Nic1 Inactivation Enables Stable Isotope Labeling with 13C615N4-Arginine in Schizosaccharomyces pombe*

Stable Isotope Labeling by Amino Acids (SILAC) is a commonly used method in quantitative proteomics. Because of compatibility with trypsin digestion, arginine and lysine are the most widely used amino acids for SILAC labeling. We observed that Schizosaccharomyces pombe (fission yeast) cannot be labeled with a specific form of arginine, 13C615N4-arginine (Arg-10), which limits the exploitation of SILAC technology in this model organism. We hypothesized that in the fission yeast the guanidinium group of 13C615N4-arginine is catabolized by arginase and urease activity to 15N1-labeled ammonia that is used as a precursor for general amino acid biosynthesis. We show that disruption of Ni2+-dependent urease activity, through deletion of the sole Ni2+ transporter Nic1, blocks this recycling in ammonium-supplemented EMMG medium to enable 13C615N4-arginine labeling for SILAC strategies in S. pombe. Finally, we employed Arg-10 in a triple-SILAC experiment to perform quantitative comparison of G1 + S, M, and G2 cell cycle phases in S. pombe.

Stable Isotope Labeling by Amino acids in Cell culture (SILAC) 1 is one of the most widely used methods in quantitative proteomics (1). It involves in vivo metabolic labeling of cell cultures (or small organisms) with different versions of stable isotope-labeled amino acids (2). To maximize the number of peptides that can be quantified after proteome digestion with trypsin, proteins are usually differentially labeled with different forms of lysine and arginine (3): L-lysine (Lys-0) and L-arginine (Arg-0); 2 H 4 -lysine(Lys-4) and 13 C 6 -arginine (Arg-6); or 13 C 6 -15 N 2 -lysine (Lys-8) and 13 C 6 -15 N 4 -arginine (Arg-10). The availability of multiple forms of labeled lysine and arginine support the application of SILAC in duplex (comparison of two states) or triplex (comparison of three states) formats. Efficient anabolic pathways mean that lysine and arginine are not essential for growth of wild type yeast cells. Auxotrophic mutants that are defective in these pathways can be used to switch yeast to an absolute dependence upon the provision of these amino acids in the external medium. Consequently, mutations in arginine and lysine biosynthesis pathways can be used to drive the complete labeling of all tryptic peptides with specific forms of these amino acids (4,5). SILAC has been used in quantitative proteomics in several yeast species, but most widely in Saccharomyces cerevisiae (6) (budding yeast) and Schizosaccharomyces pombe (fission yeast). S. pombe is extensively exploited to study cell cycle control (7), heterochromatin (8), and differentiation (9) and is increasingly the subject of large-scale quantitative proteomic studies (10,11).
A major challenge that is faced when using SILAC in fission yeast, is metabolic conversion of arginine to other amino acids such as proline, glutamine, and lysine (5). This partial labeling of additional amino acids after the conversion event produces spectra with complex isotope clusters that makes the downstream analysis challenging and error-prone. Inactivation of the "arginine conversion pathway" by removal of the orthinine transferase, Car2, effectively overcomes this problem to support the use of arginine labeling in SILAC-based experiments (5). Although this exploitation of the car2.⌬ mutation now enables SILAC technology in fission yeast, the choice of amino acids that can be employed remains limited. Only one form of heavy arginine (R6) is currently used alongside three forms of heavy lysine (Lys-4, Lys-6, and Lys-8) (5,12). Surprisingly, we could not find any studies that use arginine (Arg-10) in fission yeast, even though this is a widely exploited reagent for labeling other cell types (1).
Here, we show that labeling of fission yeast with Arg-10 leads to a general misincorporation of the stable isotope label that prevents the identification of labeled peptides. We hypothesize that successive arginase and urease activities catabolize the guanidinium group of Arg-10 to 15 N 1 -labeled ammonia. This labeled ammonium is then used as a general precursor for amino acid biosynthesis. Disruption of Ni 2ϩ -dependent urease activity through deletion of Ni 2ϩ transporter Nic1 in ammonium-supplemented medium, blocked this recycling to support 13 C 6 15 N 4 -arginine labeling SILAC strategies. As a proof of principle we employ Arg-10 in a triple-SILAC experiment to perform quantitative comparison of G1 ϩ S, M, and G2 cell cycle phases in S. pombe.

EXPERIMENTAL PROCEDURES
Yeast Strains-Standard Yeast Culture-Strains used in this used study are listed Table 1. Cell culture and maintenance were according to Moreno et al. (13). The amino acid supplements were uracil, adenine, leucine, and histidine at a concentration of 200 g/ml. Yeast strains were streaked onto YES plates (Solid Yeast Extract agar with supplements) from frozen glycerol stocks at Ϫ80°C, and grown at the permissive temperature (25°C) until single colonies formed. One single colony was used to start each liquid culture. For all physiological experiments, asynchronous cultures of cells were grown in variants of Edinburgh minimal media 2 (EMM2), with appropriate supplements from log phase starter cultures (1 ϫ 10 6 -5 ϫ 10 6 cells/ml). In EMMG, the 5g/L ammonium chloride of EMM2 is replaced with 5g/L monosodium glutamate (14). In EMMGn, EMMG is supplemented with 12 mM NH 4 Cl.
S. pombe Transformation Using Lithium Acetate-Cultures that had been grown in YES (nutrient rich) to mid-log phase (4 ϫ 10 6 -8 ϫ 10 6 cells/ml) were pelleted by centrifugation at 3400 ϫ g for 2 min, washed once in water followed by resuspension in appropriate volume of lithium acetate (0.1 M, pH 4.9) to a final density of 2 ϫ 10 8 cells/ml. After 1 h at 25°C 10-15 g of DNA was added to 100 l of competent cells alongside 2.9 volumes of PEG 4000 (50% (w/v) in 0.1 M, pH 4.9 lithium acetate). After 1 h at 25°C and 15 min at 42°C cells were washed once in water and plated onto selective minimal media.
Cell Culture-EMMG media is equivalent to EMM2 (15) except that 5 g/L ammonium chloride is replaced with 5 g/L monosodium glutamate (14). EMM2 media was also prepared with a range of ammonium chloride concentrations up to 96 mM. EMMGn is EMMG containing 12 mM NH 4 Cl. To label the cells with heavy or medium isotopes of Arginine (R) and Lysine (K), starter cultures were prepared by inoculation into EMMG/EMMGn supplemented with 37.5 g/ml each of light and medium/heavy isotopic versions of arginine and lysine. After inoculation from log phase starter cultures, main cultures were grown in EMMG/EMMGn supplemented with 75 g/ml of heavy isotopic labeled R and K for at least 10 generations before sampling to ensure the maximum labeling of the proteome. Care was taken that the culture never grew beyond mid-log phase (1 -5 ϫ 10 6 ). Each culture was harvested at mid-log phase. For direct comparisons with the EMM2 experiments of Bicho et al. (5) it was necessary to make starter cultures in YES media before inoculation into EMM2, as we were repeatedly unable to get growth in main cultures of either the SILAC strain when both the starter and main culture were made in minimal selective medium.
Cell Extracts for Proteomics Experiments-5 ϫ 10 7 cells in log phase of growth were harvested at 4000 rpm for 2 min in screw cap 1.5 ml microfuge tubes. Cell pellets were washed once in ice-cold STOP buffer (10 mM EDTA, 50 mM NaF, 150 mM NaCl, and 1 mM NaN 3 ) before being flash frozen in liquid nitrogen and stored at Ϫ80°C. Proteins were extracted by breaking the cells with 200 l acid washed glass beads in the presence of 200 l of denaturation buffer (6 M Urea, 2 M Thio-urea in 10 mM Tris buffer, pH 8.0) in a ribolyser (Fast prep FP120-BIO101, ThermoSavant, Carlsbad, CA) used for 30 s at 6.5 rpm. 200 l of denaturation buffer was added to the glass beads before incubation on a rotating rocker at RT for 30 min. A hole was made in the bottom of each tube before each was placed on the top of an empty 1.5 ml microfuge tube and spun at 4800 rpm for 2 min at RT. The insoluble debris was separated by centrifugation at 13,000 rpm for 10 min at RT. The supernatant was collected in a fresh 1.5 ml microfuge tube to be processed for LC-MS/MS. Protein concentration was measured using Bradford reagent.
In-solution Protein Alkylation and Digestion-Reduction buffer [1 M dithiothreitol (DTT) in 50 mM ammonium bicarbonate (ABC)] was added to the sample to a final concentration of 1 mM DTT before a 1h incubation at room temperature (RT). After reduction, alkylation buffer [550 mM iodoacetamide (IAA) in 50 mM ABC] was added to a final concentration of 5.5 mM IAA before a further 1h incubation at RT in the dark. The sample was then digested with endopeptidase LysC (Wako Pure Chemical Industries, Japan, 1:100 enzyme:protein) at RT for 3 h. After digestion with LysC at high urea concentration, the sample was diluted to a final concentration of 1.2 M urea/0.4 M thiourea with MiliQ water and digested with trypsin (Promega, Fitchburg, MA, 1:100 enzyme:protein) at RT overnight. The samples were acidified to inactivate trypsin by adding 1% TFA or fractionated using Off-Gel isoelectric focusing as described below (17) and desalted using C18 StageTips (18).
Off-Gel Isoelectric Focusing-Off-Gel separation was performed as described by Krug et al. (17). Briefly, peptide samples were sep-arated into 12 fractions using 13 cm Immobiline DryStrips with a pH 3-10 gradient (GE Healthcare) with the 3100 Off-Gel fractionator (Agilent, Santa Clara, CA). A maximum of 50 A were applied until 20 kVH were reached. Peptide fractions were acidified with 10 l of 30% ACN, 5% acetic acid, and 10% TFA in water.
MS Analysis-Desalted peptide fractions were separated using a nano-LC (Proxeon Biosystems) coupled to either an Orbitrap Elite or a LTQ Orbitrap XL (Thermo Fisher Scientific) mass spectrometer (MS). Peptides were separated on a 15 cm PicoTip fused silica emitter with a 75 m ID (New Objective) packed, in-house, with ReproSil-Pur C18-AQ 3 m resin (Dr. Maisch GmbH). Peptides were loaded on the column in Buffer A (0.5% acetic acid) at 700 nL/min using a maximum pressure of 280 Bar. Peptides were then eluted using a 57 or 87 min segmented gradient of 5-50% Buffer B (80% ACN in 0.5% acetic acid) at a flow rate of 200 nL/min. The mass spectrometers were operated in the positive ion mode. Spectra were recorded with datadependent mode using a dynamic exclusion of 90 s (LTQ Orbitrap XL) or 60 s (Orbitrap Elite). Resolution was set at 60,000 (LTQ Orbitrap XL) or 120,000 (Orbitrap Elite) with an accumulation target value of 1E6 charges. For peptide analysis the 10 (LTQ Orbitrap XL) or 20 (Orbitrap Elite) most intense peaks were selected from the survey scan (300 -2000 Thompson) and fragmented with collision induced dissociation at a target value of 5000 charges and recorded in the linear ion trap.
MS Data Processing-The MS data were analyzed with the Max-Quant Software Package version 1.3.0.5 (19) and searched using the Andromeda peptide search engine (20). The MS/MS spectra were searched against a database containing the proteome sequence of the S. pombe downloaded from PomBase (03.01.2013). The database contained 5145 entries from the S. pombe proteome, 248 frequently observed contaminants (downloaded from http://www. maxquant.org/contaminants.zip), as well as the reverse version of all sequences. Mass tolerance was set to 6 ppm at the MS and 0.5 Da at the MS/MS level. A minimal length of seven amino acids and full tryptic enzyme specificity were required. Two enzyme missed cleavages were allowed. Oxidation of methionine and protein N-terminal acetylation were defined as variable modifications and carbamidomethylation of cysteines was used as fixed modification. The false discovery rate was set to 1% on the peptide and protein level. In quantitative measurements, SILAC amino acids Lys4, Lys8, Arg6, and/or Arg10 were defined as fixed modifications. Significantly regulated protein groups were defined as those with calculated "Significance B" lower than 0.01.

RESULTS AND DISCUSSION
To assess the feasibility of using arginine (Arg-10) for quantitative proteomics of fission yeast, we performed SILAC labeling of S. pombe cells as described by Bicho et al. (5). We used arg3.D4 lys3.131 car2::natMX6 background and refer to this genotype as "SILAC" in the following text. Consistent with previous observations (5), the SILAC strain grows poorly in the EMM2 medium that uses ammonium as a nitrogen source (Fig. 1A). Ammonium ions inhibit one of two arginine uptake pathways, to severely compromise the growth of arginine auxotrophs such as the SILAC strain (23). Substitution of ammonium for monosodium glutamate in the medium EMMG enables both arginine import pathways to be used. Mid-log phase growth of the SILAC strain in EMMG was comparable to wild type growth in the same medium (Fig. 1A, 1B). We therefore grew cells in EMMG rather than the reduced ammonium EMM2 used by Bicho et al. EMMG is widely used throughout the fission yeast community and, beyond an impact upon TOR growth controls (24), no differences from growth on EMM2 have been observed. However, as an additional check we asked whether the inclusion of the SILAC mutations affected the expression of mitotic phenotypes arising from incapacitation of both polo kinase and the kinesin 5 motor protein? Shifting a plo1.ts41 cut7.24 SILAC strain to 36°C (25,26) led to an efficient mitotic arrest, indicating that neither the switch of nitrogen source, nor the incorporation of SILAC mutations, had any impact upon the degree of mitotic arrest invoked by these mutations (Fig. 1C).
While performing a standard triple SILAC experiment using "light" (Lys-0 and Arg-0), "medium" (Lys-4 and Arg-6), and "heavy" (Lys-8 and Arg-10) amino acids in EMMG we noticed that almost all heavy (Lys-8 and Arg-10)-labeled peptides had complex isotope clusters that effectively prevented identification of heavy peptides and proteins. The presence both arginine-and lysine-containing peptides in these complex clusters ( Fig. 2A) is indicative of a global conversion of the stable isotope label of either, the heavy arginine and/or lysine. To characterize the labeling problem in greater depth, we analyzed the light, medium-heavy, and heavy labeled samples separately. The identification rate for light and medium-heavy labeled samples was normal, whereas it was significantly reduced in samples from heavy labeled cells. To pinpoint the heavy amino acid responsible for the conversion, we measured samples labeled with a single heavy label (Arg-0/Lys-8 or Arg-10/Lys-0). Only the sample containing heavy arginine (Arg-10/Lys-0) presented the complex isotope clusters that reduced identification rates at both protein and peptide level. Thus, Arg-10 is incompatible with SILAC labeling approaches currently used for fission yeast analyses ( Fig. 2B; a summary of the proteins identified in each experiment is presented in supplementary Tables S1 to S15).
We reasoned that this extensive mislabeling of multiple amino acids probably arose from catabolism of guanidine group of heavy arginine (Arg-10) to heavy urea. Catalysis by Car1 and Aru1 would convert the guanidine group into urea that could be further catabolized by ureases into ammonia (Fig. 1D) (5). The 15 N 1 -ammonia derived from 15 N 2 -urea would then be incorporated into multiple amino acids, each containing different numbers of 15 N atoms to alter the isotope distributions of all peptides. Unlike the medium-heavy 13 C 6 -arginine (Arg-6), only heavy 15 N 4 13 C 6-arginine (Arg-10) has isotopically labeled nitrogen and so only samples labeled with this form of arginine (Arg-10) are vulnerable to this urease catabolism problem. Consistently, conversion was not observed in measurements of samples containing only medium-heavy amino acids arginine (Arg-6) and lysine (Lys-4), (Fig. 2B).
Ureases require Ni 2ϩ as the major cofactor for their catalytic activity. S. pombe encodes one characterized urease (Ure1) (27) alongside several other proteins that are predicted to exercise urease activity. Nic1 is the sole plasma membrane Ni 2ϩ transporter and supports urease activity of fission yeast (28). The observations of Eitinger et al. (28) led us to reason that the global reduction of the urease activity, arising from abolition of Nic1 function, would reduce the amino acid mislabeling we observed following growth of the SILAC strain on heavy arginine (Arg-10). We therefore deleted nic1 ϩ in the SILAC strain to generate "SILACn," before repeating the SILAC labeling as described above. As reported previously, nic1 ϩ ablation had no major impact upon the growth rate of otherwise wild type strains (Fig. 1B). Most importantly, the growth rate of SILAC and SILACn strains were identical (Fig.  1B), indicating that removal of nic1 ϩ is no more detrimental to cell physiology than the established mutations that are currently exploited for SILAC labeling in fission yeast (5).
Introduction of nic1.⌬ increased the identification of arginine (Arg-10)-labeled peptides, suggesting that urease-dependent arginine catabolism was indeed a key factor in arginine (Arg-10) mislabeling (supplemental Fig. S1). However, the number of peptides that could be identified in samples la-beled with heavy amino acids was below that obtained in samples containing light amino acids. Furthermore, the isotopic distribution of peptides was unusual, suggesting a partial incorporation of 15 N atoms. This incomplete suppression suggested either, that residual traces of Ni 2ϩ persist in nic1.⌬ cells, or that the urea is catabolized to ammonium ions by Ni 2ϩ independent urease/pathways.
The increased identification rate of (Arg-10) labeled peptides in the SILACn background indicated that, although 15 Nammonium has not been completely eliminated by removal of Nic1, its levels had been severely reduced. We therefore asked whether we could out compete the ability of the cells to incorporate these low levels of 15 N-ammonium ions in amino acid anabolic pathways, by supplementing the EMMG growth medium with levels of unlabeled 14 N-ammonium ions that are tolerated by SILAC strains (5). The SILACn strain grew with a doubling time of 5.5 h in EMMG media that contained 12 mM NH 4 Cl (EMMGn) supplemented with standard "light" forms of both lysine and arginine. This doubling rate is close to the 5 h of the SILAC strain and 4.5 h of wild type prototrophs in the same medium. When stable isotope-labeled amino acids were used as the sole source of lysine and arginine the mid-log phase, the doubling rate increased to 7 h at 25°C. For reasons that are not clear, the original SILAC strain (nic1 ϩ ) grew very poorly in EMMGn (supplemental Fig. S2). Impor-  tantly, however, the regular isotope pattern of the heavy-labeled peaks established that growth of SILACn cells in EMMGn successfully resolved the mislabeling problem (compare Fig. 2A top spectra with bottom spectra). This improved detection rate enabled us to identify comparable numbers of heavy labeled peptides to those that are routinely detected in cells labeled with light or medium amino acids (Fig. 2B).
Fission yeast is widely exploited for analysis of the cell division cycle. We therefore asked whether the combination of the SILACn background and EMMGn medium would impact upon the ability to use some well characterized mutations to synchronize progression through the cell division cycle. Removal of the function of the START transcription factor Cdc10 through the cdc10.v50 mutation reversibly arrests cell cycle progression in G1, because cells are unable to transcribe genes encoding DNA replication factors (29). As S phase is initiated immediately after anaphase chromosome segregation, the vast majority of cells in an asynchronous cdc10.v50 culture complete cytokinesis before accumulating at the G1 execution point (30). Transient cell cycle arrest at the cdc10.v50 execution point is therefore a commonly employed approach to synchronize progression through S phase, immediately after the block to Cdc10 function that is triggered by the return to the permissive temperature. We asked whether this cdc10.v50 "arrest release" approach could be used to synchronize S phase progression in a SILACn background. An early log phase cdc10.v50 SILACn culture was shifted to 36°C before returning to 25°C. FACs profiles indicated a proficient accumulation of 1N cells and an equally proficient reappearance of 2N cells following return to 25°C. (supplemental Fig. S3A). Temperature dependent inactivation of Kinesin 5 through temperature shift of the cut7.24 mutation removes the ability of the two half spindles to interdigitate. Consequently, the bipolar chromosome attachments that are required to silence the spindle assembly checkpoint (SAC) surveillance system are not established. This results in a persistent activation of the spindle assembly checkpoint that blocks cell cycle progression and so causes cells to accumulate in mitosis with monopolar spindles (26). The cut7.24 mutation completely blocked spindle formation in EMMGn, irrespective of the presence of the SILACn mutations (supplemental Fig. S3B). Finally, we tested the combined impact of the SILACn mutations and growth in EMMGn upon the widely exploited ability of the cdc25.22 mutation to reversibly arrest cell cycle progression at the G2/M boundary. Such cdc25.22 "arrest release" cultures are highly synchronized with respect to mitotic progression (5,21). cdc25.22 SILACn cells were grown to early log phase (1 ϫ 10 6 ) in EMMGn at 25°C before a transient shift to 36°C for 4.25 h. The absence of the septa that identify cells undergoing cytokinesis indicated an efficient cell cycle arrest at the G2/M boundary (data not shown). The transient appearance of replication and 4n peaks in FACS analysis of DNA content 90 -150 min after return to the permissive temperature of 25°C indicated an equally efficient return to cycle after release (supplemental Fig. S4). We conclude that neither the inclusion of the SILACn mutations nor growth in EMMGn has a deleterious impact upon the ability to manipulate cell cycle progression to generate populations at discrete cell cycle or mitotic stages with the cut7.24, cdc10.v50, or cdc25.22 mutations. The slightly slower growth rate of SILACn strains compared with WT is common among yeast auxotrophs. This reduced growth rate will have no major impact upon the interpretation of triple SILACn experiments because all three strains under query will harbor the same SILACn genetic background.
To validate this approach we performed a proof-of-principle triplex-SILAC measurement based on the duplex-SILAC cdc25.22 arrest release experiment performed by Bicho et al.  (5). In this experiment, we grew cdc25. 22 SILACn cells in either EMMGn supplemented with arginine and lysine for "light," "medium-heavy," or "heavy" labeling. Cells were transiently arrested at the G2/M boundary before return to 25°C, to release the arrest.  Table I. In 24 LC-MS runs, we recorded 747,658 MS/MS spectra that identified a total of 25,548 nonredundant peptides from 3606 S. pombe proteins, at estimated false discovery rate (FDR) of 1.30% at the protein level (supplemental Table S14). As shown in Table I, the use of Arg-10 did not influence peptide identification and the number of quantified M/L and H/L pairs were very similar, demonstrating that Arg-10 can be routinely used in SILAC measurements in the cdc25. 22 SILACn strain. Of the 3075 quantified proteins, 2430 (79%) were quantified in both replicates and correlation was high (R 2 ϭ 0.89), in spite of the fact that the labels were inversed in replicate analyses (supplemental Fig. S5A, 5B). We next compared our dataset to the data published by Bicho et al. (5) and observed good correlation (R 2 ϭ 0.59) between proteins that were significantly regulated in both datasets. Importantly, this correlation was confirmed in the replicate measurement with inverted SILAC labels (Supplemental Fig. S5C, 5D). Thus, Arg-6 and Arg-10 labeling was robust. Despite a relative paucity of significantly regulated proteins in both datasets, we could confirm several findings observed by Bicho et al. (5,31). In addition to several differentially regulated metabolic enzymes, we confirmed significant changes in the levels of polo kinase Plo1, condensin subunit Cnd2, and the HIRA histone chaperone complex hip3 in comparisons between the G2/ G1ϩS datasets. We also detected significant regulation of the Cdc14 related cell cycle phosphatase Clp1/Flp1, the membrane anchor for the Tea1 cell morphogenesis complex, Mod5, and the Greatwall kinase family member Ppk18, in both the M/G2 and G2/G1ϩS datasets (Fig. 3).
We conclude that removal of Nic1 from cells grown in EMMGn finally renders fission yeast fully compatible with SILAC labeling, to greatly expand the options for interrogation of the S. pombe proteome. The ability to routinely conduct triple SILAC analysis in this genetically malleable model organism will rapidly advance our quantitative understanding of many core principles of molecular cell biology.
□ S This article contains supplemental TablesS1 to S15 and Figs. S1 to S5.
ʈ To whom correspondence should be addressed: Proteome Center Tuebingen, Interfaculty Institute for Cell Biology, University of Tuebin-FIG. 3. The proteome of cell cycle populations generated by the cdc25. 22 arrest-release approach. Log 2 transformed SILAC ratios of protein groups plotted against their respective log 10 transformed intensity. On the left a comparison between G1S and G2 while a comparison between M and G2 is on the right. Significantly regulated proteins are marked in red; proteins discussed in the text are circled.