analysis of all protein phosphatase genes in aspergillus ...gene deletion analysis of all protein...

EUKARYOTIC CELL, Apr. 2009, p. 573–585 Vol. 8, No. 41535-9778/09/$08.00�0 doi:10.1128/EC.00346-08Copyright © 2009, American Society for Microbiology. All Rights Reserved.

Analysis of All Protein Phosphatase Genes in Aspergillus nidulansIdentifies a New Mitotic Regulator, Fcp1�†

Sunghun Son and Stephen A. Osmani*Department of Molecular Genetics, The Ohio State University, Columbus, Ohio 43210

Received 16 October 2008/Accepted 22 January 2009

Reversible protein phosphorylation is an important regulatory mechanism of cell cycle control in whichprotein phosphatases counteract the activities of protein kinases. In Aspergillus nidulans, 28 protein phos-phatase catalytic subunit genes were identified. Systematic deletion analysis identified four essential phos-phatases and four required for normal growth. Conditional alleles of these were generated using the alcApromoter. The deleted phosphatase strain collection and regulatable versions of the essential and near-essential phosphatases provide an important resource for further analysis of the role of reversible proteinphosphorylation to the biology of A. nidulans. We further demonstrate that nimT and bimG have essentialfunctions required for mitotic progression since their deletions led to classical G2- and M-phase arrest.Although not as obvious, cells with AnpphA and Annem1 deleted also have mitotic abnormalities. One of theessential phosphatases, the RNA polymerase II C-terminal domain phosphatase Anfcp1, was further examinedfor potential functions in mitosis because a temperature-sensitive Anfcp1 allele was isolated in a genetic screenshowing synthetic interaction with the cdk1F mutation, a hyperactive mitotic kinase. The Anfcp1ts cdk1F doublemutant had severe mitotic defects, including inability of nuclei to complete mitosis in a normal fashion. Theseverity of the Anfcp1ts cdk1F mitotic phenotypes were far greater than either single mutant, confirming thesynthetic nature of their genetic interaction. The mitotic defects of the Anfcp1ts cdk1F double mutant suggestsa previously unrealized function for AnFCP1 in regulating mitotic progression, perhaps counteracting Cdk1-mediated phosphorylation.

In many cellular processes reversible phosphorylation/de-phosphorylation reactions play important roles by regulatingprotein activity and subcellular localization. As enzymescounteracting the function of protein kinases, protein phos-phatases have been firmly established as key coordinators ofdiverse biological events. Three main criteria are used toclassify protein phosphatases: sequence homology, struc-ture, and catalytic mechanism concerning substrate specific-ity. According to these features, protein phosphatases aredivided into the three main groups: classical serine/threo-nine phosphatases, protein tyrosine phosphatases, and theaspartate based catalysis protein phosphatase (reviewed inreference 33).

The classical Ser/Thr phosphatases can be further subdi-vided into the phosphoprotein phosphatase (PPP) family andthe protein phosphatase Mg2�/Mn2�-dependent (PPM) fam-ily. All members of the PPP family share high sequence simi-larities and have in common a conserved phosphoesterase cat-alytic motif (PP2Ac), although their biological functions rangewidely from mitotic regulation (PP1) (8), immune response(PP2B) (4), DNA damage response (PP4) (15), and blue-lightsignaling in plants (PP7) (32). PP2C type protein phosphatasesmake up the PPM family and are characterized by structural

similarity of their catalytic domains to PPPs despite the lack ofany sequence similarity (33).

The protein Tyr phosphatase (PTP) superfamily is distin-guished by its CX5R catalytic signature motif shared by all ofits members. It consists of the classical PTPs, the dual-speci-ficity phosphatases (DSPs), the Cdc25 type phosphatases, andthe low-molecular-weight phosphatases. The classical PTPsconsist both of receptor and nonreceptor PTPs, with functionsrelated to cell adhesion and cell-cell signaling. While the DSPsbelong to the PTP superfamily, they can not only dephosphor-ylate tyrosine residues but also show phosphatase activitiestoward Ser/Thr residues. Of the DSPs, the function of Cdc14(mitotic exit) (56) and PTEN (PIP3 signaling) (17) are rela-tively well understood, while little is known about many othermembers. The Cdc25 type phosphatases are well known fortheir essential function in regulating mitotic entry via the ac-tivation of cyclin-dependent kinase 1 (Cdk1) (13). Little isknown about the function of the low-molecular-weight phos-phatases.

The aspartate-based catalysis phosphatases are defined bythe shared signature catalytic motif of DXDXT/V (CPDs). Itsfounding member is the FCP1 phosphatase with an essentialfunction in coordinating the phosphorylation status of theRNA polymerase II largest subunit C-terminal domain (RNAPII CTD) and thus regulating the activity of the transcriptionmachinery (24).

In addition to these main protein phosphatase families,SSU72 and the tyrosine phosphatase families are consideredseparate types of protein phosphatases, although both sharesignificant sequence similarities with the PTPs (12, 31).

Like many other biological events, numerous regulatorymechanisms concerning the eukaryotic cell cycle have been

* Corresponding author. Mailing address: Department of MolecularGenetics, The Ohio State University, 804 Riffe Bldg., 496 W. 12thAve., Columbus, OH 43210. Phone: (614) 247-6791. Fax: (614) 247-6845. E-mail: [email protected].

† Supplemental material for this article may be found at http://ec.asm.org/.

� Published ahead of print on 30 January 2009.

573

on May 25, 2021 by guest

http://ec.asm.org/

Dow

nloaded from

https://crossmark.crossref.org/dialog/?doi=10.1128/EC.00346-08&domain=pdf&date_stamp=2009-04-01

http://ec.asm.org/

shown to involve reversible protein phosphorylation/dephos-phorylation. During mitosis, the function of protein kinasessuch as Cdk1, Aurora, NIMA, and Polo-like kinases are ofcritical importance for proper mitotic progression (16, 30, 40).As enzymes counteracting the effects of kinases, protein phos-phatases have naturally been considered as logical candidatesto have important cell cycle functions as well. As of now, Cdc25and Cdc14 of the PTP superfamily and PP1 and PP2A of theclassical Ser/Thr phosphatase family are the protein phos-phatases whose mitotic functions have been most studied.

Identified for its essential function in mitotic entry (9, 46),Cdc25 functions to activate Cdk1 by removal of the inhibitoryTyr15 phosphorylation of Cdk1. The activity and localizationof Cdc25 itself is regulated by phosphorylation as well. It hasbeen shown that Cdc25 is activated by Cdk1 mediated phos-phorylation, forming a positive-feedback loop to facilitaterapid entry into mitosis (18). Phosphorylation by Polo-likekinases is important in determining the subcellular localiza-tion of Cdc25 (54).

The DSP family member Cdc14 is known mostly for its roleof regulating later stages of mitosis. In the budding yeast Sac-charomyces cerevisiae, Cdc14p remains sequestered in nucleoliduring interphase but is released in anaphase via the FEAR(for Cdc fourteen early anaphase release) and MEN (for mi-totic exit network) signaling cascades. Once released Cdc14plocalizes to the spindle and spindle pole body to promotemitotic exit by reversing the phosphorylation of mitotic pro-teins carried out by Cdks during earlier phases of mitosis (53,56). The first evidence for PP1 functioning in mitotic regula-tion were presented from genetic analysis in Aspergillus nidu-lans and Schizosaccharomyces pombe showing that PP1 wasrequired for the completion of mitosis (6, 8, 37). Further stud-ies in Drosophila melanogaster confirmed the essential functionof PP1 in mitotic progression (2). Since the M-phase arrestphenotypes of different PP1 mutations in different organismsshow a high degree of heterogeneity, it is generally believedthat PP1 has multiple targets during mitosis, which is alsoconsistent with its diverse localization to chromatin, nuclearlamin, nucleolus, and spindle pole body during different stagesof mitosis (11, 55).

PP2A is another classical Ser/Thr phosphatase that has con-firmed mitotic functions. Although (along with PP1) the mi-totic function of PP2A in S. pombe was identified as early as1990 (23), studies in recent years point to a specific role ofPP2A in controlling sister chromatid cohesion by interactingwith shugosin proteins at the centromere (52).

In order to analyze the protein phosphatase genes in A. nidu-lans for potential mitotic functions, we carried out a systematicgene deletion analysis of all protein phosphatase catalytic subunit-encoding genes. In addition to this genomic approach, a novelgenetic strategy, engaging the function of the mitotic kinase Cdk1(known as NimXcdc2 in A. nidulans) (42), was set up in the formof a synthetic interaction screen to isolate mutations that, com-bined with a hyperactive cdk1F mutant (61), cause increasedtemperature sensitivity. This screen identified a conditional mu-tant allele of the RNAP II CTD phosphatase, Anfcp1, and thefunction of this gene was examined for its potential role in mitosisin combination with regulated Cdk1 phosphorylation.

MATERIALS AND METHODS

General A. nidulans methodologies. Standard growth and classical geneticmethodologies for A. nidulans were in essence as previously described (44). Astrain list is provided (see Table S1 in the supplemental material). Endogenousgene deletions and green fluorescent protein (GFP) tagging was carried out asdescribed previously (35, 59). Identification and analysis of essential genes viaheterokaryon rescue was completed according to the method described previ-ously (41).

Genetic screen. For the genetic screen, A. nidulans strain FRY24 was used inwhich a nimXCDK1F mutant allele had been inserted at the chromosomal locus ofnimXCDK1 to form a tandem repeat of the wild-type nimXCDK1 and mutantnimXCDK1F alleles, linked together by the auxotrophic marker pyr4. The screenconsisted of two successive selection procedures after the initial introduction ofnew mutations via UV irradiation (administered at 3.15 � 104 �J/cm2 on 500,000viable spores on 100 plates to yield �5,000 viable spores/plate at a survival rateof 1.08%). After random UV mutagenesis, surviving cells were tested for initialtemperature sensitivity at 42°C. Strains that showed initial temperature sensitivephenotypes were selected for subsequent eviction of the nimXCDK1F allele. Byutilizing the toxicity that hydroxyurea (HU) and 5-fluoorotic acid (5�-FOA) conferupon cells expressing nimXCDK1F and pyr4, respectively, a double counterselectionwas carried out on the �5,000 surviving strains by growth on medium containing 8mM HU and 0.1% 5�-FOA to recover cells that had evicted both the nimXCDK1F

allele and the pyr4 gene via homologous recombination between the nimXCDK1F andthe wild-type nimXCDK1 alleles. After verifying the eviction of nimXCDK1F, themutant strains were analyzed for any change to their initial temperature-sensitivephenotypes within the range of 32 to 42°C. Loss of the temperature sensitivity dueto nimXCDK1F eviction confirmed that the newly introduced mutation andnimXCDK1F were genetically interacting in a synthetic manner. Mutations identifiedas showing genetic interactions with nimXCDK1F in this way were further analyzedgenetically to verify whether they were single mutations, and their allelic relationshipwas examined. Mendelian segregation of the mutant phenotypes was used to confirmthe strains contained single mutations. Allelic relationships of the different muta-tions were verified by crossing the different mutants with each other, whereupon theoutsegregation of a wild-type phenotype in the F1 progeny determined the given twomutations to have occurred on different genes. Dominance of the mutations weredetermined by generating diploid strains of wild-type and mutant haploid strains.Only recessive mutations were selected for molecular cloning purposes. Single,recessive mutations were consequently cloned by complementation via transforma-tion with an AMA1 plasmid-based A. nidulans genomic DNA library (38). A singlemutant allele of Anfcp1 was recovered and sequenced.

Conditional growth test conditions. Strain with null allele of the nonessentialphosphatase genes were tested on MAGUU plates for their response to thefollowing conditions: benomyl (04 and 0.8 �g/ml), sucrose (0.5, 1, and 1.5 M),hydroxyurea (8 and 16 mM), NaCl (0.5, 1, 1.5 M), methyl methanesulfonate(0.025 and 0.05%), 1,2,7,8-diepoxyoctane (0.01 and 0.02%), camptothecin (25�g/ml), and temperature (20 and 32 to 42°C).

Affinity purifications using the S-Tag affinity peptide. We have developed theS-Tag affinity purification system for single step protein affinity purification of A.nidulans proteins endogenously tagged with the S-Tag peptide (21). Details havebeen published elsewhere (29a). Notably, however, the standard protocol wasmodified by the addition of 320 mM ammonium sulfate (replacing 300 mMNaCl) and 20% glycerol to the extraction buffer. Purified proteins were run onstandard Laemmli–sodium dodecyl sulfate-polyacrylamide gel electrophoresis(SDS-PAGE) gels using a 3% stacking and 6% separating gel made from a 40%(wt/vol) solution of acrylamide and bis-acrylamide (37.5:1).

In vitro phosphatase reactions and analysis. After purification of protein viaS-Tag affinity purification, beads were treated with 40 U of �-phosphatase (NewEngland Biolabs) with or without phosphatase inhibitors (1 mM sodium vana-date plus 50 mM sodium Fluoride) after being washed in 2� phosphatase bufferthat included MnCl2 and protease inhibitors. The beads were mixed and incu-bated at 30°C for 30 min (with occasional mixing). Sample buffer was added, andthe supernatant was collected and analyzed by SDS-PAGE. For Coomassie bluestaining, gels were fixed in 50% ethanol and 10% acetic acid overnight, rinsedthree times for 5 min with distilled H2O, and stained with Bio-Rad Bio-SafeCoomassie blue solution for 1 h. Background staining was removed by using awater wash for 30 min. Standard methodologies were used for silver staining.

Microscopy and image capture. Fixed samples were examined by using anE800 microscope (Nikon, Inc.) with DAPI (4�,6�-diamidino-2-phenylindole), flu-orescein isothiocyanate, or Texas Red filters (Omega Optical, Inc.). Image cap-ture was performed by using an UltraPix digital camera (Life Science Resources,Ltd.). For live cell microscopy, strains were germinated in 3 ml of minimal mediain 35-mm glass bottom petri dishes (MatTek Cultureware). Visualization was

574 SON AND OSMANI EUKARYOT. CELL


http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

performed by using a Nikon Eclipse TE300 (Nikon, Inc.) inverted microscope inconjunction with an Ultraview spinning disk confocal system (Perkin-Elmer) andan Orca ER digital camera (Hamamatsu). The system is fitted with a Bioptechsdelta T4 culture dish heater and objective heating collar for precise temperaturecontrol. Image capture on both microscopes was performed by using Ultraviewimage capture software (Perkin-Elmer). Image analysis was completed by usingAdobe Photoshop 6.0 (7).

RESULTS

Identification of 28 protein phosphatase catalytic subunitgenes in A. nidulans. The protein sequence of all protein phos-phatase catalytic subunit genes of the budding yeast S. cerevi-siae were used for BLAST searches to determine sequencehomologues in the Aspergillus Comparative Database of theBroad Institute (http://www.broad.mit.edu/annotation/genome/aspergillus_group/MultiHome.html). In addition, the catalyticdomains of different classes of protein phosphatases were usedfor BLAST searches. A total of 28 different genes predicted toencode protein phosphatase catalytic subunits were thus iden-tified (Table 1). Of the 28 predicted genes, 4 had been iden-tified in previous studies. Originally isolated through geneticscreens (34), nimT and bimG were shown to be required forproper mitotic entry and progression, respectively (8, 36). Inaddition, pphA is thought to regulate hyphal morphogenesis(26), while cnaA function had been implicated in the regula-tion of the G1/S transition of the cell cycle (45), althoughrecent studies indicate it to be nonessential (50).

Generation of protein phosphatase gene deletion strains. Toexamine the function of the protein phosphatase catalytic sub-units, 27 of the 28 identified genes were deleted in a systematicmanner (we were unable to generate the cnaA-null allele). Thenull alleles were generated by replacing coding sequences withgene deletion constructs obtained by fusion PCR (59). Theefficacy of homologous recombination of these constructs intothe genome was enhanced by utilizing an A. nidulans recipientstrain of a �Ku70 genetic background (�nkuA) (35). A. nidu-lans is capable of forming heterokaryons that maintain twotypes of genetically unlike nuclei within a common cytoplasm.When an essential gene is deleted, the primary transformantcolonies maintains a heterokaryotic state in which the essentialgene function is provided by the undeleted nuclei, while thefunction of the nutritional marker is provided by the deletednuclei. Since conidiospores (conidia) of A. nidulans are uninu-cleate the heterokaryotic state cannot be propagated throughthese spores. Therefore, conidia of the primary transformantsof essential gene deletions are unable to propagate whenstreaked on nutritionally selective growth medium. Of the 27A. nidulans protein phosphatase genes investigated, deletion of4 protein phosphatase genes resulted in the consistent gener-ation of heterokaryons, confirming their cellular functions tobe essential (41). A representative heterokaryon rescue anal-ysis data set of bimG gene deletion is shown in Fig. 2A. Thediagnostic PCR results, using primers targeting sequences out-side the gene deletion construct, confirm the formation of

TABLE 1. Protein phosphatase catalytic subunit genes in A. nidulansa

A. nidulans genea S. cerevisiae-human gene e valuebGene deletion

DomainA. nidulans S. cerevisiae

AN0103 Pph3p-Pp4 3e–56 Viable Viable PP2AcAN0164 Ppg1p-Pp2Ac 2e–109 Sick Viable PP2AcAN0410 bimG Glc7p-Pp1a e–170 Lethal Lethal PP2AcAN0504 sitA Sit4P-Pp2Ac e–110 Viable Viable PP2AcAN3793 Ppz1p-Pp1b 2e–145 Viable Viable PP2AcAN6391 pphA Pph21p-Pp2a e–148 Lethal Viable PP2AcAN8820 cnaA* Cmp2p-Ppp3Ca 3e–173 Sick Viable PP2AcAN10281 Ppt1p-Ppp5 5e–110 Viable Viable PP2AcAN0914 Ptc6p-Ppm1k 3.3e–16 Viable Viable PP2CcAN1358 Ptc2p-Ppm1a e–75 Viable Viable PP2CcAN1467 Ptc7p-Pptc7 2e–16 Viable Viable PP2CcAN2472 Ptc2p-Ppm1k 4e–87 Viable Viable PP2CcAN5722 Ptc5p-Ppm2c 9e–65 Viable Viable PP2CcAN6892 Ptc1p-Ppm1a 4e–44 Viable Viable PP2CcAN1343 Nem1p-Dullard 5e–94 Sick Viable CPDcAN2902 Fcp1p-Fcp1 3e–61 Lethal Lethal CPDcAN10077 Psr1p-Ctdsp2 7e–61 Viable Viable CPDcAN0129 Pps1p-Dusp2 4e–45 Viable Viable DSPcAN4419 Yvh1p-Dusp12 e–34 Sick Viable DSPcAN4544 Msg5p-Dusp9 9e–12 Viable Viable DSPcAN5057 Cdc14p-Cdc14c 2e–117 Viable Lethal DSPcAN10138 Sdp1p-Styx 0.71 Viable Viable DSPcAN4896 Ptp1p-Ptpu2 e–31 Viable Viable PTPcAN6982 Pyp1-Ptpn1 e–34 Viable Viable PTPcAN10570 Ltp1p-Acp1 3e–20 Viable Viable LMWPcAN3810 Ssu72p-Ssu72 3e–48 Sick Lethal SSU72AN3941 nimT Mih1p-CDC25 2e–27 Lethal Viable CDC25AN4426 Siw14p-Lkhp9428 7e–31 Viable Viable Y-phosphatase

a That is, the designations given by the A. nidulans Genome database. Gene names are given for already-cloned genes. *, Not deleted in this study but previouslyidentified and characterized (45, 50).

b Values are given for comparison to the S. cerevisiae protein at the NCBI.

VOL. 8, 2009 ASPERGILLUS PROTEIN PHOSPHATASES 575


http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

heterokaryons containing both parental and deleted alleles.The deletion strains of the other 23 phosphatase genes wereconfirmed to be viable homokaryons, demonstrating the func-tion of these protein phosphatases to be nonessential. No con-ditional phenotypes associated with these gene deletions wereidentified under various experimental growth conditions re-garding genotoxic stress, high osmolarity, and high and lowgrowth temperatures. Nevertheless, null alleles of four nones-sential protein phosphatases showed significant growth defects.The deletion strains of Anppg1 (AN0164), Annem1 (AN1343),Anssu72 (AN3810), and Anyvh1 (AN4419) exhibited defectivegrowth phenotypes readily distinguishable by the small sizes oftheir colonies (Fig. 1A). When the diameter of their colonieswere measured daily over 5 days, all four showed reduced

colony growth rates compared to the control wild-type strainsand other deleted phosphatase mutants (data not shown).

Generation of conditional alleles of protein phosphatasegenes utilizing the alcA promoter system. As a tool to moreeasily study the protein phosphatase genes whose deletionsproved to be lethal, or result in severe growth defect pheno-types, strains containing conditional alleles of these genes weregenerated. In A. nidulans, the targeted regulation of gene ex-pression can be achieved by placing the gene of interest underthe control of the promoter of the alcohol dehydrogenase Igene alcA (57). As part of the alcohol utilization pathway, alcAgene expression is highly induced in response to ethanol andthreonine and suppressed in the presence of glucose (3, 28,29), while fructose, lactose, or glycerol allows a moderate level

FIG. 1. Colony growth of all nonessential phosphatase gene deletion strains. (A) While the colony sizes of all other gene deletion strains arecomparable to the wild-type control strains SO451 and TNO2A7, including strains SS140 (�AN5057), SS150 (�AN4544), SS159 (�AN0129), SS154(�AN10077), SS157 (�AN10138), SS148 (�AN5722), SS133 (�AN2472), SS141 (�AN1467), SS164 (�AN0914), SS167 (�AN4896), SS143(�AN6982), SS147 (�AN1358), SS143 (�AN6892), SS135 (�AN3793), SS170 (�AN0504), SS125 (�AN0103), SS137 (�AN4426), SS130(�AN10570), and SS129 (�AN10281), colonies of the �Anyvh1 (SS162), �Annem1 (SS132), �Anppg1 (SS153), and �Anssu72 (SS160) strains aresmaller. The numbers refer to the AN numbers designated by the Broad Institute (http://www.broad.mit.edu/annotation/genome/aspergillus_group/MultiHome.html). (B) The deletion phenotype of the indicated genes can be mimicked utilizing alcA promoter driven conditional alleles. Whentheir gene expression is suppressed on glucose medium, colony growth is impaired similar to the gene deletion strain colonies: alcA-ppg1 (SS087),�ppg1 (SS153), alcA-nem1 (SS099), �nem1 (SS132), alcA-ssu72 (SS122), �ssu72 (SS160), alcA-yvh1 (SS078), and �yvh1 (SS162). Three controlsare the nonessential phosphatase AN0129 gene under alcA control, a wild-type (growth control strain SO451), and alcA-regulated nimT, which isan essential gene (positive control strain SS124).



http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

of alcA expression (10, 43). In the present study, the alcApromoter constructs were endogenously inserted immediatelyupstream of the phosphatase gene loci, generating strains inwhich the only copy of the given phosphatase is under thecontrol of the alcA promoter. When the growth characteristicsof these strains were compared to the corresponding genedeletion strains, the suppression of the genes on glucose-con-taining medium proved to be effective enough for the growthpatterns exhibited by the alcA inducible strains to mimic that ofthe gene deletion strains. When grown on suppressing mediumalc::Annem1 strains showed growth defects closely resemblingthose of the corresponding gene deletion strain (Fig. 1B).Although the growth defects exhibited by the alc::Anppg1,alc::Anyvh1, and alc::Anssu72 strains were not as severe as inthe deletion strains, they still produced colonies of smaller sizethan wild-type strains (Fig. 1B). Suppression of the essentialprotein phosphatases was highly effective since it yielded com-plete growth arrests of the alc::Anfcp1 and alc::bimG strains,and only minicolonies were formed for the alc::nimT andalc::pphA strains (Fig. 2B).

Germination and growth defects resulting from essentialprotein phosphatase gene deletions. To examine the growthcharacteristics of cells with essential protein phosphatase genedeletions, conidia from the primary gene deletion transfor-mants were streaked on YAG medium plates and grown atroom temperature for 3 days. Since the nutritional markerused for the transformation was pyrGAF, YAG plates are nu-tritionally selective for undeleted conidia that carried the non-functional pyrG89 mutant allele. Not meeting the nutritionalrequirement, it is known that these conidia fail to project agerm tube and do not grow beyond swelling of the conidios-pores. In contrast, the conidia which are pyrG� but deleted ofthe essential phosphatase genes would grow up to the limitimposed by the lack of the essential phosphatase gene func-tion, enabling the assessment of the terminal growth pheno-types caused by the deletions. Deletion of bimG showed themost severe growth defect as the �bimG spores on YAGmedium failed to project germ tubes and nuclei failed todivide (Fig. 3A). Deletion of Anfcp1 was slightly less severeas �Anfcp1 spores were able to germinate and project very shortgerm tubes (Fig. 3A). The nimT gene deletion had a lessdetrimental effect on initial germination and polarized germ tubeprojection (Fig. 3A). Similarly, �pphA spores were also able togrow into germlings comparable to �nimT spores (Fig. 3A).

Nuclear defects in cells with the essential phosphatase genesdeleted. To determine potential nuclear and mitotic defectsassociated with the deletion of essential phosphatase genes,spores of these deletion transformants were suspended intoYG liquid medium and inoculated on coverslips, grown for18 h at room temperature, DAPI stained, and imaged by flu-orescence microscopy (as an exception, spores taken from thebimG deletion transformant were grown in nonselectiveYGUU medium since �bimG pyrG� spores and bimG pyrG89spores were more difficult to distinguish in selective YG me-dium). Previous studies conducted with the bimG11 condi-tional mutant allele had revealed that loss of BIMG activity ledto mitotic arrest of the cell with condensed DNA (8). Whenvisualized by DAPI staining the nuclei of �bimG spores werearrested with DNA remaining condensed, a phenotype consis-tent with the observations made in bimG11 mutants (Fig. 3A).

DAPI staining of �nimT cells revealed their nuclei showing aG2-phase arrest with nuclei unable to enter mitosis with un-condensed DNA, also confirming the previous conclusionsabout nimT function drawn from studies of the nimT23 tem-perature-sensitive mutant (36). Although germlings of �pphAcells were capable of at least three rounds of mitoses, theirnuclei had irregular morphologies and distribution along thegerm tube, indicative of potential mitotic defects (Fig. 3A).�Anfcp1 cells showed DAPI staining patterns that made it

FIG. 2. Essential function of protein phosphatase genes analyzed uti-lizing heterokaryon rescue and alcA-driven alleles. (A) Conidia from sixprimary transformants obtained after transformation with a bimG dele-tion construct were replica plated on selective (YAG) and nonselective(YAGUU) plates and incubated to allow colony formation. The inabilityof �bimG conidia and untransformed conidia to grow on selective YAGmedium and the ability of untransformed conidia to grow on nonselectiveYAGUU media confirms that each primary transformant colony is het-erokaryotic, indicating that bimG is essential (41). Heterokaryon forma-tion was verified using diagnostic PCR analysis, showing that the wild type(lane 1) contains just the wild-type allele and that the heterokaryon con-tains both wild-type and deletion alleles (lane 2). (B) Deletion phenotypesof essential phosphatase genes, as indicated, can be reconstituted bysuppressing their expression using the alcA promoter such that growth canoccur when alcA is induced (ethanol media) but not when alcA is sup-pressed using glucose media. Strains SS120 (alcA-fcp1), SS124 (alcA-nimT), SS121 (alcA-bimG), and SS123 (alcA-pphA) are represented. Thisboth confirms that these genes are essential and provides strains forfurther functional analysis.



http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

difficult to distinguish individual nuclei, possibly due to mitoticdefects.

Cytological abnormalities in nonessential protein phos-phatase deletions with growth defects. For the analysis of the

nonessential phosphatase gene deletions that cause a declinein colony size, clonal conidium stocks were generated for eachof the deletion strains, as well as the alcA-regulated strains.After 6.5 and 9 h of growth in YGUU-rich liquid medium at

FIG. 3. Growth arrest and nuclear defects of essential phosphatase gene deletions. (A) The terminal growth patterns of essential phosphatase genedeletions were recorded after 3 days growth of conidia from heterokaryons of the indicated gene deletions on selective plates and then photographed atlow power. �bimG conidiospores show virtually no polarized growth even after 3 days. While �Anfcp1, �pphA, and �nimT conidia are able to germinate,the projected germ tubes fail to grow to form colonies. To the right, representative germlings of the indicated deletions are shown after DAPI stainingto reveal nuclear number and morphology. �bimG and �nimT cells showed late G2- and M-phase arrest phenotypes, respectively, consistent withpreviously known loss-of-function mutant phenotypes. The nuclei of �Anfcp1 cells appear severely fragmented, and �pphA cells exhibit irregulardistributions of nuclei along the germ tube with indications of defects in proper nuclear segregation. (B) Representative germlings, after DAPI staining,of gene deletions that cause growth defects but not lethality, as indicated, after 6.5 h (top panels) and 9 h (bottom panels) of growth. Strains are SS162(�yvh1), SS153 (�ppg1), SS132 (�nem1), and SS160 (�ssu72) are represented. The R153 wild-type strain is shown for comparison. Bar, �5 �m.



http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

32°C, the cells were fixed and stained with DAPI to visualizeDNA. Control wild-type cells (strain R153) grown under thesame conditions reached, on average, the 4- and 16-nucleusstages, respectively, after 6.5 and 9 h of growth (Fig. 3B), andthe growth and cytological defects exhibited by the phos-phatase null mutants were compared to them. Deletion ofAnppg1 and Anssu72 appeared to slightly affect the overallearly growth rate without causing any pronounced nucleardefects. At time points of 6.5 and 9 h of growth, both �Anppg1and �Anssu72 cells had fewer nuclei, as well as shorter germtubes compared to the wild-type cells, indicating an overallslowdown of growth as a consequence of the gene deletions(Fig. 3B). Although less severe, these phenotype were alsoobserved in the alc::Anppg1 and alc::Anssu72 cells grown un-der suppressing conditions (the same YGUU rich liquid me-dium used is glucose-based and thus suppresses alcA-promoter-driven expression) (data not shown). In the Annem1-nullmutants, although the overall growth rate of the cells was alsoaffected, signs of DNA fragmentation were the most significantdeletion phenotype. The nuclei were small in size and of ir-regular shape, with some appearing to have condensed DNA.A feature more pronounced in the cells that had grown longer(9 h) was the presence of fiber-like structures positively stainedby DAPI that appeared to be interlinking the nuclei (Fig. 3B).These phenotypes persist in the alc::Annem1 strain grown un-der suppressing conditions, although their growth rate resem-bles that of the wild-type strain (data not shown). Taken to-gether, these phenotypes suggested defects in mitosis, leadingto fragmented DNA masses and nuclei that failed to fullyseparate from each other. The most severe growth defect phe-notype among the nonessential phosphatase gene deletionstrains was shown by the �Anyvh1 strain. After 6.5 h of growth,most of the spores had failed to germinate, and many of thosecells had not undergone their first mitosis either (Fig. 3B).Even the cells that had managed to initiate germination weresmall in size, not only in terms of the length of the germ tubebut also in the extent of how much the spore had swollen priorto germination. When �Anyvh1 cells were grown for 9 h, thegerm tubes were still very short in length (Fig. 3B).

Genetic interaction of an Anfcp1 mutation with cdk1F. Thesystematic gene deletion analysis of the protein phosphatasecatalytic subunits revealed the function of bimG, nimT, pphA,and Anfcp1 to be essential in A. nidulans. The severity of nucleardefects, suggestive of mitotic defects, observed in �Anfcp1 cellswas interesting, and the results of a novel forward geneticscreen provided insight into the potential role of Anfcp1 inregulating the cell cycle and mitosis. Cdk1 is one of the mainregulators determining entry into mitosis. Activation of Cdk1,which triggers the G2/M transition, hinges on the removal ofinhibitory phosphorylation from a tyrosine residue in the ATP-binding domain, a mechanism conserved in organisms rangingfrom fission yeast to humans. However, unlike fission yeast, inA. nidulans this mechanism is not the only means throughwhich entry into mitosis is regulated, since mutants that lackthe inhibitory phosphorylation (Tyr15) (cdk1F, tyrosine 15-to-phenylalanine mutation) still progress through the cell cycle inthe absence of genotoxic stresses. However, this allele doescause sensitivity to DNA-damaging agents and slowed S phasecaused by low levels of HU (60, 61). With the aim of identifyingadditional mitotic regulatory genes that do not function

through Tyr15 phosphorylation of Cdk1, a genetic screen wascarried out looking for mutations that show genetic interac-tions with the cdk1F mutant allele. As a result, a loss-of-function conditional mutation of Anfcp1 was isolated, whichshowed a synthetic temperature sensitivity phenotype in com-bination with cdk1F.

The genetic screen utilized an A. nidulans strain that hasboth a wild-type cdk1 and a cdk1F mutant allele linked intandem with the nutritional marker pyr4. As a hyperactivekinase mutant, the cdk1F mutant allele is dominant over thewild-type counterpart while not causing temperature sensitiv-ity. After UV mutagenesis, mutant strains showing tempera-ture sensitivity were processed to define mutations that dependon the cdk1F mutation for their temperature-sensitive pheno-type. This was done by using 5�-FOA to select against pyr4 andHU to select against cdk1F allowing selection for loop-out ofthe cdk1F allele via homologous recombination between thewild-type cdk1 and mutant cdk1F alleles. In this way alleleswithout cdk1F were selected. Strains obtained in this way weretested for loss of their temperature-sensitive phenotype. Lossof temperature sensitivity as a consequence of cdk1F evictionconfirmed that the newly introduce mutations and the cdk1Fmutation genetically interact in a negative manner, in combi-nation causing increased temperature sensitivity (Fig. 4A).

One of the mutant alleles isolated in this screen wasAnfcp1ts. When the mutant allele was cloned and sequenced, amutation that converted a conserved leucine residue, immedi-ately upstream of the “DXDXT” metal-assisted phosphotrans-ferase signature motif, to a serine residue was identified (Fig.4B). Designated Anfcp1ts, this mutant allele causes a certainlevel of temperature sensitivity by itself, but in combinationwith the cdk1F mutant causes a stronger temperature-sensitivephenotype, underscoring the genetic interaction of Anfcp1 andcdk1 (Fig. 4A).

AnFCP1 is a protein phosphatase which dephosphorylatesRNA polymerase II. Both in S. cerevisiae and in humans, theprotein phosphatase Fcp1 had been shown to dephosphorylatethe C-terminal domain of the largest subunit of RNA polymer-ase II (1, 24). To verify whether this function of AnFCP1 isconserved, its putative target RNAP II was endogenouslytagged C terminally with the affinity S-Tag. This was done sinceantibodies targeting specific serine residues within the Y-S-P-T-S-P-S heptapeptide repeats of the RNAP II CTD were in-effective, presumably because these repeats are only looselyconserved in A. nidulans (51; data not shown). The S-Tagversion of RNAP II is functional since RNAP II is essential(data not shown) and strains carrying the S-Tagged versionwere viable and displayed no growth defects. Using the S-Tagprotein purification system (21), RNAP II was affinity purifiedfrom both wild-type and Anfcp1ts strains incubated at the per-missive (32°C) and restrictive (42°C) temperatures of theAnfcp1ts conditional mutation. RNAP II purified from thewild-type strain separated into subspecies with a size range of�200 to 250 kDa at both 32 and 42°C. Unlike in the wild-typestrain, RNAP II purified at 42°C from the Anfcp1ts strainshowed the loss of subspecies with molecular masses lowerthan �250 kDa, suggesting the loss of hypo- and unphosphor-ylated isoforms of RNAP II CTD (Fig. 4C). An in vitro phos-phatase treatment confirmed that the differences in electro-



http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

phoretic mobility shown by different RNAP II isoforms wereindeed due to differential phosphorylation (data not shown).

Synthetic interaction of Anfcp1ts and cdk1F. As a hyperactiveprotein kinase and an inactivated protein phosphatase pairedtogether led to a synthetic mutant phenotype (enhanced tem-perature sensitivity), we hypothesized that this interaction maybe mediated by an abnormally elevated phosphorylation stateof a common protein substrate shared by cdk1 and AnFCP1. Inaddition to being the substrate of the phosphatase Fcp1,RNAP II CTD had been shown in previous studies to be

phosphorylated by Cdk1 in vitro (14) and undergo mitoticspecific hyperphosphorylation (58). This suggests RNAP IICTD as a candidate common substrate of Fcp1 and Cdk1. Toexamine this possibility, endogenously S-tagged RNAP II wasaffinity purified from cdk1F, Anfcp1ts, and Anfcp1ts cdk1F mu-tants at various temperatures. In the cdk1F mutant strain,RNAP II CTD phosphorylation remained constant compara-ble to wild-type strains at both 32 and 42°C, a finding consistentwith the fact that the cdk1F mutant strain does not exhibit atemperature sensitive phenotype. To compare the phosphory-

FIG. 4. Genetic interaction between the Anfcp1 phosphatase and the activated cdk1F allele of the Cdk1 mitotic kinase. (A) Colony growth at32°C (left plate) and 38.5°C (right) of strains containing the following mutations: 1, fcp1ts cdk1F (SS101); 2, cdk1F (SS105); 3, fcp1ts (SS002); 4,wild type (TNO2A7); 5, fcp1ts cdk1F cdk1; and 6, cdk1F cdk1 (FRY24). (B) Domains of AnFcp1 and the mutation causing the fcp1ts allele, whichchanges the indicated conserved leucine to a serine. An, A. nidulans; Sce, S. cerevisiae; Spo, S. pombe; Hu, human. (C) S-Tagged RNAP II waspurified from a strain with the fcp1TS mutation (SS031) or a wild-type strain (SS025) grown at 32 or 42°C. Purified proteins were separated viaSDS-PAGE, and the gel was stained with Coomassie blue. The region of the gel corresponding to RNAP II is shown.



http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

lation status of Anfcp1ts mutant and Anfcp1ts cdk1F double-mutant strains, cultures were grown at the permissive temper-ature of 32°C and subsequently shifted to temperatures of 37 to40°C. Although this was the range of temperature where thesynthetic effect of the two mutations was most pronounced(i.e., the Anfcp1ts cdk1F double-mutant strain exhibits a moresevere level of temperature sensitivity compared to the singlemutant strains), the difference of the phosphorylation statusesdiffered only marginally. A similar result was obtained whenthe cultures of the two strains were grown entirely at 37°C(data not shown). These data indicate that the synthetic inter-action of Anfcp1 and cdk1 are at best only partially mediatedthrough the phosphorylation status of RNAP II.

Localization of RNAP II through the cell cycle. RNAP II wasendogenously tagged with GFP and its localization examinedthroughout the cell cycle by time-lapse confocal microscopy. Inwild-type cells, RNAP II located to nuclei, with a partial ex-clusion from the nucleolus indicated by a darker area withineach nucleus (Fig. 5, �1.5�). At certain time points, the fluo-rescence intensity within nuclei dissipated evenly into the en-tire cell, remained dispersed for 4 to 5 min, and reconcentratedback into the nuclei that in the mean time had doubled innumbers. The doubling of the number of nuclei and the timeframe of the RNAP II staying dispersed indicated that whileRNAP II localized to the nuclei during interphase, it dispersedthroughout the cell when mitosis started and relocalized todaughter nuclei once mitosis was completed (Fig. 5). Since A.nidulans is known to undergo a partially open mitosis, themitotic specific dispersal of RNAP II indicates that it does notremain associated with nuclear components during mitosis.

To determine whether the genetic interaction of the cdk1Fand Anfcp1ts mutations affected mitosis, endogenously GFP

tagged RNAP II was followed in the wild type and in cdk1F,Anfcp1ts, and Anfcp1ts cdk1F mutants growing at 38°C. Con-cerning the length of germlings, the number and distribution ofthe nuclei and the duration of mitoses (i.e., dispersal of RNAPII::GFP from nuclei, due to NPC partial disassembly) (7), thewild-type and cdk1F strains did not appear to have a notabledifference, a finding consistent with the fact that at 37°C nei-ther of the strains showed visible growth defects. The Anfcp1ts

strain showed a slightly longer duration of RNAP II-GFP dis-persed though no major cytological defects were observed. Themost striking phenotypes were displayed by Anfcp1ts cdk1Fdouble mutants. Many of these cells displayed nuclei that werenear to each other with nuclei appearing to be joined (Fig. 6D).Time lapse imaging revealed (Fig. 7A) that such cells hadnuclei that were difficult to define from each other and whichchanged in shape and conformation through time. In fact,nuclei often appeared to merge into one continuous nucleuscontaining multiple nucleoli (see Fig. 7A at 48�, for example).These surprising changes can be seen clearly in a full-timecourse video file (see Video 1S in the supplemental material).When RNAP II::GFP was seen to disperse from such appar-ently joined nuclei, indicating entry into mitosis, the number ofnuclei was not always increased when RNAP II::GFP returnedto nuclei after completion of mitosis (Fig. 7B). Additionally,we observed some nuclei in the double mutant that appearedto complete mitosis, generating two daughter nuclei capable ofimporting RNAP II-GFP but which then remerged in an ap-parent reversal of mitosis (Fig. 7C). At time zero� (Fig. 7C),the cell contains two nuclei and a micronucleus. At 8� the cellenters mitosis and RNAP II-GFP starts to disperse from nucleito throughout the cell. At 22� the cell exits mitosis and daugh-ter nuclei reimport RNAP II-GFP, and at 24� four nuclei areapparent. However, as the cell progresses further into the cellcycle these four nuclei merge into two nuclei, each with a singlenucleolus. The micronucleus can also be seen. Thus, althoughthis cell undergoes a mitotic event the net result is regenera-

FIG. 5. RNA polymerase II changes location during mitosis.Endogenously GFP-tagged RNAP II (strain SS185) was imaged byusing time-lapse confocal microscopy as cells passed through mito-sis. RNAP II-GFP is predominantly nuclear during interphase,while it disperses into the cytoplasm during mitosis and is reim-ported to daughter nuclei during exit from mitosis and entry intoG1. Bar, �5 �m.

FIG. 6. The synthetic effects of cdk1F and Anfcp1TS are manifestedas severe mitotic defects. Representative fields of cells each expressingendogenously tagged RNAP II at the beginning of time lapse confocalmicroscopy. (A) Wild type (SS185); (B) cdk1F (SS177); (C) Anfcp1TS

(SS174); (D) cdk1F Anfcp1TS (SS190). Cells were grown and imaged at38°C. Bar, �5 �m.



http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

tion of the same nuclear configuration (compare the 0� to 40�images in Fig. 7C). Finally, cells were observed in which theRNAP II::GFP signal remained dispersed after entering mito-sis for an extended period of time, presumably due to theirinability to exit mitoses.

In nuclei where the RNAP II::GFP signals dispersed duringmitosis and reconcentrated to the nuclei upon its completion,it took significantly longer for this cycle to be completed (com-pare the timing of RNAP II-GFP dispersal in Fig. 5 to 7 forexample). To rule out the possibility that this may merely be

due to an overall slowed-down cell cycle, the frequency ofdispersed RNAP II::GFP signals in all four strains was mea-sured in fixed cells grown at 37°C (Table 2). The increasedpercentage of cells with dispersed RNAP II::GFP (Table 2)support the findings of longer mitoses in the Anfcp1ts andespecially the Anfcp1ts cdk1F mutants. The severity of thecytological defects observed in the Anfcp1ts cdk1F double-mutant cells might help explain the basis of the synthetic na-ture of the genetic interaction between the Anfcp1ts and thecdk1F mutations. While the Anfcp1ts and cdk1F single mutants

FIG. 7. Nuclear defects defined by location of RNAP II-GFP during live cell confocal microscopy in a fcp1ts cdk1F double mutant (strainSS190). (A) Cell containing an ill-defined interphase nuclear cluster that fails to enter mitosis over a 102-min period. At times, individual nucleiseem apparent, but these then merge to form one large nucleus containing multiple nucleoli, revealed as darker nuclear shadows. A video file (seeVideo S1 in the supplemental material) displays the full time course collected at 1.5-min intervals over 102 min. The playback rate is 2 frames/s(fps). (B) Cell containing a single nucleus that undergoes mitosis, releasing RNAP II-GFP throughout the cell. Upon exit from mitosis and nuclearreimport of RNAP II-GFP, no clear daughter nuclei are generated during this abortive mitosis. A video file (see Video S2 in the supplementalmaterial) displays the time course collected at 1.5-min intervals over 42 min. The playback rate is 2 fps. (C) Cell containing two nuclei and amicronucleus, presumably as a result of a previous defective mitosis. During mitosis, RNAP II-GFP is released from nuclei and then initiallyreimported to four apparent nuclei upon mitotic exit. However, the paired nuclei from each of the two parental nuclei then reverse mitosis andbecome two individual nuclei again, each containing a single nucleolus. A video file (see Video S3 in the supplemental material) displays the fulltime course collected at 2-min intervals over 56 min. The playback rate is 2 fps. Bar, �5 �m.



http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

showed moderate or no notable defects compared to wild-typecells, the Anfcp1ts cdk1F double mutant exhibited phenotypesthat went far beyond the simple addition of that of the twosingle mutant strains.

DISCUSSION

Essential protein phosphatases and cell cycle regulation.This study showed the phosphatase genes bimG, nimT, Anfcp1,and pphA to encode proteins with essential cellular functions.The fact that two of these genes were identified previously ingenetic screens for cell cycle mutants (34) and that overexpres-sion of a dominant-negative form of PphA affects growth andmitosis (26) and the identification of AnFCP1 as a potentialmitotic regulator in the screen described in this study clearlydemonstrate the essential importance of phosphatase-medi-ated protein dephosphorylation in regulating cell cycle pro-gression. The terminal phenotypes displayed by cells withbimG and nimT deleted confirmed their functions to be criticalin mitotic regulation, which is consistent with the findings fromstudies of the conditional mutant alleles bimG11 and nimT23.While Anfcp1 and pphA deletions did not yield phenotypes asclearly demonstrative of mitotic defects as bimG and nimTdeletions, their defective nuclear DNA morphologies suggestedthat there are potentially important functions of these genes inmitotic regulation as well. The availability of the conditionalAnfcp1ts mutant allele, as well as strains in which Anfcp1 andpphA are under the inducible promoter alcA, should prove to beuseful in further studies of the functions of these essential genes.

One noticeable essential phosphatase, the Cdc14 phos-phatase, that is required for the regulation of S. cerevisiaemitosis is not essential in A. nidulans. In fact, the null Ancdc14strain grows and develops normally. This suggests that Cdc14phosphatases might not be essential components of mitoticregulation in all organisms, and recent deletion studies indicateCdc14B is dispensable for human mitotic regulation (5).

Gene deletion phenotypes of nonessential protein phos-phatases. The gene deletion of four protein phosphatases didnot cause lethality but had physiological effects detrimentalenough to cause strong growth defects. Of these four phos-phatase genes, AnSSU72 is functionally linked to the essentialphosphatase AnFCP1 since the physiological target of bothphosphatases had been shown to be the C-terminal domain ofRNAP II (27). This may reflect the importance of maintaininga proper phosphorylation status of RNAP II CTD for the cell

to engage in proper transcriptional activity, by itself an activityabsolutely vital for cellular survival. Unlike AnSSU72, thephosphatase AnNEM1 is linked to AnFCP1 by belonging tothe same protein phosphatase family that shares the CPD typecatalytic domain. Defined first by the discovery of Fcp1, theCDP family is made up of protein phosphatases that engage inaspartate-based catalysis of target proteins and have in com-mon the signature catalytic motif of DXDXT/V. While onlythree different CPD phosphatases are present in A. nidulans,one of them is essential (Anfcp1) and one causes severe growthdefects upon deletion (Annem1), indicating the importance ofthe biological function conserved in this phosphatase family.Within the context of cell cycle regulation, Annem1 is of addedinterest since its deletion not only caused an overall decreaseof cellular growth but was also marked by noticeable nuclearstructure phenotypes, a strong indication that it may play animportant role in mitotic regulation and/or nuclear structure.This finding is consistent with the functions of Nem1p (S.cerevisiae) and Dullard (humans), which are known to regulatenuclear membrane biogenesis (22, 47, 49).

Genetic interaction of Anfcp1 and cdk1 and RNAP II CTDphosphorylation. The identification of Anfcp1ts to cause in-creased temperature sensitivity in combination with cdk1F re-vealed a genetic interaction between Anfcp1 and cdk1. The factthat a hyperactive kinase (cdk1F), together with an inactivatedphosphatase (Anfcp1ts), caused a synthetic mutant phenotypesuggests the effects of those two mutations may converge intoa hyperphosphorylated state of a common protein substrate.While potential targets of cdk1 phosphorylation are numerous,the only substrate of protein phosphatase Fcp1 convincinglyshown thus far is the RNAP II CTD. In contrast to our expec-tations, the overall level of RNAP II CTD phosphorylation didnot show a significant elevation when endogenously S-taggedRNAP II was purified and analyzed by SDS-PAGE underconditions where the synthetic temperature sensitivity of theAnfcp1ts cdk1F double mutant is most pronounced. Althoughthis result shows that the bulk amount of RNAP II CTDphosphorylation does not proportionately mediate the syn-thetic temperature sensitivity of the Anfcp1ts cdk1F doublemutant, it should be noticed that the A. nidulans RNAP IICTD contains 44 serine-proline dipeptide sequences, all po-tential targets of phosphorylation. It is possible that the syn-thetic phenotype may be attributed to the phosphorylation ofspecific RNAP II residues which are not readily detectable bymobility shifts during SDS-PAGE. Notably, we were unable to

TABLE 2. Frequency of RNAP II::GFP dispersal in the wild-type, cdk1F, Anfcp1TS, and Anfcp1TS cdk1F strains at 37°Ca

Expt

No. of cells (%)

Wild type cdk1F Anfcp1ts Anfcp1ts cdk1F

Dispersed Total Dispersed Total Dispersed Total Dispersed Total

1 8 234 9 220 43 726 76 7972 19 515 22 530 46 571 46 5433 31 812 31 826 30 851 81 822

Total 58 (3.7) 1561 62 (3.9) 1,576 119 (5.5) 2,148 203 (9.4) 2,162

a Cells of individual strains were grown for 8 h at 37°C and fixed. The cells in which the RNAP II::GFP signal was dispersed were counted, as well as the total numberof cells. The experiments were repeated three times independently. Student t test analysis indicated that only the difference between the wild-type and the Anfcp1ts

cdk1F strains was significant (P 0.0005).



http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

detect an obviously hyperphosphorylated form of RNAP IIduring mitosis, whereas such a form has been shown to existduring higher eukaryotic mitosis (58).

Localization of RNAP II during A. nidulans mitosis. Sincemitosis in A. nidulans is partially open (7, 39), nuclear proteinsthat are not attached to the nuclear interior during mitosis, onchromosomes, for example, can disperse into the cytoplasm viapassive diffusion once nuclear pores open at the beginning ofeach mitotic phase. In contrast, proteins such as histones,which remain associated with chromatin during mitosis, are notreleased from nuclei during mitosis. The predominant nuclearlocalization of endogenously GFP-tagged RNAP II in growingcells confirmed that it remains in the nucleus during interphasesince it is engaged in transcriptional activities. The completedispersal of RNAP II during mitosis suggests that its associa-tion with DNA may be actively disrupted during mitosis, ef-fectively disabling it from continuing transcription. During thetranscription cycle, it had been shown that the CTD of RNAPII has to be completely dephosphorylated for the RNAP II toform a preinitiation complex and reengage in transcriptionalactivity (25, 48). In mammalian cells the RNAP II CTD ishyperphosphorylated during mitosis, and this is thought to beregulated by Cdk1-mediated phosphorylation (58). Althoughin A. nidulans this mitosis-specific CTD hyperphosphorylationevent could not be observed, the fact that RNAP II completelydisperses during mitosis suggests that there still may be amitotic specific phosphorylation pattern of RNAP II CTD in A.nidulans capable of actively dissociating RNAP II from chro-matin and thus contributing to its mitosis-specific dispersal.The work described here has demonstrated that cdk1F shows astrong genetic interaction with the RNAP II CTD phosphatasegene Anfcp1. Taken together with the fact that Cdk1 is a majormitotic protein kinase which in other systems is capable ofphosphorylating RNAP II CTD in vitro, the Cdk1 kinase couldbe considered a compelling candidate to promote such a mi-tosis-specific phosphorylation event of RNAP II CTD. Intrigu-ingly, the mitotic hyperphosphorylation of RNAP II CTD inhigher eukaryotes is dependent on the Pin1 peptidyl-prolylisomerase (58). Pin1 has been shown to both inhibit Fcp1 andactivate Cdk1, thus promoting mitotic hyperphosphorylation ofRNAP II CTD. It will thus be interesting to investigate the po-tential role of Pin1 (PinA in A. nidulans [20]) in mediating thegenetic interactions and mitotic phenotypes (see below) we haveuncovered while studying Fcp1 and Cdk1. One prediction is thatincreased expression of PinA might phenocopy the fcp1ts cdk1Fdouble mutant (by inhibiting Fcp1 and activating Cdk1).

Mitotic defects in the Anfcp1ts cdk1F double mutant. Whenthe localization pattern of RNAP II was examined in theAnfcp1ts, cdk1F, and Anfcp1ts cdk1F strains at 38°C and com-pared to wild-type cells, the Anfcp1ts cdk1F double mutantsrevealed by far the most severe defects. The various pleiotropicdefects observed in the Anfcp1ts cdk1F double mutantsstrongly indicates that the synthetic interaction between theAnfcp1ts and cdk1F mutation may impact cells in ways thatmight go beyond the immediate disruption of the RNAP II-driven transcription cycle. Although the RNAP II CTD is thusfar the only known substrate of FCP1, it cannot be ruled outthat FCP1 and Cdk1 may share other common substrates.Since the effects of mutations of Anfcp1 and cdk1 would be-come amplified in a pleiotropic manner if they were to share

many substrates, this could explain why the difference ofRNAP II CTD phosphorylation between cdk1F, Anfcp1ts, andAnfcp1ts cdk1F mutants is only moderate compared to theseverity of the cytological defects observed in the double mu-tant. Alternatively, the effects of the double mutant on thefidelity of transcription could potentially preferentially affectgene expression specific to cell cycle transitions, thus causingthe mitotic defects observed. It is also possible Cdk1 and FCP1do not directly share a common substrate but function in dif-ferent cellular pathways. The interaction between Anfcp1ts andcdk1F is genetic in nature, and Anfcp1ts and cdk1F could stillshow their synthetically interacting effects if the separate bio-logical pathways in which they function merge onto commoncellular targets that provide important activities concerningmitotic regulation. In fact, reports suggesting the involvementof Fcp1 in DNA damage responses (19) indicate that Fcp1 mayfunctionally relate to Cdk1 in a manner more directly linked tomitotic regulation. However, we could not detect obviousDNA damage response defects in the fcp1ts allele.

Overall, the severe mutant phenotypes observed in the dou-ble mutant compared to that exhibited by the Anfcp1ts andcdk1F single mutants confirms the relevance of the geneticinteraction between Anfcp1ts and cdk1F and provides impor-tant evidence that the conserved Fcp1 phosphatase might playa role in mitotic regulation by counteracting Cdk1-mediatedphosphorylation.

ACKNOWLEDGMENTS

We thank Berl Oakley and members of his lab, especially TanyaNayak, for generously providing constructs used in generating theriboB::alcA insertion cassette, the TNO2A7 strain, and helpful discus-sions. We also thank all members of the Osmani lab for help andsupport, especially Hui-Lin Liu for having made available the S-Tagaffinity purification system extensively used in this study and AyshaOsmani for her expertise and steady moral support throughout theentire process of this work.

This study was supported by a grant from the National Institutes ofHealth (GM042564) to S.A.O.

REFERENCES

1. Archambault, J., R. S. Chambers, M. S. Kobor, Y. Ho, M. Cartier, D.Bolotin, B. Andrews, C. M. Kane, and J. Greenblatt. 1997. An essentialcomponent of a C-terminal domain phosphatase that interacts with tran-scription factor IIF in Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA94:14300–14305.

2. Axton, J. M., V. Dombradi, P. T. Cohen, and D. M. Glover. 1990. One of theprotein phosphatase 1 isoenzymes in Drosophila is essential for mitosis. Cell63:33–46.

3. Bailey, C., and H. N. Arst, Jr. 1975. Carbon catabolite repression in As-pergillos nidulans. Eur. J. Biochem. 51:573–577.

4. Beals, C. R., N. A. Clipstone, S. N. Ho, and G. R. Crabtree. 1997. Nuclearlocalization of NF-ATc by a calcineurin-dependent, cyclosporin-sensitiveintramolecular interaction. Genes Dev. 11:824–834.

5. Berdougo, E., M. V. Nachury, P. K. Jackson, and P. V. Jallepalli. 2008. Thenucleolar phosphatase Cdc14B is dispensable for chromosome segregationand mitotic exit in human cells. Cell Cycle 7:1184–1190.

6. Booher, R., and D. Beach. 1989. Involvement of a type 1 protein phosphataseencoded by bws1� in fission yeast mitotic control. Cell 57:1009–1016.

7. De Souza, C. P., A. H. Osmani, S. B. Hashmi, and S. A. Osmani. 2004. Partialnuclear pore complex disassembly during closed mitosis in Aspergillus nidu-lans. Curr. Biol. 14:1973–1984.

8. Doonan, J. H., and N. R. Morris. 1989. The bimG gene of Aspergillusnidulans, required for completion of anaphase, encodes a homolog of mam-malian phosphoprotein phosphatase 1. Cell 57:987–996.

9. Fantes, P. 1979. Epistatic gene interactions in the control of division infission yeast. Nature 279:428–430.

10. Fillinger, S., C. Panozzo, M. Mathieu, and B. Felenbok. 1995. The basal levelof transcription of the alc genes in the ethanol regulon in Aspergillus nidulansis controlled both by the specific transactivator AlcR and the general carboncatabolite repressor CreA. FEBS Lett. 368:547–550.



http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

11. Fox, H., P. C. Hickey, J. M. Fernandez-Abalos, P. Lunness, N. D. Read, andJ. H. Doonan. 2002. Dynamic distribution of BIMG(PP1) in living hyphae ofAspergillus indicates a novel role in septum formation. Mol. Microbiol. 45:1219–1230.

12. Ganem, C., F. Devaux, C. Torchet, C. Jacq, S. Quevillon-Cheruel, G.Labesse, C. Facca, and G. Faye. 2003. Ssu72 is a phosphatase essential fortranscription termination of snoRNAs and specific mRNAs in yeast. EMBOJ. 22:1588–1598.

13. Gautier, J., M. J. Solomon, R. N. Booher, J. F. Bazan, and M. W. Kirschner.1991. cdc25 is a specific tyrosine phosphatase that directly activates p34cdc2.Cell 67:197–211.

14. Gebara, M. M., M. H. Sayre, and J. L. Corden. 1997. Phosphorylation of thecarboxy-terminal repeat domain in RNA polymerase II by cyclin-dependentkinases is sufficient to inhibit transcription. J. Cell Biochem. 64:390–402.

15. Gingras, A. C., M. Caballero, M. Zarske, A. Sanchez, T. R. Hazbun, S.Fields, N. Sonenberg, E. Hafen, B. Raught, and R. Aebersold. 2005. A novel,evolutionarily conserved protein phosphatase complex involved in cisplatinsensitivity. Mol. Cell Proteomics 4:1725–1740.

16. Glover, D. M., M. H. Leibowitz, D. A. McLean, and H. Parry. 1995. Muta-tions in aurora prevent centrosome separation leading to the formation ofmonopolar spindles. Cell 81:95–105.

17. Goberdhan, D. C., N. Paricio, E. C. Goodman, M. Mlodzik, and C. Wilson.1999. Drosophila tumor suppressor PTEN controls cell size and number byantagonizing the Chico/PI3-kinase signaling pathway. Genes Dev. 13:3244–3258.

18. Hoffmann, I., P. R. Clarke, J. J. Marcote, E. Karsenti, and G. Draetta. 1993.Phosphorylation and activation of human cdc25-c by cdc2-cyclin B and itsinvolvement in the self-amplification of MPF at mitosis. EMBO 12:53–63.

19. Jeong, S. J., H. J. Kim, Y. J. Yang, J. H. Seol, B. Y. Jung, J. W. Han, H. W. Lee,and E. J. Cho. 2005. Role of RNA polymerase II carboxy-terminal domainphosphorylation in DNA damage response. J. Microbiol. 43:516–522.

20. Joseph, J. D., S. N. Daigle, and A. R. Means. 2004. PINA is essential forgrowth and positively influences NIMA function in Aspergillus nidulans.J. Biol. Chem. 279:32373–32384.

21. Kim, J. S., and R. T. Raines. 1993. Ribonuclease S-peptide as a carrier infusion proteins. Protein Sci. 2:348–356.

22. Kim, Y., M. S. Gentry, T. E. Harris, S. E. Wiley, J. C. Lawrence, Jr., and J. E.Dixon. 2007. A conserved phosphatase cascade that regulates nuclear mem-brane biogenesis. Proc. Natl. Acad. Sci. USA 104:6596–6601.

23. Kinoshita, N., H. Ohkura, and M. Yanagida. 1990. Distinct, essential roles oftype 1 and 2A protein phosphatases in the control of the fission yeast celldivision cycle. Cell 63:405–415.

24. Kobor, M. S., J. Archambault, W. Lester, F. C. Holstege, O. Gileadi, D. B.Jansma, E. G. Jennings, F. Kouyoumdjian, A. R. Davidson, R. A. Young, andJ. Greenblatt. 1999. An unusual eukaryotic protein phosphatase required fortranscription by RNA polymerase II and CTD dephosphorylation in Sac-charomyces cerevisiae. Mol. Cell 4:55–62.

25. Komarnitsky, P., E. J. Cho, and S. Buratowski. 2000. Different phosphory-lated forms of RNA polymerase II and associated mRNA processing factorsduring transcription. Genes Dev. 14:2452–2460.

26. Kosmidou, E., P. Lunness, and J. H. Doonan. 2001. A type 2A proteinphosphatase gene from Aspergillus nidulans is involved in hyphal morpho-genesis. Curr. Genet. 39:25–34.

27. Krishnamurthy, S., X. He, M. Reyes-Reyes, C. Moore, and M. Hampsey. 2004.Ssu72 Is an RNA polymerase II CTD phosphatase. Mol. Cell 14:387–394.

28. Kulmburg, P., M. Mathieu, C. Dowzer, J. Kelly, and B. Felenbok. 1993.Specific binding sites in the alcR and alcA promoters of the ethanol regulonfor the CREA repressor mediating carbon catabolite repression in Aspergil-lus nidulans. Mol. Microbiol. 7:847–857.

29. Kulmburg, P., D. Sequeval, F. Lenouvel, M. Mathieu, and B. Felenbok. 1992.Identification of the promoter region involved in the autoregulation of thetranscriptional activator ALCR in Aspergillus nidulans. Mol. Cell. Biol. 12:1932–1939.

29a.Liu, H. L., C. P. De Souza, A. H. Osmani, and S. A. Osmani. 2009. The threefungal transmembrane nuclear pore complex proteins of Aspergillus nidulansare dispensable in the presence of an intact An-Nup84-120 complex. Mol.Biol. Cell 20:616–630.

30. Llamazares, A., A. Moreira, A. Tavares, C. Girdam, B. A. Spruce, C. Gonza-lez, R. E. Karess, D. M. Glover, and C. E. Sunkel. 1991. polo encodes aprotein kinase homolog required for mitosis in Drosophila. Genes Dev.5:2153–2165.

31. Meinhart, A., T. Kamenski, S. Hoeppner, S. Baumli, and P. Cramer. 2005.A structural perspective of CTD function. Genes Dev. 19:1401–1415.

32. Moller, S. G., Y. S. Kim, T. Kunkel, and N. H. Chua. 2003. PP7 is a positiveregulator of blue light signaling in Arabidopsis. Plant Cell 15:1111–1119.

33. Moorhead, G. B., L. Trinkle-Mulcahy, and A. Ulke-Lemee. 2007. Emergingroles of nuclear protein phosphatases. Nat. Rev. Mol. Cell. Biol. 8:234–244.

34. Morris, N. R. 1976. Mitotic mutants of Aspergillus nidulans. Genet. Res.26:237–254.

35. Nayak, T., E. Szewczyk, C. E. Oakley, A. Osmani, L. Ukil, S. L. Murray, M. J.Hynes, S. A. Osmani, and B. R. Oakley. 2006. A versatile and efficient genetargeting system for Aspergillus nidulans. Genetics 172:1557–1566.

36. O’Connell, M. J., A. H. Osmani, N. R. Morris, and S. A. Osmani. 1992. Anextra copy of nimEcyclinB elevates pre-MPF levels and partially suppressesmutation of nimTcdc25 in Aspergillus nidulans. EMBO J. 11:2139–2149.

37. Ohkura, H., N. Kinoshita, S. Miyatani, T. Toda, and M. Yanagida. 1989. Thefission yeast dis2� gene required for chromosome disjoining encodes one oftwo putative type 1 protein phosphatases. Cell 57:997–1007.

38. Osherov, N., J. Mathew, and G. S. May. 2000. Polarity-defective mutants ofAspergillus nidulans. Fungal. Genet. Biol. 31:181–188.

39. Osmani, A. H., J. Davies, H. L. Liu, and S. A. Osmani. 2006. Systematicdeletion and mitotic localization of the nuclear pore complex proteins ofAspergillus nidulans. Mol. Biol. Cell 17:4946–4961.

40. Osmani, A. H., S. L. McGuire, and S. A. Osmani. 1991. Parallel activation ofthe NIMA and p34cdc2 cell cycle-regulated protein kinases is required toinitiate mitosis in Aspergillus nidulans. Cell 67:283–291.

41. Osmani, A. H., B. R. Oakley, and S. A. Osmani. 2006. Identification andanalysis of essential Aspergillus nidulans genes using the heterokaryon rescuetechnique. Nat. Protoc. 1:2517–2526.

42. Osmani, A. H., N. van Peij, M. Mischke, M. J. O’Connell, and S. A. Osmani.1994. A single p34cdc2 protein kinase (nimXcdc2) is required at G1 and G2 inAspergillus nidulans. J. Cell Sci. 107:1519–1528.

43. Pateman, J. A., C. H. Doy, J. E. Olsen, U. Norris, E. H. Creaser, and M.Hynes. 1983. Regulation of alcohol dehydrogenase (ADH) and aldehydedehydrogenase (AldDH) in Aspergillus nidulans. Proc. R. Soc. Lond. B Biol.Sci. 217:243–264.

44. Pontecorvo, G., J. A. Roper, L. M. Hemmons, K. D. MacDonald, and A. W.Bufton. 1953. The genetics of Aspergillus nidulans, p. 141–238. In M. Deme-rec (ed.), Advances in genetics, vol. 5. Academic Press, Inc., New York, NY.

45. Rasmussen, C., C. Garen, S. Brining, R. L. Kincaid, R. L. Means, and A. R.Means. 1994. The calmodulin-dependent protein phosphatase catalytic sub-unit (calcineurin A) is an essential gene in Aspergillus nidulans. EMBO J.13:3917–3924.

46. Russell, P., and P. Nurse. 1986. cdc25 functions as an inducer in the mitoticcontrol of fission yeast. Cell 45:145–153.

47. Santos-Rosa, H., J. Leung, N. Grimsey, S. Peak-Chew, and S. Siniossoglou.2005. The yeast lipin Smp2 couples phospholipid biosynthesis to nuclearmembrane growth. EMBO J. 24:1931–1941.

48. Schroeder, S. C., B. Schwer, S. Shuman, and D. Bentley. 2000. Dynamicassociation of capping enzymes with transcribing RNA polymerase II. GenesDev. 14:2435–2440.

49. Siniossoglou, S., H. Santos-Rosa, J. Rappsilber, M. Mann, and E. Hurt.1998. A novel complex of membrane proteins required for formation of aspherical nucleus. EMBO J. 17:6449–6464.

50. Soriani, F. M., I. Malavazi, M. E. Silva Ferreira, M. Savoldi, M. R. ZeskaKress, M. H. Souza Goldman, O. Loss, E. Bignell, and G. H. Goldman. 2008.Functional characterization of the Aspergillus fumigatus CRZ1 homologue,CrzA. Mol. Microbiol. 67:1274–1291.

51. Stiller, J. W., and M. S. Cook. 2004. Functional unit of the RNA polymerase IIC-terminal domain lies within heptapeptide pairs. Eukaryot. Cell 3:735–740.

52. Tang, Z., H. Shu, W. Qi, N. A. Mahmood, M. C. Mumby, and H. Yu. 2006.PP2A is required for centromeric localization of SgoI and proper chromo-some segregation. Dev. Cell 10:575–585.

53. Torres-Rosell, J., F. Machin, A. Jarmuz, and L. Aragon. 2004. Nucleolarsegregation lags behind the rest of the genome and requires Cdc14p activa-tion by the FEAR network. Cell Cycle 3:496–502.

54. Toyoshima-Morimoto, F., E. Taniguchi, and E. Nishida. 2002. Plk1 pro-motes nuclear translocation of human Cdc25C during prophase. EMBORep. 3:341–348.

55. Trinkle-Mulcahy, L., P. D. Andrews, S. Wickramasinghe, J. Sleeman, A.Prescott, Y. W. Lam, C. Lyon, J. R. Swedlow, and A. I. Lamond. 2003.Time-lapse imaging reveals dynamic relocalization of PP1gamma throughoutthe mammalian cell cycle. Mol. Biol. Cell 14:107–117.

56. Visintin, R., K. Craig, E. S. Hwang, S. Prinz, M. Tyers, and A. Amon. 1998.The phosphatase Cdc14 triggers mitotic exit by reversal of Cdk-dependentphosphorylation. Mol. Cell 2:709–718.

57. Waring, R. B., G. S. May, and N. R. Morris. 1989. Characterization of aninducible expression system in Aspergillus nidulans using alcA and tubulin-coding genes. Gene 79:119–130.

58. Xu, Y. X., Y. Hirose, X. Z. Zhou, K. P. Lu, and J. L. Manley. 2003. Pin1modulates the structure and function of human RNA polymerase II. GenesDev. 17:2765–2776.

59. Yang, L., L. Ukil, A. Osmani, F. Nahm, J. Davies, C. P. De Souza, X. Dou,A. Perez-Balaguer, and S. A. Osmani. 2004. Rapid production of genereplacement constructs and generation of a green fluorescent protein-taggedcentromeric marker in Aspergillus nidulans. Eukaryot. Cell 3:1359–1362.

60. Ye, X. S., R. R. Fincher, A. Tang, K. O’Donnell, and S. A. Osmani. 1996. TwoS-phase checkpoint systems, one involving the function of both BIME andTyr15 phosphorylation of p34cdc2, inhibit NIMA and prevent prematuremitosis. EMBO J. 15:3599–3610.

61. Ye, X. S., R. R. Fincher, A. Tang, and S. A. Osmani. 1997. The G2/M DNAdamage checkpoint inhibits mitosis through Tyr15 phosphorylation ofp34cdc2 in Aspergillus nidulans. EMBO J. 15:101–112.



http://ec.asm.org/

Dow

nloaded from

http://ec.asm.org/

analysis of all protein phosphatase genes in aspergillus ...gene deletion analysis of all protein...

Documents