09 June 1999 MIM upd 1a global.704
MOLECULAR INTERACTION MAP ANNOTATIONS
ATM and c-Abl
A1 ATM is present in a complex that includes c-Abl and Rad51 (Yuan et al., 1998).
A2 ATM phosphorylates c-Abl on Ser465 and thereby activates the kinase (Baskaran et al., 1997).
A3 Phosphorylation by ATM activates c-Abl (Baskaran et al., 1997). c-Abl tyrosine kinase activity is stimulated in response to ionizing radiation, ara-C, camptothecin, or etoposide (Yuan et al., 1996; Yuan et al., 1998).
A4 c-Abl binds to p53 in response to ara-C or MMS (Yuan et al., 1996); this interaction does not require c-Abl kinase activity. Moreover, c-Abl kinase activity does not require p53. Binding of c-Abl to p53 inhibits the Mdm2-mediated degradation of p53 (Sionov et al., 1999).
A5 c-Abl tyrosine-phosphorylates the C-terminal domain (CTD) of RNA polymerase II (Km=0.5mM) (Baskaran et al., 1997; Baskaran et al., 1993); the c-Abl SH2 domain is a specificity determinant for this reaction (Duyster et al., 1995).
A6 The c-Abl C-terminal region can bind the Crk SH3 domain; this interaction may link c-Abl function to the state of the cell surface with respect to integrins and focal adhesions (Gotoh and Broxmeyer, 1997).
A7 c-Abl binds and tyrosine-phosphorylates paxillin in an adhesion-dependent manner (Lewis and Schwartz, 1998).
A8 c-Abl tyrosine kinase activity is blocked by pRb which binds to the c-Abl kinase domain (Welch and Wang, 1995).
A9 Phosphorylation of pRb disrupts the c-Abl:pRb complex and releases active c-Abl (Welch and Wang, 1995).
DNA strand break processing
B1 Ku is a tight heterodimer consisting of Ku70 and Ku80/86
B2 Ku loads onto dsDNA ends and can diffuse along the DNA in an energy-independent fashion (deVries et al., 1989) . Ku can localize internally on dsDNA as well as at dsDNA ends (Yaneva et al., 1997) . Ku binds to ssb (Blier et al., 1993) and has helicase activity, but ssb-bound Ku does not activate DNA-PK (Smider et al., 1998) . Ku can also bind to hairpin-ended DNA without activating DNA-PK (Smider et al., 1998) .
B3 Ku binds to the C-terminal region of DNA-PK (amino acids 3002-3850) near the protein kinase domain (Jin et al., 1997) . DNA-PK can bind weakly and transiently to dsDNA ends without Ku (Lieber et al., 1997; West et al., 1998; Yaneva et al., 1997) . In the presence of Ku, however, the binding is stronger and more stable. DNA-PK and Ku localize adjacent to each other at dsDNA ends. DNA-PK does not bind detectably to Ku in the absence of DNA. There is about 5 times more Ku than DNA-PK in mammalian cells.
B4 The kinase activity of DNA-PK is stimulated by binding to dsDNA ends; however, the stimulation is greater in the complex with Ku (West et al., 1998; Yaneva et al., 1997) .
B5 DNA-PK phosphorylates itself, thereby blocking its interaction with Ku:DNA complex and inhibiting its kinase activity; it also phosphorylates Ku70 > Ku80, but without effect (Chan and Lees-Miller, 1996) .
B6 Autophosphorylated DNA-PK dissociates from Ku:DNA (Chan and Lees-Miller, 1996) .
B7 The SH3 domain of c-Abl binds to the C-terminal region of DNA-PK (amino acids 3414-3850) and may compete with Ku for binding to the same region (Jin et al., 1997; Kharbanda et al., 1997). c-Abl does not bind directly to Ku; however, ionizing radiation induces the association of Ku with c-Abl:DNA-PK complex (Kharbanda et al., 1997).
B8 c-Abl phosphorylates DNA-PK in the C-terminal region (amino acids 3414-3850) (Jin et al., 1997) . c-Abl-dependent phosphorylation of DNA-PK is stimulated by ionizing radiation (Kharbanda et al., 1997). c-Abl does not phosphorylate Ku (Kharbanda et al., 1997).
B9 Phosphorylation of DNA-PK by c-Abl dissociates DNA-PK from Ku:DNA (Jin et al., 1997; Kharbanda et al., 1997) .
B10 HMG1 or 2 compete with Ku for binding to DNA-PK and stimulate DNA-dependent kinase activity in vitro in the absence of Ku (Yumoto et al., 1998).
C1 The Cyclin D1 promoter is activated by E2F4, but it is repressed by E2F1 via pRb (Watanabe et al., 1998). In pRb-deficient cells, E2F1 stimulates this promoter. An Sp1 site close to the E2F element also participates in the regulation. (Overexpression of pRb can increase the expression Cyclin D1 by an unknown mechanism (Watanabe et al., 1998).)
C2 The Cyclin E and A genes (but not the Cyclin D gene) are strongly activated by E2F1 (DeGregori et al., 1995; Shan et al., 1996). Further details about cyclin E promoter regulation have recently been reported (Le Cam et al., 1999), as follows. In addition to a constitutively occupied E2F1-Sp1 site immediately upstream of the cyclin E transcription start region, there is downstream a cell cycle-regulated site (termed CERM) that may function as a cyclin E-repressor module. The CERM contains a variant E2F-recognition element, and binds a complex (termed CERC) consisting of E2F4, DP1, and either p130 or p107, as well as an unidentified necessary component.
C3 Cdk4 and Cdk6 bind exclusively to D-type cyclins.
C4 Cdk2 binds to cyclins E or A.
C5 Cdk1 (Cdc2) binds cyclins A or B.
C6 Cdk1 and 2 bind the small protein, Cks1 (Jackman and Pines, 1997). Cks1 binds at the C-terminal region of Cdk which is distinct from the region that binds cyclins. Cks1 may be involved in the dephosphorylation of Cdk Tyr15.
C7 Low concentrations p21Cip1, p27Kip1, or p57Kip2 promote the binding of Cyclin D to Cdk4 (LaBaer et al., 1997), although high concentrations are inhibitory.
C8 p16ink4a competes with Cyclin D1 for binding to Cdk4 (Hang et al., 1998).
C9 Cyclin D1 has a short half-life (<30 min), regardless whether free or Cdk4-bound. Rapid degradation of Cyclin D1 requires phosphorylation at threonine-286 (kinase unknown, but not Cdk2 or Cdk4); degradation is by way of the ubiquitin-proteasome pathway (Diehl et al., 1997) .
C10 D-type cyclins can bind the myb-like protein DMP1 (Hirai and Sherr, 1996). The binding does not require Cdk4/6 (Inoue and Sherr, 1998).
C11 DMP1 binds consensus sequences CCCG(G/T)ATGT and activates transcription (Hirai and Sherr, 1996).
C11a DNA binding is inhibited when DMP1 is bound to cyclin D (Inoue and Sherr, 1998). Cyclin D binds at the DNA-binding domain of DMP1 immediately adjacent to the myb repeats (Inoue and Sherr, 1998).
C11b DMP1 activates transcription from the p19ARF promoter and induces cell cycle arrest and p21Cip1 accumulation in a p19ARF- and p53-dependent manner (Inoue et al., 1999). Another DMP1-regulated gene is CD13/aminopeptidase N, which is activated cooperatively by DMP1 and c-Myb; its activation by DMP1 is inhibited by cyclin D independent of Cdk4/6 .
C12 When phosphorylated by Cyclin D-dependent kinases, DMP1 activates transcription (Hirai and Sherr, 1996).
C13 Cyclin E has a half-life of about 30 min. It is degraded by way of the ubiquitin-proteasome pathway subsequent to phosphorylation (possibly autophosphorylation) at threonine 380 (Clurman et al., 1996; Won and Reed, 1996) . However, it can be stabilized by binding to Cdk2.
C14 Cyclin H:Cdk7 (also known as CAK for Cdk-activating kinase) phosphorylates a site on the T-loop of Cdk’s (Thr-161 in human Cdk1, Thr-160 in Cdk2, Thr-172 in Cdk4, and Thr-170 in Cdk7 itself) and thereby causes the loop to be displaced so as to allow access to the catalytic site (Morgan, 1995). CAK readily phosphorylates Cdk2 monomer which however remains inactive (Fisher and Morgan, 1994). This phosphorylation is required for the formation of stable Cyclin A:Cdk1 dimer, but not for the formation of other Cyclin:Cdk dimers (Desai et al., 1995; Ducommun et al., 1991). CAK is localized to the nucleus (Tassan et al., 1994).
C15 Cyclin B binds only to Cdk1. Cyclin B1 is retained in the cytoplasm (by means of a cytoplasmic retention sequence) until the time of mitotic prophase when it is abruptly transported into the nucleus (Pines and Hunter, 1994). Cyclin B1 is localized to microtubules and centrosome, while Cyclin B2 localizes to the Golgi complex (Jackman et al., 1995).
C16 Cdk1 is phosphorylated at Tyr15 by Wee1 and at Thr14 by Myt1 ((Liu et al., 1997) and references cited therein). Phosphorylations of Cdk's are facilitated by Cyclin binding and stabilize the Cyclin-Cdk complex (Jackman and Pines, 1997). Human myt1 phosphorylates and inactivates Cdk1 associated with cyclins A or B, but does not phosphorylate Cdk2 or Cdk4 complexes (Booher et al., 1997) (unlike wee1 which can phosphorylate both Cdk1 and Cdk2). Myt1 is membrane-bound to endoplasmic reticulum and Golgi complex. Phosphorylation of Thr-14 and Tyr-15 occurs when the Cyclin B: Cdk1 complex assembles in the cytoplasm (Pines and Hunter, 1994). Wee1 however is in the nucleus; hence another kinase may be operating in the cytoplasm (Matsuura and Wang, 1996). Thr-14 phosphorylation preceeds Tyr-15 phosphorylation (Liu et al., 1997).
C17 Phosphorylation of Thr14 or Tyr15 in Cdk1 reduces kinase activity 10-fold; phosphorylation of both sites reduces activity 100-fold (Liu et al., 1997).
C18 Dephosphorylation of Cdk1 Thr14 and Tyr15 sites is carried out by Cdc25C which must itself be activated by phosphorylation in its N-terminal domain (Jackman and Pines, 1997). Cdc25C activity is high in mitosis during mitosis and low during interphase.
C19 Cdk2 is regulated by phosphorylation at Tyr-15, but there is much less phosphorylation at Thr-14 ((Booher et al., 1997) and references cited therein).
C20 Cdk4 may be inhibited by tyrosine phosphorylation (Terada et al., 1995), but cannot be phosphorylated at the position corresponding to Thr-14 because there is an Ala here rather than Thr.
C21 p27Kip1 can be phosphorylated by Cyclin E or Cyclin A dependent kinases and thereby may be targeted for degradation (Sheaff et al., 1997).
C22 p21Cip1-induced inhibition of Cyclin-Cdk complexes requires the binding of more than one p21Cip1 molecule (Zhang et al., 1994).
C23 Cyclin A:Cdk2, in normal human fibroblasts, exists in complex with p21Cip1 bound to PCNA (Zhang et al., 1993).
C24 Raf1 can bind and activate Cdc25A (Galaktionov et al., 1995; Weinberg, 1995) , perhaps by phosphorylation.
C24a Raf1 is activated by Ras in a complex manner involving phosphorylations, as well as positive and negative effects of 14-3-3 interactions (Roy et al., 1998; Thorson et al., 1998; Tzivion et al., 1998) (details not included in this version of the map).
C25 Cyclin A:Cdk2 is bound to p45Skp2 in complex with p19Skp1 in many transformed cells (Bai et al., 1996; Zhang et al., 1995).
C26 p19Skp1 binds to an F-box motif in p45Skp2 (Bai et al., 1996).
C27 p45Skp2 inhibits Cdk2 kinase activity and blocks phosphorylation of Cdk2 by Wee1 or CAK (Yam et al., 1999). Similarly to p21Cip1, two molecules of p45Skp2 seem to be required to inhibit Cdk2. Binding to p45Skp2 is mutually exclusive with p21cip1. (See Figure for a more detailed model.)
C28 p45Skp2 (Ser76) can be phosphorylated by Cyclin A:Cdk2 (Yam et al., 1999).
C29 p19Skp1, through its F-box motif, may link Cyclin A to the ubiquitin-proteasome protein degradation device (Bai et al., 1996).
C30 Gadd45 binds to Cdk1 and inhibits Cdk1 activity, probably by displacing Cyclin B1 (Zhan et al., 1999). Gadd45 may in this way contribute to the G2 delay response to some types of stress.
C31 Cyclin D:Cdk4 phosphorylates pRb at a subset of sites, P(D), but this does not suffice to abrogate the inhibition of E2F .
C32 Cyclin E:Cdk2 phosphorylates pRb at additional sites, P(E), after the Cyclin D:Cdk4-specific sites have been phosphorylated . (Cyclin E, in addition to acting on pRb, has actions that can induce S phase independent of pRb (Lukas et al., 1997).)
C33 Hyperphosphorylated pRb, resulting from combined phosphorylation by Cyclin D:Cdk4 and Cyclin E:Cdk2, abrogates the binding of pRb to E2F heterodimers .
C34 p21Cip1 binds Gadd45 (Kearsey et al., 1995).
C35 Cdc25A may be transcriptionally activated by c-Myc: the Myc:Max heterodimer binds to elements in the Cdc25A gene and activates its transcription (Galaktionov et al., 1996).
C36 Cdc25C is activated by hyperphosphorylation of the N-terminal domain (Gabrielli et al., 1997) which can be phosphorylated by Cyclin B:Cdk1 (Hoffman et al., 1993).
C37 The Cdc25C N-terminal domain can also be phosphorylated by Plk1 (Hamanaka et al., 1994). During mitosis, Plk1, by way of its polo-boxes, localizes progressively to the centromeres, spindle poles, centrosomes, and spindle midzone/midbody (Glover et al., 1998; Lee et al., 1998).
C38 Chk1 binds and phosphorylates Cdc25A, B, and C in vitro (Sanchez et al., 1997). Cdc25C becomes phosphorylated at Ser216.
C39 Cdc25C Ser216 also binds and is phosphorylated by C-TAK1 (Peng et al., 1998). Cdc25C Ser216 is phosphorylated throughout interphase, but not in mitosis (Peng et al., 1997).
C40 Ser216-phosphorylated Cdc25C is recognized and bound by 14-3-3 protein family members (Peng et al., 1998; Peng et al., 1997). Ser216 phosphorylation and 14-3-3 binding probably sequesters Cdc25C and thus prevents it from interacting with Cdk1 in vivo (Peng et al., 1997). 14-3-3s is localized to the cytoplasm and may be the means by which Cdc25C is sequestered outside of the nucleus (Hermeking et al., 1997).
C41 Cyclin B is degraded late in mitosis through the ubiquitin-protein ligase (E3) activity of the anaphase-promoting complex (APC) which is probably activated by phosphorylation by Plk1 (Glover et al., 1998).\
C42 CycB:Cdk1 phosphorylates and inactivates the promoter selectivity factor SL1 (Heix et al., 1998). This explains in part the silencing of rRNA synthesis during mitosis.
C43 p16 associates with TFIIH and RPase II CTD, and inhibits the phosphorylation of the CTD by TFIIH (Serizawa, 1998).
E1 DP1 or 2 form stable heterodimers with E2F1, 2, 3, 4, 5, or 6 (interactions of E2F complexes have recently been reviewed (Dyson, 1998; Helin, 1998) ).
E2 Unphosphorylated (or hypophosphorylated) pRb can bind to DP complexes of E2F1, 2, or 3, and to a lesser degree of E2F4. Binding of pRb-family members is mediated by a short highly conserved domain in the C-terminal region of E2F proteins. The E2F-binding site is in the C-terminal region of pRb.
E3 E2F4:DP can bind to p107 or p130 or, to a lesser extent, pRb.
E4 E2F5:DP binds only to p130.
E5 E2F1, 2, 3, 4, or 5 complexes with DP1 or 2 can bind to E2 promoter elements, although there may be differences in preferences for variations of the E2 consensus sequence.
E6 DP complexes of E2F1-5 can stimulate promoters containing E2 elements via a potent transactivation domain in the C-terminal region of the E2F component .
E7 E2F-DP dimers, complexed with pRb, p107, or p130, can bind and inhibit E2 promoter elements (Dyson, 1998; Mayol and Grana, 1998). In quiescent cells, the predominant complexes contain E2F4 and p130.
E8 E2F6:DP complexes bind to a variation of the E2 consensus sequence, possibly competing with other E2F complexes (Cartwright et al., 1998) .
E9 In contrast to other E2F species, E2F6:DP directly represses transcription. E2F6 lacks a transactivation domain; it has instead a repression domain in its C-terminal region (Gaubatz et al., 1998) . E2F6:DP represses a subset of E2F responsive genes (Cartwright et al., 1998) . E2F6:DP does not bind to pRb, p107, or p130 (Trimarchi et al., 1998) .
E10 E2F4 is protected against proteasomal degradation when associated with p130 (Hateboer et al., 1996) .
E11 Upon stimulation of quiescent cells by growth factors, p130 becomes hyperphosphorylated and incapable of binding E2F (the responsible kinase is unidentified) (Mayol and Grana, 1998) . Hyperphosphorylated p130 is unstable. Upon growth factor stimulation of quiescent cells, p130 declines late in G1 and is replaced by p107 which is absent in quiescent cells (Mayol and Grana, 1998) .
E12 p130 may associate with HBP1, a transcription factor involved in cell cycle exit during differentiation (Tevosian et al., 1997) .
E13 Histone deacetylase (HDAC1) binds to pocket protein family members pRb, p107, and p130 and is thereby recruited to E2F complexes on promoters (Ferreira et al., 1998) . The binding is via an IXCXE motif in HDAC1 which can bind to the C-terminal region of p130 (Stiegler et al., 1998) and presumably to the LXCXE site on the B-box of pRb (Lee et al., 1998) .
E14 The interaction with HDAC1 enhances transcriptional repression by pocket proteins (Brehm et al., 1998; Ferreira et al., 1998; Luo et al., 1998) .
E15 E2F-regulated genes include many that are involved in cell cycle progression and control. Individual genes are differently regulated. DHFR is activated via the E2F transactivation domain, whereas B-myb, Cyclin E, E2F-1, E2F-2, and Cdc2 are regulated via the repression domain of pRb-family proteins (Dyson, 1998).
E16 pRb binds and activates C/EBP (CCAAT/enhancer-binding proteins) (Chen et al., 1996; Chen et al., 1996) . The binding of C/EBP to its DNA recognition elements is enhanced.
E17 Hypophosphorylated pRb binds c-Jun, JunD, and JunB (Nead et al., 1998) . This enhances the binding of the Jun family members to c-Fos and stimulates transcriptional activation by the Fos:Jun complexes. A region (amino acids 612-657) in the B-pocket of pRb and a region in the C-pocket can independently bind c-Jun. The binding site in c-Jun is in the leucine zipper region.
E18 pRb binds c-Abl via the pRb C-pocket (residues 768-785 and 825-840). pRb can bind c-Abl and E2F simultaneously.(Welch and Wang, 1995; Whitaker et al., 1998). The c-Abl-binding C-pocket and the E2F-binding C-terminal domain of pRb are distinct from each other.
E19 pRb binds Mdm2 via the pRb C-terminal 44 residues (Tan and Wang, 1998; Xiao et al., 1995). These C-terminal residues are not required for the growth suppressive effect of pRb.
E20 CycA:Cdk2 binds to E2F1, 2, or 3 at a site near the N-terminal region, as a consequence of which both the E2F and DP component are phosphorylated. Phosphorylation of either impairs the binding between the E2F and DP monomers (Dynlacht et al., 1997; Krek et al., 1995; Xu et al., 1994).
E21 The p107 promoter contains E2F recognition elements and can be repressed by pRb or p107 (Zhu et al., 1995) .
E22 Raf1 can bind pRb or p130 which are not thereby dissociated from E2F complexes, although promoter inhibition is reversed (Wang et al., 1998) . There was no detectable binding to p107. Binding to pRb is mediated by the N-terminal 28 amino acids of Raf1. The kinase activity of Raf1 was required to reverse the pRb-mediated promoter repression (Wang et al., 1998) , but the phosphorylation sites on pRb remain to be described, and therefore are not indicated in the diagram.
E23 Sp1 cognate elements are found in the promoter regions of several S-phase genes that also contain E2F elements, including DHFR, c-myc, thymidine kinase, cyclin E and E2F1 (Datta et al., 1995). Binding of Sp1 and E2F to the chromatin-organized thymidine kinase promoter was cooperative (Karlseder et al., 1996).
E24 Sp1 and E2F1 bind to each other (Karlseder et al., 1996; Lin et al., 1996; Watanabe et al., 1998). Sp1 binds to the N-terminal region of E2F1; this region is also present in E2F2 and E2F3, but not in E2F4 and E2F5; accordingly, Sp1 can bind E2F1-3, but not E2F4 or 5 (Karlseder et al., 1996). The Sp1-binding region of E2F1 may overlap the cyclin A-binding region. It is however separated from the transactivation and pRb-binding regions which are near the E2F1 C-terminus. E2F1-binding requires the C-terminal region of Sp1 where Zn-fingers are located (Karlseder et al., 1996; Lin et al., 1996). (Sp1 may function as a higher-order complex (see (Karlseder et al., 1996)).) The E2F1-binding region of Sp1 is phosphorylated by an Sp1-associated kinase when quiescent cells are induced to proliferate (Black et al., 1999).
E25 Sp1 and E2F binding sites are both essential for activation of the murine thymidine kinase promoter (Karlseder et al., 1996). The promoter was activated when the Sp1 and E2F sites were separated by 6 or 10 base pairs, but not when they were separated by 20 base pairs. The DHFR promoter was strongly activated by Sp1 alone, but hardly at all by E2F1 alone (co-transfection in insect cells) (Lin et al., 1996). E2F1, however, enhanced the activating ability of Sp1, even in the absence of a functional E2F binding site on the promoter.
E26 Sp1 binds to p107 (within the first 385 amino acids of p107) which is separate from the p107 pocket region that binds E2F4 (Datta et al., 1995).
E27 p107 inhibits Sp1-dependent transcription. Binding of p107 to Sp1 seems to inhibit the binding of Sp1 to DNA (Datta et al., 1995).
Chromatin and acetylase box
H1 Histone deacetylase (HDAC1) removes acetyl groups from histones, thereby making nucleosomes compact and inhibitory to transcription. (I.e., HDAC1 removes acetyl groups that inhibit the inhibitory effect of compact nucleosomes on transcription. Thus there is an odd number (3) of negative effects, which resolves to a net negative effect.)
H2 Gadd45 binds to core histones in chormatin or nucleosomes whose structure has been loosened by acetylation or UV radiation (Carrier et al., 1999).
H3 p300 and CBP have intrinsic histone acetyl transferase activity (Ogryzko et al., 1996).
H4 p300 binds PCAF ((Grossman et al., 1998) and references cited therein).
H5 p300 binds to the transactivation domain of E2F1 (Lee et al., 1998). E2F1 and p53 may be reciprocally regulated by their mutual dependence on coactivation by limiting amounts of p300 (Lee et al., 1998).
H6 The p300 C-terminal region can bind Cyclin E:Cdk2 (Perkins et al., 1997).
H7 p300, via its Cys/His-rich region C/H3, associates with RPase II via the intermediacy of RNA helicase A (RHA) which can bind both RPase II and the C/H3 domain (Nakajima et al., 1997).
M1 c-Myc and pRb compete for binding to AP2 (Batsche et al., 1998) .
M2 AP2 and Max compete for binding to c-Myc (Batsche et al., 1998) . AP2 and Myc associate in vivo via their C-terminal domains (Gaubatz et al., 1995) .
M3 The E-cadherin promoter is regulated via AP2 recognition elements (Batsche et al., 1998; Hennig et al., 1995; Hennig et al., 1996) .
M4 c-Myc and pRb enhance transcription from the E-cadherin promoter in an AP2-dependent manner in epithelial cells (mechanism unknown) (Batsche et al., 1998) . Activation by pRb and c-Myc is not additive, suggesting that they act upon the same site, thereby perhaps blocking the binding of an unidentified inhibitor. No c-Myc recognition element is required for activation of the E-cadherin promoter by c-Myc. Max blocks transcriptional activation from the E-cadherin promoter by c-Myc, presumably because it blocks the binding between c-Myc and AP2.
N1 XPC forms a tight complex with HR23B, a homolog of yeast Rad23. HR23B is present in large excess over XPC (Sugasawa et al., 1997).
N2 The XPC:HR23B complex may be the primary recognizer of a variety of DNA lesions and the initiator of the nucleotide excision repair of non-transcribed DNA regions (Sugasawa et al., 1998). XPC is required to open the DNA to allow access of other repair factors, such as XPA and RPA, to the vicinity of the lesion (Evans et al., 1997). XPC is not required for transcription-coupled repair, perhaps because the lesion-containing DNA region is opened by the encounter with the transcription machinery (Mu and Sancar, 1997; Sugasawa et al., 1998).
N3 XPC is necessary to promote the stable binding of XPA to UV-damaged DNA (Li et al., 1998). XPA binds to DNA and preferentially at sites of bulky damage produced, for example, by UV, cisplatin, or N-AAF. However, the association constant of XPA for UV-irradiated DNA is only several fold above that for unirradiated DNA, suggesting that other factors (such as XPC) may be required for effective lesion recognition (Jones and Wood, 1993; Sugasawa et al., 1998).
N4 RPA binds directly to XPA via the C-terminal region of RPA2 (He et al., 1995; Stigger et al., 1998) .
N5 XPA binds ERCC1 (residues 93-120) (Li et al., 1994) (dissociation constant 2.5x10-7 M (Saijo et al., 1996)).
N6 The XPF C-terminal region (residues 814-905) binds to the C-terminal region of ERCC1 (residues 224-297) (de Laat et al., 1998).
N7 RPA binds XPG and ERCC1:XPF, the nucleotide excision repair endonucleases (He et al., 1995; Matsunaga et al., 1996) (reviewed by (Wold, 1997) ).
N8,9 The ERCC1:XPF heterodimer incises the damaged DNA strand 15-24 nucleotides to the 5’ side of the lesion (Mu et al., 1995). XPG and ERCC1:XPF cut on the 3’ and 5’ side of the lesion, respectively (de Laat et al., 1998) (and references cited therein). RPA binding enhances the activity of XPG and ERCC1:XPF (Matsunaga et al., 1996). RPA is required for the nucleotide excision process (Moggs et al., 1996; Mu et al., 1996) .
N10 RPA binds single-strand regions at locally unwound intermediates in nucleotide excision repair (Evans et al., 1997). The RPA:XPA complex binds cooperatively to DNA damage sites (He et al., 1995).
N11 Rad51-coated ssDNA, together with Rad52 and RPA, stimulate strand exchange in homologous recombination (Bauman and West, 1997; New et al., 1998) (see also S14).
N12 Rad51 binds to c-Abl directly (Yuan et al., 1998). c-Abl phosphorylates Rad51 Tyr54, as a consequence of which the binding of Rad51 to ssDNA is inhibited (Yuan et al., 1998). Ionizing radiation induces c-Abl-dependent phosphorylation of Rad51 (Yuan et al., 1998).
N13 XPG binds to and further stabilizes the open DNA complex with XPA, RPA, and TFIIH; at the same time, XPC:HR23B dissociates from the DNA complex (Wakasugi and Sancar, 1998). XPG and XPC:HR23B appear to be mutually exclusive in binding to complexes containing TFIIH. XPC and XPA, as well as ATP hydrolysis, are required for the stable binding of TFIIH to DNA lesions (Li et al., 1998). RPA (at least the p34 subunit) is present together with XPA and TFIIH in the open complex (Li et al., 1998).
N14 PARP binds to double-stranded DNA termini (Chen et al., 1994), as well as single-strand breaks (Zhan et al., 1994).
N15 DNA-PK is ADP-ribosylated by PARP in vitro and the protein kinase activity (including p53 and RPA substrates) is stimulated thereby (Ruscetti et al., 1998) . (PARP is phosphorylated by DNA-PK (Ruscetti et al., 1998) , but the consequences of this are unknown -- therefore this reaction is not included in the chart.)
N16 PARP is phosphorylated at a serine and a threonine site by PKC a and b in vitro, as a consequence of which PARP binding to DNA is weakened and PARP activity reduced (Bauer et al., 1992) .
N17 PARP binds and stimulates DPase a (Simbulan et al., 1993). PARP was found to be physically associated with the catalytic subunit of the DPase a-primase tetramer during S or G2 phase of the cell cycle, but not in G1 (Dantzer et al., 1998). PARP has been found associated with an active multiprotein replication complex consisting of approximately 40 proteins, 15 of which were found to be poly(ADP-ribosyl)ated, including DPase a, DPase d, topoisomerase I, and PCNA (Simbulan-Rosenthal et al., 1998).
N18 PARP binds XRCC1 by way of their respective BRCT (BRCA1 C-terminus) modules (Masson et al., 1998). This interaction inhibits PARP’s catalytic activity. The PARP:XRCC1 complex is not dissociated by 1 M NaCl.
N19 XRCC1 binds DNA ligase III by way of BRCT modules (Masson et al., 1998). XRCC1 has 2 BRCT modules: one at residues 314-402 binds PARP; the other is at the C-terminus (residues 538-633) and binds DNA ligase III at a C-terminal BRCT domain (residues 841-922) (Masson et al., 1998).
N20 XRCC1 binds DPase b (Masson et al., 1998). XRCC1 can bind simultaneously to PARP, DNA ligase III, and DPase b to form a multi-protein assembly that may function in the short patch pathway of base-excision repair (Masson et al., 1998).
N21 BRCA1 and BRCA2 co-localize with Rad51 in discrete nuclear foci (Chen et al., 1998; Scully et al., 1997). BRCA1 residues 758-1064 bind Rad51 in vitro. In response to DNA damage or drug-induced DNA synthesis inhibition, BRCA1 and Rad51 relocalize to PCNA-containing replication structures; this is accompanied by phosphorylation of BRCA1 (dependent neither on ATM nor on DNA-PK) (Chen et al., 1998; Scully et al., 1997).
P1 Casein kinase 1d and e have been reported to phosphorylate serines 4,6, and 9 in vivo (Knippschild et al., 1997). p53 up-regulates casein kinase 1d (but not 1e) (Knippschild et al., 1997) . Phosphates at N-terminal amino acids 1-15 turn over rapidly (McKendrick & Meek, unpublished, cited by (Meek, 1998) ); it is not known which protein phosphatase is responsible.
P2 ATM phosphorylates Ser15 (but not Ser37) in vitro (Banin et al., 1998; Canman et al., 1998). IR or UV induce Ser15 phosphorylation (Siliciano et al., 1997). In ATM-defective cells, the Ser15 phosphorylation response to IR is delayed, but this response to UV is normal. IR-induced post-translational modification may stimulate p53 transcriptional activity (Siliciano et al., 1997). ATM communicates radiation-induced DNA damage to p53 (Kastan et al., 1992). ATM phosphorylates p53 at Ser15 in response to ionizing radiation (Shieh et al., 1997; Siliciano et al., 1997) . ATM-dependent phosphorylation of Ser15 was observed in response to double-strand breaks produced by microinjected HaeIII, and was required for p53 stabilization (Nakagawa et al., 1999). The ATM-related protein, ATR, can phosphorylated p53 at both Ser15 and Ser37 in response to ionizing radiation or UV (Tibbetts et al., 1999).
P3 Casein kinase 1d phosphorylate Thr18, but only after Ser15 has been phosphorylated (E. Appella, personal communication).
P4 Thr18 phosphate turns over rapidly through the action of an unidentified phosphatase.
P5 Phosphorylation of Ser18 prevents stable binding of p53 to Mdm2 (E. Appella, personal communication). The ability of Mdm2 to inhibit p53-dependent transactivation was reported to be impaired by phosphorylation of Ser15 and Ser37 (Shieh et al., 1997), but this may occur indirectly via stimulated phosphorylation of Ser18 (E. Appella, personal communication). Recent evidence implicates phosphorylation of Ser20 in the inhibition of Mdm2 binding and consequent p53 degradation (Shieh et al., 1999; Unger et al., 1999). These phosphorylation sites are within the region required for Mdm2 binding. Phosphorylation of Ser20 is rapid after ionizing radiation, but delayed after UV (Shieh et al., 1999).
P6 DNA-PK phosphorylates Ser15 and 37 (Shieh et al., 1997). Phosphorylation of p53 by DNA-PK is heavily dependent on the presence of DNA double-strand ends (Shieh et al., 1997).
P7 JNK1 (SAPK1a) phosphorylates only serine 33 ( (Milne et al., 1995) ; Milne & Meek, unpublished, cited by (Meek, 1998). The SAPK-like kinase that phosphorylates murine serine 34 (corresponding to human serine 33) is activated following treatment of cells with UV (Milne et al., 1995). (The kinase is inactivated by a phosphatase -- not shown.)
P8 T73 and T83 were phosphorylated by activated recombinant p42-MAP kinase but not by inactive MAP kinase or by the activating protein, MAP kinase (Milne et al., 1994) . Evidence suggested that this may occur in vivo.
P9 CycA:Cdk2 and CycB:Cdc2 (but not CycE:Cdk2 or CycD:Cdk4) phosphorylate serine 315 in vitro (Wang and Prives, 1995) .
P10 PKC phosphorylates S378 (Takenaka et al., 1995). However, the phosphorylation of p53 by PKC in vitro may not correlate with the ability of PKC to enhance transcriptional activation by p53 (Youmell et al., 1998).
P11 The Cdk7:CycH:p36 (CAK) component of TFIIH phosphorylates p53 in vitro, enhancing its sequence-specific DNA binding activity (Lu et al., 1997). The phosphorylation was within C-terminal region 311-393, which includes serines 371, 376, 378, and 392 as potential sites. The relevant site may be Ser392. This kinase can also phosphorylate Ser33, at least in vitro (Ko et al., 1997).
P12 Ser376 becomes dephosphorylated in response to ionizing radiation, probably by activation of an unidentified phosphatase by ATM (Waterman et al., 1998).
P13 Phosphorylation of Ser378 creates a recognition site for the binding of 14-3-3 (Waterman et al., 1998). It is not known which 14-3-3 family members are involved.
P14 14-3-3 probably binds as a homodimer.
P15 Phosphorylation of Ser376 blocks the Ser378-binding site for 14-3-3.
P16 Binding of 14-3-3 at the Ser378 site enhances the sequence-specific DNA binding of p53 (Waterman et al., 1998), probably by blocking the non-specific DNA binding.
P17 Ser392 phosphorylation increases 10-fold the association constant for tetramer formation by the p53 tetramerization domain (Sakaguchi et al., 1997). Ser315 phosphorylation largely reversed the effect Ser392 phosphorylation. Ser392 phosphorylation activates p53 for specific DNA binding and transcription (Hao et al., 1997; Hupp and Lane, 1994), probably due to stabilization of p53 tetramers. Murine Ser389 (corresponding to human 392) is phosphorylated in response to UV, but not IR or etoposide (Kapoor and Lozano, 1998; Lu et al., 1998).
P18 At the normally low p53 concentration in cells, p53 should be largely monomeric (Sakaguchi et al., 1997). At elevated p53 concentrations and/or increased association constant due to Ser392 phosphorylation, a sharp transition to the tetramer may be expected.
P19 CK2 phosphorylates murine Ser389 (human 392) (Kapoor and Lozano, 1998).
P20 PCAF acetylates p53 Lys320 (Sakaguchi et al., 1998).
P21 p300 acetylates p53 Lys382 (Sakaguchi et al., 1998).
P22 Acetylation of Lys320 or Lys382 enhances sequence-specific DNA binding (Gu and Roeder, 1997; Sakaguchi et al., 1998)
P23 Acetylation of p53 by PCAF is strongly inhibited when Ser378 is phosphorylated (K. Sakaguchi and E. Appella, personal communication).
P24 p300 coactivates p53-activated promoters, including p21, Bax, and Mdm2 (Avantaggiati et al., 1997; Lill et al., 1997; Thomas and White, 1998). It is not known whether histone acetylation contributes to this coactivation. p300, and not its close relative CBP, is involved in this action (Yuan et al., 1999).
P25 The p300 C-terminal region binds the transactivation domain in the N-terminal region of p53 (Grossman et al., 1998; Lee et al., 1998; Sakaguchi et al., 1998). This interaction is involved in the coactivation of p53 by p300, perhaps via acetylation of p53 and/or histones.
P26 A p53 N-terminal region (distinct from that which binds Mdm2) binds to DP1 and competes with E2F1 for binding to DP1 (Sorensen et al., 1996).
P27 p53 C-terminal domain binds the TFIIH-associated DNA helicases, XPD and XPB (ERCC2 and 3), as well as CSB which is involved in strand-specific DNA repair (Lu et al., 1997; Wang et al., 1995).
P28 p53 N-terminal peptide (residues 17-27) binds as amphipathic a-helix in a hydrophobic cleft in the Mdm2 N-terminal region (residues 17-125) (Kussie et al., 1996).
P29 Mdm2 inhibits p53-mediated transactivation (Momand et al., 1992). The binding of Mdm2 to the p53 transactivation domain blocks its interaction with TAFII70 and TAFII31 (discussed by.(Freedman and Levine, 1998)). Thus, in addition to inhibiting gene activation by p53, Mdm2 can suppress the basal transcription of p53-dependent genes.
P30 p53 is short-lived (t1/2 = 5-40 min, depending on cell type) (Reihsaus et al., 1990). p53 degradation occurs by way of ubiquitination and proteasome (Maki et al., 1996).
P31 Mdm2 stimulates p53 degradation (Bottger et al., 1997; Haupt et al., 1997; Kubbatat et al., 1997; Midgley and Lane, 1997), apparently by functioning as an E3 ubiquitin ligase that specifically recognizes p53 (Fuchs et al., 1998; Honda et al., 1997). Mdm2 contains a nuclear export signal, as well as a nuclear import signal; the ability of Mdm2 to shuttle between nucleus and cytoplasm is required for Mdm2-mediated p53 degradation (Freedman and Levine, 1998; Freedman and Levine, 1999; Roth et al., 1998). Mdm2 transports p53 from nucleus to cytoplasm where p53 is degraded (Tao and Levine, 1999).
P32 The Mdm2-induced ubiquitination of p53 is reduced when p53 is phosphorylated at the DNA-PK sites (Ser15 and 37) (Honda and Yasuda, 1999).
P33 The C/H1 domain near the N-terminus of p300 binds both Mdm2 and the core domain of p53 (Grossman et al., 1998). p300 binds independently to both p53 and Mdm2. Much of the Mdm2 in the cell may exist in stable complex with p300. The site of Mdm2 binding is in the vicinity of the p300 C/H1 domain (residues 342-414) (Grossman et al., 1998). The C/H1 region of p300 may function as a platform that positions these molecules to facilitate the Mdm2-dependent ubiquitination p53 (Grossman et al., 1998). (See also P31.)
P34 The p19ARF exon 1b-encoded N-terminal domain binds to the Mdm2 C-terminal region (Zhang et al., 1998). p19ARF binds to Mdm2 in a region that overlaps the p300-binding domain, suggesting that p19ARF may compete with p300 for Mdm2 binding (Pomerantz et al., 1998).
P35 The Mdm2 N-terminus (residues 1-220) associates with E2F1:DP1 and/or pRb and enhances the transcriptional activity of E2F1:DP1 (Martin et al., 1995; Xiao et al., 1995). (There is a sequence resemblance between p53 and E2F1 that may correspond to a common binding site.)
P36 TBP binds to an acidic domain in central Mdm2 (residues 221-272) (Leveillard and Wasylyk, 1997; Thut et al., 1997).:
P37 The C-terminal region of Mdm2 (residues 432-489) binds to an HMG-like region of TAFII250 and activates the Cyclin A promoter (Leveillard and Wasylyk, 1997).
P38 Caspase-3 cleaves Mdm2 at a specific site between the central and C-terminal domains (Chen et al., 1997; Erhardt et al., 1997).
P39 The Mdm2 gene can be activated via a p53-responsive intronic promoter (Juven et al., 1993; Zauberman et al., 1995). This may be part of a negative feedback control of p53 actions (Wu et al., 1993). Mdm2 appears to be expressed in all tissues at all times (Montes et al., 1996).
P40 p19ARF can bind directly to p53 as well as to Mdm2 (Kamijo et al., 1998). Ternary complexes can form with Mdm2 as the bridging molecule.
P41 p19ARF inhibits the Mdm2-dependent ubiquitination and degradation of p53 (Honda and Yasuda, 1999; Pomerantz et al., 1998; Stott et al., 1998). p19ARF and Mdm2 co-localize in nucleoli (Pomerantz et al., 1998). Ectopic expression of p19ARF extended the half-life of p53 from 15 to ~75 min (Kamijo et al., 1998). Zhang et al. reported that p19ARF promotes Mdm2 degradation in HeLa cells (Zhang et al., 1998). Co-expression of p19ARF reduced the half-life of ectopically expressed Mdm2 from 90 min to 30 min. In murine embryonic fibroblasts, however, p19ARF caused endogenous Mdm2 to accumulate (Kamijo et al., 1998; Sherr, 1998), and this action appears to be independent of p53. One way or another, p19ARF antagonizes Mdm2 function by a mechanism that does not require increased Mdm2 degradation (Sherr, 1998).
P42 E2F1 directly activates the p19ARF gene promoter and induces the expression of p19ARF protein (Bates et al., 1998). A potential E2F1 binding site (GCGGGAAA) was noted upsteam from the p19ARF initiation codon. This may be how E2F1 can stimulate p53-mediated apoptosis (discussed by Prives, 1998 (Prives, 1998).
P43 p53 binds to a consensus site in the p21 gene in response to ionizing radiation (Chin et al., 1997; el-Deiry, 1998).
P44 p53 binds to the Gadd45 gene at a site in the third intron where there is a single p53-binding consensus sequence (p53 does not bind to the Gadd45 promoter region) (Kastan et al., 1992). p53 is necessary for induction of the Gadd45 gene in response to ionizing radiation (but not for induction in response to MMS, UV, or starvation medium). Ionizing radiation induced binding of p53 to its consensus site in the Gadd45 gene (Chin et al., 1997). However, p53 can participate in transcriptional induction of the GADD45 promoter in the absence of direct DNA binding (Zhan et al., 1998).
P45 p53 induces the 14-3-3s gene; there is a p53-binding element 1.8 kb upstream from the transcription start (el-Deiry, 1998; Hermeking et al., 1997). 14-3-3s is strongly induced by ionizing radiation, and ectopic expression of 14-3-3s causes a G2-arrest response similar to that produced by p53, suggesting that the 14-3-3s family member is involved in the suppression of Cdc25C (Hermeking et al., 1997).
P46 p53 binds PARP in vitro and in vivo (Vaziri et al., 1997); the p53 N-terminal region (residues 1-72) is mainly responsible. p53 can bind a DNA consensus sequence and PARP simultaneously. p53 can be ADP-ribosylated in vitro, and this p53 modification prevents binding to its consensus sequence (Wesierska et al., 1996).
P47 p53 binds BRCA1; this binding enhances the transcriptional activity of p53 (Ouchi et al., 1998; Zhang et al., 1998). The interacting regions map to BRCA1 residues 224-400 (where there is a BRCT domain) and to the p53 C-terminal region.
P48 c-Fos is transcriptionally activated by p53 through a p53-recognition element located in the first intron of the c-fos gene (Elkeles et al., 1999).
P49 DNA-PK can phosphorylate Mdm2 at Ser17 in vitro and thereby inhibit the binding of Mdm2 to p53 (Mayo et al., 1997).
P50 A functional p53 site resides within intron 2 of KARP1 (Myung et al., 1998). KARP1 mRNA is induced by IR in a p53 and ATM-dependent fashion, and KARP1 protein accumulates after IR. KARP1 is transcribed from the same locus as Ku80. Transcription of Ku80 however is not p53 dependent and not induced by IR . (Ku80 transcription begins within exon 3 of KARP1, the p53 site being approximately equidistant between the KARP1 and Ku80 transcription start sites.) Constitutive expression of Ku80 is normally greater than that of KARP1 by 1-2 orders of magnitude, but p53 greatly increases KARP1 protein levels without much change in Ku80 (Mysung et al., 1997; Myung et al., 1998).
P51 Bax appears to be transcriptionally activated by p53, although this seems to occur only in apoptosis-competent cells (Selvakumaran et al., 1994; Selvakumaran et al., 1994; Zhan et al., 1994).
P52 HMG1 binds p53 and enhances p53 binding to DNA, as well as p53-mediated transcriptional activation (Jayaraman et al., 1998). This may be due in part to HMG1-induced DNA bending and assembly of a nucleoprotein structure containing p53.
R1 p19Skp1 is a subunit of CBF3 which binds centromere DNA (Connelly and Hieter, 1996).
R2 DNA polymerase a-primase may be regulated by phosphorylation of its p68 subunit by Cyclin A:Cdk2 which reduces and Cyclin E:Cdk2 which increases its DNA initiation activity (Voitenleitner et al., 1999).
R3 PCNA (as a trimer) forms a topological ring around dsDNA and functions as a sliding clamp (Tinker et al., 1994) .
R4 Clamp-loader RF-C loads PCNA onto the DNA (Kelman, 1997). RF-C interacts with the C-terminal side of PCNA (Mossi et al., 1997).
R5 PCNA binds to Pol d and enhances the processivity of DNA replication (reviewed by Kelman (Kelman, 1997)).
R6 p21Cip1 (C-terminal domain) binds PCNA and inhibits replication, perhaps by dissociating PCNA from Pol d (reviewed by Kelman (Kelman, 1997)). Assembly of PCNA around DNA by RF-C is not substantially inhibited. The dissociation constant of the p21:PCNA complex is 2.5 nM (Gibbs et al., 1997).
R7 PCNA binds to flap-endonuclease FEN-1 (DNase IV) (Kelman, 1997) . The PCNA:FEN-1 complex processes branched DNA intermediates (Wu et al., 1996).
R8 PCNA binds DNA ligase I; the complex may function to join Okazaki fragments during DNA replication (Levin et al., 1997).
R9 PCNA binds repair endonuclease XPG (Gary et al., 1997).
R10 PCNA binds Gadd45 (Chen et al., 1995; Smith et al., 1994). This interaction enhances nucleotide excision repair. PCNA binds Gadd45 and p21Cip1 competitively.
R11 Cyclin D binds PCNA directly and inhibits PCNA-dependent replication (Matsuoka et al., 1994; Pagano et al., 1994).
R12 p53 binds and transiently transactivates the PCNA promoter (when p53 reaches high levels, transactivation seems to stop) (Morris et al., 1996; Shivakumar et al., 1995; Xu and Morris, 1999).
S1,2 RPA is a tight heterotrimer of p70RPA1, p34RPA2, and p14RPA3 subunits. The RPA2:RPA3 dimer binds to the C-terminal region of RPA1, whereas neither RPA2 nor RPA3 form soluble dimers with RPA1 (reviewed by (Wold, 1997) ). Direct binding has been demonstrated between RPA1 and RPA2, and between RPA2 and RPA3, but not between RPA1 and RPA3 (Kim et al., 1996; Lin et al., 1996) . RPA functionality probably occurs only with intact RPA trimer; binding to RPA2:RPA3 may be required for proper folding of RPA1 (Henricksen et al., 1994; Stigger et al., 1994) .
S3 RPA binds tightly to ssDNA (association constant of 109-1011 M-1) and can also bind to sites of DNA damage where short ssDNA segments may appear due to disturbed base pairing (reviewed by (Wold, 1997) ). Binding may occur to bubbles as small as 4 nt. RPA binds preferentially to pyrimidine-rich sequences. The binding tends to be cooperative. Binding to ssDNA makes RPA a better substrate for phosphorylation (Blackwell et al., 1996; Gomes et al., 1996) . The strongest and probably initial site of ssDNA binding is to the RPA1 central domain (Kim et al., 1996) ; subsequent binding to weaker sites on RPA2 and RPA3 extends the binding from an 8-nucleotide to a longer region (Bochkarev et al., 1997) .
S4 RPA binds DNA polymerase a via the N-terminal half of RPA1, and stimulates the activity of DNA polymerase a/primase and DNA polymerase d (Braun et al., 1997; Wold, 1997). This is at least in part due to stabilization of ssDNA segments during replication, and is helped by the binding of RPA to DNA polymerase a . Although ssDNA binding is primarily via RPA1, stimulation of replication requires the intact RPA trimer.
S5 The C-terminal region of RPA2 also binds the DNA base-excision enzyme, U-glycosylase (Nagelhus et al., 1997).
S6 The N-terminal region of RPA1 binds to the N-terminal region of p53 (amino acids 41-73) (Wold, 1997).
S7 The binding of RPA to p53 inhibits the binding of RPA to ssDNA (Dutta et al., 1993) and the binding of p53 to its promoter sites (Miller et al., 1997).
S8 Human RPA2 can be phosphorylated at several Ser/Thr sites within the N-terminal 33 amino acids (Henricksen et al., 1996) . At least 5 phosphorylation states have been resolved (Zernik-Kobak et al., 1997). In unirradiated cells, RPA2 is phosphorylated primarily at Ser-23 and Ser-29 (Dutta and Stillman, 1992; Pan and Hurwitz, 1993). DNA damage by UV, ionizing radiation, camptothecin, or etoposide induces phosphorylation of additional sites (Shao et al., 1999; Zernik-Kobak et al., 1997). These hyperphosphorylations are attributable to DNA-PK (Shao et al., 1999) and independently to ATM (Gately et al., 1998) .
S9 Cyclin A (but not Cyclin E) binds to RPA2 (Gibbs et al., 1996).
S10 Cyclin A:Cdk2 and Cyclin A:Cdc2 phosphorylate RPA2; the major site for Cdc2 has been mapped to Ser-29 (Niu et al., 1997) .
S11 DNA-PK phosphorylates RPA2 at Thr-21, followed by Ser-33 (Niu et al., 1997) .
S12 ATM (or an ATM-associated kinase other than DNA-PK or ATR) hyperphosphorylates RPA2 on serine and threonine residues in response to DNA damage; the reaction is dependent upon both ssDNA and linear dsDNA (Gately et al., 1998) .
S13 DNA-damage-induced hyperphosphorylation of RPA2 prevents the binding of RPA to p53 (Abramova et al., 1997) .
S14 Human RPA binds Rad52; this interaction may be essential for homologous recombination (Park et al., 1996). The C-terminal domain of RPA2 may be involved in an interaction with Rad52 (Stigger et al., 1998) .
S15 RPA also binds Rad51, but more weakly than Rad52 (Golub et al., 1998); the binding involves the C-terminal region of RPA1 (residues 168-327). In addition, Rad51 and 52 may bind to each other (Milne and Weaver, 1993) [[look up]], suggesting that RPA, Rad51, and Rad52 function in concert in homologous recombination (see also N8). RPA co-localizes with Rad51 in nuclear foci induced by DNA damage (Golub et al., 1998).
S16 DNA-PK can bind to RPA1 (Shao et al., 1999) .
S17 Phosphorylation of RPA2 by DNA-PK impairs the binding of RPA to DNA-PK (Shao et al., 1999) .
S18 KARP1 stimulates DNA-PK activity and may function in complex with DNA-PK and Ku (Mysung et al., 1997; Myung et al., 1998) (see also P50).