Use of Titanium Dioxide To Find Phosphopeptide ... - ACS Publications

Dec 14, 2010 - Here, we report a novel way to identify protein changes occurring during sperm development through the epididymis. Phosphopeptides were...
0 downloads 0 Views 5MB Size
ARTICLE pubs.acs.org/jpr

Use of Titanium Dioxide To Find Phosphopeptide and Total Protein Changes During Epididymal Sperm Maturation Mark A. Baker,†,‡,* Nathan D. Smith,† Louise Hetherington,† Matthias Pelzing,§ Mark R. Condina,§ and R. John Aitken†,‡ †

The ARC Centre of Excellence in Biotechnology and Development, Reproductive Science Group and ‡Priority Research Centre in Reproductive Science, School of Environmental and Life Sciences, University of Newcastle, Callaghan, NSW, 2308, Australia § Bruker Biosciences, PTY LTD, 28 Albert St, Preston, VIC 3072, Australia

bS Supporting Information ABSTRACT: Although the overall performance of modern mass spectrometers has increased, proteomic analysis of complex samples still requires prefractionation either at the protein or peptide level to allow for in-depth analysis of normal cellular function. Here, we report a novel way to identify protein changes occurring during sperm development through the epididymis. Phosphopeptides were first enriched from either the rat caput or caudal regions of the epididymides using TiO2, and the profiles then quantitatively compared. We show that 77 TiO2-enriched peptides become significantly modified in the epididymis, equating to 53 proteins. Through the use of immunoblot analysis, we confirmed that three proteins, ornithine-decarboxylase antizyme 3, heat-shock protein 90R, and testis-lipid binding protein, undergo major protein loss during epididymal passage. Many other proteins, including t-complex protein 10 and Spata18 show testis unique expression, appear to undergo phosphorylation during this same time frame. These data provide mechanistic insight into the means by which spermatozoa acquire functionality during epididymal transit. KEYWORDS: sperm phosphorylation, epididymal maturation, mass spectrometry, proteomics, titanium dioxide, label-free quantitation, ornathine decarboxylase, heat shock protein 90

ithout doubt, the field of proteomics has driven many technological advances, including that pertaining to the mass spectrometry. The latest models are capable of up to 5 log orders of protein sensitivity (attomole) and scanning speeds that can perform over 10,000 MS2 events in a single typical LC-MS run. Despite this, the greatest single problem facing proteomic analysis of any given sample is complexity. Highlighting this fact, the blood proteome has an estimated 6000 proteins and a reported dynamic range of 1012, making it impossible to characterize all the proteins on a single, one-dimensional (c-18) run.1 Hence, there is a need for prefractionation of complex samples at either the protein or peptide level. On a more reduced scale, the sperm proteome is thought to contain around 2500 proteins, with a dramatically reduced dynamic range compared to the blood proteome. 2,3 Furthermore, there is a paucity of data concerning the protein remodelling events that take place during sperm maturation. Interestingly, following their release into the seminiferous tubules, spermatozoa pass into the rete testes as morphologically differentiated but functionally incompetent cells. Functional competence is subsequently acquired as spermatozoa pass through a tightly coiled, highly differentiated duct known as the epididymis. It is only during this part of the spermatozoon’s maturation that the potential to exhibit a set of

W

r 2010 American Chemical Society

specific functions necessary for fertilization is acquired, including the capacities for progressive motility and interaction with the oocyte.4 An intriguing facet of this functional transformation is that it occurs in the complete absence of contemporaneous gene transcription and protein translation on the part of the spermatozoa. Thus, post-translational modifications (PTMs) are thought to be the major means by which spermatozoa acquire functionality during epididymal transit.5-7 In principle, these PTMs could be either components of an intrinsic maturation program orchestrated by the spermatozoa themselves or driven by external factors present in epididymal plasma. In this context, more than 100 proteins have been found by 2D-PAGE in the epididymal fluids of species such as boar and stallion.14,15 A number of these proteins have been characterized and found to be conserved across species, including angiotensin converting enzyme, antitrypsin, clustrin C2 and C3, galactosidase, glutathione peroxidase, cholesterol transfer protein, lactoferrin, hexosaminidase, prostaglandin D2 synthase, RNase like protein,8-10 and retinoic acid binding protein.11 Such high levels of conservation suggest that these proteins play essential roles in Received: July 13, 2010 Published: December 14, 2010 1004

dx.doi.org/10.1021/pr1007224 | J. Proteome Res. 2011, 10, 1004–1017

Journal of Proteome Research sperm maturation.10,12-17 From the spermatozoa’s perspective some proteins are known to become processed during epididymal maturation by proteolytic cleavage including the MDC (metalloprotease disintegrin cysteine-rich) family of proteins (fertilin-R, fertilin-β, tMDC I, tMDC II, and tMDC III),18,19 testase 1 (ADAM 24),20 basigin,21 and PH20.22-25 Protein phosphorylation/dephosphorylation is another form of PTM that is clearly involved in epididymal sperm maturation. For example, inhibition of the serine/threonine phosphatase isoform 1, gamma 2, stimulates motility in otherwise normally quiescent caput spermatozoa,26 suggesting this enzyme inhibits sperm motility during epididymal transit. Using difference in 2D-gel electrophoresis (DIGE) analysis, around 60 protein spots were found to change from the caput to cauda regions. One of these, the β-subunit of the mitochondrial ATPase, was found to undergo significant serine phosphorylation during epididymal transit.27 In a similar manner, epididymal maturation results in spermatozoa acquiring the potential to initiate a cAMPdependent tyrosine phosphorylation cascade that is centrally involved in the induction of sperm capacitation.28 These isolated examples apart, very little is known about the protein remodeling events taking place at the level of mammalian spermatozoa themselves as they acquire functionality within the epididymis. The present study was therefore designed to prefractionate peptides (subpeptidome) using TiO2 as a novel approach both to elucidate the protein phosphorylation changes associated with epididymal sperm maturation and to determine total protein changes. Our strategy involved running TiO2-enriched phosphopeptides from tryptic extracts of rat spermatozoa recovered from the caput and caudal regions of the epididymis. These peptides were then analyzed using nanoflow-LC with direct injection into an ion trap mass spectrometer. Survey scans from these files were then used to quantitatively plot the relative difference in phosphopeptide expression using software programs. As a result of this analysis we have found 77 statistically significant changes that occur during epididymal maturation that have, in turn, provided unique insights into the fundamental mechanisms that control the functional transformation of mammalian spermatozoa as they transit the epididymis.

’ EXPERIMENTAL SECTION Materials

Unless otherwise stated, all chemicals were purchased from Sigma-Aldrich at the highest research grade available. Albumin and ammonium persulfate were obtained from Research Organics (Cleveland, OH); D-glucose, sodium hydrogen carbonate, sodium chloride, potassium chloride, calcium chloride, potassium orthophosphate, and magnesium sulfate were all analytical reagent grade, purchased from Merck (BDH Merck, Kilsyth, VIC, Australia); ultrapure water, 2,5-dihydroxybenzoic acid (DHB), sodium sulfate, and zinc sulfate were from Fluka (Castle Hill, NSW, Australia); chloroform, methanol, and formaldehyde were purchased from Fronine (Riverstone, NSW, Australia) at the highest purity available. Tris was from ICN Biochemicals (Castle Hill, NSW, Australia), and acrylamide was from Biorad (Castle Hill, NSW, Australia). The 96-well plates were from Greiner BioOne (Interpath Services, Heidelberg West, VIC, Australia). Quantum Scientific (Pierce, Milton, QLD, Australia) supplied the phosphatase inhibitors and BCA assay kit, TiO2 was collected from a disassembled column

ARTICLE

(Titansphere, GL Sciences Inc., Tokyo), and the trypsin was from Promega (Annandale, NSW, Australia). Preparation of Rat Spermatozoa

Institutional and NSW State Government ethical approval was secured for the use of Wistar rats in this research program. Adult rats (∼8 weeks) were asphyxiated, and the epididymides were removed. Pure suspensions of spermatozoa were obtained from the caudal region of the epididymis by back-flushing, as previously described.29,30 The rat spermatozoa were gently isolated into BWW media and allowed to disperse for 10-15 min. The samples were then centrifuged (800  g, 3 min) to pellet the spermatozoa and separate these cells from any residual epididymal plasma. Caput-derived spermatozoa were isolated by finely slicing zones 1-3 with a surgical-grade scalpel. The sperm where then gently teased out into BWW media. Due to epithelial cell contamination, the sample was then put through 30% Percol solution as described elsewhere.27 The pellet was inspected, and only samples that contained a maximum of 1 round cell per 100,000 spermatozoa were used. Approximately 400 μL of lysis buffer consisting of 4% (w/v) CHAPS 7 M urea, 1/100 dilution of Halt phosphatase inhibitor (Pierce), and 2 M thiourea in water was added to 1  107 sperm cells. The thiourea aided in the removal of proteins from both caput and caudal epididymal origin, since cysteine cross-linking does occur during transit. For this reason, silver stains and immunoblot analysis against R-tubulin were used to show equal protein amounts. The sample was lysed for 1 h at 4 C with constant rotation and then centrifuged (16,000  g, 15 min, 4 C), and the supernatant was transferred to a new Eppendorf tube. A protein estimation was subsequently performed using a 2D quant kit (G.E. Healthcare, Castle Hill, NSW, Australia) and 150 μg of protein precipitated using methanol/chloroform as described elsewhere.31 Trypsin was then added in a ratio of 50:1 (protein/trypsin) in 25 mM ammonium bicarbonate containing 1 M urea, with constant shaking overnight at 37 C. Protein Estimation

Before methanol/chloroform precipitation, an aliquot corresponding to an equal amount (approx 5 μg) of protein from each sample was precipitated and resuspended in 1% (w/v) SDS and run into a 10% SDS-PAGE gel. The gel was silver-stained as previously described30 and visually inspected to ensure the quantification had been accurate and that extraction was equal between the more readily soluble caput spermatozoa compared to the caudal-derived ones. Finally, to ensure that the proteins in the original sample were fully digested, we loaded equal amounts of tryptic lysate from each sample onto a 10% SDS-PAGE. These samples were placed into bromophenyl-blue-free loading buffer and run halfway through a SDS-gel in order to visualize the peptides following silver staining. Phosphopeptide Enrichment

Purification and enrichment of phosphopeptides from the tryptic digest was performed by a method similar to that previously described.32 Tryptic peptides were diluted 5-fold in DHB buffer [350 mg/mL DHB, 80% (v/v) ACN, 2% (v/v) TFA] and applied to TiO2 beads (200 μg) pre-equilibrated in 50% ACN. The sample was then washed once in DHB buffer, before being washed three times with wash buffer [80% ACN (v/v), 2% TFA (v/v)] to remove the DHB. The sample was then directly eluted into a 96-well plate using elution buffer (elution buffer consisted of 25 μL, 2.5% ammonium hydroxide, pH g 10.5). The sample 1005

dx.doi.org/10.1021/pr1007224 |J. Proteome Res. 2011, 10, 1004–1017

Journal of Proteome Research was immediately neutralized with 0.3 μL of formic acid (1.2% final concentration). All buffers used ultrapure water and were made fresh on the day of experimentation. The eluates, typically in 25 μL, were then dried in a vacuum concentrator and resuspended in 10 μL of 0.1% (v/v) formic acid. Mass Spectrometry-Linear Ion Trap

Separation of tryptic peptide mixtures was achieved by nanoscale reversed-phase high pressure liquid chromatography (HPLC; nanoAcquity; Waters), in combination with online electrospray ionization (ESI)-MS. Primarily, the mass spectrometric analysis was performed on the LTQ linear ion trap system (Thermo-Finnigan, San Jose, CA), and a 3D-ion trap (Bruker Daltonic) was used to supplement the data (see below). Prior to mass-spectral analysis, a nanoflow uHPLC system (Waters, Rydalmere, NSW, Australia) was used for HPLC separation, employing a linear gradient of 2-40% buffer B (ACN, 0.1% (v/v) formic acid [FA]) over 50 min. The C18 column system consisted of a trap (300 μm i.d. 5 mm length) and a separation column (75 μm inner diameter  150 mm length, 1.7 μM particle). For online coupling, a nano ion spray source was used, and in the case of the linear ion trap this was equipped with a New Objective ESI needle (10 μm silica tip). The micro-TOF Q was run at 1 Hz with a scan range of 50-2200 Da. In both cases, we have targeted only the masses of those peptides seen to change in the Q-TOF quantitation. The needle voltage was 1.8 kV in positive ion mode. The scan cycle consisted of a survey scan (mass range 500-2000 amu) followed by MS/MS of the six most intense signals in the spectrum with an exclusion list for ion signals set to 25 s after one occurrence. For CID analysis, we used normalized collision energies set to 26; q = 0.18 with an activation time set to 30 ms and the isolation width set to m/z 1.0. Experimental Conditions LC-ESI-CID/ETD-MS/MS. To provide a larger list of phosphorylation sites, the same TiO2enriched spermatozoa samples were analyzed using a nanoHPLC system coupled to a 3D-ion trap. For all experiments, an Ultimate3000 (Dionex, Castle Hill, Sydney) equipped with a ternary low pressure mixing gradient pump (LPG-3600) with a membrane degasser unit (SRD-3600), a temperature controlled pulled-loop autosampler (WPS-3000T), and a temperature controlled column oven with flow manager (FLM-3100) were used. The LC experiments were performed using the “direct injection” setup under the following conditions: Nanocolumn C18 Pep. Mobile phase A: Map100, 75 μm i.d.  150 mm, 3 μm, 100 Å 100% water þ 0.1% FA. Mobile phase B: 20/80 water/ACN þ 0.08% FA. Flow rate nanocolumn: 600 nL/min. Gradient: 4-45% B over 45 min, 90% B for 5 min, 4% B for 30 min. Loopsize: 2 μL. Injection volume: 2 μL (FullLoop) by P User Defined Program. The oven temperature was set to 40 C For the CID/ETD experiments an amaZon ETD Ion Trap (Bruker Daltonik GmbH, Bremen, Germany) was used equipped with an online-nanosprayer spraying from a 0.090 mm o.d. and 0.02 mm i.d. fused silica capillary to avoid any loss of phosphopeptides. A detailed description of the ETD setup of the ion trap instrument including the generation of the reagent anion of fluoranthene was given previously.33 Fine tuning using the smart parameter setting option (SPS) for 900 m/z, compound stability 60%, and trap drive level at 100% in normal mode resulted in the following mass spectrometric parameters: dry gas temperature, 180 C; dry gas, 4.0 L min-1; nebulizer gas, 0.4 bar; electrospray voltage, 4500 V; high-voltage end-plate offset, -200 V; capillary

ARTICLE

exit, 140 V; trap drive: 57.4; funnel 1 in, 100 V out 35 V and funnel2 in, 12 V out 3.3 V; ICC target, 500 000; maximum accumulation time, 50 ms. The sample was measured with the “Enhanced Scan Mode” at 8100 m/z per second (which allows mono isotopic resolution up to four charge stages) polarity positive, scan range from m/z 100 to 3000, 5 spectra averaged, and rolling average of 2. The ETD reaction time was set to 100 ms using a reactant ICC of 500 000 allowing a maximum accumulation time for the reactant ion of 20 ms. The “Smart Decomposition” was set to “auto”. Acquired ETD/CID spectra were processed in DataAnalysis 4.0, and deconvoluted spectra were further analyzed with BioTools 3.2 software and submitted to Mascot database search (Mascot 2.2.04, IPI_rat database (v3.57, 39 873 sequences; 20 327 636 residues, release date 25/03/2009); peptide mass tolerance ( 0.3 Da, fragment mass tolerance ( 0.4 Da; enzyme specificity trypsin with 2 missed cleavages considered. The following variable modifications have been used: Deamidated (NQ), Oxidation (M), Phosphorylation (STY). DeCyderMS Analysis

To identify which proteins become modified during capacitation, the DeCyderMS peptide profiles from all fractions were compared to one another. For this purpose, the raw data files are imported into the program, which then converts the information into a virtual peptide image. All of the images are then loaded into the Pepdetect module of DecyderMS. This program first identifies individual peptides by placing a box around them, then the ion counts for each peptide are integrated, and the Log2 of this number is reported. The process is then repeated for each replicate sample analyzed. This information is then uploaded into the Pepmatch module, which matches the same peptides within the different virtual images, based on m/z and retention time. It should be noted that because the data were collected in an ion trap mass spectrometer, the tandem MS can be used to ensure correct matching of peptides (i.e., MS/MS spectra are the same). In the Pepmatch module, each virtual image is assigned a group (in this case, group 1 being caput spermatozoa and group 2 cauda spermatozoa), and a group-to-group comparison is performed. The integrated peptide counts are then used for statistical analyses of probability that any given change is due to chance. After this analysis, false discovery rates can then be determined from the data. For the complete data set, we have performed n = 8 biological replicates for each sample and 3 technical replicates on one sample to ensure reproducibility. DecyderMS Parameters

LTQ files of the survey scan performed in profile mode were imported into DecyderMS (v 2.0). Following visual interpretation of a complete DecyderMS image, the image was cropped before the appearance of the first peptide and following the appearance of the last eluted peptide. Typical cropped images covered retention times of 25-55 min. These images were then used for peptide analysis in the Pepdetect module. The standard filters were used with the following modifications: signal-tobackground threshold set to 10 with background subtracted, quantification checked, charge assignment was required, with limited charge assignment from two peaks; peptides below a signal-to-noise ratio of 4 and an unspecified charge, together with overlapping peptides, were removed. A typical run consisted of a peak width of 0.7 min. Following detection, every peptide was manually inspected for accuracy of charge and correct peak assignment. In cases of incorrect charge assignment, the m/z

1006

dx.doi.org/10.1021/pr1007224 |J. Proteome Res. 2011, 10, 1004–1017

Journal of Proteome Research was manually modified. Peptides were manually adjusted in cases of incorrect detection or added in cases where the program had clearly missed the peptide. Once the set of assignments had been made for all images, they were imported into the Pepmatch module, the retention times were aligned using the standard parameters, and the peptides were matched using the aligned retention times. Matching tolerances were 1.5 min elution time and m/z of 0.5 Da. Statistical Analysis

Following importation into DecyderMS, the program normalized the data on the basis of total peptide counts to compensate for intensity differences between the elution profiles. The peptide counts were based upon the intact peptide, as opposed to spectral counts that use the daughter ions for quantification. The program integrates the designated area under the curve based upon the MS-scan for each individual peptide and assigns a Log2 intensity value (note: if a peptide has multiple charge states, the sums of all of the states are collated). Peptides within different runs were matched on the basis of m/z, charge, and elution times. Peptide matching was manually confirmed using MS/MS spectral comparison. Student's t tests were then performed for each group (n = 8, caput vs cauda). For statistical evaluation, we only took into account those peptides that were matched in at least 14 of the 16 profiles. As label-free quantification based on MS-survey scan is a relatively new concept, there is no standard protocol by which the data should be interpreted. Therefore, we have reported 77 peptides with t test results indicating that a change occurred during epididymal maturation with a probability of p < 0.05. Also, we applied three FDR algorithms as described;34 these included B-H,35 B-L,36 and Bonferroni34 corrections. Those peptides significant for each correction are clearly labeled. Mass Spectrometric Data Interpretation

The derived mass spectrometry data sets were converted to generic format (*.dta) files using the Bioworks Browser (v3.3.1). These files were then searched against the IPI database (v3.36 containing 42,688 proteins) using the Bioworks (v3.3.1) search algorithm, TurboSequest (v3.3.1; Thermo-Finnigan). The species subset was set at Rattus norvegicus, and the number of allowed trypsin missed cleavages was set to 2.0, while oxidation of methionine and phosphorylation of serine, threonine, and tyrosine were selected. Peptide tolerance was set to 1.0, and intensity threshold was set to 100. The parent ion selection was set to 1.4 with fragment ion set to 0.7. For the differential analysis of phosphopeptides no filters were set but rather, manually validated to ensure the correct assignment of the y and b ion series in the case of CID or z and c ions for ETD and to ensure that the sequence was not redundant, thus justifying final inclusion in our list. The Sequest and MASCOT data for all of the peptides, together with the number of images the peptides were found in, the mass and retention times, and the actual spectra for all identified peptides (p < 0.05) are given in Supporting Information. To determine that a true phosphorylation site was observed, we looked for the characteristic neutral loss for serine/threonine residues. Unambiguous phosphorylation was allocated based on consecutive y or b ion series within which the phosphorylated amino acid could be readily detected. In cases of phosphorylation ambiguity, where the spectra did not permit assignment, we have shown the potential phospho-sequence in parentheses.

ARTICLE

For replicate analyses, on the 3D-trap system, the individual search results were combined into protein compilations using the ProteinExtractor functionality of Proteinscape 2.1.0 573 (Bruker Daltonics, Bremen, Germany), which conserved the individual peptides and their scores, while combining them to identify proteins with much higher significance than achievable using individual searches. To exclude false positive identifications, peptides with Mascot scores below 25 (which was chosen on the basis of manual evaluation of the MS/MS data of peptides with scores below this number) were rejected, unless part of a peptide pair in which one peptide pair had a score above 30. In addition, any protein with a score of 50 or less was instantly rejected. Finally, a false discovery threshold for protein scores of