Article pubs.acs.org/jpr
Metal Affinity Enrichment Increases the Range and Depth of Proteome Identification for Extracellular Microbial Proteins Korin E. Wheeler,*,†,‡ Brian K. Erickson,‡,§,∥ Ryan Mueller,⊥ Steven W. Singer,†,# Nathan C. VerBerkmoes,§ Mona Hwang,† Michael P. Thelen,† and Robert L. Hettich*,§ †
Physical and Life Sciences Directorate, Lawrence Livermore National Laboratory, Livermore, California, United States Chemical Sciences Division, Oak Ridge National Laboratory, Tennessee 37831, United States ∥ Graduate School of Genome Science & Technology, University of Tennessee, Knoxville, Tennessee, United States ⊥ Department of Earth and Planetary Science, University of California, Berkeley, California, United States §
S Supporting Information *
ABSTRACT: Many key proteins, such as those involved in cellular signaling or transcription, are difficult to measure in microbial proteomic experiments due to the interfering presence of more abundant, dominant proteins. In an effort to enhance the identification of previously undetected proteins, as well as provide a methodology for selective enrichment, we evaluated and optimized immobilized metal affinity chromatography (IMAC) coupled with mass spectrometric characterization of extracellular proteins from an extremophilic microbial community. Seven different metals were tested for IMAC enrichment. The combined results added ∼20% greater proteomic depth to the extracellular proteome. Although this IMAC enrichment could not be conducted at the physiological pH of the environmental system, this approach did yield a reproducible and specific enrichment of groups of proteins with functions potentially vital to the community, thereby providing a more extensive biochemical characterization. Notably, 40 unknown proteins previously annotated as “hypothetical” were enriched and identified for the first time. Examples of identified proteins includes a predicted TonB signal sensing protein homologous to other known TonB proteins and a protein with a COXG domain previously identified in many chemolithoautotrophic microbes as having a function in the oxidation of CO. KEYWORDS: proteomics, mass spectrometry, microbes, metal affinity, enrichment, extracellular proteome, IMAC
■
INTRODUCTION Characterization of the entire spectrum of expressed proteins for microorganisms growing in natural environments is essential for a comprehensive understanding of the ecophysiology of microbial populations. The identification of lowabundance proteins and other proteins that are difficult to detect (such as membrane proteins) remains a formidable technical challenge, in part due to dominant, high-abundance proteins in complex proteomes. In particular, abundant proteins can interfere with more minor proteins in both chromatographic separations and mass spectrometric measurement. For example, minor protein biomarkers and sensor proteins are generally difficult to identify. Affinity chromatography has been established as an effective method to increase the depth of proteomic identification.1 Recent approaches use combinatorial libraries2,3 to enrich for low abundance proteins in model systems. More targeted studies have included heparin chromatography for enrichment of brain signaling proteins4 and the use of lectin affinity chromatography to capture the glycoproteome.5 However, all of these approaches suffer problems related to robustness across experiments and selectivity/sensitivity of enrichment. © 2011 American Chemical Society
Metals provide a unique ability to bind a variety of ligands, including proteins. Recent work has demonstrated that immobilized metal affinity-chromatography (IMAC)6 coupled with mass spectrometry selectively enriches proteins,7−10 including those with post-translational modifications such as phosphorylation.11 Thus far, IMAC studies coupled with proteomic approaches have primarily utilized a single metal affinity column in proteomic enrichment approaches.7,12,13 Enrichments utilizing one metal are widely accepted, but present several limitations, including ambiguous specificity and the pervasive masking of low abundance proteins by ubiquitous proteins. Although IMAC enables the enrichment of proteins through metal-affinity, it is not specific for physiologically active metal binding (due to protein perturbations from enrichment conditions), nor does it work well for strongly bound metal− protein complexes.10 The specificity of biomolecules for metals is approximately dictated by the Lewis acidity of the metal.14 Additionally, the strength of metal binding generally follows the Irving-Williams series,15 which states that biomolecules will nonspecifically bind to metals higher in the series (i.e., greater Received: July 22, 2011 Published: December 22, 2011 861
dx.doi.org/10.1021/pr200693u | J. Proteome Res. 2012, 11, 861−870
Journal of Proteome Research
Article
Leptospirillum group II bacteria, Leptospirillum group III bacteria, and various Archaea comprise 43, 28 and 29% of the population, respectively.18 The extracellular proteomic fraction was obtained by thawing 50 mL of biofilm, suspending it in 110 mL 0.2 M H2SO4 (pH 1.1), followed by homogenization in a glass tube with vigorous strokes of a dounce homogenizer. The resulting homogeneous cell suspension was stirred for 2 h at 4 °C and then centrifuged at 24000× g for 12 min. The supernatant from this procedure was used as the extracellular protein wash for subsequent chromatography and proteomic experiments. The extracellular protein wash was then precipitated in 95% ammonium sulfate. After centrifugation at 24000× g for 12 min, the solid protein pellet was dialyzed at 4 °C in 20 mM 4Morpholineethanesulfonic acid (MES), pH 5 until all of the protein pellet was solubilized (18 h), then for another 24 h in 20 mM MES, pH 5 with 5 mM EDTA to remove residual metal contamination. The final two dialysis solutions (6 h each) brought the extracellular fraction back to the conditions of the column binding buffer (20 mM MES, pH 5). All protein samples used in this study originated from the same extracellular protein wash. Immobilized metal affinity chromatography (IMAC) was used to enrich proteins with metal binding ligands. The column was run at pH 5 to ensure deprotonation of Asp, Glu and phosphorylated residues and to increase binding to the metalated IMAC resin. Although running the column at a lower pH would better replicate physiological conditions of the AMD solution, it would have decreased proteome enhancement by decreasing nonspecific, yet favorable electrostatic interactions. A 5 mL HiTrap Chelating HP column (GE Healthcare) was equilibrated with 0.1 M metal in buffer A (20 mM, pH 5 MES), followed by loading of the metal salt in the same buffer. Metals tested included ZnSO4, CuSO4, Fe2(SO4)3, CoSO4, MgSO4, NiSO4, and MnSO4. Following metal loading, the column was rinsed with 50 mL binding buffer to ensure that only bound metal was left on the column. A total of 2.5 mg extracellular protein was then added to the metal loaded column and rinsed with 50 mL binding buffer. Eluent with protein from the buffer rinse were labeled as the “unbound” column fraction. The column was then rinsed with buffer B, 20 mM MES with 0.05 M EDTA and 0.5 M NaCl at pH 5. The eluent after washing with buffer B was labeled the “bound” column fraction. Each column run was repeated for a total of three technical replicates. Binding of proteins to the IMAC column was monitored by absorbance at 280 nm, 1D SDS PAGE, and Bradford protein assays. After initial analysis, the protein column fractions were precipitated with 100 μL 100% trichloroacetic acid solution for every 900 μL of protein solution (10% final TCA concentration). The sample was then incubated at 4 °C for 1 h. Samples were centrifuged (10 min at 25000× g) and supernatant was removed. The protein pellet was washed with 4 °C methanol and air-dried. Samples were stored at −80 °C until trypsin digestion.
negative hydration energy). Despite these limitations, the use of IMAC columns provides efficient and reproducible enrichment of selective proteins, highlighting the potential application of this approach for achieving deeper proteomic coverage. The coupling of affinity-based protein purification with highperformance liquid chromatography and high-throughput mass spectrometry (MS) brings together powerful tools to increase the dynamic range of proteomic identifications. This combination is aptly suited for extensive protein purification, identification, and characterization. The ability of MS to provide high confidence protein identification, relative quantification, and details of protein modification make this technique well suited for providing an unparalleled depth of proteomic coverage.16,17 In this study, a library of biologically active metals, Cu(II), Co(II), Mn(II), Mg(II), Ni(II), Zn(II), and Fe(III), IMAC enrichment was employed to enhance the depth of extracellular proteomic coverage of an extremophile microbial community. Biofilms collected from acid mine drainage (AMD) represent a characterized model system for investigations into community structure and metabolism.18−20 Extracellular (secreted or matrix-associated) proteins likely mediate interactions between the microorganisms and their immediate environment, characterized by extremely low pH and metal rich conditions. Ironoxidizing bacteria within the community maintain metal homeostasis within submolar concentrations of iron (0.1−0.3 M, depending on sampling location)21 and millimolar concentrations of copper, zinc and arsenic.21 Extracellular proteins can function in nutrient transport, interorganism communication, and defense mechanisms. This study provides another puzzlepiece in the emerging picture of the AMD microbial community system, complementing details already known about extracellular proteins,22−24 community membership,19 growth state dependence,25 and strain variation.26,27 In practice, the IMAC enrichment could not be conducted at the low pH of the AMD system. Therefore, the proteins cannot be guaranteed to retain their physiologically native structures. However, the goal of this work was to deepen the overall proteome detection; thus, any denaturing of the proteins, caused by the elevated pH, may serve to enhance the differential metal affinity enrichments through increased residue−solvent accessibility. Ultimately, the coupling of IMAC enrichment and MS analysis serves to increase the dynamic measurement by detecting lower abundance proteins, including proteins of unknown function, and may open a new route for targeted studies of the role of these proteins in metabolic functions.
■
EXPERIMENTAL PROCEDURES
Sample Preparation and Immobilized Metal Affinity Chromatography (IMAC)
Biofilm samples were collected from the surface of drainage solutions in the Richmond Mine at Iron Mountain in northern California by Prof. Jillian Banfield and co-workers (University of California-Berkeley) and provided for this study. As soon as collected, biofilms were immediately frozen in dry ice and stored in a −80 °C freezer until further processing. Samples used in this study were collected from the site location “AB Muck” in May 2007.19 The pH of the mine solution at the sampling site was 1.0, at a temperature of 41 °C. Visual examination and fluorescence in situ hybridization analysis indicated that the sample represents a mature growth stage biofilm (growth stage 2, GS2) with a diverse microbial community, for example,
Proteome Characterization via Multidimensional Liquid Chromatography−Tandem Mass Spectrometry
Each of the extracellular pellets (∼1−2 mg protein material) were resuspended, denatured, and reduced in 2 mL of 6 M guanidine-HCl, 10 mM DTT, at 60 °C for 1 h. Samples were diluted 6-fold in 50 mM Tris-HCl/10 mM CaCl2 (pH 7.8), sequencing-grade trypsin (Promega, Madison, WI) was added at ∼1:100 (w/w), and digestions were performed with gentle 862
dx.doi.org/10.1021/pr200693u | J. Proteome Res. 2012, 11, 861−870
Journal of Proteome Research
Article
shaking at 37 °C for 18 h. This was followed by a second addition of trypsin at 1:100 and an additional 5 h incubation. The samples were then treated with 20 mM DTT for 1 h at 37 °C as a final reduction step, and immediately desalted with Sep-Pak Plus C18 (Waters, Milford, MA). All samples were concentrated and solvent exchanged into 0.1% formic acid in water by centrifugal evaporation to ∼1 mg/mL starting material, filtered, aliquoted, and frozen at −80 °C until LC− MS/MS analysis. Mass spectrometric measurements of all preparations were completed on an LTQ-XL mass spectrometer (Thermo Fisher Scientific, San Jose, CA). The mass spectrometer was coupled online with an Ultimate HPLC (LC Packings, a division of Dionex, San Francisco, CA). The system utilized a 2D nano-LC tandem mass spectrometry (MS/MS) setup with an approximate flow rate of 200−300 nL/min at the nanospray tip. The split-phase columns were prepared in-house and consisted of strong cation-exchange material (Luna SCX 5 μ 100 Å Phenomenex, Torrance, CA) and C18 reverse phase (RP) material (Aqua C18 5 μ 200 Å Phenomenex). For all samples, ∼100− 300 μg of protein was loaded off-line onto the back of the multiphase column. The loaded RP-SCX column was then positioned on the instrument behind a ∼15 cm C18 RP column (Jupiter C18 5 μ 300 Å Phenomenex) also packed via pressure cell into a Pico Frit tip (100 μm with 15 μm tip New Objective, Woburn, MA). All samples were analyzed via an 8 h, 4-step, 2D analysis. During the entire chromatographic processes, the LTQ mass spectrometer was operated in a data-dependent MS/MS mode as previously described.22 All samples were analyzed in duplicate on the LTQ.
abundance in the unbound fraction was less than the abundance in the bound fraction, or the protein was undetected in the unbound fraction. The protein was classified as “unbound” if both of the following conditions were satisfied with both replicates: (1) the abundance in the unbound fraction was nonzero (indicating detection) and (2) the abundance in the bound fraction was less than the abundance in the unbound fraction, or the protein was undetected in the bound fraction. All other proteins were labeled “false”, indicating these thresholds were not met for IMAC bound or unbound designation. The results were clustered using Cluster version 3.0 (http:// bonsai.ims.u-tokyo.ac.jp/∼mdehoon/software/cluster/ software.htm#ctv) and visualized as heat maps with TreeView software. Proteins of unknown function exhibiting differential enrichment on a specific IMAC column were batch submitted to Pfam for domain and motif analysis.31 Pfam was executed with the following parameters: merged global and local strategy and an E-value cutoff of 1.0. The resulting identifications were then further filtered at an E-value Co(II) > Ni(II) > Zn(II) > Fe(III). Somewhat surprisingly, this trend is inconsistent with Hard Soft Acid Base theory, used to predict metal binding preferences and the Irving-Williams series for divalent metals (Mn(II) < Co(II) < Ni(II) < Cu(II) > Zn(II)) generally used to predict the strength of metal binding for divalent metals. Additionally, the trend does not overlay with the metal’s abundance in the AMD environment. Of the 485 nonredundant protein identifications, each was classified within three general categories: bound to a specific IMAC column (differential IMAC enrichment), bound to multiple IMAC columns (universal IMAC enrichment), or nonbound to any metals. The majority of identified proteins (295, 61%) can be classified as universally IMAC enriched proteins; 137 (28%) classified as differentially IMAC enriched; and the remaining 53 proteins (11%) classified as preferentially nonbound for any of the IMAC columns (Figure 3).
2D-LC−MS/MS Measurements
In total, 485 nonredundant proteins (these are proteins who may share some peptides, but do not have 100% sequence identity to any other proteins) were identified by mass spectrometric measurements across the 32 IMAC preparations. This value includes replicate determinations of the bound and unbound fractions from each of the seven metals, as well as the control column (peptide sequences, SEQUEST scores and protein sequence coverage can be found within Tables SM1− 33, Supporting Information). The total number of identified proteins correlated well with the expected number of proteins in the extracellular fraction identified in a previous, nonenriched analysis (531 extracellular proteins).22 A pooled average of 370 nonredundant proteins was identified in each column. The identified proteins were often found across multiple IMAC columns, with an overlap ranging from 13% to 93% among the columns. Reproducibility among replicates was very good, at an average of 91%, with several replicates exhibiting a deviation as low as 0.5 or 1%. One exception must be noted for the cobalt-IMAC, which had a higher standard deviation of 27% for the unbound and 48% for the bound column fractions. It appears that a remarkably high number of proteins were identified in one of the two mini-MudPIT measurements (472 identified proteins compared to an average of 354 in other samples), potentially inflating the relative error for the Co-IMAC. In every column except Fe(III), there were a total of 2−5 times as many proteins identified within the IMAC unbound fraction as compared to the bound fraction (Figure 2).
Amino Acid Specificity of IMAC Enrichment
Clearly, IMAC enrichment of proteins is influenced by exposed ligands in the proteins that are available for metal binding, but the identification of proteins enriched by IMAC in this study is not necessarily indicative of physiological metal binding. It has been previously shown that IMAC purification can be influenced by biases in the residue content of proteins (particularly histidine). In an effort to determine if any amino acid over-representation exists in the proteins identified through the IMAC-MS/MS analysis, the residue content of the 485 nonredundant proteins were evaluated (Table 2). Among proteins identified as Zn(II), Cu(II), and Mg(II) binding, a slight increase in the frequency of histidine occurrence was observed; this is in contrast to Co(II), Mn(II), and Ni(II) binding proteins, with a slight decrease in histidine content. Note that while Ni(II) and Co(II) IMAC columns are commonly used to purify His-tagged proteins, these multiple repetitive histidine residues in the tags do not represent a feature in naturally expressed proteins, and thus suggest that the results above cannot be directly compared to the His-tag case. The residue content of identified proteins was further analyzed by sequence alignment of each protein against the rest and consistently resulted in no discernible bias among the amino acid content. Pfam Analysis of Differentially Enriched Proteins of Unknown Function
Figure 2. Distribution of identified proteins among the bound and unbound fractions. The number of proteins identified by MS proteomics within the unbound (gray) and bound (blue) chromatographic fractions. Error bars indicate standard deviation from average.
Among the identified IMAC enriched proteins, it is apparent that a significant subset have unknown functions (35%), and of these, many were found to have differential IMAC enrichment 864
dx.doi.org/10.1021/pr200693u | J. Proteome Res. 2012, 11, 861−870
Journal of Proteome Research
Article
Table 1. Number of Proteins Labeled as Uniquely IMAC Bound or Unbound after Analysis of MS Data from Chromatographic Fractions Unbound Bound Total
Fe(III)
Co(II)
Mn(II)
Ni(II)
Cu(II)
Mg(II)
Zn(II)
control
65 72 137
67 210 277
49 274 323
59 177 236
57 229 286
45 158 203
88 165 253
0 55 55
Thus, the 116 newly identified proteins represent a ∼20% increase in proteomic measurement depth of the extracellular fraction. The ability to bind and enrich proteins allows for deeper coverage of the proteome and, additionally, verification of numerous proteins that were previously only predicted to exist. These newly identified proteins of unknown functions, formerly annotated as hypothetical, are likely of relatively low abundance due to their identification only with IMAC enrichment. The 116 proteins originate from each of the major microbes within the community and comprise a variety of functional annotations. The ability to identify and characterize a significant number of additional proteins in the extracellular space highlights the benefits of an expansive enrichment and MS detection strategy. The newly identified proteins have a wide variety of predicted functions. As reported previously, over 57% of the extracellular proteome consists of proteins of unknown function.22 It is not surprising then, that a significant portion (35%) of the newly identified proteins reported here have no known function. Thus, a subset of the novel proteins previously annotated as hypothetical can now be reannotated as “proteins of unknown function” due to their definitive measurement. The remaining proteins have expected extracellular functional annotations, including secretion/efflux/transport, cytochromes, dehydrogenases/reductases, and kinases. Among the newly identified proteins, neither physicochemical (molecular weight or pI) biases nor functional bias appear within the subgroup. The IMAC enrichment pattern is similar to that for the entire extracellular proteome identified. Indeed, the binding pattern for the newly identified proteins clearly shows that the vast majority bound to the cobalt and manganese loaded IMAC columns (Figure SM1, Supporting Information). In fact, in combination with the copper column, which has a very different protein enrichment pattern, these cobalt and manganese IMAC columns enriched for all but 13 of the 116 (89%) newly identified proteins (Figure SM1, Supporting Information).
Figure 3. General classification of metal binding. The identified proteins can be generally classified as differentially enriched (bound only to one specific metal), promiscuous (bound to multiple metals), or no metal binding. The majority (61%) of identified proteins were classified as promiscuous.
patterns. A portion of the proteins with an unknown function were identified as binding to a specific metal. These proteins contain no reasonable sequence homology to any currently characterized proteins. Forty-five proteins of unknown function were chosen for more detailed analysis because they exhibited selective binding for one specific metal. Functional domains and motifs were computationally predicted by batch submission of these 45 unique protein sequences to Pfam.31 Fifty-six unique Pfam identifications were predicted, representing multiple domains and motifs, with some proteins containing multiple, high scoring hits.
■
DISCUSSION
Novel Protein Identification
Analysis of IMAC Enrichment
Previous MS analysis of the extracellular fraction of the AMD community has reproducibly identified between 500−600 proteins.22 Among the 485 extracellular proteins identified here, 116 had not been identified in the prior MS measurements.19,22,25
The 485 identified proteins were further classified by their affinity for the seven different metals: universally enriched, universally enriched without metal-binding, and differential enrichment across the metals tested.
Table 2. Average Amino Acid Abundance within Protein Groups that Bound to Specific Metals, or Groups of Metals (averages)
total AA
His
Cys
Met
acidic
basic
hydrophobic
hydrophilic
all metals All (including nonmetal bound control column) divalent metals Zn(II), Cu(II) and Mg(II) Co(II), Mn(II), and Ni(II) Cu(II) Ni(II) Co(II) Zn(II) Mn(II)
228.51 220.50 232.92 206.42 186.80 244.50 339.60 232.50 234.64 263.69
1.90 1.63 2.15 2.45 1.22 1.91 1.74 1.25 2.22 2.32
0.91 1.09 0.70 1.24 0.75 0.38 0.43 0.51 1.14 1.08
2.46 2.39 2.82 2.10 3.35 2.55 3.04 2.97 2.86 3.37
17.28 17.98 16.92 12.96 18.40 18.16 25.60 19.91 19.49 22.87
29.48 28.63 32.23 31.12 26.62 31.47 30.04 28.87 33.58 33.84
44.57 42.65 44.25 43.37 41.73 42.53 44.14 42.60 42.40 42.86
55.43 57.35 55.75 56.63 58.27 57.47 55.86 57.40 57.60 57.14
865
dx.doi.org/10.1021/pr200693u | J. Proteome Res. 2012, 11, 861−870
Journal of Proteome Research
Article
Table 3. Identified Flagellar Proteins and Their Pattern of IMAC Enrichmenta cell motility proteins
Co(II)
Cu(II)
Fe(III)
Mg(II)
Mn(II)
Ni(II)
Zn(II)
control
Epl_15865_87_COG1681 Archaeal flagellins LII_11111_14_Putative flagellin LII_11111_17_Flagellar hook-associated protein (FlgL) LII_11111_21_Putative flagellin LII_11111_26_Flagellar basal body rod protein LII_11277_262_Probable flagellar hook protein FlgE LII_11277_263_Probable flagellar hook capping protein FlgD LII_8241_208_Probable flagellar hook protein (FlgE) LII_8241_209_Probable flagellar hook capping protein (FlgD) LII_8241_641_Putative flagellar basal body rod protein LII_8241_645_Putative flagellin LII_8241_649_Flagellar hook-associated protein (FlgL) LII_8241_652_Putative flagellin LIII_8063_25_flagellin domain protein LIII_8063_31_flagellin domain protein LIII_8063_36_flagellar basal-body rod protein FlgG LIII_9612_10_flagellin domain protein Unasn_10454_1_COG1749 Flagellar hook protein FlgE Unasn_4203_2_COG1344 Flagellin and related hook associated proteins
0 −1 0 −1 −1 −1 −1 −1 −1 −1 −1 0 −1 −1 0 −1 −1 1 −1
0 −1 0 −1 0 −1 −1 −1 0 0 −1 0 −1 −1 −1 1 −1 0 −1
0 −1 1 −1 0 −1 −1 −1 −1 0 −1 0 −1 −1 0 0 −1 0 −1
1 −1 1 −1 −1 −1 0 −1 0 −1 −1 1 −1 −1 −1 0 −1 1 −1
1 −1 0 −1 −1 −1 −1 −1 −1 −1 −1 0 −1 −1 −1 0 −1 0 −1
0 −1 −1 −1 −1 −1 −1 −1 −1 −1 −1 −1 −1 0 −1 0 −1 0 0
−1 −1 0 −1 −1 −1 −1 −1 −1 −1 −1 0 −1 −1 −1 −1 −1 0 −1
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
a
Proteins identified preferentially within the bound fraction of an IMAC column are indicated with a 1; proteins preferentially identified within the unbound fraction are shown with a −1; proteins either unidentified within a column or those with no bound/unbound preference are shown with a 0.
With the IMAC enrichment procedure, universally IMAC enriched proteins dominated the data set at 61%. The most predominant group of universally IMAC enriched proteins were ribosomal proteins and nearly all proteins associated with ribosomal structural roles in the database were identified as universally IMAC enriched.33 Although there is no evidence of the involvement of metal ions in peptide bond formation, metals are believed to play an important structural role in rRNA folding and stabilization of the rRNA structures34 Magnesium is the only divalent metal known to be abundant in the ribosome, along with a lesser abundance of zinc; however, the bounty of ribosomal phosphates may lead to nonspecific interactions with cationic metals. Conversely, the remaining 190 (39%) proteins represent differentially enriched proteins, with specific affinity for one metal (137, 28%) or nonbinding proteins (53, 11%). The enrichment of these proteins provides enhanced opportunities for further characterization. For example, eleven flagella proteins, the majority of flagellar proteins in the protein sequence database (11/18, 61%), demonstrate clear nonmetal binding, or show universal enrichment within the unbound fraction of the IMAC column (Table 3). It is noteworthy that flagellar proteins, designed to be exposed to the metal rich AMD environment, have no metal affinity. The IMAC enrichment data for each protein was clustered to reveal trends within the pattern of proteomic IMAC enrichment among the library of seven metals and enrichment for each individual protein (Figure 4). After clustering, the eight IMAC columns were divided into four distinct groups: (1) the control with no metal; (2) Fe(III), the only trivalent metal; (3) Co(II), Mn(II), Ni(II); and (4) Cu(II), Mg(II), Zn(II). As an outlier from the binding trends for the other six metals, the cluster of enrichment results from the ferric IMAC column demonstrated preferential Fe(III) binding for 7 proteins. Although the majority of iron in the AMD solutions is Fe(II), an abundance of Fe(III) is observed (with a [Fe2+] / [Fe3+] ratio between 9 and 2, depending on location) due to the
Figure 4. Heat map representation of protein enrichment among the seven metals. Metals are clustered from the left: Zn(II), Cu(II), Mg(II), Ni(II), Co(II), Mn(II), Fe(III), Control. Proteins identified as metal-bound are shown in yellow, metal-unbound in blue, and nonenriched in black. Differential enrichment (specific binding) falls into six major clusters, identified to the right.
submolar Fe concentration in these solutions. Immobilized Fe(III) is known to weakly interact with the carboxylic and phenolic groups and strongly bind phosphate groups, with chelation within a four-membered ring complex.35 The majority of proteins selectively enriched by Fe(III) were proteins of unknown function; however, one is a putative type I cytochrome. This observation is consistent with previous studies demonstrating nonspecific binding of Fe(III) to cytochromes through exposed protein surface interactions.6 Thus, the 866
dx.doi.org/10.1021/pr200693u | J. Proteome Res. 2012, 11, 861−870
Journal of Proteome Research
Article
Figure 5. Distribution of protein function among reductive, nonreductive and Ni(II)/Co(II) metals. The majority of the identified proteins have an unknown function. Three functional categories (COG) are widely represented across the seven metals: post-translational modification (O), energy production (C), and defense mechanisms (V). Among the redox, nonredox and Ni/Co metals, it appears that the distribution of proteins across functions remains consistent.
phosphate groups were likely to have also preferentially bound to other metal loaded IMAC columns.
by each metal except Fe(III). TonB was not identified in either control column’s bound replicates, illustrating its marked enrichment in IMAC columns loaded with divalent metals. TonB proteins are components of the TonB-dependent transport (TBDT) mechanism, which is generally located within the outer membrane and critical for the uptake of low abundance iron from the environment in gram-negative bacteria.36,37 BLAST analysis38 of this protein against the NCBI nonredundant database revealed a C-terminal region of the protein that exhibits sequence similarity (E-value
