Vermeulen et al. 10.1073/pnas.0508573102. |
Fig. 4. Representative surface-enhanced laser desorption/ionization (SELDI)-TOF MS spectra of exposed and unexposed subjects showing the three identified proteins [m/z 4,057 (Protein A), 7,763 (Protein B), and 9,287 (Protein C)]. The x axis depicts m/z. The y axis depicts the average intensity of ion peaks.
Fig. 5. Amino acid sequence of platelet factor (PF)4 and connective-tissue-activating peptide (CTAP)-III, with fragments identified by tandem MS underlined.
Table 4. Demographic characteristics of individuals included in the discovery and validation sets
Parameter | Discovery set | Validation set | ||
Control (n = 10) | Exposed (n = 10) | Control (n = 10) | Exposed (n = 10) | |
Sex | ||||
Male | 6 | 5 | 4 | 4 |
Female | 4 | 5 | 6 | 6 |
Current alcohol | ||||
Yes | 4 | 5 | 3 | 4 |
No | 6 | 5 | 7 | 6 |
Current smoking | ||||
Yes | 5 | 3 | 4 | 4 |
No | 5 | 7 | 6 | 6 |
Recent infection | ||||
Yes | 1 | 0 | 0 | 1 |
No | 9 | 10 | 10 | 9 |
Age, years [Mean (SD)] | 38.4 (5.7) | 32.7 (6.9) | 32.0 (8.2) | 38.7 (3.9) |
BMI, kg/m2 [Mean (SD)] | 23.2 (2.08) | 22.8 (3.06) | 23.4 (4.36) | 22.5 (2.73) |
Benzene air level, ppm* | N.A. | 37.9 (22.5) | N.A. | 31.3 (24.3) |
BMI, body mass index; N.A., not applicable.
*Based on mean individual exposure estimates in the last month before blood collection. Mean exposures in last 3 months were 31.0 (14.4) and 29.5 (18.8) for the exposed subjects in the discovery and validation sets, respectively
Table 5. Amino acid sequences and homology of peptides detected from the tryptic digest of the 7,763-Da (n = 7) and 9,287-Da (n = 2) proteins
Protein m/z | Peptide m/z | Mowse score | Sequence | Protein match | Mowse score* indicating: | |
Significant homology | Identity or extensive homology | |||||
7,763 | 944.5 | 17 | TTSQVRPR | PF4 precursor | >14 | >21 |
| 1,039.6 | 52 | HITSLEVIK | PF4 precursor | N.A. | >21 |
| 1,404.8 | 77 | KICLDLQAPLYK | PF4 precursor | >27 | >29 |
| 1,520.8 | 86 | AGPHCPTAQLIATLK | PF4 precursor | >20 | >22 |
| 1,551.7 | 35 | EAEEDGDLQCLCVK | PF4 precursor | >19 | >36 |
| 2,223.9 | 58 | EAEEDGDLQCLCVKTTSQVR | PF4 precursor | >24 | >51 |
| 2,477.2 | 22 | EAEEDGDLQCLCVKTTSQVRPR | PF4 precursor | >14 | >41 |
9,287 | 1,724.8 | 52 | GKEESLDSDLYAELR | PBP precursor | >27 | >32 |
| 1,127.6 | 14 | KICLDPAPR | PBP precursor | >13 | >21 |
N.A., not applicable (no score for significant homology was given).
*Mowse Score = –10 × Log (P), where P is the probability that the match is a random event (P < 0.05).
Supporting Materials and Methods
Identification of Protein Biomarkers. Purification. Samples with relatively high and low levels of the 7.7- and 9.3-kDa markers were used for purification. The 9.3-kDa protein was purified in a multistep process using anion exchange chromatography and cut-off membrane fractionation. One-hundred microliters of each serum sample was mixed with 150 l of U9 buffer prior to binding to 0.2 ml of Biosepra Q HyperD F resin (Pall). As in the discovery study, proteins were eluted in a stepwise pH gradient by using aliquots of each elution buffer with pH values of 9, 7, 5, 4, and 3. Finally, each column was washed with 33% isopropanol/17% acetonitrile/0.1% trifluoroacetic acid solution to elute proteins still bound to the resin. The fraction containing the 9.3-kDa protein was then applied to a YM-30 Microcon filtration unit (Millipore) to remove high-molecular-mass proteins. Acetonitrile (30%)/0.1% trifluoroacetic acid was added to reduce protein–protein interactions, the filtrate collected, and the process repeated. Filtrates containing the 9.3-kDa protein were pooled, dried in a SpeedVac (Labconco, Kansas City, MO), and, finally, purified by using SDS/PAGE.
During development of the ProteinChip immunoassay for the 9.3-kDa biomarker, the 7.7-kDa protein was observed to bind with high affinity to spots with immobilized control rabbit IgG and to remain bound under mild wash conditions. This unique and unexplained affinity for nonspecific rabbit IgG was utilized to purify the 7.7-kDa protein in a single step. Rabbit IgG (Sigma,) was coupled to interaction discovery mapping affinity beads (Ciphergen Biosystems, Fremont, CA). One-hundred microliters of each serum sample was diluted with 400 µl of PBS, mixed with 50 µl of IgG beads, and incubated for 1 h at room temperature in an end-over-end mixer. The supernatants were aspirated, and the beads were washed three times with 400 µl of PBS. Bound proteins were eluted with 250 µl of 50% acetonitrile/0.5% trifluoroacetic acid. The eluted fractions were dried in a SpeedVac (Labconco) and subjected to reduction with DTT and alkylation with iodoacetamide prior to gel electrophoresis.
Samples were loaded on NuPAGE 16% Tricine gels (Invitrogen). The gels were run according to the manufacturer’s instructions and stained by using the Colloidal blue Staining kit (Invitrogen). Protein bands, corresponding to the biomarkers detected by surface-enhanced laser desorption/ionization (SELDI)-TOF MS profiling, were excised for further processing and identification. Gel pieces containing no protein were processed alongside the protein bands as a negative control.
Tryptic digestion and protein identification.
The excised gel bands were destained by incubating successively with methanol/acetic acid, acetonitrile/ammonium bicarbonate (pH 8), and acetonitrile solutions and dried in a SpeedVac (Labconco). The dried gel pieces were rehydrated with 25 mM ammonium bicarbonate (pH 8) containing 10 ng/µl modified Trypsin (Roche Applied Science, Indianapolis, IN) and incubated for 16 h at 37°C. One microliter of the supernatant containing the peptides from the protease digestion was applied to a NP20 ProteinChip Array (Ciphergen Biosystems) and 1 µl of the energy-absorbing molecule saturated α-cyano-4-hydroxycinnamic acid was applied to each spot.Peptides from the digests were subjected to sequence analysis by tandem MS. The single MS spectra and collision-induced dissociation (CID) spectra from MS/MS analysis were acquired on a tandem TOF mass spectrometer equipped with a Ciphergen Biosystems PCI-1000 ProteinChip Interface. First, spectra were collected from 0.8 to 3 kDa in single-MS mode. After reviewing the spectra, specific ions were selected and introduced intos the collision cell for CID fragmentation. The CID spectral data was submitted to the database-mining tool mascot (Matrix Science, London, U.K.) for identification.