Systematic protein array pixelation ofthe human plasma proteome

The profiling of human plasma proteome began with major protein depletion from a total of 193 mL (14.5 mg) plasma (BDCA02-Heparin) using the MARS antibody column (Fig. 3A). Following depletion, 2.4 mg of unbound proteins were recovered, indicating that the six targeted proteins constituted approximately 83% of plasma proteins in this sample. Analysis ofthe bound fraction by 2-DE showed that the bound proteins were the six targeted proteins with no apparent evidence of other proteins [9].

The depleted plasma was then fractionated by MicroSol-IEF into seven fractions (Fig. 3B). Based on the 1-D gel analysis ofthe MicroSol-IEF fractions, the plasma proteins were well distributed throughout the seven fractions although the terminal fractions (F1 and F7) have the least amount of proteins as judged by the staining intensity (Fig. 3B). Many protein bands (including nondepleted abundant proteins) were present only in a specific fraction, indicating that MicroSol-IEF effectively separated proteins based on their pi. For example, a major protein with apparent MW of approximately 25 kDa was located almost exclusively in F4 (pH 5.4-5.9) with minor amounts found in more acidic fractions as observed by 1-D gel (Fig. 3B). Subsequent analysis by MS/MS identified this protein as apolipo-protein A-I precursor with calculated MW of 30.8 kDa and pi of 5.6. The good agreement with the observed values, confirmed the effectiveness ofthe pi and MW separation in this strategy. Hence, MicroSol-IEF not only further reduced the complexity ofthe plasma proteome, but also confined most remaining abundant proteins into specific fractions. This allows higher amounts of samples to be analyzed in downstream processes and permits us to dig deeper into the plasma proteome for lower abundance proteins.

Following MicroSol-IEF fractionation, the seven fractions were further separated by 1-D SDS-PAGE for a total distance of 4 cm (Fig. 3B). The gel lanes were then sliced and analyzed as uniform 2 mm pixels for a total of 140 pixels. Each pixel was digested in-gel with trypsin and analyzed by LC-ESI-MS/MS on an LCQ Deca XP1 mass spectrometer. In order to obtain a better correlation between the observed and the calculated MW ofthe proteins identified, the amount of sample loaded on the gel was limited to avoid overloading and to provide the optimal resolution of protein bands. In addition, the edge ofthe gel lanes where some degree of vertical smearing is frequently observed was excluded when the lane was cut. Depending upon protein concentration, between 1.3 and 2.8% (average 1.9%) of each Micro-Sol-IEF fraction was loaded onto 1-D gels used for pixelation. This average amount is equivalent to approximately 3.7 mL (278 mg) ofthe original plasma sample. Following tryptic digestion ofthe pixels, only 16.7% of each digestion mixture was analyzed by LC-ESI-MS/MS analysis. Hence, an amount equivalent to 0.6 mL (45 mg) ofthe original plasma sample was actually consumed in the final analysis.

From the LC-ESI-MS/MS analysis ofthe 140pixels, a 2-D array ofthe human plasma proteome was generated (Fig. 4A). Each pixel in the array has a distinct range of MWand pi as shown and contains a group ofidentified proteins (from 3 to

Tab. 1 Number of nonredundant proteins identified from human plasma/serum using different filters

Sample Filter Number of nonredundant proteinsc)

Sample Filter Number of nonredundant proteinsc)

Tab. 1 Number of nonredundant proteins identified from human plasma/serum using different filters

Total

>3

2

1

Plasma

Sfa)

744

140

45

559

Serum

Sfa)

4377

365

387

3625

Plasma

HUPOb)

575

138

36

401

Serum

HUPOb)

2890

297

223

2370

Plasmacommond)

HUPOb)

319

132

29

158

Serumcommone)

HUPOb)

319

178

33

108

Combined''

HUPOb)

3146

316

251

2579

Without Igg)

HUPOb)

3104

308

241

2555

a) Filter used: XCorr

> 1.9 (z = 1), 2.3 (z

= 2), 3.75 (z =

3)

and ACn > 0.1

or Sf > 0.7

b) Filter used: XCorr

> 1.9 (z = 1), 2.2 (z

= 2), 3.75 (z =

3)

and ACn > 0.1

and RSp < 4

c) >3, 2, and 1 indicate the number of

unique peptides per protein

d) Proteins in plasma that are also identified in the serum data set e) Proteins in serum that are also identified in the plasma data set f) Both plasma and serum data sets were combined for analysis.

g) Combined data set with immunoglobulin entries removed d) Proteins in plasma that are also identified in the serum data set e) Proteins in serum that are also identified in the plasma data set f) Both plasma and serum data sets were combined for analysis.

g) Combined data set with immunoglobulin entries removed

36 proteins) defined by one or more peptides that passed the XCorr/ACn/Sf criteria. Each pixel was assigned a name in the format Fx-y, where x is the MicroSol-IEF fraction (1-7), and y is the MW fraction from 1 (largest) to 20 (smallest). In general, the number of proteins identified in the pixels corresponds roughly to the staining density of the gel (Figs. 3B, 4A). A total of 744 nonredundant proteins defined by 3235 nonredundant peptides were identified from all the 140 pixels. Of these, 185 proteins (24.9%) were identified by at least two different peptides (high-confidence) whereas the majority (75.1%) was single-peptide proteins (Tab. 1).

A unique feature of this method is that the 2-D array can also be used to display the distributions of specific proteins that provide insight into their MW, pi, and the presence ofalternate forms ofeach protein such as alternate splices and proteolytic fragments (Fig. 4B-D). Of course due to the fact that many plasma proteins are heterogeneously modified such as by glycosylation and proteolytic processing, the observed MW and pi are not expected to closely match the values derived from amino acid sequences. Due to the high sensitivity of the mass spectrometer, the high- and moderate-abundance proteins (mg/mL - mg/mL) were commonly found in more than one pixel. Since the relative abundance of a specific protein can be roughly determined from the number of unique peptides identified [26], the primary position of an abundant protein is determined by the pixel containing the maximum number of peptides. The distribution of three proteins with varying abundance (apolipoprotein B-100, 720 mg/mL; ceruloplasmin, 210 mg/mL; metal-loproteinase inhibitor 1, 14 ng/mL [27]) is shown in Fig. 4. The distribution of

Fig. 4 Distributions of identified plasma proteins in the 2-D protein array. (A) Heat map showing the number of proteins identified that passed the XCorr/ACn/Sf criteria for each pixel in the analysis ofthe human plasma sample. Redundant proteins among pixels were not eliminated in this data set. Total number of proteins identified was 2255. Total number of nonredundant proteins was 744, which were defined by 3235 nonredundant peptides. (BD) Heat maps showing the distributions of peptides identified for apolipoprotein B-100, ceruloplasmin, and metalloproteinase inhibitor 1.

Fig. 4 Distributions of identified plasma proteins in the 2-D protein array. (A) Heat map showing the number of proteins identified that passed the XCorr/ACn/Sf criteria for each pixel in the analysis ofthe human plasma sample. Redundant proteins among pixels were not eliminated in this data set. Total number of proteins identified was 2255. Total number of nonredundant proteins was 744, which were defined by 3235 nonredundant peptides. (BD) Heat maps showing the distributions of peptides identified for apolipoprotein B-100, ceruloplasmin, and metalloproteinase inhibitor 1.

apolipoprotein B-100 in the array indicated that the protein was present in at least two major forms (Fig. 4B). Both major forms were larger than 200 kDa; the smaller form (in F4-2 and F5-2) had a pi in the range of pH 5.4-6.4, whereas the pi of the larger form (in F3-1) is between pH 4.9 and 5.4. The observed MWofthe protein and the multiple forms observed are consistent with the calculated MW of 515.6 kDa, and the reported forms of the protein such as B-74, B-48, and B-26 with apparent MW of 400, 259, and 140 kDa, respectively [28]. The observed pi of the protein is slightly lower than the theoretical pi of 6.6, which could be caused by heterogeneous modifications such as glycosylation [28]. The moderate-abundance protein, ceruloplasmin, was found mainly in F4-5 which is consistent with the calculated MWof 122.2 kDa and the theoretical pi of 5.4 (Fig. 4C). Unlike high- and moderate-abundance proteins, low-abundance proteins were usually identified by a single peptide that was found in only one or two pixels. For example, metallopro-teinase inhibitor 1 precursor identified by the single-peptide GFQALGDAADIR is found only in pixel F6-15 at ~30 kDa and pi between 6.4 and 8.1 (Fig. 4D). These values are close to the expected MW of 23.2 kDa and pi of 8.5 for the protein. The MS/MS spectrum of this peptide was verified by manual inspection (see also Fig. 7). Hence, the MW and pi values derived from the 2-D array can be used to reinforce the protein identifications made by SEQUEST, especially for proteins identified by a single peptide, which is the group of proteins that predominates in most shotgun proteomics approaches.

Was this article helpful?

0 0

Post a comment