Mutant construction

C dclSp possesses a num ler of structural motifs (Fig. 4.3). These include the highly conserved W alker A and W alker B box m otifs w hich are essential for stable binding and hydrolysis o; nucleotides in a num ber of proteins (Walker et al., 1982). The W alker A m otif contiins the consensus sequence GxxGxGKT (as described in C hapter 2). Structural amlysis has show n that this m otif forms a loop that binds nucleotides. The W alker B m otif contains the consensus sequence DExD. The conserved residues play a role in coordinating the m agnesium ion (Guenther et al., 1997; Story and Steitz, 1952). M any of the proteins that contain the Walker A motif are proteins in w hich hydrolysis is coupled to the binding of another molecule, (DNA, RNA, protein). In the RecA protein, ATP and ADP have been found to stabilize different conformations (Fujita et al., 1997; Story and Steitz, 1992). The conformation stabilized by ATP has a m uch higher affinity for DNA. This provides an indication as to how NTP b inding and hydrolysis could relate to cdclSp function. ATP b in d in g could be req u ired for cdclS p b in d in g to DNA and hydrolysis could then result in the release of cdclSp from chrom atin, although w hether this occurs before initiation of replication, acts as the trigger for initiation or occurs later is currently unknown.

The W alker A motif was m utated so the conserved lysine residue was changed to the negatively charged glutam ate residue. This am ino acid substitution had previously been show n to have the m ost dram atic phenotype w hen m ade in the Saccharomyces cerevisiae homologue Cdc6p (Perkins and Diffley, 199S; W ang et al., 1999; W einreich et al., 1999). The conserved glutam ate of the W alker B motif was m utated to a glycine. This m utant h ad originally been identified as a dom inant negative w hen overexpressed in budding yeast (Perkins and Diffley, 199S).

The major role of cdclSp is thought to be to load the MCM family of proteins onto chrom atin (Nishitani et al., 2000). For this to occur cdclSp m ust be localized in the nucleus. Im m unofluoresence data show s th at the overexpressed cdclS protein localizes to the nucleus (N ishitani and N urse, 1995) an d a m ore recent study reports that a m yc-tagged cdclSp expressed at w ildtype levels is located in the nucleus (N ishitani et al., 2000). C dclSp contains two putative bipartite nuclear

Chapter 4 M utational an a ly sis of cdc18

localization signals (Fig. 4.3) which could be responsible for the correct localization of the protein and hence have a direct effect on the ability of cdclSp to perform its designated function. Bipartite nuclear localization signals (NLS) contain two basic residues (lysine or arginine) followed by a spacer of 10 or 11 amino acids followed

by at least 3 basic residues out of the next 5 (Robbins et al., 1991). Cdc2p consensus p h o sp h o ry latio n sites are often found close to b ip artite n uclear localization se q u en c es, in d ic a tin g th a t lo ca liz atio n m ay be in flu e n c e d b y cdc2p p h o sp h o ry latio n (Robbins et al., 1991; Takei et al., 1999). M utating the basic residues of the proxim al arm of the bipartite NLS to neutral residues has been found to exclude the cytokine interleukin-5 from the nucleus (Jans et al., 1997). I therefore m utated the two putative nuclear localization signals as indicated:

NLSl RKRxxxxxxxxxxxKRxK AAAxxxxxxxxxxxKRxK

NLS2 KKxxxxxxxxxxKxxKR AAxxxxxxxxxxKxxKR

I also combined the two m utants in case the sequences acted redundantly.

Finally a m u ta n t w as constructed in w hich all 6 of the cdc2p consensus phosphorylation sites, S/T PxK /R , were m utated so the S /T residue was changed to alanine. Stepwise mutagenesis was perform ed to construct both the NLS1+NLS2 and P l-6 m utants. The plasm ids were sequenced after each round of mutagenesis to confirm that the expected m utation had been successfully introduced w ithout the acquisition of new m utations. Phosphorylation is know n to be required to target cdclSp for degradation via the SCF pathw ay (Baum et al., 1998; Jallepalli et al., 1997; Kominami an d Toda, 1997). I was therefore interested to see w hether a stable cdclSp w ould interfere w ith the norm al alternation of S and M phase.

Chapter 4 M utational an a ly sis of cdc18

T able 4.1

C onstruction o f m utants

M utant Motif Amino acid substitution

WA GxxGxGKT K-^E

WB DExD E-»G

NLSl RKRxxxxxxxxxxxKRxK RKR^AA A

NLS2 KKxxxxxxxxxxKxxKR KK->AA

P l-6 S/T PxK /R S/T-^A

The m utants were constructed by site-directed m utagenesis of the pJK148-gcdcl8- 3HA plasm id as previously m entioned, and were then transform ed into the 81- cdcl8 S/O strain. Transform ants w ere selected by conversion of the 81-cdcl8 S/O strain to leucine prototrophy. To increase the probability of integration at the leul locus, the plasm id can be linearisesd by digesting w ith a restriction enzyme that cuts w ithin the leu l gene. H ow ever, due to the presence of the genomic cd cl8 fragm ent in the plasm id there was no restriction site available that was present only once in the leul gene and not elsewhere in the plasmid. I w as therefore unable to im prove the chance of integration at the leul locus in this way. The presence of the 5.2kb of genomic DNA in the plasm id also provides a large region of homology w ith the cells genomic DNA and produces a second site w here hom ologous recom bination w ould result in integration of the plasm id. These factors m ay have b een resp o n sib le for the low frequency of in te g ratio n at the le u l locus (ap p ro x im ately 1:50 in teg ran ts w ere in te g rate d at the le u l locus). M any transform ants w ere screened by Southern blotting to ensure that the plasm ids w ere integrated at the leul locus.

DNA w as prepared from transform ants and a B am H l digest w as perform ed. This w as followed by Southern blotting. Blots w ere probed w ith cdcl8 and leul (Fig. 4.4). A B am H l digest produces a ban d of 9.1 kb at the cdcl8 locus which should rem ain unchanged. The leul locus corresponds to a 14.8 kb band. On integration of the pJK148-gcdcl8-3HA plasm id at the leul locus, this 14.8kb band is converted to four bands following Bam H l digestion (Fig. 4.4 A). A leul probe m ade w ith only

Chapter 4 M utational an a ly sis of cdc18

the open reading frame recognises a 4.2kb and 17.7kb band (Fig. 4.4 B top panel). W hen the vector pJK148 is integrated as a control, the leul probe still recognises bands of the same size. This is because the cloning of gcdclS into pJK148 deleted a B am H l site l.lk b dow nstream of the leul gene. W ithout the presence of the gcdcl8

fragm ent, this B am H l site still exists, th u s b an d s of 4.2kb an d 17.7kb are recognised, as seen on the leul Southern (Fig. 4.4 B, top panel, lane 2) b u t the 1.4kb an d 2.9 kb b an d s are not pro d u ced as they are specific to the genomic cdcl8 fragment. A cdcl8 probe therefore recognises the 9.1kb ban d at the endogenous locus and the 2.9kb band w hen the m utant is integrated at leul (Fig. 4.4 B, bottom panel, lanes 3 ,4 and 6 to 9).

The exception to this band pattern occurred w hen the W alker B (WB) m utant was integrated (Fig. 4.4 B, lane 5). The leul probe recognised bands of the correct size, as did the cdcl8 probe for the 2.9kb band. This indicated that the pJK148-gcdcl8- 3HA-WB plasm id was integrated correctly at the leul locus. However, there was a change in band size at the endogenous cdcl8 locus, from 9.1 kb to approxim ately 7kb. This could m ean one of two things, either a new B am H l site had been introduced or 2kb of DNA had been deleted w ithin the original B am H l fragment. A 2kb deletion could affect w hether the w ildtype gene w as present and w hether the prom oter w as present that allows the gene to be sw itched off by thiamine. I have therefore tried to remake the strain although so far to no avail. This has been attrib u ted to the low frequency of integration at the le u l locus as previously described. In the m eantim e I have perform ed a characterization of the 81-cdcl8 S/O -W B strain and show n that the full length, w ildtype c d c l8 protein is both produced and switches off in response to thiamine, as expected.