The kinase responsible for Thr161‐Thr160 phosphorylation and activation of cdc2/cdk2 (CAK:cdk‐activating kinase) has been shown previously to comprise at least two subunits, cdk7 and cyclin H. An additional protein co‐purified with CAK in starfish oocytes, but its sequencing did not reveal any similarity with any known protein. In the present work, a cDNA encoding this protein is cloned and sequenced in both starfish and Xenopus oocytes. It is shown to encode a new member of the RING finger family of proteins with a characteristic C3HC4 motif located in the N‐terminal domain. We demonstrate that the RING finger protein (MAT1: ‘menage à trois’) is a new subunit of CAK in both vertebrate and invertebrates. However, CAK may also exist in oocytes as heterodimeric complexes between cyclin H and cdk7 only. Stable heterotrimeric CAK complexes were generated in reticulocyte lysates programmed with mRNAs encoding Xenopus cdk7, cyclin H and MAT1. In contrast, no heterodimeric cyclin H‐cdk7 complex could be immunoprecipitated from reticulocyte lysates programmed with cdk7 and cyclin H mRNAs only. Stabilization of CAK complexes by MAT1 does not involve phosphorylation of Thr176, as the Thr176–>Ala mutant of Xenopus cdk7 could engage as efficiently as wild‐type cdk7 in ternary complexes. Even though starfish MAT1 is almost identical to Xenopus MAT1 in the RING finger domain, the starfish subunit could not replace the Xenopus subunit and stabilize cyclin H‐cdk7 in reticulocyte lysate, suggesting that the MAT1 subunit does not (or not only) interact with cyclin H‐cdk7 through the RING finger domain.

Transitions of the cell cycle are controlled by cyclin-dependent protein kinases (cdks) whose phosphorylation on the Thr residue included in the conserved sequence YTHEVV dramatically increases the activity. A kinase responsible for this specific phosphorylation, called CAK for cdk-activating kinase, has been recently purified from starfish and Xenopus oocytes and shown to contain the MO15 gene product as a catalytic subunit. In the present paper, we have cloned the human homolog of Xenopus p40MO15 by probing a HeLa cell cDNA library with degenerate oligonucleotides deduced from Xenopus and starfish MO15 sequences. Human and Xenopus MO15 displayed a strong homology showing 86% identity with regard to amino acid sequences. Northern blot analysis of RNA extracts from a series of human tissues as well as from cultured rodent fibroblasts revealed a unique 1.4 kb MO15 mRNA. No variation in the amount of MO15 transcript or protein was found along the entire course of the fibroblast cell cycle. Fluorescence in situ hybridization on human lymphocyte metaphases showed two distinct chromosomal locations of human MO15 gene at 5q12-q13 and 2q22-q24. By using gene tagging and mammalian cell transfection, we demonstrate that the KRKR motif located at the carboxy terminal end of MO15 is required for nuclear targeting of the protein. Mutation of KRKR to NGER retains MO15 in the cytoplasmic compartment, whilst the wild-type protein is detected exclusively in the nucleus. Interestingly, we demonstrate that the nuclear targeting of MO15 is necessary to confer the protein its CAK activity. In contrast to the wild-type, the NLS-mutated MO15 expressed in Xenopus oocytes is unable to generate CAK as long as the nuclear envelope is not broken. The nuclear localization of both the MO15 gene product and CAK activity may imply that cdks activation primarily occurs in the cell nucleus.

p40MO15, a cdc2‐related protein, is the catalytic subunit of the kinase (CAK, cdk‐activating kinase) responsible for Thr161/Thr160 phosphorylation and activation of cdk1/cdk2. We have found that strong overexpression of p40MO15 only moderately increases CAK activity in Xenopus oocytes, indicating that a regulatory CAK subunit (possibly a cyclin‐like protein) limits the ability to generate CAK activity in p40MO15 overexpressing oocytes. This 36 kDa subunit was microsequenced after extensive purification of CAK activity. Production of Xenopus CAK activity was strongly reduced in enucleated oocytes overexpressing p40MO15 and p40MO15 shown to contain a nuclear localization signal required for nuclear translocation and generation of CAK activity. p40MO15 was found to be phosphorylated on Ser170 and Thr176 by proteolytic degradation, radiosequencing of tryptic peptides and mutagenesis. Thr176 phosphorylation is required and Ser170 phosphorylation is dispensable for p40MO15 to generate CAK activity upon association with the 36 kDa regulatory subunit. Finally, Thr176 and Ser170 phosphorylations are not intramolecular autophosphorylation reactions. Taken together, the above results identify protein‐protein interactions, nuclear translocation and phosphorylation (by an unidentified kinase) as features of p40MO15 that are required for the generation of active CAK.

In multidrug-resistant (MDR) derivatives of the mouse lymphoid tumor P388, the emergence of MDR is associated with overexpression and transcriptional activation of the mdr3 gene, either in the absence of (P388/VCR-10) or concomitant with (P388/ADM-2) gene amplification. In both instances, Northern (RNA) blotting analyses have suggested the presence of altered mdr3 transcripts in these cells, possibly originating from novel transcription initiation sites. The mechanisms underlying mdr3 overexpression in these cells have been investigated. In P388/VCR-10 cells, Southern blotting analyses together with genomic DNA cloning and nucleotide sequencing have demonstrated the presence of an intact mouse mammary tumor virus (MMTV) within the boundaries of intron 1 of mdr3. cDNA cloning and nucleotide sequencing indicated that this integration event results in the synthesis and overexpression of a hybrid MMTV-mdr3 mRNA which initiates within the U3 region of the 5' long terminal repeat (LTR) of the provirus. Consequently, this mRNA lacks the normal exon 1 of mdr3 but contains (i) MMTV LTR-derived sequences at its 5' end, (ii) a novel mdr3 exon, mapping within the boundaries of intron 1 downstream of the MMTV integration site and generated by alternative splicing, and (iii) an otherwise intact 3' portion of mdr3 starting at exon 2. A similar type of analysis of P388/ADM-2 cells revealed that mdr3 overexpression in these cells is associated with the integration of an intracisternal A particle (IAP) within an L1Md repetitive element, immediately upstream of mdr3. The IAP insertion results in the overexpression of hybrid IAP-mdr3 mRNA transcripts that initiate within the 3' LTR of the IAP and which contain IAP LTR-derived sequences at the 5' end spliced 14 nucleotides upstream of the normal exon 1 of mdr3. Taken together, these results indicate that independent retroviral insertions were the initial mutagenic event responsible for mdr3 overexpression and survival during drug selection of these cell lines. Amplification of the rearranged and activated mdr3 gene copy occurred during further selection for high-level drug resistance in P388/ADM-2 cells.

We have produced human cyclin A in Escherichia coli and investigated how it generates H1 kistone kinase activity when added to cyclin-free extracts prepared from parthenogenetically activated Xenopus eggs. Cyclin A was found to form a major complex with cdc2, and to bind cdk2/Eg1 only poorly. No lag phase was detected between the time when cyclin A was added and the time when H1 histone kinase activity was produced in frog extracts, even in the presence of 2 mM vanadate, which blocks cdc25 activity. Essentially identical results were obtained using extracts prepared from starfish oocytes. We conclude that formation of an active cyclin A-cdc2 kinase during early development escapes an inhibitory mechanism that delays formation of an active cyclin B-cdc2 kinase. This inhibitory mechanism involves phosphorylation of cdc2 on tyrosine 15. Okadaic acid (OA) activated cyclin B-cdc2 kinase and strongly reduced tyrosine phosphorylation of cyclin B-associated cdc2, even in the presence of vanadate. 6-dimethylaminopurine, a reported inhibitor of serine-threonine kinases, suppressed OA-dependent activation of cyclin B-cdc2 complexes. This indicates that the kinase(s) which phosphorylate(s) cdc2 on inhibitory sites can be inactivated by a phosphorylation event, itself antagonized by an OA-sensitive, most likely type 2A phosphatase.We also found that cyclin B- or cyclin A-cdc2 kinases can induce or accelerate conversion of the cyclin B-cdc2 complex from an inactive into an active kinase. Cyclin B-associated cdc2 does not undergo detectable phosphorylation on tyrosine in egg extracts containing active cyclin A-cdc2 kinase, even in the presence of vanadate. We propose that the active cyclin A-cdc2 kinase generated without a lag phase from neo-synthesized cyclin A and cdc2 may cause a rapid switch in the equilibrium of cyclin B-cdc2 complexes to the tyrosine-dephosphorylated and active form of cdc2 during early development, owing to strong inhibition of the cdc2-specific tyrosine kinase(s). This may explain why early cell cycles are so rapid in many species.

Purified cyclin B-cdc2 kinase has been shown previously to trigger cyclin degradation in interphase frog extracts by initiating a cascade of reactions that includes cyclin ubiquitinylation and ends with proteolysis. However, cyclin A-cdc2 kinase was not assayed in these early experiments. Here we have shown that full-length recombinant human cyclin A failed to induce cyclin degradation when it was added to frog extracts free of cyclin B, although it formed an active kinase complex with Xenopus cdc2. A highly purified kinase complex containing a truncated human cyclin A and starfish cdc2 also failed to switch on the cyclin degradation pathway. In contrast, both recombinant cyclin B and highly purified cyclin B-cdc2 kinase readily triggered degradation of both cyclins B and A in frog extracts. Whilst free cyclin A had no inhibitory effect, cyclin A-cdc2 kinase delayed degradation of both cyclins A and B induced by cyclin B-cdc2 kinase. The finding that cyclin A-cdc2 kinase cannot turn on, and even delays, cyclin destruction may be essential to prevent premature inactivation of MPF (maturation-promoting factor) before complete condensation of chromosomes and formation of the metaphase spindle.

Truncated cyclin A and cyclin B lacking the N-terminal domain comprising the 'destruction box' escape from proteolysis and arrest cells at metaphase. Mutation of a conserved arginine residue of the destruction domain makes cyclin B resistant to proteolysis. Here we show that mutation of the same residue also makes cyclin A resistant to proteolysis, in either of two situations in which the cyclin degradation pathway is turned on: (i) in Xenopus extracts of activated eggs where the degradation pathway has been permanently turned on by adding a recombinant undegradable cyclin B in which the arginine residue of the destruction box has been substituted by alanine; (ii) in extracts of metaphase II-arrested oocytes after Ca2+-dependent inactivation of the cytostatic factor (CSF).

Exit from metaphase of the cell cycle requires inactivation of MPF, a stoichiometric complex between the cdc2 catalytic and the cyclin B regulatory subunits, as well as that of cyclin A‐cdc2 kinase. Inactivation of both complexes depends on proteolytic degradation of the cyclin subunit, yet cyclin proteolysis is not sufficient to inactivate the H1 kinase activity of cdc2. Genetic evidence strongly suggests that type 1 phosphatase plays a key role in the metaphase‐anaphase transition of the cell cycle. Here we report that inhibition of both type 1 and type 2A phosphatases by okadaic acid allows cyclin degradation to occur, but prevents cdc2 kinase inactivation. Complete inhibition of type 2A phosphatase alone is not sufficient to prevent cdc2 kinase inactivation following cyclin proteolysis. We show further that residue 161 of cdc2 is phosphorylated in active cyclin A or cyclin B complexes at metaphase, whilst unassociated cdc2 is not phosphorylated. Proteolysis of cyclin releases a free cdc2 subunit, which subsequently undergoes dephosphorylation and then migrates more slowly than its Thr161 phosphorylated counterpart in Laemmli gels. Removal of phosphothreonine 161 requires cyclin proteolysis. However, it does not occur even after cyclin proteolysis, when both type 1 and type 2A phosphatases are inhibited. We conclude that both cyclin degradation and dephosphorylation of Thr161 on cdc2, catalysed at least in part by type 1 phosphatase, are required to inactivate either cyclin B‐ or cyclin A‐cdc2 kinases and thus for cells to exit from M phase.

Neutral endopeptidase (EC 3.424.11, NEP) is a membrane-bound zinc-metallopeptidase. The substrate specificity and catalytic activity of NEP resemble those of thermolysin, a bacterial zinc-metalloprotease. Comparison of the primary structure of both enzymes suggests that several amino acids present in the active site of thermolysin are also found in NEP. Using site-directed mutagenesis of the cDNA encoding the NEP sequence, we have already shown that His residues 583 and 587 are two of the three zinc ligands. In order to identify the third zinc ligand, we have substituted Val or Asp for Glu616 or Glu646. Val616 NEP showed the same kinetic parameters as the non-mutated NEP. In contrast, the mutant Val646 NEP was almost completely devoid of catalytic activity and unable to bind the tritiated inhibitor [3H]N-[2(R,S)-3-hydroxyaminocarbonyl-2-benzyl-1-oxypropyl]gl ycine, the binding of which is dependent on the presence of the zinc ion. Replacing Glu for Asp at position 646 conserved the negative charge, and the mutant enzyme exhibited the same Km value as the non-mutated enzyme, but kCat was decreased to less than 3% of the value of the non-mutated enzyme. When compared to the non-mutated enzyme Asp646 NEP showed a higher susceptibility to chelating agents, but bound the tritiated inhibitor with the same affinity. Taken together, these observations strongly suggest that Glu646 of NEP is the third zinc-coordinating residue and is equivalent to Glu166 in thermolysin.

We have isolated cDNAs encoding kinases from a murine pre-B cell line by screening a lambda gt11 cDNA expression library with anti-phosphotyrosine antibodies. One cDNA was identified to encode the previously isolated tyrosine kinase c-lyn. Among the remaining clones, we have characterized a cDNA encoding a novel kinase which we have designated TIK. Sequence analysis of this cDNA indicates that the TIK enzyme lacks the features thought to be conserved among protein tyrosine kinases. Although isolated on the basis of its reactivity with the anti-phosphotyrosine antibody, the TIK enzyme was found to have only serine and threonine kinase activity. The amino-terminal portion of the TIK protein contains a cdc2 phosphorylation consensus sequence. Three mRNA transcripts derived from the TIK gene are detected in a variety of adult murine tissues.

Exit from M phase, which requires cyclin degradation, is prevented from occurring in unfertilized eggs of vertebrates arrested at second meiotic metaphase due to a cytostatic factor recently identified as p39mos, the product of the proto‐oncogene c‐mos. Calpain can destroy both p39mos and cyclin in vitro in extracts prepared from metaphase‐arrested Xenopus eggs, but only when free Ca2+ concentration is raised to the millimolar range. When free Ca2+ concentration is raised for only 30 s to the micromolar range, as occurs in physiological conditions after fertilization, cyclin degradation is induced, byt p39mos is not degraded. Cyclin proteolysis at micromolar free Ca2+, is not inhibited by calpastatin, and therefore does not involve calpain. A cyclin mutant modified in the destruction box is found to be resistant at micromolar, but not millimolar free Ca2+, suggesting that the ubiquitin pathway mediates cyclin degradation at micromolar Ca2+ concentration whereas calpain is involved at the millimolar level. A synthetic peptide which binds Ca(2+)‐calmodulin with high affinity suppresses cyclin degradation at micromolar but not millimolar free Ca2+, and this only when it is present in the extract during the first 30 s after raising free Ca2+ concentration. The inhibition of the cyclin degradation pathway by the Ca(2+)‐calmodulin binding peptide can be overcome by adding calmodulin. These results strongly suggest that a Ca(2+)‐calmodulin process is required as an early event following fertilization to release the cyclin degradation pathway from inhibition in metaphase‐arrested eggs. In contrast, p39mos degradation is not required.

Cells in M phase contain a dominant factor, called M-phase-promoting factor or maturation-promoting factor (MPF), which is able upon transfer to drive G2 cells into M phase (Masui and Markert 1971; Kishimoto et al. 1982) or to induce prematurely some cytological features of M phase in G1 cells (Johnson and Rao 1970). Purification of MPF activity from vertebrate and invertebrate oocytes at meiotic metaphase led to isolation of a single component, identified as cdc2 kinase. It corresponds to the major M-phase-specific H1 histone kinase, not only in oocytes and early embryos, but also in yeast (for review, see Dorée 1990; Nurse 1990). In starfish at first meiotic metaphase, cdc2 kinase can be purified as a heterodimer comprising one molecule of a single cyclin-B species and one molecule of cdc2 (Labbé et al. 1989b). In Xenopus oocytes at second meiotic metaphase, purified MPF was found to contain both cdc2 and two...

We report the cloning and functional analysis of a complete clone for the third member of the mouse mdr gene family, mdr3. Nucleotide and predicted amino acid sequence analyses showed that the three mouse mdr genes encode highly homologous membrane glycoproteins, which share the same length (1,276 residues), the same predicted functional domains, and overall structural arrangement. Regions of divergence among the three proteins are concentrated in discrete segments of the predicted polypeptides. Sequence comparison indicated that the three mouse mdr genes were created from a common ancestor by two independent gene duplication events, the most recent one producing mdr1 and mdr3. When transfected and overexpressed in otherwise drug-sensitive cells, the mdr3 gene, like mdr1 and unlike mdr2, conferred multidrug resistance to these cells. In independently derived transfected cell clones expressing similar amounts of either MDR1 or MDR3 protein, the drug resistance profile conferred by mdr3 was distinct from that conferred by mdr1. Cells transfected with and expressing MDR1 showed a marked 7- to 10-fold preferential resistance to colchicine and Adriamycin compared with cells expressing equivalent amounts of MDR3. Conversely, cells transfected with and expressing MDR3 showed a two- to threefold preferential resistance to actinomycin D over their cellular counterpart expressing MDR1. These results suggest that MDR1 and MDR3 are membrane-associated efflux pumps which, in multidrug-resistant cells and perhaps normal tissues, have overlapping but distinct substrate specificities.

The gene responsible for multidrug resistance (mdr), which encodes the P-glycoprotein, is a member of a multigene family. We have identified distinct mdr gene transcripts encoded by three separate mdr genes in the mouse. Expression levels of each mdr gene are dramatically different in various mouse tissues. Specific mdr RNA transcripts of approximately 4.5, 5, and 6 kilobases have been detected. Each of the mdr genes has a specific RNA transcript pattern. These results should be considered in relation to understanding the normal physiological function of the mdr multigene family.

Analogies in the sequences of two related zinc metallopeptidases, the bacterial thermolysin (316 amino acids) and the recently cloned neutral endopeptidase 24.11 ("enkephalinase", 749 amino acids), have been demonstrated by a hydrophobic cluster analysis method derived from the Lim theory. Two sequence alignments are proposed for the entire primary structure of thermolysin and the C-terminal part of endopeptidase 24.11. Except for an arginine residue, all the amino acids involved in the active site of thermolysin have been retrieved in both models of endopeptidase 24.11 within conserved clustered structures. The first model is characterized by a deletion of the Ca2+-binding coil present in thermolysin and the second by replacement of this coil by two alpha-helices. In both models an Arg residue can be located in the active site of the neutral endopeptidase.

Direct comparison of the primary structure of neutral endopeptidase (NEP, EC 3.4.24.11) with that of thermolysin, a bacterial metalloendopeptidase with a similar specificity, has revealed very few similarities between the two sequences, except for two conserved short segments. In thermolysin, these segments contain several of the residues involved in catalysis, including two zinc coordinating histidines (His-142 and His-146) and a third histidine (His-231) involved in stabilizing the transition state through hydrogen bonding. The role of the corresponding histidines in NEP (His-583, His-587 and His-637) was explored by site-directed mutagenesis of NEP cDNA and expression of the mutated cDNA in COS-1 cells. Substitution of either His-583 or His-587 of NEP for Phe completely abolished the activity and Zn-directed inhibitor recognition of the recombinant enzyme, suggesting that these residues play a role similar to His-142 and His-146 of thermolysin as zinc ligands. In contrast, substitution of His-637 for a phenylalanine residue was without effect on enzyme activity.

Neutral endopeptidase (EC 3.4.24.11) is an integral membrane protein found in the plasma membrane of many cell types. The cDNA coding for the complete primary structure of neutral endopeptidase has recently been cloned and sequenced (Devault, A. Lazure, C., Nault, C., Le Moual, H., Seidah, N. G., Chretien, M., Kahn, P., Powell, J., Mallet, J., Beaumont, A., Roques, B. P., Crine, P., and Boileau, G. (1987) EMBO J. 6, 1317-1322). Comparison of the sequence of neutral endopeptidase with that of thermolysin, a bacterial Zn-metalloendopeptidase, suggests that Glu-584 in neutral endopeptidase probably corresponds to Glu-143 in thermolysin, which is an essential amino acid involved in catalysis. To test directly the importance of Glu-584 in the catalytic activity of neutral endopeptidase by site-directed metagenesis, we have constructed an expression vector in which the rabbit kidney cDNA encoding the entire neutral endopeptidase sequence is introduced downstream from the SV40 virus early promotor. After transfection in COS-1 monkey kidney cells, this vector was found to promote the expression of a protein with biochemical and catalytic properties identical to kidney neutral endopeptidase. Oligonucleotide-directed mutagenesis of Glu-584 to either valine or aspartic acid completely abolished the enzymatic activity of the recombinant protein without changing its affinity for the substrate-related tritiated inhibitor [3H]N-[(2R,2S)-3-hydroxyamino-carbonyl-2-benzyl-1-oxopropyl]-glycine. This observation clearly identifies Glu-584 as one of the important residues responsible for the catalytic activity of the enzyme.