Proteome Resolution by Two-Dimensional Gel Electrophoresis Varies

Proteome Resolution by Two-Dimensional Gel Electrophoresis Varies with the Commercial Source of IPG Strips Regan C. Taylor†,‡ and Jens R. Coorssen*,†,‡,§,| Hotchkiss Brain Institute, and Departments of Physiology & Biophysics, Biochemistry & Molecular Biology, and Cell Biology & Anatomy, University of Calgary Faculty of Medicine, 3330 Hospital Drive NW, Calgary AB, T2N 4N1, Canada Received June 17, 2006

By facilitating reproducible first dimension separations, commercial immobilized pH gradient (IPG) strips enable high throughput and high-resolution proteomic analyses using two-dimensional gel electrophoresis (2DE). Amersham, Biorad, Invitrogen, and Sigma all market linear pH 3-10 IPG strips. We have applied optimized 2DE protocols with both membrane and soluble brain protein extracts to critically evaluate all four products. Resolved protein spots were quantitatively evaluated after carrying out these protocols using IPG strips from the four companies. Biorad and Amersham IPG strips resolved a high number of membrane and soluble proteins, respectively. Furthermore, Amersham IPG strips eluted the largest amount of protein into the second dimension gels and had the most protein remaining in the strip after 2DE. Biorad and Amersham IPG strips maintained a consistent linear pH 3-10 gradient, whereas those from Invitrogen appeared nonlinear or “compressed” within the central pH region. The gradient range within Sigma IPG strips appeared to be slightly less than pH 3-10, due to one extended pH unit within the gradient. Overall, all four commercially available IPG strips have the ability to resolve both membrane and soluble brain proteomes. The difference is that Amersham and Biorad do so more consistently and with better spot resolution. It appears that the physical/chemical nature of commercially available IPG strips can vary considerably, leading to marked differences in subsequent protein resolution in 2DE. These differences likely reflect variations in the uptake of proteins into the strips, and differences in the focusing and elution of proteins from the first to the second dimension. These differences would appear, in part, to underlie some inter-lab variations in the effective resolution of proteomes. Keywords: proteomics • brain • immobilized pH gradient • isoelectric focusing

Introduction

refined 2DE protocols do effectively enable the high-resolution analyses of membrane proteomes.3-6

Despite the emergence of complementary methods, the resolution of proteins using two-dimensional gel electrophoresis (2DE) is the most common practice in current proteomics and biomedical research. 2DE has thus far proven to be the most straightforward and widely applied large-scale separation technique to isolate specific proteins and compare proteome differences with high resolution. Therefore, 2DE is likely to remain the primary approach in proteomics for the immediate foreseeable future due to its maturity, adaptability, highresolution separations, and cost effectiveness.1,2 Furthermore, although 2DE has been extensively used to analyze cytosolic proteins, it has been somewhat dogmatically considered unsuited to the resolution of membrane proteins. Nonetheless,

Isoelectric focusing (IEF), for the first dimension of standard 2DE, is routinely carried out in immobilized pH gradient strips (IPGs). Developed by Go¨rg and colleagues in the 1980s, IPGs have revolutionized modern 2DE, almost completely replacing tube gels for the first dimension of separation.7,8 Adoption of IPGs has thus contributed substantially to the widespread use of 2DE as a routine and reproducible proteomic analysis method. The defining technology of the IPG revolves around ampholytes, which are low molecular weight molecules that act as either an acid or a base. These ampholytes are covalently bound to the acrylamide gel matrix, generating the pH gradient. With a standardized pH gradient, IEF can be used to separate proteins to their corresponding isoelectric point within the IPGs gel matrixes. Ideally, this technology provides for substantially more effective and reproducible IEF than that generally afforded by tube gels.

* To whom correspondence should be addressed. Tel.: (403) 220-2422. Fax: (403) 283-7137. E-mail: [email protected]. † Hotchkiss Brain Institute. ‡ Department of Physiology & Biophysics. § Department of Biochemistry & Molecular Biology. | Department of Cell Biology & Anatomy. 10.1021/pr060298d CCC: $33.50

 2006 American Chemical Society

Moreover, the process of casting IPGs readily lends itself to automation, mass production, and commercialization. IPG strips are ideal for large-scale commercial preparation as they Journal of Proteome Research 2006, 5, 2919-2927

2919

Published on Web 10/07/2006

research articles

Taylor and Coorssen

Table 1. IPG Strips Used during the First Dimension of 2DE Experiments company

lot no.’s tested

catalog no.

price per 12 (CND $)

Amersham Biorad Invitrogen Sigma

309652, 310015 708003B 1313702, 1314332 015K1072, 025K1415

17-6001-11 163-2099 ZM0018 I-2531

$114 $93 $97 $99

are easily integrated with current “best quality” industry practices. While some users do cast their own IPGs,9 for convenience and high reproducibility most groups purchase precast IPGs that are commercially available from a number of companies.4-6,10,11 However, it is quite likely that the different companies produce IPGs by somewhat different methods. Understandably, protocols are proprietary. Thus, with different types and/or combinations of ampholytes, preparation techniques, and practices, there arises the distinct possibility of differences in quality and resolution between commercial IPGs. We hypothesize that these differences could contribute to routinely observed differences in protein resolution and spot patterns between laboratories, which are already complicated by differences in sample handling, extraction, and electrophoresis protocols. Enhancing quality and optimizing the resolution of proteins during 2DE is paramount to effective proteomic analyses. Currently, 2DE is maximally capable of resolving and detecting hundreds to thousands of proteins in a single separation. Ultimately both the resolution and the detection of proteins must be enhanced to more fully address underlying biological complexity, and identify critical components. Moreover, poorly resolved or unresolved protein species can confound mass spectrometric analysis, often resulting in negative data or even false positive identifications. Only highly resolved proteins can be most effectively identified by mass spectrometry.12 Protein resolution in 2DE is dependent on many factors, one of which is the overall ability of IPGs to absorb proteins into the first dimension gel matrix. Potential problems include proteins that do not effectively enter the IPG gel matrix, and/ or those that may not be efficiently transferred into the second dimension separating gel. The efficiency of these processes is, in part, related to both the physical properties of the gel strip and plastic backing, and the chemical makeup of the gel itself. Additionally, it was previously demonstrated that sample handling and protein extraction procedures had significant effects on resolution in the first dimension of 2DE.5,6 It is therefore reasonable to expect some differences in final protein resolution depending on the composition, preparation, and quality of different commercially available IPG strips. Here, we assess IPGs from four major commercial suppliers in terms of proteins resolved using an optimized 2DE protocol. Amersham Biosciences, Biorad Laboratories, Invitrogen, and Sigma all produce linear, 7 cm long, pH 3-10 IPGs (Table 1); the linearity of the advertised immobilized gradients suggests that the strips from each company should focus nearly identical proteomes from the same protein isolates. We applied a refined 2DE protocol to resolve murine brain proteomes, in parallel, on these different commercially available IPGs, followed by automated digital image analysis to quantitatively compare the resulting 2DE protein maps. We have identified some significant differences in the relative performance of the tested IPGs and confirmed these using 2DE of protein standards. Both quantitative and qualitative differ2920

Journal of Proteome Research • Vol. 5, No. 11, 2006

ences, arising from the different properties of these commercially available IPGs, are discussed.

Materials and Methods Reagents. All chemicals used are of analytical grade or higher quality. Urea, thiourea, pH 3-10 ampholytes, tributylphosphine (TBP), HEPES, sodium orthovanadate, staurosporine, cantharidin components of the broad spectrum protease inhibitor cocktail,13 and two packages of 12 7-cm IPGs (pH 3-10, linear) were purchased from Sigma (St. Louis, MO). pI standards, 30% acyrlamide/bisacyrlamide solution, low melting agarose, broad range (3-10) ampholyte solutions, tris-glycine SDS running buffer, Sypro Ruby, and 7 cm IPGs (pH 3-10, linear) were obtained as one large lot from Biorad Laboratories (Hercules, CA); we have extensive experience with Biorad IPGs4-6 and are thus confident in the lot-to-lot reproducibility from this supplier. The other IPG strips were obtained as two packages of 12 each from both Invitrogen (Carlsbad, CA) and Amersham Biosciences (Uppsala, Sweden). All IPGs used in this study (Table 1) were linear pH 3-10 of 4% T, 3% C composition. Narrow range ampholytes (pH 2.5-4, 3.5-5, 5-7, 7-9, and 8-9.5) were from Fluka (Buchs, Switzerland), CHAPS from Anatrace (Maumee, OH), lysophosphatidylcholine (LPC) from Avanti Polar Lipids (Alabaster, Alabama), and thiourea from Fisher Scientific (Hampton, NH). TEMED, glycerol, PBS, dithiothreitol (DTT), and 40% acrylamide solution were from Bio Basic Inc. (Markham, Ontario). Isolation of Soluble and Membrane Protein Extracts by Frozen Disruption. The sample tissue that was disrupted, lysed, and fractionated into total membrane and soluble fractions came from the same ∼5 mm cubed cortex area in the mouse brain. This specific sample was used for all IPGs tested. Using automated frozen disruption, as previously described,6 a mouse brain sample was powdered in the frozen state using a Mikro Dismembrator (Braun AG). After powdering (40 Hz frequency for 60 s), the sample was placed into 900 µL of hypotonic lysis buffer (20 mM HEPES, 1 mM sodium orthovanadate, 4 µM cantharidin, 4 µM staurosporin, protease inhibitor cocktail, and 2 mM DTT).6 The uniform homogenate was placed on ice and vortexed intermittently for 3 min, and then 2 × PBS was added to restore isotonicity. The solution was then aliquoted and centrifuged at 180 000g for 1.5 h at 4 °C using an Optima-Max E ultracentrifuge with a TLS-55 rotor (Beckman-Coulter, Fullerton, CA). The membrane pellet was resuspended in IEF buffer 1 (8 M urea, 2 M thiourea, 3% CHAPS, 1% LPC, protease inhibitor cocktail, and 2 mM DTT).4 To fully suspend the pellet, the extraction was carried out with periodic vortexing and pipetting for 1 h at 4 °C. The soluble protein fraction was concentrated using microscale centrifugal concentration devices (Pall Life Sciences, East Hills, NY) at 14 000g for 1.5 h at 4 °C; a buffer exchange with 4 M urea followed to wash residual PBS through the column. The soluble protein retentate in 4 M urea was combined 1:1 with IEF buffer 1. Protein Quantification. The EZQ Protein Quantification Kit (Molecular Probes, Eugene, OR) was used to assay the total protein in both the membrane and the soluble protein fractions, as described previously.5 Fluorescence was measured using the ProXpress Proteomic Imaging System (Perkin-Elmer, Boston, MA) and analyzed using ImageQuant 5.2 software (Molecular Dynamics, Sunnyvale, CA). IEF. IEF was carried out essentially as previously described.4-6 Briefly, the samples were adjusted to 1 mg/mL total protein with IEF buffer, combined with a broad range of soluble

research articles

Proteome Resolution by 2D Gel Electrophoresis

ampholytes.4 Sequential disulfide reduction and alkylation was carried out using 45 mM DTT/2.3 mM TBP and 230 mM acrylamide, respectively.14,15 IPGs were passively hydrated with the reduced and alkylated samples for 12 h at 25 °C. IEF parameters were as previously described, and all gels were focused for 37 500 V h.4 IPGs from each commercial source were used in parallel and placed randomly in the IEF cell for focusing. Both soluble and membrane protein isolates were resolved in parallel. 2DE and Imaging. Equilibration and second dimension SDSPAGE were carried out as previously described.4 The second dimension was carried out in mini format (9 cm) and consisted of uniform 12.5% T, 2.6% C resolving gels overlaid with 5% T, 2.6% C stacking gels. After equilibration, electrophoresis was carried out at 125 V for 10 min and 90 V to completion, at 4 °C,16 using the MiniProtean II electrophoresis system (Biorad). Second dimension gels were fixed and stained with Sypro Ruby (BioRad) as previously described.4 Following a 15 min destain and three 5 min water washes, the gels were imaged using the ProXpress imager. Normalization of images was maintained by imaging each gel to a uniform total fluorescence that was just below full saturation. Testing for Appropriate pH Gradients. A small volume (2 uL) of a commercially available set of pI standards (BioRad) was combined with IEF buffer and resolved as described above on the different commercially available IPGs. Image Analysis. The outside edges of all of the images were identically cropped to establish a standardized area for analysis. This defined area, from molecular weights of 15-100 kDa and from pI 4-9, contained the specific protein spots A-E as markers. These parallel image areas were then analyzed using Progenesis Workstation 2005 (Nonlinear Dynamics, Newcastle, UK). The standardized areas of interest from all gels were matched, warped, and normalized using automated protocols, and stain speckles were then filtered. Total spot numbers and volumes within the normalized area were determined from the automated analyses. Protein spots A-E were previously assigned and covered an effective pI range across the gels; x-coordinate pixel values from Progenesis Workstation were converted to mm (10 pixels per mm) in the analyses of vertical and horizontal protein migration distances. The horizontal pixel distances between five sets of two proteins each were easily matched in all gels using the Progenesis software. The same spots A-E were used to identify protein volume differences as those used in migration distance analyses. Progenesis provided a quantitative analysis of the protein abundances. For pI marker data, differences in migration distances were determined between the spots corresponding to trypsin inhibitor, actin, BSA, and conalbumin charge isomers (including type 1 conalbumin, the most basic spot in the isomer string). Protein Assay of IPG Strips after 2DE Analysis. All IPG strips were restained for 2 h in Sypro Ruby, followed by 5 min for destaining and a 5 min water wash. Each strip was powdered in the Mikro Dismembrator, and the resulting sample was solubilized in 200 µL of water. The protein assay was carried out in 384-well plates, and fluorescence was measured using the Wallac Victor2 multi-label HTS counter (Perkin-Elmer, Boston, MA). 40 µL of solution was added to each of three wells in a 384-well plate, and Sypro Ruby fluorescence was assessed at 485 nm (ex) and 595 nm (em); the background fluorescence determined for a parallel set of IPGs that had not been loaded with protein was subtracted. Measurements were also taken

from a set of protein dilutions in 50% Sypro Ruby and 0.1% SDS, which produced a linear relationship (n ) 3, R2 > 0.99).

Results and Discussion Properties of IPG Strips and Analysis of Focusing Capacity. The dimensions of the IPGs used were 70 × 3 × 0.5 mm. The Amersham product sheet specifically noted the width of the gel after hydration as 3.0 mm, whereas the other companies recorded the width of the strip backing as 3.3 mm. By direct measurement of the strips, it is clear that the gel portions of the IPGs are all ∼3 mm in width. The Biorad catalog states that gel length tolerances of 70 ( 2 mm are maintained to allow for consistent pI separations. All companies advertise a 70 mm long linear gradient from pH 3-10, and any change in length within 2 mm will likely produce only minimal distortion in the gradient. Significant differences in the total number of protein spots resolved by 2DE were found after using these different, commercially produced IPGs. Aliquots from an extract of total mouse brain membrane proteins were analyzed in parallel on the commercially supplied IPGs from the four different companies (Figure 1). IEF on Amersham, Biorad, and Sigma IPGs, followed by SDS-PAGE, resolved a total of 378 ( 6, 434 ( 43, and 382 ( 20 membrane proteins (n ) 3), respectively (Figure 3). In contrast, identical extracts analyzed using Invitrogen IPGs resulted in significantly fewer proteins resolved by 2DE (322 ( 14, n ) 3; p < 0.05). In the low molecular weight range (below 25 kDa), Amersham, Biorad, Invitrogen, and Sigma IPGs resolved a total of 140 ( 19, 136 ( 13, 117 ( 13, and 124 ( 17 membrane proteins (n ) 3), none of which were significantly different from each other (p < 0.05). Aliquots from an extract of total mouse brain soluble proteins were also analyzed in parallel on the IPGs supplied by the four different companies (Figure 2). IEF on Amersham and Biorad IPGs, followed by SDS-PAGE, resolved a total of 534 ( 23 and 482 ( 49 soluble proteins (n ) 3), respectively (Figure 3). Although there were significantly fewer proteins resolved using Sigma IPGs (449 ( 50, n ) 3) relative to those from Amersham, the IPGs from Invitrogen yielded the lowest number of resolved proteins (407 ( 61, n ) 3), being significantly different from both Amersham and Biorad IPGs (p < 0.05). In the low molecular weight range (below 25 kDa), Amersham, Biorad, Invitrogen, and Sigma IPGs resolved a total of 147 ( 29, 135 ( 25, 132 ( 54, and 154 ( 38 soluble proteins (n ) 3), none of which were significantly different from each other (p < 0.05). The quality of the IPG determines the focusing and ultimate resolution of proteins. Therefore, the total number of proteins resolved by 2DE, following IEF on the IPGs from the four companies, indicate significant differences in quality and reproducibility. Overall, Biorad and Amersham strips resolved high numbers of proteins in a reproducible manner, although the consistency of resolution was somewhat lower with the Biorad IPGs (e.g., higher relative error). Conversely, 2DE using Invitrogen IPGs always had the lowest total number of resolved proteins, and the linearity of the gradient appeared different from all of the other IPGs tested. Therefore, the consistently lower resolution of proteins seen when using Invitrogen IPGs is potentially due to the nature of the pH gradient. An additional factor affecting protein resolution could be protein losses that have been noted during common 2DE procedures.17 However, Zhou and colleagues also point out that these general protein losses do not invalidate comparisons between conditions that are subjected to identical handling. Journal of Proteome Research • Vol. 5, No. 11, 2006 2921

research articles

Taylor and Coorssen

Figure 1. Total mouse brain membrane proteome by 2DE following IEF in IPGs from four different companies. The protein spots labeled A-E were used to analyze vertical and horizontal migration distances, and protein volumes.

Proteins that are lost between the first and second dimension are more likely to be those that remain in the IPG strip because of adsorption to the IPG gel matrix or proteins that could not focus into the gel of the strip. Thus, to identify factors influencing the differences in resolution noted above, protein volumes within the second dimension and first dimension gels after 2DE were compared. Significant differences in the total abundances of protein in the standardized area resolved by 2DE were found after using the commercially produced IPGs. From 2DE analyses of brain soluble proteins, Biorad, Invitrogen, and Sigma IPGs resolved total volumes in relative units (RU) of 646.3 ( 113.5, 676.4 ( 70.7, and 538.7 ( 52.6 RU × 106, (n ) 3), respectively. In contrast, identical extracts analyzed using Amersham IPGs resulted in significantly higher total protein volumes resolved by 2DE (1060.3 ( 18.0 RU × 106, n ) 3; p < 0.05). There were no significant differences found between the commercial IPGs with regard to total protein abundances in analyses of brain membrane proteins (data not shown). To further identify differences in specific protein abundances, the volumes in spots A-E were analyzed. Significant differences in the total abundances of protein in spots A-E resolved by 2DE were found after using the commercially 2922


produced IPGs. From 2DE analyses on brain membrane proteins, Amersham and Invitrogen IPGs yielded subsequent total spot volumes of 453.5 ( 18.1 and 427.8 ( 52.5 RU × 104, (n ) 3), respectively (Figure 4). There were significantly lower volumes resolved from Biorad (352.5 ( 29.6 RU × 104, n ) 3) and Sigma (326.4 ( 18.6 RU × 104, (n ) 3) IPGs relative to those from Amersham (p < 0.05). From 2DE analyses of brain soluble proteins, Biorad, Invitrogen, and Sigma IPGs resolved total volumes of 643.6 ( 144.3, 580.2 ( 141.0, and 563.0 ( 165.0 RU × 104, (n ) 3), respectively (Figure 4). In contrast, identical extracts analyzed using Amersham IPGs resulted in significantly higher total protein volumes resolved by 2DE (1009.0 ( 82.0 RU × 104, n ) 3; p < 0.05). On the basis of the data above (Figures 1-4), we sought to quantify the proteins that were not eluted out of the IPGs themselves during the second dimension of electrophoresis. The homogenized gels from IPGs used in the experiments described above were thus analyzed for total Sypro Ruby fluorescence (Figure 5). From the IPGs focused with membrane proteins, those from Amersham had significantly more fluorescence (8017 ( 1248 arbitrary fluorescence units; AFU) remaining after the second dimension of electrophoresis than did those from Biorad (6629


research articles

Figure 2. Total mouse brain soluble proteome by 2DE following IEF in IPGs from four different companies. The protein spots labeled A-E were used to analyze vertical and horizontal migration distances, and protein volumes.

Figure 3. Number of resolved mouse brain membrane and soluble proteins using commercially produced IPG strips. The “*” indicates a significant difference from all other companies; “†” indicates a significant difference from Amersham and Biorad; “‡” indicates a significant difference from Amersham only (p < 0.05).

( 838 AFU); both of these were significantly higher than the fluorescence from the Invitrogen gels (3884 ( 1268 AFU) and Sigma IPG gels (2368 ( 663 AFU, n ) 6; p < 0.05). From the

Figure 4. Total spot volume of resolved mouse brain membrane and soluble proteins (labeled A-E in Figures 1 and 2, respectively) using commercially produced IPG strips. The “*” indicates a significant difference from all other companies; “†” indicates a significant difference from Biorad and Sigma (p < 0.05).

gel strips focused with soluble proteins, those from Amersham with a fluorescence of 6391 ( 1669 AFU were significantly greater than those from Invitrogen with 3973 ( 611 AFU, which is greater than strips from either Biorad (3302 ( 326 AFU) or Journal of Proteome Research • Vol. 5, No. 11, 2006 2923

research articles

Taylor and Coorssen

focused the greatest number of proteins in the first place. Thus, it would appear that these strips not only elute out a higher volume of proteins during the second dimension of separation, but also retain a higher volume of proteins within the strips. By the same logic, Invitrogen strips should also focus and elute a comparable volume of proteins, yet the spot numbers following 2DE are not consistent with this interpretation. Rather, the lower number of resolved spots in gels derived from the Invitrogen IPGs is due to the compression in the central region of the pH gradient in these strips. In contrast, Sigma IPGs appear to have a linear gradient, suggesting that, at least in the case of soluble proteins, these strips absorb and/or focus proteins less consistently than do IPGs from the other companies. This is likely due to a combination of the physical and chemical nature of the strips.

Figure 5. Relative total amount of unresolved mouse brain membrane and soluble proteins remaining in different commercially produced IPG strips after second dimension SDS-PAGE. The “*” indicates a significant difference from all other companies; “†” indicates a significant difference from Invitrogen and Sigma; “‡” indicates a significant difference from Biorad and Sigma (p < 0.05).

Sigma (2064 ( 831 AFU, n ) 6; p < 0.05). The background Sypro Ruby fluorescence measured using unloaded strips from Amersham, Biorad, Invitrogen, and Sigma IPGs was 869 ( 208, 674 ( 219, 848 ( 256, and 1125 ( 269 AFU (n ) 3), none of which were significantly different from each other (p < 0.05). In terms of qualitative evaluation, it was generally agreed that IEF on Amersham IPGs resulted in greater vertical streaking within the resulting 2DE gels. This may represent a higher capacity for these particular IPGs to absorb and focus proteins, which would subsequently lead to a higher volume of proteins eluting into the second dimension gel. As a higher volume of protein moves through the second dimension gel, greater streaking would be expected, particularly as complete protein denaturation cannot be ensured as in standard SDS-PAGE. An interpretation of the protein volume data is that Amersham IPGs retain the most protein following the second dimension of separation because these strips also absorbed and

Gradient Linearity and Uniformity of Protein Spot Locations. At this point, two fundamental properties of the IPGs required consideration: the first was the physical nature of the gel (and its adherence to a support backing) that can affect how proteins enter and later elute from the gel, and the second was the chemical buffer system used to create an immobilized linear pH gradient that can effectively focus the proteins to their pI. The nature both of the gel and of the gradient determines the extent to which proteins are focused and resolved by 2DE. The vertical and horizontal components measured between spots on the gels were used to assess the pH gradients of the gels. Five randomly selected spots were used to assess vertical and horizontal protein migration distances between the four distinct spot pairs (Table 2). There were no significant differences in any vertical migration distances (data not shown) presumably because all IPGs were placed on identical second dimension gels composed of a 5% stacking gel overlaid on a 12.5% separating gel for SDS-PAGE.4 However, there were significant differences in horizontal migration distances between proteins. As an example, within the soluble component, the horizontal measurement of migration distances between the x-coordinate values for points B and C when using Amersham, Biorad, and Sigma IPGs was 6.2 ( 0.4, 5.6 ( 0.4, and 5.3 ( 0.9 mm (n ) 3), respectively. While these values are not different from each other, the distances are all significantly greater than for those

Table 2. Average Horizontal Distances (mm) between Selected Points on the Gels (n ) 3)a company

AE

AB

BC

CD

DE

Amersham (soluble) Biorad (soluble) Invitrogen (soluble) Sigma (soluble)

57.8 ( 1.3

27.1 ( 1.1*

6.2 ( 0.4

9.4 ( 0.7

15.1 ( 0.8

52.6 ( 2.1

20.4 ( 0.3

5.6 ( 0.4

10.0 ( 0.3

16.5 ( 2.1

41.1 ( 5.2*

16.0 ( 0.3*

3.8 ( 0.1*

8.2 ( 0.6*

13.1 ( 4.4

52.2 ( 3.3

22.3 ( 0.5

5.3 ( 0.9

9.8 ( 0.6

14.5 ( 1.4

Amersham (membrane) Biorad (membrane) Invitrogen (membrane) Sigma (membrane)

43.4 ( 2.4

10.7 ( 0.8

3.7 ( 0.8

19.1 ( 1.1

9.9 ( 0.5

41.4 ( 1.1

9.0 ( 1.2

3.3 ( 0.2

18.0 ( 0.3

11.1 ( 0.3

30.4 ( 0.4*

6.5 ( 0.2*

2.0 ( 0.2*

13.4 ( 0.2*

8.4 ( 0.5

41.6 ( 2.3

9.4 ( 1.0

3.6 ( 0.6

19.1 ( 0.8

9.5 ( 0.8

a See Figures 1 and 3 for the proteins labeled A-E in each of the membrane and soluble protein 2DE data sets. The “*” indicates a significant difference from the corresponding horizontal distances measured on 2DE maps that had been resolved in parallel using the IPG strips from all other companies, as indicated (p < 0.05).

2924


research articles


Figure 6. pI marker images for each of the four companies. Spots 1a, 1b, and 1c are the conalbumin charge isomers, 4 is bovine serum albumin (BSA), 5 is actin, and 6 is trypsin inhibitor. Invitrogen IPGs yield compressed horizontal distances between type 1 conalbumin isomers and trypsin inhibitor.

measured on the gels that had been resolved using Invitrogen IPGs (3.8 ( 0.1 mm, n ) 3; p < 0.05). Within the membrane component, the horizontal measurement of pixel distances between the x-coordinate values for points B and C when using Amersham, Biorad, and Sigma IPGs is 3.7 ( 0.8, 3.3 ( 0.2, and 3.6 ( 0.6 mm (n ) 3), respectively. Again, these values were not different, whereas this horizontal spacing was significantly smaller on the gels that had been resolved using Invitrogen IPGs (2.0 ( 0.2 mm, n ) 3; p < 0.01). As a whole, the data in Table 2 indicate that the pH gradient for Invitrogen IPGs is nonlinear, being somewhat more compressed in the central region. Accordingly, migration distances of spots D and E are not found to be significantly different on any of the commercially produced IPGs tested. This would suggest that the pH gradient in the Invitrogen IPGs is compressed in the central region, but more linear toward the ends of the strips. To further test for differences in the pH gradients of the different commercially prepared IPGs, pI markers were focused and subsequently resolved by second dimension SDS-PAGE (Figure 6). Spots labeled 1a, 1b, 1c, 4, 5, and 6 were selected to assess horizontal protein migration distances (Table 3). Again, there were significant differences in these inter-protein migration distances.

Table 3. Average Horizontal Distances (mm) between Selected Marker Proteins after 2DE (n ) 3)a company

6-1a

6-5

5-4


40.2 ( 1.6 39.6 ( 1.3 31.8 ( 2.7* 38.6 ( 0.4

9.8 ( 0.5 10.6 ( 0.3 8.2 ( 0.3* 11.3 ( 0.3*

9.2 ( 0.3 8.8 ( 0.3 7.2 ( 0.2* 9.1 ( 0.4

company

4-1a

1c-1b

1b-1a


19.4 ( 0.2 19.9 ( 0.4 16.1 ( 0.3* 20.8 ( 0.3

2.4 ( 0.2 2.4 ( 0.1 2.3 ( 0.1 2.3 ( 0.3

2.9 ( 0.1 2.9 ( 0.1 2.7 ( 0.1 2.7 ( 0.1

a See Figure 6 for the proteins labeled 1a-6 in the pI protein marker 2DE data set. The “*” indicates a significant difference from the corresponding horizontal distances measured on 2DE maps that had been resolved in parallel using the IPG strips from all other companies, as indicated (p < 0.05).

The difference in migration distance between points 1a (type 1 conalbumin) and 4 (BSA), as measured on gels derived from Amersham, Biorad, and Sigma IPGs, is 19.4 ( 0.2, 19.9 ( 0.4, and 20.8 ( 0.3 mm, respectively, which are all significantly greater than the protein separation found when using Invitrogen IPGs for the first dimension (16.1 ( 0.3 mm, n ) 3; p < Journal of Proteome Research • Vol. 5, No. 11, 2006 2925

research articles 0.01). Horizontal protein migration distances in 2DE when using Invitrogen IPGs are all significantly lower except for the distances between the conalbumin charge isomers. This is fully consistent with the above data, showing that no significant differences in protein migration distances exist toward the basic end of the proteome, even when using Invitrogen IPGs. Between points 5 (actin) and 6 (trypsin inhibitor), the distance when using Amersham and Biorad IPGs was 9.8 ( 0.5 and 10.6 ( 0.3 mm (n ) 3), respectively. While these values are not different from each other, the distances are both significantly lower than for those measured on the gels that had been resolved using Sigma IPGs (11.3 ( 0.3 mm, n ) 3; p < 0.05). This suggests that the gradient between ∼pH 4-5 is longer in the Sigma IPGs than are the other pH units throughout the rest of the commercially produced IPGs. Overall, Invitrogen IPGs appeared to have a compressed gradient within the middle of the pH range (∼pH 4-7). This contributes substantially to a lower resolution of distinct proteins within the second dimension gel. Without reproducible linear gradients from pH 3-10, use of these strips to resolve the full proteome of a given sample is not as effective as with the other commercially available IPGs. The compact gradient results in lower resolution of individual spots, which hinders effective proteome analyses.

Conclusion All four commercially available IPG strips have the ability to resolve both membrane and soluble brain proteomes. However, it is clear that there are significant differences between the IPGs made by these companies. It appears that the physical/ chemical nature of commercially available IPGs can vary considerably, leading to marked differences in subsequent protein resolution by 2DE. In the membrane component gels, significant differences were found, as Amersham, Biorad, and Sigma IPGs each resolved more protein spots than did those from Invitrogen (Figure 2). In those gels used to analyze soluble proteins, Amersham and Biorad strips resolved more proteins than Invitrogen strips, but only Amersham more so than Sigma strips (Figure 4). No significant trend was found in the basic, acidic, or low molecular weight ranges of the gels. Overall, Amersham IPGs were effective at focusing and eluting proteins when used in 2DE. As noted above, the streaking did appear higher in 2DE when Amersham IPGs were used. This could be due to the high volume of proteins that the strips elute into the second dimension gels. While the proteins migrate through the polyacrylamide matrix, the high volumes of proteins create some additional streaking. Biorad IPGs resolved large numbers of proteins (not significantly different from Amersham), especially in the analyses of brain membrane proteomes. Considering the resolution of pI markers, Biorad IPGs also appeared to produce an enhanced resolution at the basic pH extreme. Thus, with the lowest cost (Table 1), these IPGs do have appeal for use in high throughput proteomics. Invitrogen strips tend to focus proteins well but the gradient is nonlinear, being compressed in the central region from pH 4-7, which generally leads to poorer spot resolution. Overall, the horizontal pixel distances between spots on the second dimension gels showed a marked decrease when using the Invitrogen IPGs to focus proteins (Tables 2 and 3). Notably, Invitrogen does have the most effective packaging for easy removal of IPGs directly off the supporting case. 2926


Taylor and Coorssen

In the gels used to analyze brain proteomes and pI markers, proteins that were resolved from Sigma IPGs appeared to be shifted to the right side of the second dimension separating gel, relative to the protein spot patterns arising from all other IPGs tested. This can most easily be seen in the pI marker gels, as the proteins from Sigma IPGs were consistently resolved near the edge of the separating gel (Figure 6). The pH gradient appears to be longer than those produced by the other companies between pH 4-5, causing the rest of the gradient to be shifted to the right. This does not provide an explanation for why the total spot numbers were not always as high as Amersham and Biorad; the apparent right shifting would have no effect on spot numbers as the areas of interest selected for all gels were equal in molecular weight and isoelectric point dimensions. Another issue concerning the physical nature of the IPGs became evident in the Sigma IPGs after staining. IPG gels from Sigma strips frequently start to peel off at the basic end during the imaging process (unpublished observations). Although this does not affect the focusing ability or overall resolution of proteins, it does suggest a lack of consistency of the gels to remain adhered properly to the backing. IPGs from different companies thus lead to unique protein resolutions through 2DE. There are physical and chemical factors inherent to the IPGs from particular companies that promoted effective focusing and ultimate resolution of proteins in the second dimension. Also, IPGs are expected to have a reproducible ability to resolve a particular proteome. The IPGs should have stable plastic backings to which the gel adheres, and a linear immobilized pH gradient within the gel. Ampholytes used to create the gradient vary between companies; Amersham, for example, uses mixtures of Pharmalyte and Ampholine within the gels. Like commonly used ampholytes in any IPG strip, Ampholine consists of polyamino-polycarboxylic acids with a high buffering capacity. Despite possible variations in the carrier ampholytes within the gels, companies that are claiming a 3-10 linear pH gradient over a 7 cm strip should, in theory, have comparable resolving potential. Notably, our buffer system also contains a low concentration of ampholytes that is designed to optimize protein movement and solubility,4 reducing the chances of protein precipitation during IEF.18 Although Invitrogen IPGs do seem to focus a high volume of proteins, their resolving potential is hindered by the compressed gradient within the IPG strip. Invitrogen IPGs would thus be more effective if the gradient was linear; proteins seem to efficiently elute out of the gel, but the spots are not resolved well due to the compression of the gradient in the central region (Figures 1, 2, and 6). These differences would appear, in part, to underlie some inter-lab variations in the effective resolution of proteomes. While Amersham and Biorad strips qualitatively and quantitatively resolve proteins better, Sigma and Invitrogen strips nonetheless also effectively represent the bulk of the resolvable proteome. In conclusion, however, the highest resolution of the brain (and other) proteomes by 2DE is dependent upon the commercial source of the IPGs used.

Acknowledgment. We would like to thank Dr. V. Wee Yong for supplying mouse brain tissue. We would also like to thank Roby Butt and Matthew Churchward for helpful discussions concerning 2DE protocols and image analysis. J.R.C. acknowledges support from the Canadian Institutes of Health Research, the Alberta Heritage Foundation for Medical Re-

research articles


search, the Canada Foundation for Innovation, the Calgary Health Region, and the Alberta Network for Proteomics Innovation.

References (1) Rabilloud, T. Two-dimensional gel electrophoresis in proteomics: old, old fashioned, but it still climbs up the mountains. Proteomics 2002, 2, 3-10. (2) Gorg, A.; Weiss, W.; Dunn, M. J. Current two-dimensional electrophoresis technology for proteomics. Proteomics 2004, 4, 3665-85. (3) Wallin, E.; von Heijne, G. Genome-wide analysis of integral membrane proteins from eubacterial, archaean, and eukaryotic organisms. Protein Sci. 1998, 7, 1029-38. (4) Butt, R. H.; Coorssen, J. R. Postfractionation for enhanced proteomic analyses: routine electrophoretic methods increase the resolution of standard 2D-PAGE. J. Proteome Res. 2005, 4, 982-91. (5) Churchward, M. A.; Butt, R. H.; Lang, J. C.; Hsu, K. K.; Coorssen, J. R. Enhanced detergent extraction for analysis of membrane proteomes by two-dimensional gel electrophoresis. Proteome Sci. 2005, 3, 5. (6) Butt, R. H.; Coorssen, J. R. Pre-extraction sample handling by automated frozen disruption significantly improves subsequent proteomic analyses. J. Proteome Res. 2006, 5, 437-48. (7) Bjellqvist, B.; Ek, K.; Righetti, P. G.; Gianazza, E.; Gorg, A.; Westermeier, R.; Postel, W. Isoelectric focusing in immobilized pH gradients: principle, methodology and some applications. J. Biochem. Biophys. Methods 1982, 6, 317-39. (8) Gorg, A.; Postel, W.; Westermeier, R.; Gianazza, E.; Righetti, P. G. Gel gradient electrophoresis, isoelectric focusing and twodimensional techniques in horizontal, ultrathin polyacrylamide layers. J. Biochem. Biophys. Methods 1980, 3, 273-84. (9) Poland, J.; Cahill, M. A.; Sinha, P. Isoelectric focusing in long immobilized pH gradient gels to improve protein separation in proteomic analysis. Electrophoresis 2003, 24, 1271-5.

(10) Hoving, S.; Gerrits, B.; Voshol, H.; Muller, D.; Roberts, R. C.; van Oostrum, J. Preparative two-dimensional gel electrophoresis at alkaline pH using narrow range immobilized pH gradients. Proteomics 2002, 2, 127-34. (11) McDonough, J.; Marban, E. Optimization of IPG strip equilibration for the basic membrane protein mABC1. Proteomics 2005, 5, 2892-5. (12) Yergey, A. L.; Coorssen, J. R.; Backlund, P. S., Jr.; Blank, P. S.; Humphrey, G. A.; Zimmerberg, J.; Campbell, J. M.; Vestal, M. L. De novo sequencing of peptides using MALDI/TOF-TOF. J. Am. Soc. Mass Spectrom. 2002, 13, 784-91. (13) Coorssen, J. R.; Blank, P. S.; Tahara, M.; Zimmerberg, J. Biochemical and functional studies of cortical vesicle fusion: the SNARE complex and Ca2+ sensitivity. J. Cell Biol. 1998, 143, 1845-57. (14) Herbert, B. R.; Molloy, M. P.; Gooley, A. A.; Walsh, B. J.; Bryson, W. G.; Williams, K. L. Improved protein solubility in twodimensional electrophoresis using tributyl phosphine as reducing agent. Electrophoresis 1998, 19, 845-51. (15) Herbert, B.; Galvani, M.; Hamdan, M.; Olivieri, E.; MacCarthy, J.; Pedersen, S.; Righetti, P. G. Reduction and alkylation of proteins in preparation of two-dimensional map analysis: why, when, and how? Electrophoresis 2001, 22, 2046-57. (16) Coorssen, J. R.; Blank, P. S.; Albertorio, F.; Bezrukov, L.; Kolosova, I.; Backlund, P. S., Jr.; Zimmerberg, J. Quantitative femto- to attomole immunodetection of regulated secretory vesicle proteins critical to exocytosis. Anal. Biochem. 2002, 307, 54-62. (17) Zhou, S.; Bailey, M. J.; Dunn, M. J.; Preedy, V. R.; Emery, P. W. A quantitative investigation into the losses of proteins at different stages of a two-dimensional gel electrophoresis procedure. Proteomics 2005, 5, 2739-47. (18) Khoudoli, G. A.; Porter, I. M.; Blow, J. J.; Swedlow, J. R. Optimisation of the two-dimensional gel electrophoresis protocol using the Taguchi approach. Proteome Sci. 2004, 2, 6.

PR060298D

Journal of Proteome Research • Vol. 5, No. 11, 2006 2927

Proteome Resolution by Two-Dimensional Gel Electrophoresis Varies

Recommend Documents