Rapid and Sensitive Method for Detection of Y402, H402, I62, and V62 Variants of Complement Factor H in Human Plasma Samples Using Mass Spectrometry

Abstract

purpose. Variations in the complement factor H (CFH) gene are tightly associated with age-related macular degeneration (AMD) across diverse populations. Of the many nonsynonymous coding variants in CFH, two are most strongly associated with increased risk of AMD: isoleucine 62 to valine (I62V) and tyrosine 402 to histidine (Y402H). Detection of these variations in a patient’s blood is important for a risk assessment of AMD and disease prognosis. However, traditional methods of genetic analysis cannot be used for measuring CFH allotypes in some sources of human plasma and other biological fluids not containing DNA. The purpose was to develop a protein-based method of detecting CFH allotypes.

methods. A combination of a single-step affinity enrichment of CFH, gel separation, and mass spectrometry identification of the CFH peptides spanning amino acids at positions 62 and 402 was used to identify individual CFH allotypes.

results. The CFH isoforms V62, I62, H402, and Y402 were reliably detected based on identification of tryptic peptides with masses of 1148.59 Da, 1162.60 Da, 2031.88 Da, and 2057.88 Da, respectively, using MALDI-TOF-TOF. The presence or absence pattern of these peptides in mass spectra of different CFH samples robustly correlated with all nine genotypes of CFH, as a result of variations at positions 62 and 402.

conclusions. A rapid and sensitive method has been developed for detection of V62, I62, H402, and Y402 variants of CFH in human plasma samples using mass spectrometry. This method can be used in clinical laboratories equipped with a basic inexpensive mass spectrometer capable of performing peptide fingerprinting.

Complement factor H (CFH) is an abundant plasma glycoprotein that is the main soluble inhibitor of the alternative complement pathway. Variations in the CFH gene have been implicated as one of the strongest genetic risk factors for early-stage age-related macular degeneration (AMD), accounting for up to 50% of the population-attributable risk percentage. ^{1 2 3 4} Of the many nonsynonymous coding variants in CFH, two are strongly associated with increased AMD risk: isoleucine 62 to valine (I62V) and tyrosine 402 to histidine (Y402H). ⁴ It has been also reported that the same variations in the CFH gene are associated with dense deposit disease (DDD; known formerly as membranoproliferative glomerulonephritis type II), suggesting that dysregulation of the complement alternative pathway is important in this disease pathophysiology, as well. ^{4 5} Identification of persons carrying CFH variants associated with these diseases would allow clinicians to provide an accurate risk assessment and develop strategies to reduce effects of other extrinsic factors. The ability to identify CFH variants in human blood or plasma also facilitates purification of CFH variants for biochemical studies aimed at understanding how these amino acid substitutions effect a change in biochemical properties of the CFH that contribute to risk for AMD and DDD. 

Current methods of detecting CFH allelic variants rely primarily on PCR-based DNA genotyping of patient’s blood, whereas testing of CFH protein variants in human plasma is still in its earliest stage of development. Recently, a method was described for measurement of CFH Y402H variants in plasma using variant-specific monoclonal antibodies. ⁶ This method is limited by availability of the monoclonal antibody and allows testing for only a single CFH variant, Y402H. Herein, we describe a mass spectrometry (MS)-based method for rapid and sensitive detection of Y402, H402, I62, and V62 variants of CFH in human plasma samples. For detection by this method, plasma CFH is first enriched on a heparin-agarose column, separated on a polyacrylamide gel, in-gel cleaved with trypsin, and finally analyzed by MS. The reproducibility and specificity of this method was validated with DNA-genotyped plasma samples and demonstrated 100% accuracy in identifying all nine genotypes of CFH resulting from combinations of the polymorphisms at amino acid positions 62 and 402. 

Methods

Plasma Samples

The plasma samples used in this study were obtained from the cohort of genotyped blood samples at Duke University ² and the University of Iowa ⁴ to establish the initial association of factor H and the risk of developing AMD. 

Plasma Factor H Enrichment and PAGE Separation

Immobilized heparin-agarose (50 μL; Pierce, Rockford, IL) was added to a column (Handi-spin; Pierce) and the beads washed twice with 300 μL of 10 mM sodium phosphate buffer pH 7.4. Plasma (30 μL) was diluted with 90 μL of the same buffer before it was added to the beads. The columns were capped and rotated at room temperature for 15 minutes and centrifuged, and the beads washed three times. CFH was eluted with 40 μL buffer containing 500 mM NaCl. In some cases, it was further enriched by gel-filtration on an FPLC column (16/10 Superose; GE Healthcare, Piscataway, NJ). Typically, 1 μL of crude plasma, 20 μL of eluate from a heparin-Sepharose column, and 0.5 to 1 μg of CFH collected in gel filtration eluate were applied to a 10% Bis-Tris Precast gel (Criterion XT; Bio-Rad, Hercules, CA) for SDS-PAGE. Known amounts of commercially available CFH protein standard (Complement Technology, Tyler, TX) were run on the same gel. After electrophoresis, the gel was washed in water and stained with Coomassie blue reagent (Gel Code Blue Stain; Pierce) for 1 hour, followed by a 30-minute rinse in water. Bands, running at the same position as the CFH standard, were cut from the gel, reduced with 10 mM DTT, modified at cysteine residues with iodoacetoamide and trypsinized according to the manufacturer’s protocol (In-Gel Tryptic Digestion Kit; Pierce). 

MS Analysis

The extracted peptides were dissolved in 5 μL of 50% acetonitrile and 0.1% trifluoroacetic acid containing 5 mg/mL α-4-hydroxycinnamic acid (Sigma-Aldrich) as a matrix for MALDI MS. The peptide–matrix mixture (0.5 μL) was loaded onto a 192-spot MALDI plate and analyzed on a MALDI-TOF-TOF system (4700 Proteomic Analyzer; Applied Biosystems, Foster City, CA). A combination of peptide mass fingerprinting and precursor fragmentation analysis was performed by using the Mascot algorithm to search the NCBInr human database (http://www.ncbi.nlm.nih.gov/sites/entrez?db=protein; provided in the public domain by National Center for Biotechnology Information, Bethesda, MD). Search parameters were the following: precursor tolerance, 50 ppm; MS/MS fragment tolerance, 0.5 Da; fixed modification, carbimidomethyl; variable modification, methionine oxidation; and number of missing cleavages, 1. 

Results

MS provides a rapid and reliable method of identifying proteins and their isoforms. It does not require extensive purification of proteins of interest from biological samples, but rather simple enrichment procedures, thanks to the ability of this method to identify hundreds of proteins in crude mixtures. To establish the applicability of MS to identification of the CFH isoforms in human blood, we obtained plasma samples containing four different combinations of the amino acids at positions 62 (I62V) and 402 (Y402H) in CFH. The four combinations we used (IIYY, VVHH, VVYY, and VIYH) were each enriched, either on heparin-agarose or on heparin-agarose followed by gel filtration chromatography. Typically, these procedures yielded between 0.1 and 1 μg of CFH from 30 μL of plasma, which ran as a distinct band of ∼190 kDa (Fig. 1) . 

The identity of the 190-kDa band as CFH was established using the MALDI-TOF-TOF analysis, which gave an identification confidence interval of 100% for control commercial CFH standards and the CFH samples enriched on heparin-agarose columns or both heparin-agarose and gel filtration columns. However, when analysis of plasma samples was performed without any CFH enrichment, confident identification of CFH protein could not be consistently achieved. Typical sequence coverage for CFH identification ranged from 30% to 50% for individual samples, with the number of mass values matched to predicted peptides ranging from 33 to 49 of a total of 79 theoretical peptides. The exclusion of predicted CFH tryptic peptides with masses beyond the MS identification range of 800 to 4000 Da decreased the number of predicted tryptic peptides to 70. In addition, the CFH glycoprotein contains eight N-glycosylation sites, ⁷ four of which are located within the tryptic peptides in the mass range of 800 to 4000 Da. Consistently, none of these peptides was identified by MS. Therefore, the total number of theoretically identifiable peptides was 66 and the identification rate was up to 74%. The absence of mass peaks corresponding to the remaining 17 predicted CFH tryptic peptides can be explained by not yet identified posttranslational modifications, weak ionization, or cleavage constraints for trypsin. Of interest, protein scores for CFH enriched by either one or two column steps were similar (data not shown), despite a noticeable enrichment of the protein after two column steps compared with the one column procedure (Fig. 1) . The intensities of mass peaks were also comparable for CFH samples after a one- or two-step purification (data not shown). Therefore, the single step purification strategy on heparin-agarose was sufficient for reliable CFH detection by MS analysis. 

Fortunately, both peptides containing the amino acid residues varying among the CFH isoforms displayed intense mass peaks, whose identities were further confirmed by the MS/MS analysis. Figure 2illustrates the identification of the V62 and I62 CFH variants based on the MS/MS spectra of the peptides SLGNV ⁶² IMVCR (1148.59 Da; Fig. 2A , peptide 1a) and SLGNI ⁶² IMVCR (1162.60 Da; Fig. 2B , peptide 1b). Similarly, the identities of the H402 and Y402 CFH variants were established based on the mass peaks of peptides CYFPYLENGYNQNH ⁴⁰² GR (2031.88 Da; Fig. 3A , peptide 2a) and CYFPYLENGYNQNY ⁴⁰² GR (2057.88 Da; Fig. 3B , peptide 2b). Because all four peptides contain cysteine residues, they were found to be predominantly modified with iodoacetamide (whose mass is included in these peptide masses). In addition, small fractions of peptides 1a and 1b (<25%) were also found to be derivatized with oxygen at the methionine residue (not shown). However, this oxidation did not significantly reduce the main mass peaks of peptides 1a and 1b, which where used for CFH isoform identification. 

To determine whether we could distinguish the Y402, H402, V62, and I62 variants of CFH in human plasma samples, we analyzed gel bands corresponding to CFH from genotyped human plasma samples (Fig. 1) . Figure 4shows mass peaks 1148.59 Da (referred to from now on as peak 1148) and 1162.60 Da (peak 1162) corresponding to peptides 1a and 2b, respectively, which were observed in the samples with the genotypes VV62 (Fig. 4A) , II62 (Fig. 4B)and VI62 (Fig. 4C) . The appearance of 1148 peak corresponded to the VV or VI genotype, whereas the appearance of peak 1162 correlated with the genotype corresponding to CFH 62VI or 62II. The peak intensity and signal-to-noise ratio for both peaks were sufficiently high to allow a reliable detection of these peaks in all the samples tested. Peak 1148 was absent in samples with a II62 genotype (Fig. 4B) , and peak 1162 was absent in samples with VV62 genotype (Fig. 4A) . Figure 5shows mass peaks of 2031.88 (peak 2031) and 2057.88 (peak 2057) Da, corresponding to peptides 2a and 2b, respectively. The peptide 2a, peak 2031 was detected in samples with genotypes corresponding to the CFH HH402 and HY402 variants (Figs. 5A 5C) , while the peptide 2b, peak 2057 was detected in samples that were YY402 and HY402 (Figs. 5B 5C) . The peptide 2a was absent in samples with the YY variant and the peptide 2b was absent in samples with the HH variant. 

To establish the detection reliability of the CFH isoforms (V62, I62, H402, and Y402) based on identification of the peptides 1a, 1b, 2a, and 2b, we analyzed human plasma samples of all nine CFH genotypes (VVHH, VIHH, IIHH, VVHY, VIHY, IIHY, VVYY, VIYY, and IIYY). In each case, the appearance of the 1148, 1162, 2031, and 2057 mass peaks reflected the presence of V, I, H, and Y variants of CFH, respectively (Table 1) . For example, the presence of peaks 1148 and 2031 in the absence of peaks 1162 and 2057 corresponded with the VVHH CFH allotypic variant, whereas the presence of peaks 1148, 1162, and 2031 and the absence of peak 2057 corresponded to the VIHH variants. We also conducted a masked analysis for detection of unknown genotypes of CFH in nine human blood samples and found a 100% correlation between the peptide mass peak patterns described herein and the corresponding patient’s CFH genotype. 

Discussion

MS is a rapidly developing analytical tool for protein identification in complex biological samples. The high resolution and sensitivity of this method provides extensive sequence coverage for proteins of interest, facilitates the analysis of posttranslational modifications, and allows identification of sequence variations that often have important physiological consequences. Recent studies have highlighted the utility of the MS-based approaches for detection of microheterogeneity and allotypic variations of plasma proteins including hemoglobin, C-reactive protein, transferrin and Cu/Zn-superoxide dismutase. ^{8 9} This established MS as a reliable alternative to DNA genotyping in identification of protein isoforms, which is particularly useful when biological fluids (e.g., plasma or serum) available for the analysis do not contain DNA. 

We have developed a rapid and sensitive method for detection of the I62V and H402Y variants of CFH in human plasma samples based on a combination of crude fractionation, gel separation, and MS. This method requires only few microliters of plasma and is suitable for analyzing multiple samples. The analysis of many plasma samples with known genotypes (both masked and unmasked) based on the presence or absence of four characteristic peptides allowed us to identify all nine CFH isoforms with 100% accuracy. We should stress that, although the MS/MS analysis was used to confirm the identity of each peptide, MS analysis alone was sufficient for reliable peptide identification. Therefore, a basic inexpensive mass spectrometer capable of performing only peptide fingerprinting (MS analysis of peptides without their fragmentation) could be used for this assay. This makes it feasible for use in many clinical and analytical laboratories. 

Currently, there are several potentially important applications for this assay. First, it can be used to test nongenotyped human plasma samples that are available in large amounts from commercial sources and cannot be genotyped due to complete removal of cells and, therefore, DNA. These plasma samples provide an excellent source from which each CFH variant could be purified for biochemical studies. Second, genotyping of CFH in blood of patients who undergo liver transplantation, unless performed on biopsy specimens, may not accurately reflect the circulating CFH protein variant, due to the relative contribution of secreted CFH synthesized in the nonhomologous transplanted organ. In this case an MS-based plasma assay could be used to determine the distribution of the CFH variants in a transplantation patient’s plasma. 

A future application of an MS-based plasma assay for CFH, which would require further modification into a quantitative method, would be to use it to determine relative amounts of CFH variants based on the relative intensity of the characteristic mass peaks. This kind of assay could be a valuable tool for population-based studies for measurement of the distribution of different CFH variants in pooled serum samples from certain cohorts of subjects and patients. 

 Supported by The Foundation Fighting Blindness (CBR); in part by core grants from the National Eye Institute (NEI; P30 EY005722) and Research to Prevent Blindness to the Duke Eye Center; National Institutes of Health Grant R24 EY017404 (GSH); and funds from NEI Grant EY012118 provided by Margaret Pericak-Vance (University of Miami) and Jonathan Haines (Vanderbilt University), which funded a portion of the samples used in this work.

 Submitted for publication August 28, 2008; revised November 4, 2008; accepted Januray 27, 2009.

 Disclosure: U. Kelly, None; C.BowesRickman, None; E.A. Postel, None; M.A. Hauser, None; G.S. Hageman, Optherion, Inc. (I); V.Y. Arshavsky, None; N.P. Skiba, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Figure 1.

 

Coomassie stained SDS-PAGE gel of various CFH preparations. Lane 1: commercial CFH standard (0.5 μg); lane 2: 1 μL of plasma obtained from an I62/Y402 individual; lane 3: 20 μL of CFH-containing eluate from heparin-agarose column (single step enrichment); and lane 4: CFH purified on a heparin-agarose column followed by a gel filtration column (0.8 μg).

Figure 2.

 

MS/MS spectra of precursors 1148.59 (A) and 1162.60 (B) confirming the sequence identity of CFH peptides SLGNVIMVCR (peptide 1a) and SLGNIIMVCR (peptide 1b), respectively.

Figure 3.

 

MS/MS spectra identifying the precursors 2031.88 (A) and 2057.88 (B) as CFH peptides CYFPYLENGYNQNHGR (peptide 2a) and CYFPYLENGYNQNYGR (peptide 2b), respectively.

Figure 4.

 

MS spectra of plasma samples collected from individuals with genotypes VV62 (A), II62 (B), and VI62 (C).

Figure 5.

 

MS spectra of plasma samples genotyped as HH402 (A), YY402 (B), and HY402 (C).

Table 1.

 

 

Correlation between CFH Genotypes and the Characteristic Mass Peaks

Table 1.

 

Correlation between CFH Genotypes and the Characteristic Mass Peaks

Genotype	Mass Peaks (Da)
		1148 Peptide 1a	1162 Peptide 1b	2031 Peptide 2a	2057 Peptide 2b
VVHH	Yes	No	Yes	No
VIHH	Yes	Yes	Yes	No
IIHH	No	Yes	Yes	No
VVHY	Yes	No	Yes	Yes
VIHY	Yes	Yes	Yes	Yes
IIHY	No	Yes	Yes	Yes
VVYY	Yes	No	No	Yes
VIYY	Yes	Yes	No	Yes
IIYY	No	Yes	No	Yes

 

The authors thank Lisa S. Hancox for coordinating the plasma acquisition from the University of Iowa.