ANÁLISIS FACTORIAL CONFIRMATORIO (VALIDACIÓN DEL

At the core of the cell cycle, controlling the checkpoints are p ro tein kinases know n as cyclin d ep en d en t kinases or cdk's. Their activity is regulated by cyclins expressed at precise periods in the cycle, by site-specific p h o sp h o ry la tio n and d ep h o sp h o ry latio n and b y inhibitory proteins. The cyclin/cdk complexes are protein kinases, able to phosphorylate other proteins in a sequence-specific fashion, and their functional targets are now being characterised.

The cyclins w ere initially identified in invertebrate eggs^^^, as proteins specifically degraded at mitosis. There are now over twelve distinct cyclins in m am m alian cells, w ith varying levels of homology. All contain a 'cyclin box' through which they contact a cdk. The cyclins confer specificity to their partner kinases, acting to activate them (when associated) and de-activate them (when degraded). They also target the kinase w ithin the cell, nam ely to the nucleus (the A-type cyclins) or to the mitotic apparatus (the B-type cyclins)^^^'^^^ w hich m ay be crucial given the sim ilar in vitro consensus target sequence for phosphorylation by m ost cyclin/cdk complexes: (K/R)-S/T-P-X-(K/R) (see Nigg, 1993 for review).

Different cdk/cyclin complexes are active at different phases of the cycle. W hen m am m alian cells are arrested in G l by the rem oval of grow th factors and then stim ulated to progress into the cycle, then the first cy clin /cd k complex to be activated is generally a D-type cyclin b o u n d to cdk4 or cd k6^^^'^^^. D-cyclins are typically expressed th ro u g h o u t the cycle, and dow n-regulated only w hen the su p p ly of grow th factors is stopped (see Sherr, 1993 for review). The function of cyclin D1 is m ost likely to be required in early to m iddle G l since cyclin D1 antibodies prevent S-phase entry^^®'^^^ w hen applied p rior to the restriction point b u t have no effect subsequently. It is likely therefore th at cyclin D1 complexes phosphorylate a substrate required for Gl-S transition, m ost probably the retinoblastom a (Rb) protein since in cells lacking Rb, D-type cyclins are not required for progression through the restriction point and p l6 (a cyclin D l/c d k inhibitor protein, see below) does not cause cell cycle arrest^^^'^^^.

Cyclin E, which like the D-type cyclins can com plem ent for G l cyclin m u tatio n s in yeast, is also expressed early after m itogen application in quiescent cells. In proliferating cells, cyclin E expression is periodic and regulates the Gl-S transition at w hich point its levels are maximal^^®'^^^. It is associated predom inantly w ith cdk-2, a kinase w hich is in a genetically different class to cdk4 or cdk6 (since these cannot complem ent cdc2 in yeast while cdk2 can)^^®"^^^.

As well as being a target for the cyclin D complexes, Rb is also a likely target for cyclin E complexes, and is phosphorylated on a num ber of cdk consensus sites during G l. W hen phosphorylated, Rb releases proteins bound to it such as E2F-1 (for review see Lam and La Thangue, 1994233 Nevins, 1994^^^), a transcription factor that regulates genes required for S-phase.

The cyclin E gene is up-regulated by E2F and so a positive feedback system operates whereby cyclin E-dependent phosphorylation of Rb releases E2F and drives cyclin E expression. In this w ay, initial p h o sp h o ry latio n of Rb by cyclin D com plexes, w hich is m itogen- d e p en d e n t, becom es m ito g en -in d ep en d en t as cyclin E expression becom es established. Inactivation or absence of Rb reduces a celTs grow th factor requirem ents b u t does not elim inate them and so there m ust be additional levels of restriction point control^^^'^^^'^^®. Proteins related to Rb and E2F-1 have since been identified, and are classified into the Rb and E2F families respectively. The other Rb-family m em bers (pl07 and pl30) which are also putative cdk/cyclin targets seem to bind to and inactivate other E2F family members such as E2F-4. It is unlikely th at the activities of the E2F proteins are controlled solely by the phosphorylation state of the Rb-family members: the phosphorylation tim ing of pl30 is subtly different to that of Rb^^^ and p l0 7 doesn't seem to be regulated by phosphorylation. It has been suggested th at p l0 7 expression levels m ay control E2F-4 availability^'^^'^'^^.

After, or just prior to, cells entering S-phase, cyclin E is degraded and cyclin A is produced, w ith w hich cdk2 now complexes. Cyclin A /c d k 2 can p h o sp h o ry late DP-1 (another E2F fam ily m em ber), abolishing DP-1/E2F-1 DNA binding which m ay be the m echanism for turning off the cyclin E autoregulatory loop. Later in the cycle, cyclin A binds to cdk-1 (originally know n as cdc-2) and is then degraded prior to division.

Cyclin B/cdk-1 controls division, and the destruction of cyclin B at the end of M-phase is required before exit to the next G l. D egradation at M -phase is conferred by a destruction box^^^ th o u g h only cdk-1 associated cyclins are destroyed^'^'^'^'^^. The mitotic cyclin destruction system is inactivated by G l cyclin/cdk activity w hich ensures th at in

each ro u n d of the cell cycle, G l cyclins m ust accum ulate before the mitotic cyclins^^^.

O ther cyclin/cdk pairs also have im portant roles related to cell cycle control, especially cdk7(M 015 or C A K )/cyclin H^47-25i

regulates the phosphorylation and hence activity of other cdk's. A num ber of other cyclins and cdk's have also been described^^^'^^. The properties of the m am m alian cdk/cyclins are sum m arised in the table below (Fig. 1.3 after Pines^^®).