abundances of 0.09% 54Fe, 7.44% 56Fe, 92.41% 57Fe, and 0.05% 58Fe was prepared by dilution of the stock solution (10 µg g-1) obtained from Spectrascan (Teknolab A.S. Dro¨bak, Norway) in a 25 mM sodium citrate/citric acid (pH 4) (Merck). For the species-specific IDA, enriched 57Fe was obtained from Cambridge Isotope Laboratories as Fe2O3 (see preparation of the isotopically enriched iron(18) Heumann, K. G.; Rottmann, L.; Vogl, J. J. Anal. At. Spectrom. 1994, 9, 13511355. (19) Hinojosa Reyes, L.; Moreno Sanz, F.; Herrero Espı´lez, P.; Marchante-Gayo´n, J. M.; Garcı´a Alonso, J. I.; Sanz-Medel, A. J. Anal. At. Spectrom. 2004, 19, 1230-1235. (20) Harrington, C. F.; Vidler, D. S.; Watts, M. J.; Hall, J. F. Anal. Chem. 2005, 77, 4034-4041. (21) Arndt, T. Clin. Chem. 2001, 47, 13-27. (22) Havlis, J.; Schevchenko, A. Anal. Chem. 2004, 76, 3029-3036.

Analytical Chemistry, Vol. 78, No. 24, December 15, 2006

8219

saturated transferrin section for more detailed information). Natural iron (as chloride, Merck) with relative abundances 5.8% 54Fe, 91.72% 56Fe, 2.2% 57Fe, and 0.28% 58Fe was also used as well as sodium bicarbonate (Merck). Human serum samples from healthy volunteers were kindly provided by Hospital Central of Asturias (Spain), Laboratory for Biochemical Analysis (Asturias, Spain) and those from alcoholic patients were obtained from Hospital de San Agustin of Avile´s (Asturias, Spain). The certified serum sample (CRM 470) was obtained form the Community Bureau of References (BCR) of the Commission of the European Communities. HPLC-ICPMS. HPLC separations were carried out using a dual-piston liquid chromatographic pump (Shimadzu LC-10AD, Shimadzu Corp., Kyoto, Japan) equipped with a sample injection valve Rheodyne, model 7125 (Cotati, CA), fitted with a 100-µL injection loop, and an anion-exchange column, Mono-Q HR 5/5 (50 × 5 mm i.d., Pharmacia, Amersham Bioscience, Spain). All the mobile phases were passed through a scavenger column (25 × 0.5 mm i.d.) placed between the pump and the injection valve in order to eliminate possible metal traces present as contamination in the mobile phases.23The scavenger column was packed with Kelex-100 (Schering) impregnated silica C18 material (20µm particle size) (Bondapack, Waters Corp., MA). Specific atomic detection of Fe in the column effluent was performed using an inductively coupled plasma mass spectrometer model 7500 from Agilent Technologies (Agilent, Tokyo, Japan) equipped with a collision cell system (ICP-(ORS)-MS). H2 has been used as collision gas at a flow of 2.5 mL min-1 in order to reduce the 40Ar16O+ interference on 56Fe+. ESI-Q-TOF. The instrument used for this study was a QStar XL model (Applied Biosystems) equipped with the ion spray source and using N2 as nebulization gas. The instrument was calibrated daily using a standard solution of poly(propylene glycol). Tf fractions from the HPLC column were reconstituted in 95% acetonitrile (Merck)/0.2% formic acid (Merck) and injected at a flow of 5 µL min-1. The scanned range goes from m/z 500 to 4000, and the applied voltage is 5500 V. The Bayesian deconvolution algorithm available in the Analyst software is applied to the intact Tf spectrum. Experimental working conditions are summarized in Table 1. Iron Saturation and Separation of Tf Isoforms. For iron saturation of human serum transferrin, 0.5 mL of the serum sample was diluted (1 + 1) in 25 mM Tris-acetic acid buffer and then incubated with 25 µL of a 10 mM Fe solution (as FeCl3) and 25 µL of a 500 mM sodium bicarbonate solution for 30 min at room temperature. The samples previously saturated with iron were subsequently filtered through 0.22-µm syringe filters and injected in the HPLC system (100 µL). The Tf isoforms were separated by means of a linear gradient of ammonium acetate (0250 mM in 45 min) buffered by 25 mM Tris-acetic acid (pH 6.5) solution. The eluent from the HPLC column was introduced online into the ICP-(ORS)-MS detector for specific iron detection. Neuraminidase Studies. Proteomics grade R(2f3,6,8,9)neuraminidase is a highly purified enzyme from A. ureafaciens that releases R(2f3)-, R(2f6)-, R(2f8)-, and R(2f9)-linked sialic acids. The relative rates of cleavage are reported to be R(2f6) 〉 (23) Soldado Cabezuelo, A. B.; Montes Bayo´n, M.; Blanco Gonza´lez, E.; Garcı´a Alonso, J. I.; Sanz-Medel, A. Analyst 1998, 123, 865-867.

8220 Analytical Chemistry, Vol. 78, No. 24, December 15, 2006

Table 1. Operating Conditions for FPLC, ICPMS, and ESI-MS.

anion exchange column mobile phases gradient injection volume flow rate detection

FPLC Parameters Mono Q HR 5/5 (50 × 5 mm i.d., Pharmacia) (A) 25 mM Tris-HAc buffer (pH 6.5) (B) A + 250 mM ammonium acetate 0%-75% B in 45 min 100 µL 1 mL min-1 UV at 280 nm

ICPMS Parameters forward power 1500 W external flow 15 L min-1 carrier gas flow 1.1 L min-1 54Fe, 56Fe, 57Fe isotope monitored integration time per isotope 0.1 s collision/reaction gas H2 flow 2.5 mL min-1 QP-bias - 11 V octapole-bias - 13 V extraction - 3.5 V ESI-MS Parameters scan type positive TOF MS ion spray voltage 5.5 kV nebulizing gas N2 injection rate 5 µL min-1 external calibration poly(propylene glycol) scan range m/z 500 - 4000 spectrum deconvolution Bayesian protein reconstruction

R(2f3) 〉 R(2f8) and R(2f9), which makes it ideal for complete nonspecific removal of sialic acid groups.24 Human serum Tf (3 mg) was prepared in the reaction buffer (50 mM sodium phosphate, pH 6.0, 3 mL). A 600-µL sample of this solution was mixed with 25 µL of the neuraminidase solution (0.05 mLU), and the reaction mixture was incubated at 37 °C. Aliquots were removed at different intervals between 0 and 15 days. The aliquots were desalted and preconcentrated by using a Centricon YM-50 Centrifugal Filter device and iron saturated, before they were analyzed by HPLC with UV detection at 280 nm. Postcolumn Isotope Dilution Analysis. For postcolumn isotope dilution analysis, a solution containing 200 ng g-1 57Fe was prepared in citrate buffer (pH 4) and continuously introduced through a T piece (at 0.1 mL min-1, using a peristaltic pump) at the end of the column, and the mixture was nebulized into the plasma. Instrumental parameters can be seen in Table 1. The intensity chromatograms (counts s-1) were converted, after adequate mathematical treatments,25 into mass flow chromatograms (ng min-1) using the on-line isotope dilution equation. The amount of Fe was obtained in each chromatographic peak by integration using Origin 7.5. Mass bias was measured daily before and after the analysis using the exponential law (k ) -0.05). For quantification of the isotopically enriched transferrin isoforms, postcolumn reverse isotope dilution analysis was conducted by adding 200 ng g-1 Fe solution (natural iron) in citrate buffer similarly to the previously described system. Preparation of the Isotopically Enriched Iron-Saturated Transferrin. For the species-specific spike, a standard of human (24) Crivellente, F.; Fracasso, G.; Valentini, R.; Manetto, G.; Riviera, A. P.; Tagliaro, F. J. Chromatogr., B 2000, 739, 81-93. (25) Sariego Mun ˜iz, C.; Marchante-Gayo´n, M.; Garcia Alonso, J. I.; Sanz-Medel, A. J. Anal. At. Spectrom. 2001, 16, 587.

Figure 1. Chromatographic separation of the transferrin isoforms by FPLC using UV detection at 280 nm after enzymatic treatment with neuraminidase at different reaction times (t ) 0 h after addition of neuraminidase and t ) 24 h). The ESI deconvoluted mass spectrum of fractions 2 and 2′ are also shown after preconcentration and cleaning. Key: 0, galactose; ], sialic acid; 9, N-acetylglucosamine; O, mannose.

serum transferrin of 0.56 mg mL-1 was prepared in Milli-Q water and then saturated with isotopically enriched 57Fe (abundances were detailed in Table 3) using conditions slightly different from those used before. In this case, a more concentrated solution of 57Fe was prepared by dissolving 0.0107 g of Fe O (Cambridge 2 3 Isotope Laboratories, Inc.) in 4.8 g of sub-boiling HNO3 and 0.2 g of sub-boiling HCl (final concentration ∼2000 µg g-1). An aliquot of this solution of 1 mL was dissolved in 10 mL of Milli-Q water and conserved as stock solution in which isotope abundances were characterized. An aliquot of this solution (0.1 mL) was diluted (1 + 1) in 250 mM ammonium acetate to obtain around pH 4.5 prior the addition of the apoprotein. This adjustment was necessary to reduce protein precipitation, which occurs when using lower pH values. Additionally, pH values above 5.5 start to produce iron precipitation as Fe(OH)3 that has to be controlled. This solution was mixed with the transferrin standard and the rest of the components necessary for saturation, as previously described. The evolution of the saturation process with time was monitored by HPLC-UV at 280 nm selective for the complex Fe-transferrin. RESULTS AND DISCUSSION Study of the Enzymatically Prepared Calibrator. A common approach to CDT analysis in human serum samples is the use HPLC or IEF with different detection systems. Some of these methods make use of an enzymatically prepared calibrator followed by iron saturation to calculate absolute concentration of

the different CDTs, mainly asialo- and disialotransferrin. For this purpose, a standard of pure transferrin is normally desialylated by enzymatic treatment with neuraminidase that releases sialic acid residues gradually until the appearance of asialo-Tf, while leaving the glycan chain attached to the protein. In order to study the possibility of obtaining a standard of S0, S1, and S2 (which are almost negligible in the commercially available transferrin standard), the previously mentioned enzymatic treatment was performed, as detailed in the Experimental Section. The HPLC-UV, measuring at 280 nm, results obtained for a commercial transferrin standard incubated with neuraminidase, at two different incubation times, and further saturated with iron, are shown in Figure 1. As can be seen in the UV chromatograms of Figure 1, the percentage of the different transferrin sialoforms changed with time during the enzymatic treatment (see reaction time 0 and 24 h, respectively) by decreasing the forms corresponding to three and four sialic acids and increasing the forms that contains one and two sialic acids. Longer incubation times (data not shown) provided mainly the forms with one and zero sialic acids. However, further structural characterization of the enzymatically produced sialoforms by electrospray mass spectrometry (ESIQ-TOF) has revealed that there are significant differences with respect to the low-sialylated forms originally present in the human serum samples.13 For this purpose, the species eluting between 10 and 17 min was collected in both chromatograms (tr ) 0 h and tr ) 24 h) and processed for ESI-Q-TOF analysis. In the rightAnalytical Chemistry, Vol. 78, No. 24, December 15, 2006

8221

Table 2. Results Obtained for the Analysis of the Serum Samples from Healthy Individuals Using Postcolumn Isotope Dilution Analysisa

S2 S3 S4 S5 S6 a

Figure 2. (A) Postcolumn isotope addition of 57Fe in citrate buffer and chromatographic separation of the transferrin isoforms corresponding to a saturated serum sample (56Fe) by FPLC-ICPMS. B) Isotope ratio 57/56 measured during the elution of Fe bound to Tf by species-unspecific IDA.

hand side of Figure 1, it is possible to observe the deconvoluted mass spectra of these two fractions: F-2 (tr ) 0 h) and F-2′ (tr ) 24 h), respectively. As can be observed, the main species observed in F-2 shows a molecular mass of 77 348 Da, which can be assigned to the structure containing a single sugar branch terminated with two sialic acid (theoretical molecular mass 77 365 Da, ∆m ) -0.02%). However, the main ion observed in F-2′ shows a molecular mass of 78 975 Da ascribable to a tetrasialo-Tf that has lost two sialic acids (584 amu) (theoretical molecular mass 78 989 Da, ∆m ) -0.02%). Although F-2 and F-2′ show similar retention times (same number of sialic acids and charges, as shown in the right-hand side of Figure 1) the deconvoluted MS spectra show that these are clearly two different species. Moreover, considering that iron saturation could be different depending on the oligosaccharide moiety,26 and since this previous step is mandatory for further separation of the sialoforms, the enzymatically prepared calibrator does not seem to be appropriate for an accurate quantification of the human serum transferrin sialoforms and particularly CDTs in real serum samples. Therefore, alternative methods are evaluated. (26) Nagaoka, M.; Maitani, T. J. Inorg. Biochem. 2005, 99, 1887-1894.

8222

Analytical Chemistry, Vol. 78, No. 24, December 15, 2006

reference values (%)

% 10 sera

% 3 injections