CHRIS NORBURY AND PAUL NURSE CELL CYCLE

Cyclins and cell cycle control The universal eukaryotic cell cycle regulator p34cdC2 may be

modulated by association with distinct cyclin protein subunits, but the molecular basis of cyclin action remains unclear.

It is becoming increasingly clear that cyclins, first de- scribed as proteins that accumulate periodically during the rapid and synchronous early cell division cycles of sea urchin embryos [ 11, have an important general role in cell ‘kycle control. Cyclins of two types, termed A and B, peak just before each M phase and decline abruptly as M phase is completed (Fig. 1). The two types, now identiIied in a variety of eukaryotes, are distinguished by the slightly earlier appearance and disappearance of the A type, by their gel mobilities and by their amino acid se- quences which seem to have been conserved separately during evolution. A third group of proteins, the products of the WHII/DAFl, CLNl and CLN2 genes of budding yeast, are classed as rnembers of the cyclin family on the basis of sequence similarity alone, The cyclin-specihc se- quences found in the three types are mainly limited to a central region of about 200 amino acids, flanked by diver- gent N- and C-terminal sections that bring the size of the protein to between 43 and 60 kD. Even within the central ‘cyclin box’ there is only about 30% identity between cy- clins of a given type, and only seven dispersed residues are perfectly conserved among the known members of the cyclin family (S. Forsburg, personal communication), so the biological functions of these proteins may not be precisely equivalent in the various species in which they have been identified.

An essential role for cyclins in the initiation of mitosis has been established. Cyclin B becomes complexed with the serine/threonine protein kinase p34cdc2 to form the heterodimer originally deiined functionally as maturation- promoting factor (MPF). This complex is now thought to be responsible for the initiation of meiotic and rn- totic M phases in all eukaryotes (reviewed in [2] ). Cyclin B is clearly required for the activation of p34d2, but it is possible that cyclins also intluence p34&2 substrate specificity and sub-celUular localization.

Cyclin A has an essential function in Drosophila [3], as embryos homozygous for null mutations in the cyclin A gene become blocked in the Interphase preceding rn- tosis 16, when the maternal store of cyclin A has been exhausted. Although this observation does not indicate which cell cycle event requires cyclin A, the cyclin A- associated kinase reaches maximal activity just before the peak of cyclin B-associated mitotic kinase in several sys- tems (see for example, [4,5]), perhaps indicating a func- tion for cyclin A in mitotic initiation. The earlier obser- vation that messenger RNA encoding either cyclin A or cyclin B is able to induce meiotic maturation in Xenopus oocytes adds weight to this suggestion.

Volume 1 Number 1 1991

Fig.1. A schematic diagram of the eukaryotic cell cycle. The pwdc* protein kinase fulfills two essential functions; at start, where entry into S phase is controlled, and at the G2/M tran- sition, mitotic initiation. These two functions can be be distin- guished at the molecular level by association of pWdC* with a Cl cyclin or cyclin B, respectively. The role of cyclin A and its associated protein kinase has yet to be established, but it has been suggested to act at mitotic initiation and at a point earlier in the cell cycle.

In addition to forming a complex that Initiates M phase, p34d-a function is also required during the Gl phase for commitment to a new round of DNA replication in yeasts, and possibly in higher eukaryotes ([6] and refer- ences therein). The existence of two such diverse roles for a single protein kinase implies that the mitotic and Gl forms of the enzyme are fundamentally different, per- haps as a result of covalent modification, different sub- cellular localization or association with specific regula- tors. In the absence of such a distinction, it would be dilhcult to imagine how the normal alternation of S and M phases might be achieved, If the Gl role for p34&2 is as universal as its role in mitosis, it could have profound consequences, as it is during the Cl phase that controls over proliferation are thought to act in most animal cells.

It has been suggested that a form of the p34Ck2 protein kinase associated with cyclin A may function at a point earlier in the cell cycle than mitosis [ 71. This suggestion is /

23

based on the observation that cyclin A-associated kinase is more active in cells blocked in S-phase than in cells blocked in mitosis. Furthermore, addition of cyclin A to an in vitro extract of Gl phase human cells is sufficient to enable the extract to replicate exogenous DNA [8]. But the physiological significance of this observation is not known, as the mitotic form of p34cdc2 (associated with cyclin B) can fully substitute for cyclin A in this assay. In addition, the kinetics of cyclin A accumulation suggest a role later in the cell cycle than Gl/S phase although it is far from clear how, or at what point cyclin A might act.

Recent work by Pines and Hunter [5] suggests that the majority of cyclin A is complexed not with p34d2 itself, but with a different protein of relative moIecular mass about 33 kD. This protein contains the PSTAIR motif, that is highly conserved among cdc2 proteins though not spe- ci6c to them, but is not recognized by antibodies raised against the very specific C-terminal region of p34Ck? The cyclin A-associated protein is able to phospholylate the same residues in histone Hl as are phosphotylated by bona $de p34&;? [4]. This cyclin A-associated protein seems therefore to belong to a family of protein kinases closely related to p34 C&a and with similar or even iden- tical specificities in vitro, but with different functions in vivo.

The products of the budding yeast WHIl/DAFl, CLNl and CW2 genes are better candidates than cyclin A for Gl-specific m80dulators of p34d2. These genes con- tribute overlapping functions in the Gl phase such that all three must be inactivated to block progression through Gl 191. In contrast to the mitotic B cyclins, which though essential for the initiation of M-phase, do not nor- mally determine its timing [2], the ‘Gl cyclins’ encoded by WHIl/DAFl and CLN2 are able to influence the rate at which budding yeast cells enter S phase. For CLN2 this rate-limiting {effect is probably mediated through the functional equivalent of p34Ck-a (in budding yeast en- coded by the CDC28 gene), as the CLN genes were iden- tified as suppressors of a cdc28 mutation, and the CLN2 product has been shown biochemically to interact di- rectly with p34ax28 [lo].

It will be important now to determine if and to what ex- tent the ‘Gl cyclins’ of budding yeast have been function-

ally conserved during eukaryotic evolution. If the hunt un- covers higher eukaryotic homologues that, like the ‘Gl cyclins’ of budding yeast, can influence the rate of entry into S phase, we will be sign&z&y closer to an under- standing of the relationship between proliferative and cell cycle controls in mammalian cells.

References 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

EVANS T, ROSENTHAL ET, YOUNGBL~M J, DISTEL D, HUNT T: Cy- clin: a protein specified by maternal mBNA in sea urchin eggs that is destroyed at each cleavage division. Cell 1983, 33384-396.

NURSE P: Universal control mechanism regulating onset of M-phase. Nature 1990, 344:503-508.

LEHNER C, ~‘FARREXL PH: Expression and function of Drosophila cyclin A during embryonic cell cycle progres- sion. Cell 1989, 56:957+68.

MINSHULL J, G~L?XEYN R, HILL CS, HUNT T: The A- and B-type eydin associated cdc2 kinases in Xelzopus turn on and off at different times in the cell cycle. E&U?0 J 1990, 92865-2875.

PINES J, HUNTER T: Human cyclin A is adenovirus ElA- associated protein p60 and behaves differently from cycIin B. Nature 1330, 346:76&763.

BLOW JJ, NURSE P: A cdc2Bke protein is involved in the initi- ation of DNA replication in Xenopus egg extracts. Cell 1990, 62:855-862.

GIORDANO A, WHYTE P, HARLow E, FRANZA BR, BEACH D, DRAETTA G: A 6OkD cdc2associated polypeptide complexes with the ElA proteins in adenovirus-infected cells. Cell 1989, 58:981-990.

D’Lkso G, MARRACCINO RL, MA&HAK DR, ROBERTS JM: Cell cycle control of DNA replication by a homologue from human cells of the p34dCz protein kinase. Science 1990, 250:786791.

RICHARDSON HE, WI’MENBERG C, CROSS F, REED Sk An essen- tial Gl function for cyclin-like proteins in yeast. Cell 1989, 59:1127-1133.

WITIENBERG C, SUGIMOTO K, REED SI: Gl-specific eycIin.s of S. cwevisiue: cell cycle periodicity, regulation by mating pheromone and association with the p34coczs protein ki- nase. Cell 1.990, 62:22>237.

Chris Norbury and Paul Nurse, ICRF Cell Cycle Group, Microbiology Unit, Department of Biochemistry, South Parks Road, Oxford OX1 3QU, UK.

24 @ 1991 Current Biology