Identification of Proteins in Slow Continuous Ultrafiltrate by Reversed-Phase Chromatography and Proteomics David M. Lefler, Roger G. Pafford, Nancy A. Black, John R. Raymond, and John M. Arthur* Medical University of South Carolina and Ralph H. Johnson VA Medical Center, Charleston, South Carolina 29425 Received August 4, 2004
Continuous modes of renal replacement therapy (CRRT) are increasingly being utilized in the intensive care unit. The removal of cytokines and other inflammatory proteins during ultrafiltration may be responsible for some of the beneficial effects of CRRT. We used proteomic tools to identify proteins found in the ultrafiltrate from a patient with acute renal failure. Identification of these proteins could help elucidate the mechanism(s) of improved outcome with continuous renal replacement therapy. Protein was loaded on a reversed-phase C4 column and eluted with stepwise isocratic flows starting with 0%, 5%, 10%, 25%, and 50% of acetonitrile. Effluent was collected, pooled, desalted, and separated by two-dimensional gel electrophoresis (2DE). Reversed-phase separation improved the resolution and the number of spots seen on the gels. Protein spots were digested with trypsin and spotted onto MALDI plates. Proteins were identified by either peptide mass fingerprinting using a MALDI-TOF mass spectrometer or by peptide sequencing using a MALDI-TOF/TOF tandem mass spectrometer. From 196 spots cut, 47 were identified, representing multiple charge forms of 10 different proteins. Proteins identified were albumin, apolipoprotein A-IV, β-2-microglobulin, lithostathine, mannose-binding lectin associated serine protease 2 associated protein, plasma retinol-binding protein, transferrin, transthyretin, vitamin D-binding protein and Zn R-2 glycoprotein. Continuous renal replacement therapy is frequently used in acutely ill patients with renal failure. Removal of proteins occurs during this process. The physiological significance of this protein removal is unclear. Identification of these proteins will lead to better understanding of the role of protein removal in continuous renal replacement therapy. Keywords: proteomics • dialysate • 2D gel • continuous renal replacement therapy
Introduction Continuous renal replacement therapies are widely used for treatment of acute renal failure in the ICU setting. Protein loss with the ultrafiltrate occurs and can be of a large magnitude. Patients requiring slow continuous renal replacement therapies are frequently acutely ill with multi organ system failure. They may be more susceptible to loss of important proteins. Conversely, they may benefit from removal of inflammatory or other harmful proteins. A relatively small number of proteins have been identified in ultrafiltrate from continuous therapies. These proteins include myoglobin,1 endothelin 1,2 factor VII and factor XIII,5 procalcitonin, interleukin-6, TNF-R, and interleukin-1 β.4 Removal of inflammatory proteins may play an important role in the improvement in clinical condition produced by CVVH in acutely ill patients.15 A number of experimental models have confirmed the association between reduced serum levels or removal in ultrafiltrate of inflammatory mediators and outcome. In pigs with pancreatititis, clinical improvement after initiation of CVVH is associated with a reduction in serum TNF* To whom correspondence should be addressed. E-mail: arthurj@ musc.edu.
1254
Journal of Proteome Research 2004, 3, 1254-1260
Published on Web 11/09/2004
R.21 In a model of septic pigs, CVVH lowered TNF-R and improved immunoresponsiveness.20 In a group of septic patients, CVVH resulted in a decrease in the serum levels of the prognostic markers procalcitonin, interleukin-6, TNF-R, and interleukin-1 β.4 Clearance of proteins during continuous dialytic therapies differs between diffusive and convective modalities. Clearance of amino acids is greater during CVVH than CVVHD.13 Loss of proteins is greater during CVVH than CVVHD.14 Reduction of selected inflammatory mediators by ultrafiltration (CVVH) has been compared to diffusive clearance (CVVHD). CVVH resulted in a greater lowering of serum TNF R, but not IL-6, IL-10 or sl-selectin.9 Protein losses during continuous dialytic therapy can be significant. Protein concentrations in ultrafiltrate have been reported as high as 77 mg/dL.8 A more recent report measured a lower amount (6 mg/dL).14 Over a 24 h period at ultrafiltration rates of a liter per hour this would result in a daily protein loss of between 1.5 and 18.5 g per day. For some proinflammatory proteins, reduction in serum levels may be largely mediated by binding of the protein to the membrane.12 A more complete knowledge of the identity of proteins lost during continuous therapies would help understand both the potentially harmful and beneficial effects of protein loss. In the 10.1021/pr0498640 CCC: $27.50
2004 American Chemical Society
Proteins in SCUF
current study, we have used proteins obtained during slow continuous ultrafiltration (SCUF) to limit the proteins to those obtained during ultrafiltration. Until recently, identification of proteins in ultrafiltrate required an educated guess of the identity of the protein and methodical investigation of one protein at a time. More recently two-dimensional gel electrophoresis (2DE) coupled to protein identification by mass spectrometry and informatic methods has provided an opportunity to identify proteins without a priori knowledge of what they may be. In 2DE proteins are separated by isoelectric point in the horizontal dimension and by molecular size in the vertical dimension. Proteins are stained and spots can be visualized across the gel. Staining intensity is proportional to the abundance of the protein for any given protein. 2DE has improved the ability to visualize and quantitate individual proteins because proteins with molecular sizes that are not distinguishable on a one-dimensional gel can be separated based on their isoelectric points into discrete spots. Several serious problems remain, however, for complex protein mixtures. One is the wide range of protein abundance in a sample. The presence of high abundance proteins makes it difficult to observe low abundance proteins. A second problem is that even with good protein separation on a two-dimensional gel, multiple proteins may be present in a single spot. In this study, we have used 2DE in combination with prefractionation of the fluid to identify proteins present in fluid from SCUF. We have used a third dimension (chromatography) of separation to attempt to improve the results for analysis of proteins in ultrafiltrate. By adding separation with reversed-phase chromatography, we have increased the number of spots seen, the ability to resolve spots within a given area of the gel and the ability to identify proteins spots. Forty-seven protein spots have been identified in SCUF fluid by peptide mass fingerprinting and eight were confirmed by sequencing. This is the first report of the application of proteomic technologies to identification of proteins in ultrafiltrate.
research articles
Figure 1. Separation of SCUF proteins by reverse phase HPLC. SCUF fluid (150 mL) was loaded (not shown) onto C4 reversedphase column and eluted with stepwise increases in acetonitrile of 5%, 10%, 25%, 50%, and 100%. Protein elution as measured by absorbance at 280 nm occurred with each increase. Proteins from the 5% and 10% acetonitrile elution were pooled together (fraction 1) and fractions collected during elution with 25% (fraction 2) and 50% (fraction 3) were pooled separately. The final (100%) elution fraction was not analyzed since very little protein was eluted. Pooled fractions were concentrated by lyophilization and analyzed by 2D gel electrophoresis.
Methods All studies were approved by the Institutional Review Board at the Medical University of South Carolina. Ultrafiltrate was collected from a patient with postoperative acute renal failure secondary to acute tubular necrosis. Ultrafiltration was done on a Prisma (Hospal) machine. Ultrafiltration rate was 100 mL/ hr, blood flow rate was 100 mL/min and the dialyzer was a Prisma M60, AN69 hollow fiber hemofilter/dialyzer. There was no replacement fluid or dialysate flow. Ultrafiltrate was collected for the study over the first 4 h of use. Chromatographic Separation. Protein separation was performed using a Biologic Duoflow HPLC (Bio-Rad, Hercules, Ca). One-hundred fifty mL of SCUF fluid was loaded onto a HiPore reversed-phase (C4) column (Bio-Rad, Hercules, Ca). The column was washed for 4 h with water at a flow a rate of 1 mL minute and a maximum pressure of 1000 mmHg. Proteins were eluted using a stepwise gradient of water (buffer A) and acetonitrile (buffer B) consisting of 0, 10, 25, and 50% buffer B at 1 mL/min. Protein elution was measured with an on-line detector monitoring absorption at 280 nm. Eluate was collected during the entire separation. Since the amount of protein collected at the step from 5 to 10% acetonitrile was small, samples collected with 5% and 10% acetonitrile were pooled (fraction 1). Protein collected at 25% (fraction 2) and 50% (fraction 3) acetonitrile were also pooled. Pooled protein fractions were snap frozen in liquid nitrogen and lyophilized
Figure 2. Broad pI range two-dimensional gel electrophoresis of proteins separated by reverse phase HPLC. Proteins were separated over a pI range of 3-10. The unfractionated sample (upper left panel) was poorly resolved. Pooled fractions showed better resolution and separation of proteins between reverse phase fractions.
in a SuperModulyo lyophilizer (ThermoSavant) at
