Identification and Characterization of Two Ca 2 1 /CaM-dependent Protein Kinases Required for Normal Nuclear Division in Aspergillus nidulans *

We utilized an expression screen to identify two novel Ca 2 1 /calmodulin (CaM)-regulated protein kinases in Aspergillus nidulans . The two kinases, CMKB and CMKC, possess high sequence identity with mammalian CaM kinases (CaMKs) I/IV and CaMKK a / b , respectively. In vitro CMKC phosphorylates and increases the activity of CMKB, indicating they are biochemical homologues of CaMKK a / b and CaMKI/IV. The disruption of CMKB is lethal; however, when protein expression is postponed, the spores germinate with delayed kinetics. The observed lag corresponds to a delay in the G 1 -phase acti- vation of the cyclin-dependent kinase NIMX cdc2 . Disruption of cmk C is not lethal, but spores lacking CMKC also germinate with delayed kinetics and a lag in the activation of NIMX cdc2 . Analysis of D cmk C suggests a role for CMKC in regulating the first and subsequent nuclear division cycles. We conclude that both CMKB and CMKC are required for the proper temporal activation of NIMX cdc2 as spores enter the cell cycle from quiescence and suggest that this relationship exists during the G 1 /S transition of subsequent cell divisions. The Similar during and Aspergillus polymerase chain reaction, and the product was ligated into the corresponding sites of pGEM-7Zf( 2 ) (Promega). The cDNA was subse-quently subcloned as a Sac I/ Xho I fragment into pET30b (Novagen), thus creating pETCMKC. The sequence of both expression vectors was confirmed by sequencing. BLR(DE3)pLysS (Novagen) bacteria were transformed with the expression plasmids, and several colonies were used to inoculate 100 ml of LB and grown overnight at 30 °C. 20 ml of the overnight culture was used to seed 1 liter of LB medium, and the culture was grown at 37 °C to an A 600 nm of 0.6; protein expression was then induced by the addition of 1 m M isopropyl B- D -thiogalactopyrano-side for 2 h. After induction, the culture was pelleted by centrifugation and resuspended in 30 ml of lysis buffer (40 m M Tris-HCl, pH 7.5, 100 m M KCl, 100 m g/ml Pefabloc, 10 m g/ml leupeptin, and 10 m g/ml aproti- nin). The bacterial pellet was lysed by mild sonication and clarified by centrifugation at 16000 3 g for 30 min. The supernatant was incubated with 0.5 ml of nickel nitrilotriacetic acid-agarose resin (Qiagen) with rocking for 1 h at 4 °C. Thebound resin was loaded onto a column and washed ml lysis buffer followed 10 ml lysis buffer supplemented m M imidazol. bound lysis m M imidazol. purified enzymes were stored glycerol All point mutations of CMKB were mega-primer mutagenesis technique and confirmed by sequencing Glutathione S -transferase-CaMKI and maltose-bind- ing protein-CaMKK b as described Ssp sequence homology, a member of a growing family of CaMKKs.

expression is repressed, the strain grows dramatically slower than control strains, and CMKA mRNA and protein are undetectable by Northern and Western analysis, respectively. However, CaM affinity chromatography yielded CaMK activity that represented about 10 percent that present in the wild-type strain. These results suggested that either CMKA repression was not absolute or that A. nidulans contains an unidentified additional CaMK.
Pharmacological evidence exists that CaMKs may function not only in G 2 but also in the G 1 phase of the cell cycle of metazoans. Specifically, the selective CaMK inhibitor, KN-93, has been shown to arrest cells in G 1 . Although these pharmacological studies have been interpreted as indicating CaMKII activity is required for G 1 progression, KN-93 inhibits all three multifunctional CaMKs, CaMKI, II, and IV (6), and therefore inhibition of any of these enzymes could be responsible for the G 1 arrest. Additionally, the nature of the KN-93-induced G 1 arrest depends upon the cell line and the experimental protocol. The drug arrests HeLa cells in G 1 with high histone H1 kinase activity, indicating that the arrest is at the G 1 /S boundary (7). However, NIH-3T3 cells arrest earlier in G 1 with low levels of cyclinD1 and an enhanced association of p27 kip1 with Cdk2 (8,9). It is unknown whether the conflicting results are due to cell line differences or represent two independent points at which a CaMK activity is required. Therefore, the putative CaMKs and their role(s) in the G 1 phase of the cell cycle have yet to be defined.
Since many Ca 2ϩ /CaM-signaling pathways appear to be functionally conserved between A. nidulans and mammalian cells, we have used the filamentous fungus to further our understanding of the essential roles of CaMKs. In the current study we present the characterization of two novel Ca 2ϩ /CaMdependent protein kinases in A. nidulans, CMKB and CMKC. In vitro, the two fungal kinases, CMKB and CMKC, function analogously to the mammalian CaMK cascade members: CaMKI/IV and CaMKK␣/␤. Disruption of cmkB is lethal, indicating CMKB is required for A. nidulans growth. Both retarded expression of CMKB and disruption of cmkC severely delay the nuclear division cycle before the initiation of NIMX cdc2 kinase activity. We propose that both CMKB and CMKC are required for the proper temporal activation of NIMX cdc2 upon re-entry into the cell cycle from a quiescent state and during the G 1 /S transition of subsequent nuclear division cycles.
Expression Cloning of CMKB and CMKC-CMKB and CMKC cDNAs were isolated by screening an A. nidulans gt11 cDNA expression library (a gift of Dr. Greg May) with a protein A-Aspergillus CaM fusion protein, as described by Stirling et al. (11). The protein A-CaM fusion protein was generated by subcloning a hemagglutinin-tagged CaM as a blunted AflIII/BamHI fragment into the blunted EcoRI and BamHI sites of pALP1 (a gift of Dr. Michael Stark), thereby creating pALP-ACaM. The protein A-CaM fusion protein was purified from cultures grown in 1 liter of Luria-Bertoni (LB) to an A 600 nm of 1.0 by phenyl-Sepharose chromatography, following the protocol for purification of bacterially expressed A. nidulans CaM (12). Approximately 4 ϫ 10 5 plaques were screened. Of the approximately 150 positive plaques identified, 40 were randomly chosen for purification. As a secondary screen, these 40 plaques were hybridized with random-primed cDNA probes of A. nidulans CMKA and CnA. Clones negative for CMKA and CnA were plaque-purified, and the plasmids were excised as described by the manufacturer (Stratagene). The inserts were then sequenced by the dideoxynucleotide termination reaction (Amersham Pharmacia Biotech).
Cloning and Expression of CMKB and CMKC-Both protein kinases were expressed as hexa-histidine-tagged fusion proteins. BamHI and EcoRI sites were added to the CMKB cDNA by polymerase chain reaction, and the resulting product was subcloned into the corresponding sites of pTrcHisB (Invitrogen), generating pTrcCMKB. SacI and KpnI restriction sites were generated on the 5Ј and 3Ј end of the CMKC cDNA by polymerase chain reaction, and the product was ligated into the corresponding sites of pGEM-7Zf(Ϫ) (Promega). The cDNA was subsequently subcloned as a SacI/XhoI fragment into pET30b (Novagen), thus creating pETCMKC. The sequence of both expression vectors was confirmed by sequencing. BLR(DE3)pLysS (Novagen) bacteria were transformed with the expression plasmids, and several colonies were used to inoculate 100 ml of LB and grown overnight at 30°C. 20 ml of the overnight culture was used to seed 1 liter of LB medium, and the culture was grown at 37°C to an A 600 nm of 0.6; protein expression was then induced by the addition of 1 mM isopropyl B-D-thiogalactopyranoside for 2 h. After induction, the culture was pelleted by centrifugation and resuspended in 30 ml of lysis buffer (40 mM Tris-HCl, pH 7.5, 100 mM KCl, 100 g/ml Pefabloc, 10 g/ml leupeptin, and 10 g/ml aprotinin). The bacterial pellet was lysed by mild sonication and clarified by centrifugation at 16000 ϫ g for 30 min. The supernatant was incubated with 0.5 ml of nickel nitrilotriacetic acid-agarose resin (Qiagen) with rocking for 1 h at 4°C. The bound resin was loaded onto a column and washed with 10 ml of lysis buffer followed by 10 ml of lysis buffer supplemented with 50 mM imidazol. The bound protein was eluted in lysis buffer containing 250 mM imidazol. The purified enzymes were stored in 40% glycerol at Ϫ80°C. All point mutations of CMKB were generated by the mega-primer mutagenesis technique and confirmed by sequencing (13). Glutathione S-transferase-CaMKI and maltose-binding protein-CaMKK␤ were bacterially expressed and purified as described previously (14 -16).
Peptide Kinase Assays-All CaM kinases were assayed at 30°C in 50 mM Tris-HCl, pH 7.5, 1 mM dithiothreitol, 10 mM MgCl 2 , 200 M ATP, 0.2 l/reaction [␥-32 P]ATP, 1 M A. nidulans CaM, 1 mM CaCl 2 , 0.1% Tween 20. The kinases were assayed using the following concentrations: 100 ng/reaction CMKB, 5 ng/reaction CaMKI, 500 ng/reaction CMKC and CaMKK␤ and 1 g/reaction crude protein extracts. CMKB was assayed using 100 M ADR1G (LKKLTRRASFSGQ) as a peptide substrate; 100 M synapsin site 1 (LRRRLSDANF) was used for CaMKI. The reactions were performed in a 30-l final reaction volume for 5 min. The kinase reactions were terminated by spotting 20 l of the reaction onto p81 phosphocellulose filters followed by extensive washing in 75 mM phosphoric acid. The specific activity was determined after liquid scintillation counting of the dried filters. To assay the ability of CMKC and CaMKK␤ to activate the downstream kinases, a 25 l reaction including both activating and target kinases but without the peptide substrate and [␥-32 P]ATP was preincubated at 30°C for 10 min. The peptide assay was then initiated by the addition of the substrate and [␥-32 P]ATP.
Disruption of cmkB and cmkC-Screening the A. nidulans chromosome specific library from the Fungal Genetics Stock Center with cDNA probes of both kinases identified cosmids containing the genomic sequences. The genomic clones of CMKB and CMKC were isolated by polymerase chain reaction of cosmids W4E06 and L4O01, respectively, taking advantage of the same oligonucleotide primers used to isolate the cDNAs. The genomic clones were subcloned into pGEM-7Zf(Ϫ) (Promega). To generate the disruption plasmid for cmkB, an approximately 600-bp ClaI/ScaI fragment of the cmkB was isolated, blunted, and subcloned into the SmaI site of pRG1 (17), thereby generating pCMKBdis. Ligation of an approximately 700 bp of blunted HindIII/ SacII fragment of cmkC into pRGI generated pCMKCdis. A. nidulans GR5 conidia were transformed with pCMKBdis and pCMKCdis by electroporation as described by Sanchez and Aguirre (18). The positive transformants were selected by growth on minimal medium in the absence of uridine and uracil to select for the presence of the pyr4 nutritional marker. Transformants of pCMKBdis were maintained as heterokaryon-generating cultures by transfer of mycelia instead of spores. Southern analysis was performed using the vegetatively growing mycelia as described by Rasmussen et al. (12). Transformants of pCMKCdis were streaked three times to ensure strain purity followed by Southern analysis.
pCMKB5Ј was generated by blunt end ligation of the BamHI/ScaI fragment of cmkB into the SmaI site of pRG1. A. nidulans GR5 germlings were transformed with the disruption plasmid and selected according to the protocol followed for disruption of cmkC.
A. nidulans Growth Assays-The nuclear number was determined by growing spores in liquid minimal medium on glass coverslips at a spore concentration of 1 ϫ 10 4 spores/ml. Germlings were collected by removing coverslips at various times followed by fixing and staining with the fluorescent dye 4Ј,6-diamidino-2-phenylindole as described by Harris et al. (19).
Antibody Generation and Western Blotting-Rabbit anti-CMKB polyclonal antibodies were generated using keyhole limpet hemocyanincoupled hexa-histidine-tagged CMKB as the antigen. The coupled protein was injected into rabbits, and antiserum was harvested by standard techniques (20). Western blots were performed using a 1:3000 dilution of the anti-CMKB antibody and a similar dilution of IgGconjugated horseradish peroxidase (Amersham Pharmacia Biotech). Relative CMKB expression was quantified using a Molecular Dynamics PhosphorImager equipped with a densitometer.
Histone H1 Kinase Assay-Extracts for histone H1 kinase assays were prepared as described by Dayton et al. (3). Histone H1 kinase activity was quantified using a Molecular Dynamics PhosphorImager following SDS-polyacrylamide gel electrophoresis separation of the reactions.

RESULTS
Cloning of CMKB and CMKC-To identify novel CaM-binding proteins in A. nidulans, we used an expression-cloning strategy similar to the modified far-Western technique described by Stirling et al. (11). A primary screen of 4 ϫ 10 5 phage plaques yielded more than 150 positives, 40 of which were chosen randomly for further characterization. To eliminate the cloning of known A. nidulans Ca 2ϩ /CaM-binding proteins and to validate the primary screen, a secondary hybridization screen was performed using the cDNAs of both the known Ca 2ϩ /CaM-dependent protein kinase (CMKA) and calcineurin (CnA). In this manner, 23 of the 40 positive clones were identified as CMKA or CnA. Of the remaining 17 plaques, 11 were predicted to encode three independent proteins, two protein kinases and a large coil/coil protein. The six remaining plaques either contained short or unique inserts.
The two protein kinases identified (CMKB and CMKC) share homology with the protein kinases of the mammalian CaM kinase cascade. The sequence data for both CMKB and CMKC have been submitted to the DDBJ/EMBL/GenBank TM data bases. The CMKB cDNA is predicted to encode a 404-amino acid protein. As shown in Fig. 1A, CMKB maintains 67% amino acid identity with a predicted CaM kinase from the plant pathogenic fungus Colletotrichum gloeosporioides (CMK) and 43% identity with CMKI of Schizosaccharomyces pombe. CMKB also shares appreciable identity with mammalian kinases, rat CaMKIV (34%) and rat CaMKI (35%). CMKB possesses a Thr residue at the identical position within the activation loop of the catalytic domain that is the site of the activating phosphorylation in mammalian CaMKs I and IV. Thus, based on the primary sequence, CMKB is an A. nidulans homologue of mammalian CaMKI and IV. The CMKC cDNA is predicted to encode a protein of 518 amino acids. CMKC is most homologous to the family of CaM kinase kinases, sharing 30 and 26% identity with rat CaMKK␣ and CaMKK␤, respectively. CMKC also retains high sequence homology with Caenorhabditis elegans CeCaMKK (24%) and S. pombe SspI (21%). Thus, based on primary sequence homology, CMKC is a member of a growing family of CaMKKs.
Biochemical Characterization of CMKB and CMKC-Mammalian CaMK␣/␤ activates CaMKI/IV by phosphorylation of a critical activation loop Thr (14,15). To determine whether CMKB and CMKC are biochemical homologues of CaMKI/IV and CaMK␣/␤, we analyzed bacterially expressed and purified CMKB and CMKC. As shown in Fig. 2A, recombinant CMKB is dependent upon Ca 2ϩ /CaM for enzymatic activity. In the presence of Ca 2ϩ /CaM, the activity of CMKB is approximately 15 nmol/min/mg using ADR1G as a peptide substrate. Preincubation of CMKB with CMKC results in a 73-fold increase in the specific activity of CMKB (Fig. 2B). The increase in protein kinase activity correlates with an increase in the phosphorylation state of CMKB (data not shown). Similar to CaMKs I and IV, mutation of the CMKB activation loop Thr to Ala (T179A) completely abrogates the ability of CMKC to activate CMKB (Fig. 2B) and prevents CMKC-dependent phosphorylation (data not shown). Furthermore, the T179D mutation partially mimics T179 phosphorylation (data not shown). Thus, like mammalian CaMKI and IV, CMKB requires both Ca 2ϩ /CaM and activation loop phosphorylation for maximal activity.
To assess whether CMKB and CMKC are functionally interchangeable with the mammalian kinases in vitro, we measured the ability of the A. nidulans kinases to activate and be activated by the mammalian cascade members CaMKI and CaMKK␤. As demonstrated in Fig. 3A, recombinant human CaMKI is activated approximately 17-fold by preincubation in the presence of rat CaMKK␤, Ca 2ϩ /CaM, and ATP. CMKC can functionally mimic CaMKK␤ in activating CaMKI 8-fold compared with the unactivated kinase. Reciprocally, rat CaMKK␤ weakly activates A. nidulans CMKB (Fig. 3B). Collectively our in vitro biochemical analyses indicate CMKB and CMKC are A. nidulans homologues of the mammalian CaM kinase cascade members CaMKI/IV and CaMKK␣/␤, respectively.
Disruption of cmkB-To assess the physiological role of cmkB, we disrupted the gene by homologous recombination. The homologous integration of the cmkB disruption vector is diagramed in Fig. 4A. We were unable to isolate any disruption strains despite screening more than 100 transformants (data not shown). Since A. nidulans normally grows as a haploid organism, the disruption of an essential gene results in a nonviable spore; thus, if cmkB is required for proliferation, no viable homologous recombinants would be identified by this method.  4. Generation of A. nidulans ⌬cmkB. A, representation of the homologous integration of pCMKBdis into the endogenous cmkB locus. The approximate locations of the SacI restriction sites used for the identification of the homologously recombined plasmid are identified. B, diagram of a heterokaryon transformation strategy. The resulting heterokaryon maintains two nuclei, one pyr4 Ϫ cmkB ϩ and the second pyr4 ϩ cmkB Ϫ . Asexual reproduction produces uninucleate spores that maintain the genotype pyr4 Ϫ cmkB ϩ or pyr4 ϩ cmkB Ϫ . To determine if cmkB is essential for growth, the spores are grown in YG in the absence of UU; under these conditions neither spore will germinate if cmkB is essential. In the presence of UU, approximately 50% of the spores should germinate, all of which should represent the parental strain. C, Southern analysis of ⌬cmkB.
Because we could not generate a viable haploid ⌬cmkB strain, we surmised that cmkB is essential for A. nidulans growth and attempted to disrupt cmkB in a heterokaryon. As diagrammed in Fig. 4B, heterokaryons are generated by transforming conidia after the first DNA synthesis or by fusion of two non-viable germinating spores, and the subsequent transformants are maintained as multinucleate mycelia. Strains in which an essential gene is disrupted are forced to grow as heterokaryons because the wild-type nucleus maintains a copy of the essential gene, whereas the disrupted nucleus provides the selectable nutritional marker (absent in wild type). Southern analysis of greater than 75 transformed strains maintained as heterokaryons revealed 3 strains (one of which is depicted in Fig. 4C) that yielded the expected Southern pattern represent- Genomic DNA was collected as described under "Experimental Procedures," and 10 mg was digested with SacI and hybridized with a full-length cDNA probe of cmkB. The endogenous gene is represented by a 4.7-kbp band, and the disrupted gene is represented by the appearance of the 6.6and 3.7-kbp fragments. D, growth of ⌬cmkB in either YG or YG ϩ UU. Spores from ⌬cmkB were grown for 10 h in either YG or YG ϩ UU, fixed, and stained with 4Ј,6-diamidino-2-phenylindole. The percentage of budded cells was determined from more than 200 spores in each condition.
ing both the wild-type cmkB locus and a cmkB locus disrupted by homologous integration of pCMKBdis. As depicted in Fig.  4B, when ⌬cmkB undergoes asexual reproduction, the resulting spores will each possess a single haploid nucleus with the following genotypes: pyr4 ϩ cmkB Ϫ or pyr4 Ϫ cmkB ϩ . In the absence of uridine and uracil, fewer than 1% of the spores underwent nuclear division, whereas in the presence of these nutritional supplements, 48% of the spores germinated. These data indicate cmkB is essential in A. nidulans, since the only viable spores are those that contain wild-type cmkB grown under conditions in which the pyr4 Ϫ genotype is not lethal. The presence of the wild-type cmkB locus and absence of the integrated pCMKBdis plasmid in these viable mycelia was confirmed by Southern blot (data not shown).
Role of CMKB during Germination-Since CMKB appears to be essential, we generated an A. nidulans strain in which CMKB expression was slightly delayed due to the disruption of the endogenous cmkB promoter (dcmkB) (Fig. 5, A and B). As demonstrated in Fig. 5C, the level of CMKB in the control strain (pAL5#10) is low at germination but after 1 h, CMKB begins to accumulate and continues to increase linearly up to 5 h, at which time a peak 5-fold increase in protein level is observed. However, CMKB expression is delayed in dcmkB relative to control by approximately 1 h. Once CMKB begins to accumulate, it does so linearly and reaches a comparable plateau after 6 h of growth. Presumably the CMKB expression in dcmkB is the result of read-through from an upstream promoter.
Using dcmkB, we next examined the physiological consequences of delayed CMKB expression. To determine the effects of altered CMKB expression on the nuclear division cycle of A. nidulans, the number of germlings with two or more nuclei was determined for dcmkB and a nutritionally complemented control (pAL5#10) as a function of time after germination. As demonstrated in Fig. 6A, pAL5#10 begins to undergo the first nuclear division before 6 h, and by 8 h after inoculation, nearly 100% of the control conidia have at least 2 nuclei. In contrast, dcmkB conidia are delayed approximately 2 h in the executing the first nuclear division. Thus, CMKB is required for the proper timing of the initial nuclear division after germination.
To determine whether the delayed expression of CMKB affected only the first nuclear division or also subsequent divisions, we examined the ability of dcmkB to continue with the second nuclear division by scoring the percent germlings that, having undergone one nuclear division, proceed to accomplish subsequent divisions as indicated by the presence of four or more nuclei. As seen in Fig. 6B, by 10 h of growth, all of the pAL5#10 spores have generated germlings with 4 or more nuclei. The absolute timing of initiation of the second nuclear division in dcmkB is delayed approximately 2 h in comparison with the controls, but the rate at which the germlings proceed with the second nuclear division is similar to that of the control. Thus, retarded CMKB expression upon germination affects only the rate of the first nuclear division cycle, not subsequent mitoses.
To more finely probe CMKB function during germination, we investigated whether the lag in dcmkB germination is before activation of NIMX cdc2 . After germination the initiation of Sphase is dependent upon and indicated by the activation of the A. nidulans single cyclin-dependent kinase, NIMX cdc2 (21,3). We assayed NIMX cdc2 activity during germination by precipitating the kinase with p13-agarose and measuring the ability of the complex to phosphorylate histone H1 in vitro. As demonstrated in Fig. 7, activity of NIMX cdc2 in the control strain is first evident after 5 h of growth and increases linearly through FIG. 6. Growth characteristics of dcmkB and ⌬cmkC. A, delayed expression of CMKB and disruption of cmkC result in a delay in the initiation of nuclear division. Percentage of germlings with two or greater nuclei of the nutritionally complemented control, pAL5#10 (circles), dcmkB (squares), and ⌬cmkC (triangles) strains. Strains were grown on coverslips, and the nuclear number was determined after staining with 4Ј,6-diamidino-2-phenylindole as described under "Experimental Procedures." Percentages were calculated from more than 500 individual germlings. B, the delay of dcmkB and ⌬cmkC growth is not restricted to the first nuclear division cycle The percentage of germlings with two or greater nuclei that also have four or greater nuclei was calculated for pAL5#10 (circles), dcmkB (squares), and ⌬cmkC (triangles). The percentages were calculated from more than 500 individual germlings at each time point. The graph is representative of multiple experiments. FIG. 7. The activation of NIMX cdc2 is delayed in both dcmkB and ⌬cmkC. Histone H1 kinase assays were performed on protein extract from pAL5#10 (black bar), dcmkB (gray bar), and ⌬cmkC (white bar) at various times after germination. The graph is representative of three independent experiments. 7 h. However, when cmkB expression is delayed in dcmkB, NIMX cdc2 activation is also postponed. Thus, although we cannot rule out the possibility that cmkB plays an important role elsewhere in the nuclear division cycle, we conclude that CMKB is required for the normal temporal regulation of NIMX-cdc2 activity before S-phase in A. nidulans.
Generation and Characterization of ⌬cmkC-The disruption scheme for cmkC was similar to the original approach utilized for CMKB. In contrast to the pCMKBdis transformation, several strains were identified from the initial transformation of GR5 with pCMKCdis that yielded the Southern banding pattern predicted for homologous recombination. Fig. 8A shows a representative Southern blot of a single strain in which the cmkC locus has been disrupted. These results indicate cmkC is not required for the viability of A. nidulans.
Since cmkC is not essential for growth of A. nidulans, we first sought to determine whether, in the absence of CMKC, we could detect the activity of other activating kinases in crude protein extract that might functionally complement the loss of cmkC. Fig. 8B demonstrates that crude protein extract from the parental GR5 strain can activate CMKB in vitro, whereas crude protein extract from ⌬cmkC cannot. Although we cannot rule out the presence of an insoluble or developmentally regulated activating protein kinase, these results suggest cmkC is the predominant activating kinase of CMKB in A. nidulans.
To assess the requirements for CMKC in the nuclear division cycle, we performed the same experiments as described for dcmkB. As demonstrated in Fig. 6A, the timing of first nuclear division is dramatically delayed in the absence of CMKC as compared with the control. Surprisingly, approximately 30% of the germlings fail to undergo a single nuclear division within 20 h post-inoculation. Thus, like CMKB, CMKC is important for proper timing of the first nuclear division after germination. cmkC disruption alters not only the kinetics of the first nuclear division but also subsequent divisions (Fig. 6B). The lag of the second nuclear division is observed by the reduced rate at which ⌬cmkC germlings containing two nuclei accumulate four nuclei in comparison with the control. The extended duration of the second nuclear division in the absence of CMKC indicates that cmkC is required for the proper temporal regulation of not only the first but also the subsequent nuclear divisions.
To determine whether, similar to cmkB, cmkC function is required for the proper timing of DNA synthesis, we followed NIMX cdc2 activation upon germination. Fig. 7 demonstrates that disruption of cmkC severely retards NIMX cdc2 activation by both delaying the onset of activation and dampening its maximal activity. The impaired activation of histone H1 kinase activity correlates with the severe delay in both the initiation of nuclear division and subsequent exponential growth. Thus, we conclude that CMKC plays a non-essential but important role during G 1 after germination. DISCUSSION Using an expression cloning strategy, we identified two novel Ca 2ϩ /CaM-dependent protein kinases in A. nidulans. CMKB and CMKC are homologues of the mammalian CaMK cascade members, CaMKI/IV and CaMKK␣/␤, respectively. Although closely related to CaMKI and IV, CMKB shares highest identity with Ca 2ϩ /CaM-dependent protein kinases of S. pombe (cmk1) and C. gloeosporioides (CMK), and each of these homologues possesses a Thr within its activation loop at the identical position as CMKB Thr-179, indicating they may also be subject to activating phosphorylation. The closest homologue of CMKC in lower eukaryotes is S. pombe ssp1. The identification of CMKB and CMKC and their predicted homologues in other fungal species suggest, similar to other signal transduction systems, the CaMK cascade is evolutionarily conserved from unicellular eukaryotes to mammals.
cmkB is the second essential CaMK identified in any system. The terminal phenotype of ⌬cmkB spores indicates CMKB is required for an early step in the nuclear division cycle. Furthermore, we demonstrate that CMKB functions before the S-phase activation of NIMX cdc2 , after germination. A. nidulans CMKA, the other essential CaMK, is required for G 2 progression (5). Thus, two distinct CaMKs are essential for progression through G 1 and G 2 in A. nidulans.
The requirement of CMKB in G 1 is consistent with previous studies of CaM and CaMKs in other systems. The CaM inhibitor W-13 arrests mammalian cell cycle progression in at least two points during re-entry from G 0 (22). Furthermore, the general CaMK inhibitor, KN-93, blocks cell proliferation before S phase in both cycling cells and those stimulated to enter the cell cycle from G 0 (7)(8)(9). Our experiments in Aspergillus, which demonstrate that disabling CMKB disrupts entry into the cell cycle, attribute to CMKB, among the CaM kinases, an essential role in the transition of spores from quiescence into S-phase. Experiments are ongoing to determine whether the homologues of CMKB, CaMKs I/IV, are essential targets of CaM during G 1 in mammalian cells.
Although CMKC is not essential for proliferation, it does FIG. 8. Generation of ⌬cmkC. A, Southern analysis of a representative ⌬cmkC. 10 g of genomic DNA was digested with PvuI and hybridized with a full-length CMKC cDNA probe. A 3.9-kbp band represents the endogenous gene, and the disrupted gene yields 7.0-and 1.9-kbp fragments. B, activation of CMKB by GR5 and ⌬cmkC crude protein extracts. CMKB was preincubated with crude protein extracts from either GR5 or ⌬cmkC and compared with unactivated CMKB. The values represent the average Ϯ S.E. of three independent experiments.
perform an important role in G 1 progression after germination. However, if CMKB and CMKC comprise a linear kinase cascade in vivo, as suggested both by our in vitro biochemical data and by the role of the two kinases at similar points in the cell cycle, it would be predicted that CMKC should also be essential for the nuclear division cycle. There are multiple explanations for this conundrum. The simplest explanation is that other protein kinases can perform the predicted function of CMKC in its absence. Although disruption of CMKC eliminates all activating activity in crude A. nidulans extract, we cannot completely eliminate the possibility that other kinases or CMKC isoforms capable of phosphorylating and activating CMKB are present in the fungus. Another possibility is that CMKB is not the relevant substrate for CMKC in vivo. The mammalian CaMKKs have been demonstrated to phosphorylate and activate both protein kinase B and AMP-activated protein kinase (23,24). Thus it is possible that in vivo CMKC possesses substrates other than CMKB that are important, but not essential, for G 1 progression. Alternatively, activation loop phosphorylation of CMKB may not be required for CMKB function in vivo. In vitro, bacterially expressed CaMKI, CaMKIV, and CMKB all possess significant kinase activity that is dramatically enhanced upon activation loop phosphorylation (25,15). Thus, in the absence of CMKC, it is possible that the reduced but not abolished CMKB activity is sufficient to slowly drive the nuclear division cycle through the G 1 in the absence of activation loop phosphorylation. Finally, Hook et al. (26) have recently provided evidence that activation loop phosphorylation may not be required for CaMKI and IV activation but may instead alter their substrate specificity. If this is the case in vivo, the two kinases may form a cascade, but CMKC would not necessarily be essential for proliferation. For example, CMKB may possess an activation-independent substrate essential for G 1 progression and an activation-dependent substrate whose phosphorylation is not strictly required for G 1 progression but that affects the kinetics of this cell cycle phase. Thus, the fact that CMKC is not essential for the nuclear division cycle in A. nidulans does not eliminate the possibility that CMKB and CMKC form a functional CaMK cascade in vivo.
To investigate if CMKB and CMKC are epistatic, we attempted to functionally complement the absence of CMKC by ectopically expressing CMKB or CMKB T179D in ⌬cmkC. CMKB T179D partially mimics T179 phosphorylation; therefore if CMKB and CMKC are members of a kinase cascade, we predicted that overexpression of wild-type CMKB or T179D would rescue the growth delay of ⌬cmkC. However, neither enzyme complements the loss of CMKC in ⌬cmkC. Furthermore, similar to S. pombe cmk1 (27), ectopic T179D expression lengthens the nuclear division cycle in both wild-type and ⌬cmkC strains. Although these experiments failed to make clear whether CMKB and CMKC are components of a linear cascade in vivo, they do demonstrate the importance of the multiple modes of regulation of CMKB. Finally, overexpression of the wild-type enzyme has no effect on the nuclear division cycle, whereas expression of the artificially activated fulllength protein is detrimental to growth.
Consistent with our observations in A. nidulans, pharmacological inhibitors of both CaM and CaMK arrest mammalian cells in G 1 before cdk activation. KN-93 arrests NIH 3T3 cells with reduced cdk4 and cdk2 activities attributed to a reduction in cyclinD1 levels and an increased association of p27 kip with cdk2/cyclinE (9). The CaM antagonist W13 also inhibits G 1 progression in normal rat kidney cells before cdk4 and cdk2 activation, but this CaM inhibition does not appear to affect cyclinD or -E protein levels. Instead Taules et al. (28) demonstrate that in the presence of W13, cdk4/cyclinD1 is exclusively cytoplasmic rather than nuclear. Thus, the authors suggest that CaM directly or indirectly modulates cdk4 activity by regulating its nuclear localization. In Aspergillus, W7 blocks germination before the activation of the S-phase cdk/cyclin complex, NIMX cdc2 /NIME cyclinB (3). Thus, NIME cyclinB synthesis and/or the subcellular localization of NIMX cdc2 /NIME cyclinB represent attractive targets of the CMKB and CMKC cascade in A. nidulans.