The role of the N-terminal acetyltransferase NatA in transcriptional silencing in Saccharomyces cerevisiae Antje Geißenhöner 111.1231.2451.367293#30481.49456#20910111.512491#18213141.6151617475#2981819305#1792021485#1551.7222324251.826272282.12.1.12.1.229302.1.32.1.431322.1.52.1.62.22.2.1332.2.234352.2.3362.2.42.2.52.2.637383393.1369#210403.23.2.141428#334423.2.2433.2.344592#659453#48446436#274473.2.43.2.548209#773.2.649404#2683.2.750113#703.33.3.151492#2523.3.252396#214533.3.3410#216543.3.4254#1243.3.555348#1123.3.656534#1113.43.4.157363#1493.4.258251#2213.53.5.1593.5.260528#315613.5.362367#2184634.16465664.267505#3246869208#1584.370714.472 References73747576777879808182838485868788899091 Abbreviations92 Curriculum vitae93 Publications94 Acknowledgements9596enTable of contentsHelp Results
Nat1 was required for repression of the HM loci, telomeres and the rDNA locus

The deletion of NAT1 was previously described to cause pronounced derepression at the natural HML locus and at marker genes inserted in subtelomeric regions (Fig. 3.1A, 3.1C) (Mullen et al. 1989) (Aparicio et al. 1991). In contrast, due to functional redundancy within the HMR-E silencer, wild-type HMR is not affected by nat1∆ unless it is weakened by the deletion of the Rap1 binding site (Stone et al. 1991). To further evaluate the role of NatA in silencing, we tested its effect on the synthetic HMR-E silencer (HMR SS ΔI). This silencer variant consists solely of minimal binding sites for ORC, Rap1 and Abf1 and lacks much of the functional redundancy of natural HMR (McNally and Rine 1991). Significantly, nat1∆ caused complete derepression at HMR SS ΔI, as monitored by the loss of mating ability due to the coexpression of a information in the MATα strain (Fig. 3.1A). This supported the notion that NatA had a function at HMR that was masked by the functional redundancy of the natural HMR.


Abb. 3.1: NatA activity was required for HM, telomeric and rDNA silencing.
(A) The deletion of NAT1 resulted in derepression of HML and HMR SS ΔI, as measured by the reduced mating ability of MATa and MATα strains, respectively. Patch-mating assays were performed with MATa strains AEY2 (WT) and AEY80 (nat1Δ), and MATαHMR SS ΔI strains AEY5 (WT) and AEY1273 (nat1Δ). (B) Silencing of MET15 inserted into the rDNA locus was impaired by nat1Δ, as indicated by the brighter colony color of strain AEY 2786 (nat1Δ) compared to AEY160 (WT) on lead indicator medium. (C) Silencing of URA3 inserted near the left telomere of chromosome VII depended on functional NatA. Serial dilutions of strains AEY1017 (WT) and AEY2371 (nat1Δ) were assayed on 5-FOA containing medium counterselecting for URA3-expressing cells.

We next asked whether NatA also functioned in rDNA silencing. To this end, we tested the effect of nat1Δ on the expression of a MET15 reporter gene integrated at the rDNA locus, whose expression can be monitored on lead indicator medium (Smith and Boeke 1997). nat1Δ strains showed a brighter colony color than wild-type strains on this medium, indicating that MET15 was derepressed by nat1Δ (Fig. 3.1B).

Together, NatA functioned in all forms of silencing in S. cerevisiae, suggesting that one or more silencing factor(s) common to all three silenced regions is the target of NatA.

Orc1 required Nα-acetylation by NatA for its function in telomeric silencing Tethering of Orc1 or Sir1 to the silencer bypassed the requirement for NatA in silencing

The involvement of NatA in all three classes of silencing in yeast indicated that one or more silencing factors common to all silenced loci depended upon Nα-acetylation for proper function. In order to narrow down the number of potential candidates, we sought to genetically characterize the precise role of nat1∆ in silencing. We first asked through which of the HMR-E silencer elements nat1∆ functioned. For these experiments, we exploited the fact that derepression at natural HMR requires the loss of at least two of the three silencer elements ORC, RAP1 and ABF1. This can be achieved either by deleting the binding site in cis, or by mutating the respective protein in trans. We reasoned that measuring the effect of nat1∆ on individual cis deletions would indicate which trans factor it affected. Interestingly, silencing was completely abrogated in nat1Δ strains with HMR-E lacking the Rap1 binding site, thereby suggesting ORC or Abf1, but not Rap1, as NatA targets (Fig. 3.2A). In contrast, nat1Δ did not cause significant derepression when the ORC or Abf1 binding sites were deleted, showing that NatA functioned via these elements (Fig. 3.2A). Since the Abf1 binding site plays a minor role in silencing and the penultimate amino acid of Abf1 is an aspartate, which makes it unlikely to be a NatA substrate, we focused on ORC and asked whether it was a target of NatA.

We therefore sought to dissect through which of the six ORC subunits NatA functioned in silencing. For these experiments, we took advantage of the fact that silencing at HMR can be achieved by replacing the ORC binding site of the synthetic HMR-E silencer by unrelated Gal4 binding sites and expressing fusions of the ORC subunits or of Sir1 to the Gal4-DNA binding domain (Fox et al. 1997). This so-called tethered silencing approach circumvents the functional complexity of silencing and allowed us to dissect the contributions of the individual ORC subunits to NAT1-dependent silencing. Tethering of Gal4-Sir1 bypasses the requirement for ORC in silencing (Fox et al. 1997), which supports the notion that ORC recruits Sir1 to the silencer. Importantly, Gal4-Sir1 mediated silencing was independent of NAT1 (Fig. 3.2B), indicating that NatA functioned upstream of Sir1, and hence through ORC, in silencing.


Fig. 3.2: The silencing function of NatA was genetically linked to ORC1.
(A) The deletion of the binding site for Rap1, but not for ORC or Abf1, from HMR-E disrupted HMR silencing in nat1Δ mutants. HMR silencing was tested by the α-mating ability of wild-type and nat1Δ strains with HMR-E lacking the binding site for ORC (AEY84, AEY2146), Rap1 (AEY81, AEY2144) and Abf1 (AEY71, AEY2148). Results from quantitative mating assays are given relative to a value of 1.0 for AEY2. (B) Tethered silencing by Orc1, but not the other ORC subunits was independent of NAT1 and required SIR1. In MATα strains AEY1275 (WT), AEY1276 (nat1Δ) and AEY 2947 (nat1Δsir1Δ), the ORC binding site of the synthetic HMR-E silencer was replaced by five Gal4-binding sites (HMR SSΔI, 5xGal4-RAP-ABF). The strains carried plasmids encoding the Gal4 DNA binding domain fused N-terminally to Orc1 (5-267aa) (pAE408), Orc2 (pAE108), Orc3 (pAE595), Orc4 (pAE597), Orc5 (pAE109), Orc6 (pAE516) and Sir1 (pAE100) and were tested for HMR silencing in patch-mating assays.

We next tested whether the tethering of individual ORC subunits required NAT1 to establish silencing. The rationale of these experiments was that if N-terminal acetylation were required for an ORC subunit, direct tethering of this subunit to the silencer by an N-terminal fusion to Gal4 would relieve its requirement for NatA. Significantly, we found that tethered silencing of all subunits except the Orc1 N-terminus (amino acids 5 to 267) was disrupted in nat1Δ strains (Fig. 3.2B), whereas tethered Orc1 was able to provide silencing in the absence of NAT1. Interestingly, this silencing still depended upon Sir1, since the NatA independent Gal4-Orc1 mediated silencing was abrogated in a sir1∆ strain (Fig. 3.2B). These observations indicated that Orc1 needed the N-terminal acetylation in order to fulfill its function in silencing and that the acetylation did not affect Orc1’s ability to interact with Sir1. Consistent with this, Orc1 carries an alanine at the penultimate position, making it a likely candidate for Nα-acetylation by NatA.

Orc1 was N-terminally acetylated by NatA

Since the above genetic experiments strongly suggested Orc1 as a silencing-relevant substrate of NatA, we directly tested whether Orc1 was N-terminally acetylated in a NatA dependent fashion. For this purpose, a fusion of the first 250 amino acids of Orc1 to the Tandem Affinity Purification (TAP) tag (Orc1-TAP) was introduced into wild-type and nat1Δ strains. The TAP tag allows the fast and simple purification of large amounts of the tagged protein by three successive steps: affinity chromatography on IgG agarose is followed by tobacco etch virus (TEV) protease cleavage and purification with calmodulin-coated beads (applied below) (Rigaut et al. 1999).

Since Nα-acetylation shifts the isoelectric point (pI) of a given protein towards a more acidic pH (Kimura et al. 2000), we used isoelectric focussing gels to determine whether nat1∆ altered the pI of Orc1-TAP. Significantly, Orc1-TAP migrated at a more basic pI when isolated from a nat1∆ strain as compared to a wild-type strain (Fig. 3.3A), suggesting that Orc1 was acetylated by NatA.

It has previously been proposed that NATs can also provide ε-N-acetylation (Polevoda and Sherman 2003a). Therefore, to test whether the IEF band shift corresponded to Nα-acetylation of Orc1, we used mass spectrometry to measure differences in acetylation in N-terminal peptides derived from Orc1-TAP that was isolated from wild-type or nat1∆ strains. Acetylation extends the mass of NAT substrates by 42 Dalton (Da), which is the size of the bound acetyl group (Polevoda and Sherman 2001). Orc1‑TAP samples purified with the TAP protocol from wild-type or nat1∆ strains were digested individually with AspN and GluC endopeptidases in order to obtain N-terminal peptides of a suitable size. We obtained a set of two different protein solutions of the wild-type and the nat1Δ derived samples, which were examined in independent experiments. In the subsequent analysis, the measured mass of the N-terminal peptide from the wild-type and the nat1Δ probe was compared to the calculated value on the basis of the amino acid sequence (Fig. 3.3B).

In the AspN as well as the GluC cleaved sample, the measured mass of the wild-type N-terminal peptide was larger by 42 Da than the calculated value (Fig. 3.3C). However, in both cases this size increase was not found in the nat1Δ strain (Fig. 3.3D). Furthermore, neither the wild-type nor the nat1Δ strain-derived N-terminal fragments matched the calculated mass of a peptide containing the initial methionine (Fig.3.4A). This supported the notion that the initiator methionine was removed from proteins with alanine at the penultimate position.

The mass 560.47 of the AspN-cleaved nat1Δ probe was assigned to the N-terminal peptide AKTLK. To further verify this assignment, the peptide was sequenced by Post-Source Decay MALDI analysis (Chaurand et al. 1999) (Fig. 3.4B). Here, the peptide was degraded into fragments containing different numbers of amino acid residues and the fragment spectrum was recorded. The joined fragment data resulted in the sequence of the complete peptide AKTLK and thus confirmed it to be the unmodified form of the N-terminal peptide of Orc1.

In summary, the mass spectrometric data demonstrated that Orc1 was N-terminally acetylated in the presence of Nat1 and not acetylated in its absence, strongly suggesting that is was a direct target of NatA.

Mass spectrometry was performed by Christoph Weise (FU Berlin).

 Unacetylated orc1 mutants displayed telomeric derepression

We next asked whether the observed N-terminal acetylation of Orc1 was of significance for its silencing function. To this aim, we generated orc1 alleles in which the penultimate amino acid was changed from alanine to valine or proline, and tested their effect on silencing. Proline as well as valine promote the cleavage of the initiator methionine, but prevent N‑terminal acetylation (Huang et al. 1987). In order to test whether the respective mutants were acetylated or not, we tested the isoelectric properties of the TAP variants Orc1‑A2P and Orc1‑A2V that were constructed analogous to wild-type Orc1‑TAP. Significantly, the isolelectric point of Orc1‑A2P‑TAP and Orc1‑A2V‑TAP was at a more basic pH than wild-type Orc1, although the calculated pI was roughly the same for all Orc1 versions (Fig. 3.3A). The shift was comparable to that of wild-type Orc1‑TAP in the nat1Δ background, showing that the mutations to valine or proline had abrogated the ability of Orc1 to be acetylated by NatA.

We then asked whether these mutations had an impact on telomeric silencing, since the deletion of NAT1 strongly affects silencing of subtelomeric genes (Fig. 3.1C).


Fig. 3.3: Orc1 was N-terminally acetylated by NatA.
(A) The isoelectric point (pI) of the Orc1 N-terminus shifted to a more basic pH either by the deletion of NAT1 or by the mutation of the penultimate residue alanine to valine or proline. Whole cell protein extracts of strains AEY2719 (WT), AEY2758 (nat1Δ), AEY3107 (orc1-A2P) and AEY3110 (orc1-A2V) were applied to IEF and SDS gels. TAP-tagged Orc1 (amino acids 1-250) was detected in subsequent immunoblots using the PAP antibody. The faster migrating band in the SDS gel was identified as Orc1 by MALDI-TOF analysis and probably is a proteolytic fragment. (B) Theoretical molecular mass of N-terminal peptides of Orc1 generated by proteolysis with AspN or GluC endopeptidase. The molecular mass as calculated using (http://us.expasy.org/tools/peptide-mass.html) increases by 42 Da due to Nα‑acetylation. (C) MALDI time-of-flight mass spectra of Orc1-TAP derived from a wild-type, but not from a nat1Δ strain, identified the mass of an acetylated N-terminal peptide of Orc1. Orc1-TAP was purified for MALDI-TOF analysis from AEY2719 (WT) and AEY2758 (nat1Δ). Data obtained from the AspN and GluC cleaved samples were consistent for each strain with minimal differences to the theoretical value due to the precision of measurements. (D) The MALDI-TOF spectrum of Orc1-TAP from the nat1Δ strain, but not from wild-type strain, contained the mass of an unacetylated N‑terminal Orc1 peptide. Analysis was performed as in Fig. 3.3C.

Fig. 3.4: The N-terminal peptide of Orc1, whose identity was verified by fragmentation, lacked the initial methionine.
(A) A mass corresponding to the Orc1 N‑terminal peptide including the initial methionine was detected neither in the wild-type nor in the nat1Δ derived probe. MALDI-TOF spectra of AspN cleaved Orc1-TAP were obtained as in Fig. 3.3C. The result was confirmed by the data of the GluC cleaved samples (not shown). (B) The sequence of the nat1Δ-derived 560.47 Da peptide corresponded to the N-terminus of Orc1. The peptide was sequenced by fragmentation in post-source decay MALDI analysis. The detected N-terminal sequence ions AK (b2=200), AKT (b3=301), AKTL (b4=414), and C-terminal sequence ions K (y1=147), LK (y2=260), TLK (y3=361) and KTLK (y4=489) added up to the amino acid sequence AKTLK of the Orc1 N‑terminus.

For this purpose, we monitored the repression of an URA3 reporter gene inserted in the subtelomeric region of chromosome VII‑L (Gottschling et al. 1990). Comparable to nat1Δ, orc1‑A2P and orc1‑A2V caused a strong derepression of the subtelomeric URA3 reporter as indicated by diminished growth on URA3-counterselective 5‑FOA medium (Fig. 3.5A). This showed that the loss of N‑terminal acetylation of Orc1 compromised its function in telomeric silencing


Fig. 3.5: Nα-acetylation of Orc1 was essential for telomeric silencing.
(A) A URA3 gene inserted near the left telomere of chromosome VII was derepressed in unacetylated orc1‑A2P and orc1‑A2V mutants. In these mutants as well as in the nat1Δ mutant, the telomeric effect was not suppressed by the overexpression of SSB1. URA3 expression was tested in serial dilution assays of strains AEY1017 (ORC1), AEY3038 (orc1‑A2V), AEY3105 (orc1‑A2P), and AEY2371 (nat1Δ) on 5-FOA containing medium. For SSB1 overexpression, strains were transformed with pAE964. (B) The loss of Nα‑acetylation of Orc1 did not impair silencing of HML and HMR SS ΔI. Patch-mating assays were performed to test HML silencing using MATa strains AEY2867 (ORC1), AEY3102 (orc1‑A2P), AEY2913 (orc1‑A2V), and AEY2912 (nat1Δ), and to test HMR SS ΔI silencing using MATα strains AEY2866 (ORC1), AEY3103 (orc1-A2P), AEY2903 (orc1-A2V), and AEY2916 (nat1Δ). (C) nat1Δ, but not unacetylated orc1, caused the slight derepression of ADE2 inserted at the HMR locus. Serial dilutions of strains AEY743 (WT), AEY3101 (orc1‑A2P), AEY2721 (orc1‑A2V) and AEY3109 (nat1Δ) were grown on medium lacking adenine.

We next tested whether this defect was suppressed by SSB1 overexpression. In the nat1Δ mutant, defective HML silencing and temperature sensitivity were suppressed by overexpression of the gene encoding the ribosome-bound chaperone Ssb1 (Gautschi et al. 2003). This Hsp70 homolog (and the 99% identical Ssb2), like NatA, is located close to the tunnel exit of the large ribosomal subunit and cross-links to a variety of nascent polypeptides (Pfund et al. 1998). Since Ssb1 is assumed to prevent misfolding of newly synthesized proteins, its ability to suppress nat1Δ defects suggested that the nat1Δ phenotype derives from disturbed protein folding rather than decreased protein stability. As shown in Fig. 3.5A, telomeric silencing was not increased upon SSB1 overexpression in the unacetylated orc1 mutants or in the nat1Δ strain.

Though we do not understand why SSB1 overexpression does not suppress the telomeric silencing defect of nat1Δ, it prompts the presumption that this silencing defect may not be the result of impaired protein folding of Orc1.

 HM silencing was not affected by the lack of N-terminal acetylation of Orc1

We next tested whether HM silencing was also impaired by the lack of N-terminal acetylation of Orc1. However, in contrast to the strong defect caused by nat1Δ,no effect was detectablein the orc1‑A2P and orc1‑A2V mutants at HML and the synthetic HMR SS ΔI (Fig. 3.5B). In addition, the mutants showed no derepression of the sensitive ADE2 reporter inserted at HMR, whereas nat1∆ caused a slight effect in this context (Fig. 3.5C). One possible explanation for this result is that HM silencing is more robust than telomeric silencing and thus is less sensitive to the orc1 mutations. Furthermore, this suggested that more NatA silencing targets exist in HM silencing.

Nα‑acetylation was not required for the protein stability of Orc1

Among the known NAT substrates, some require the Nα-acetylation for protein stability. For example, there is evidence that the half-life of non-acetylated α‑MSH in rabbit plasma is one-third of that of the acetylated form (Rudman et al. 1983).

In order to test whether Nα‑acetylation was required for the protein stability of Orc1, we compared the abundance of HA‑tagged Orc1 in wild‑type and nat1Δstrains by Western Blot analysis. Since similar amounts of Orc1 were present in whole cell protein extracts of both strains (Fig. 3.6), we concluded that it was not destabilized by the loss of Nα‑acetylation. This result was consistent with the observation of (Mayer et al. 1989) that Nα‑acetylation has no general protection function, since it does not prevent proteins from degradation by the ubiquitin system.


Fig. 3.6: Orc1 was present in equal amounts in a wild-type and a nat1Δ strain.
Whole cell protein extracts of strains AEY3068 (NAT1) and AEY3070 (nat1Δ) expressing HA-tagged ORC1 were loaded on a SDS gel as samples of 22 μg (lanes 1 and 4), 11μg (lanes 2 and 5), and 5.5μg (lanes 3 and 6) protein. HA-tagged Orc1 was detected in a subsequent Western blot using an α‑HA antibody.

NatA activity, but not Nα-acetylation of Orc1, was required for replication

Since the ORC complex functions as the eukaryotic replication initiator, we further asked whether Nα‑acetylation of Orc1 was relevant for its replication function. We therefore tested orc1‑A2P and orc1‑A2V strains for temperature sensitivity, a phenotype that is associated with replication defects in orc2‑1 and orc5‑1 mutants (Loo et al. 1995a). Both unacetylated orc1 mutants grew as well as wild-type strains and were not temperature sensitive, suggesting that replication was not affected (Fig. 3.7A). Therefore, the temperature sensitivity of the nat1Δ strain appeared to be based on other defects than the missing Nα‑acetylation of Orc1.

In order to further evaluate functional links between NatA and the ORC complex, we next investigated genetic interactions between nat1Δ and orc2‑1. Interestingly, we found that nat1∆ orc2‑1 double mutants were unable to survive. In crosses between nat1∆ and orc2‑1 strains, double mutant segregants did not grow up except for a few cases, where pinprick colonies appeared after prolonged incubation, but which were unable to form colonies when restreaked (Fig. 3.7B). In addition, the viability of orc2‑1 nat1Δ double mutants was dependent on the presence of an Orc2 encoding plasmid (Fig. 3.7C). Since orc2‑1 affects replication, our results suggested that nat1∆ compromised replication even further such that the double mutants were unable to replicate. In summary, we found that the replication function of the ORC complex, but not of its subunit Orc1, was genetically linked to NatA activity.


Fig. 3.7: nat1Δ affected the replication function of the ORC complex independently of Orc1.
(A) Unacetylated orc1 mutants were not temperature-sensitive and thereby differed from nat1Δ. Serial dilutions of strains AEY2866 (ORC1), AEY 3103 (orc1‑A2P), AEY2903 (orc1‑A2V) and AEY2916 (nat1Δ) were grown for two days on complete medium at the indicated temperatures. (B) orc2‑1 nat1Δ double mutants were not viable. orc2‑1 and orc2‑1 nat1Δ segregants from an orc2‑1 nat1Δ double heterozygous cross (AEY24 crossed with AEY1227) were grown for five days on complete medium at 23°C. (C) Viability of the orc2-1 nat1Δ double mutant was rescued by plasmid-borne ORC2. AEY3161 (orc2‑1 nat1Δ pURA3‑ORC2) transformed either with pJR1818 (pHIS3‑ORC2) (Fox et al. 1997) or with pRS313 (vector) was tested for ORC2 dependence by counterselection for pURA3‑ORC2 on 5‑FOA medium.
Figures B and C are courtesy of Ann Ehrenhofer-Murray.

Synthetic lethality between nat1∆ and SUM1-1 was suppressed by orc1∆1‑235

We next sought to determine the role of NatA in SIR independent, SUM1‑1 dependent silencing. However, in a set of genetic crosses in which nat1Δ and SUM1‑1 segregated, we observed synthetic lethality between nat1∆ and SUM1‑1 (Fig. 3.8). The segregation of the unmarked SUM1‑1 mutation was determined by following sum1∆::URA3 in the segregants from sum1∆::URA3/ SUM1‑1 heterozygous diploids. Interestingly, nat1∆ was not synthetically lethal with sum1∆ (data not shown), suggesting that the lethality was due to novel properties of the mutant Sum1‑1 protein.

Since Sum1‑1 has been shown to interact with the N-terminus of Orc1 and because NatA acetylates this very N-terminus, we hypothesized that the lethality may be connected to the lack of Orc1 acetylation. The ability of Sum1‑1 to function in silencing is abrogated by the deletion of amino acids 1 to 235 of Orc1 (Rusche and Rine 2001). Hence, we tested whether this deletion also abrogated the synthetic lethality of SUM1‑1 with nat1∆.


Fig. 3.8: SUM1-1 nat1Δ double mutants were inviable.
SUM1-1 nat1Δ segregants of tetrads dissected from a cross between SUM1‑1 (AEY1224) and nat1Δ (AEY3008) are marked by arrows. Figure is courtesy of A. Ehrenhofer-Murray.

Significantly, strains with orc1Δ1‑235 as the sole source of Orc1 that were both nat1∆ and SUM1‑1 were readily recovered from a cross and showed normal growth characteristics (data not shown; see Materials and Methods for experimental details). Thus, the synthetic lethality of nat1∆ with SUM1‑1 was abrogated by deletion of the N-terminus of Orc1.

Data from 3.2.6 and 3.2.7 are courtesy of Ann Ehrenhofer-Murray.
N-terminal deletions of Orc1 caused silencing defects distinct from those of nat1Δ HMR silencing was disrupted in N-terminally truncated orc1 mutants

In previous studies, the silencing function of Orc1 has been shown to depend on the N‑terminal region of 235 amino acids, which is capable of binding to Sir1 (Bell et al. 1995). Zhang and colleagues (2002) specified the Sir1 interacting domain of Orc1 to lie within amino acids 100 and 129. Since mutations of the penultimate amino acid of Orc1 affected telomeric silencing in our studies, we wished to determine the functional relevance of the N‑terminal 100 amino acids of Orc1. To this end, we constructed a series of orc1 mutants with N‑terminal deletions of increasing size. Strains with orc1Δ1‑10, orc1Δ1‑28, orc1Δ1‑51 and orc1Δ1‑100 as the sole source of Orc1 were obtained by inserting the respective mutant allele into the LEU2 locus of a strain whose endogenous ORC1 gene was disrupted.

Silencing in these mutants was first tested at different sensitized HMR versions. ADE2 inserted at HMR was silenced in orc1Δ1‑10 and orc1Δ1‑28 mutants, but was expressed upon the deletion of 51 or 100 residues from the Orc1 N-terminus (Fig. 3.9A). The phenotypic difference between orc1Δ1‑28 and orc1Δ1‑51 caused us to examine whether the region between 28 and 52 amino acids was of special significance for the silencing function of Orc1. Indeed, as shown in Fig. 3.9A, the deletion of this region resulted in complete derepression of HMR::ADE2.

Silencing was next tested at HMR carrying either a synthetic silencer(HMR SS ΔI) or a silencer variant further sensitized by the deletion of the Abf1 binding site (HMR SS abf1 -  ΔI). Both silencers were affected by the deletion of 28 amino acids or more of the Orc1 N-terminus (Fig. 3.9B). In addition, deleting the region of amino acids 29 to 51 also interrupted silencing at HMR SS abf1 -   Δ I.

In contrast to the HMR variants, HML silencing was not affected by any of the N-terminal Orc1 deletions (Fig. 3.9B). We expect that the more robust wild-type HMR silencers are likewise not affected. This would be in agreement with the observation of Bell et al. (1995) that deleting the N-terminal 235 amino acids of Orc1 does not affect the mating ability (and thus the natural HM silencers) of an otherwise wild-type strain.


Fig. 3.9: N-terminal truncations of Orc1 impaired HMR silencing.
(A) ADE2 inserted at the HMR locus was derepressed when the N‑terminus of Orc1 was shortened by 51 or 100 amino acids and when the region between 29 and 51 amino acids was deleted. Serial dilutions of strains AEY743 (WT), AEY2587 (Δ1‑10), AEY2589 (Δ1‑28), AEY2333 (Δ1‑51), AEY2335 (Δ1‑100) and AEY2760 (Δ29‑51) were grown on medium lacking adenine to test ADE2 expression. (B) In contrast to HML silencers, synthetic HMR silencer variants were affected by the deletion of the N-terminal 28, 51 or 100 amino acids of Orc1, and the region between amino acids 29 and 51. Patch-mating assays were performed to test silencing at HMR SS ΔI and HMR SS abf1 - ΔI using MATα strains AEY2866 and 2864 (WT), AEY2877 and 2883 (Δ1‑10), AEY2907 and 2908 (Δ1‑28), AEY2879 and 2904 (Δ1‑51), AEY2880 and 2905 (Δ1‑100) and AEY2910 (Δ29‑51). HML silencing was tested in patch-mating assays of MATa strains AEY2867 (WT), AEY2887 (Δ1‑10), AEY2937 (Δ1‑28), AEY2888 (Δ1‑51), AEY2889 (Δ1‑100) and AEY2911 (Δ29‑51).

Alpha‑factor sensitivity was reduced in N-terminally truncated orc1 mutants

Response to α-mating pheromone (α‑factor) is required for the mating ability of haploid MAT a cells and is normally characterized by arrest in late G1 and the formation of mating projections (so-called shmoos). Derepression of HML, however, generates an a/α‑diploid phenotype and therefore α‑factor resistance of haploid MAT a cells, as indicated by continued divisions in the presence of α‑factor. Thus, α‑factor sensitivity of MAT a cells can serve as a measure of the silencing status of HML (Pillus and Rine 1989).

α-factor response tests are a more sensitive way than the usual patch-mating assays to investigate HML silencing, and we therefore employed this method here to measure HML silencing in the N‑terminally truncated orc1 mutants. To this end, we examined the morphology of at least 300 individual cells of each orc1 strain after 18 hours of growth on α‑factor containing medium. As for wild‑type cells, almost all cells carrying Orc1 lacking 10 or 28 amino acids of the N‑terminus formed shmoos, indicating repression of HML (Fig. 3.10).


Fig. 3.10: α‑factor response was abrogated by nat1Δ and diminished in orc1 mutants lacking 51 or 100 amino acids of the N‑terminus or the region of amino acids 29 to 51.
100 cells per strain were analyzed individually after 18 hours of exposure to α‑factor. The ability to respond to α‑factor was measured by the formation of one mating projection per cell (shmoo), whereas α‑factor resistance was indicated by budding and subsequent colony formation. Structures emerging from alternated shmooing and budding are referred to as shmoo clusters. Results of at least three individual experiments per strain are given with respective standard deviations. MATa strains used were depicted in figures 3.5(B) and 3.9(B).

Interestingly, shmoos were also generated by all of the unacetylated orc1‑A2P and orc1‑A2V cells, suggesting tight HML repression in these mutants. However, α‑factor sensitivity was reduced by the deletion of the N‑terminal 51 or 100 amino acids or the region between amino acids 29 and 51 of Orc1. In these strains, the shmooing fraction was smaller, whereas a significant number of cells continued dividing and eventually formed colonies. Interestingly, another portion of these mutants formed shmoo-clusters. Here, shmoo formation alternated with cell divisions, indicating unstable repression of HML (Enomoto and Berman 1998).

In summary, the α‑factor response tests revealed that HML silencing was affected by increasing truncations of the N‑terminus of Orc1. As expected, the deletion of NAT1 resulted in complete α‑factor resistance, due to strong HML derepression, which was in contrast to the strong α‑factor‑response of orc1‑A2P and orc1‑A2V mutants indicative of full HML repression.

 N-terminal truncations of Orc1 enhanced the α‑factor resistance of sir1Δ

sir1Δ strains have a characteristic α‑factor response phenotype, namely a mixed population of genetically identical cells, with one portion completely repressed and the other completely derepressed at HML. Thus, SIR1 was proposed to function in establishment rather than maintenance of transcriptional repression (Pillus and Rine 1989). The finding that the ORC binding site of HM silencers is likewise involved in the establishment of silencing (Sussel et al. 1993) is in accordance with the model that Sir1 is recruited to the silencer by Orc1.


Fig. 3.11: Deletion of 51 amino acids from the Orc1 N‑terminus enhanced the α‑factor response defect of sir1Δ mutants.
The ability to respond to α‑factor was tested as described in Fig. 3.10 using MATa strains AEY2867 (WT), AEY2888 (Δ1‑51), AEY3000 (sir1Δ), AEY3002 (orc1 Δ1‑51 sir1Δ #1) and AEY3003 (orc1 Δ1‑51 sir1Δ #2).

We sought to determine whether the truncation of the very N‑terminus of Orc1 would enhance the α‑factor response defect in sir1Δ cells. To this aim, we combined the deletion of SIR1 with the orc1Δ1‑51 mutation, which had affected each of the above-tested silencers. Significantly, in two individual double mutants, we found an increased portion of colony-forming cells (Fig. 3.11), indicating further derepressionof HML. Thus, the silencing defect of sir1Δ was enhanced by the deletion of the N-terminal 50 amino acids of Orc1. This effect is surprising in light of the current view that Orc1’s sole function in silencing is to recruit Sir1. It rather suggests that Orc1 has a broader task.

 Telomeric silencing was affected by N-terminal truncations of Orc1

We next asked whether truncations within the N-terminal 100 amino acids of Orc1 had an impact on telomeric silencing. To this aim, we investigated the expression of a subtelomeric URA3 reporter gene in the different orc1 N‑terminal mutants. The deletion of 28, 51 or 100 amino acids from the N‑terminus, as well as the removal of amino acids 29 to 51, increased the expression of URA3, as indicated by diminished growth of these mutants on counterselective 5‑FOA medium (Fig. 3.12). This was the first evidence for a function of the Orc1 N‑terminus in telomeric silencing.


Fig. 3.12: N‑terminal truncations of 28, 51 and 100 amino acids, as well as removing the region of amino acids 29 to 51 of Orc1, reduced telomeric silencing.
Silencing of URA3 inserted near the left telomere of chromosome VII was tested in serial dilution assays of strains AEY1017 (WT), AEY3031 (Δ1‑10), AEY3040 (Δ1‑28), AEY3032 (Δ1‑51), AEY3034 (Δ1‑100) and AEY3036 (Δ29‑51) on 5‑FOA containing medium counterselecting for URA3 expressing cells.

Interestingly, this phenotype was weaker than that of the N‑terminally unacetylated orc1 mutants. This is surprising given that both mutations should abolish the Nα-acetylation of Orc1. In contrast, this result suggests that the N‑terminal deletions suppress the defect caused by missing Nα‑acetylation, implicating that the two types of mutations have different consequences for telomeric silencing.

 Replication was not disturbed by N-terminal truncations of Orc1

As shown above, the lack of Nα‑acetylation of Orc1 appeared to have no impact on its replication function, since unacetylated mutants grew as well as wild-type strains (Fig. 3.7A).

To determine whether this was also the case for the N‑terminal deletion mutants of Orc1, we tested their growth at different temperatures.


Fig.3.13: N-terminal deletions of up to 100 amino acids of Orc1 did not affect the temperature sensitivity of the respective mutants.
Strains AEY743 (WT), AEY2587 (Δ1‑10), AEY2589 (Δ1‑28), AEY2333 (Δ1‑51), AEY2335 (Δ1‑100) and AEY2760 (Δ29‑51) were grown for two days on complete medium at the indicated temperatures.

Significantly, none of the mutants displayed a growth defect or temperature sensitivity (Fig. 3.13), suggesting that the first 100 amino acids of Orc1 were dispensable for its function in replication. This result agreed with the notion of Bell et al. (1995) that the N-terminal 235 amino acids of Orc1 have no function in replication, since their deletion causes only a slight reduction of plasmid stability.

 The N-terminal 51 amino acids of Orc1 were required for its two-hybrid interaction with Sir1

The N‑terminus of Orc1 (amino acids 5‑235) interacts with the C‑terminus of Sir1 (amino acids 346‑678) in a two-hybrid assay (Triolo and Sternglanz 1996; Gardner et al. 1999). This interaction was interrupted when the region of amino acids 100-129 of Orc1 was substituted with the corresponding region of human Orc1, but remained intact when amino acids 21 to 35 were replaced by four alanines (Zhang et al. 2002). The latter observation implicated that the part before amino acids 100‑129 was dispensable for Orc1 to interact with Sir1.


Fig. 3.14: Deletion of the N-terminal 51 amino acids abrogated the ability of Orc1 to interact with Sir1 in a two‑hybrid assay.
(A) The reporter genes ADE2 and HIS3 were induced in two-hybrid strain AH109 by simultaneous expression of Gal4BD‑Sir1(346‑678) and Gal4AD‑Orc1(1‑235), but not Gal4AD‑Orc1(52‑235). The bait-vector pAE952 was co-transformed with a prey-vector containing either no insert (pAE953), full-length Orc1 (pAE951) or Orc1(52‑235) (pAE966). Two-hybrid interaction was tested by monitoring the expression of HIS3 and ADE2 in serial dilution assays on media lacking histidine or adenine, respectively. (B) The prey protein of Orc1(52‑235) was as abundant as that of Orc1(1‑235) in the two-hybrid strains AEY3028 (Orc1(1‑235)) and AEY3099 (Orc1(52‑235)). A SDS gel of whole cell extracts was analyzed by Western blotting with antibodies against the HA epitope that was part of the prey vector.

We therefore tested whether the deletion of the N‑terminal 50 amino acids of Orc1, which affected silencing in our experiments, would disrupt the two-hybrid interaction with Sir1.

Using Sir1(346‑678) as bait and Orc1(1-235) as prey, the two-hybrid reporter genes HIS3 and ADE2 were only expressed when Orc1 contained its N‑terminal 50 amino acids (Fig. 3.14A). To eliminate the possibility that the missing interaction was due to a lower availability of mutant Orc1, we measured its abundance in a Western blot. Here, equal levels of Orc1(1‑235) and Orc1(52‑235) prey protein were detected in the respective two-hybrid strains (Fig. 3.14B).

Although the loss of physical interaction between Sir1 and Orc1(52‑235) has to be confirmed in vivo, for example by Co-Immunoprecipitation, the disrupted two-hybrid interaction was a first indication that the N‑terminal 51 amino acid region of Orc1 is required for its binding to Sir1.
Sir3 was a substrate of NatA Sir3 was Nα-acetylated by NatA

In a previous study, (Stone et al. 2000) observed decreased telomeric silencing and an enhanced sir1Δ mating defect when the penultimate alanine of Sir3 was exchanged for a threonine. This sir3‑A2T mutation was epistatic to nat1∆, and suggested that Nα‑acetylation was required for the silencing function of Sir3. We therefore tested Sir3 directly for Nα‑acetylation by isoelectric focusing, in analogy to our experiments with Orc1. The isoelectric point of a TAP‑tagged N‑terminal peptide of Sir3 (amino acids 1 to 235) was more acidic in a wild-type strain than in a nat1Δ strain (Fig. 3.15A), suggesting that Sir3 was Nα‑acetylated by NatA.

Fig. 3.15: Sir3 was acetylated by NatA.
(A) The isoelectric point of the Sir3 N-terminus became more basic upon the deletion of NAT1. Whole cell extracts of strains AEY3171 (WT) and AEY3173 (nat1Δ) expressing TAP-tagged Sir3 peptides (amino acids 1‑235) were analyzed as described in Fig. 3.3A. (B) The silencing defect at the synthetic HMR silencer caused by the mutated penultimate amino acid of Sir3 was not enhanced by missing Nα‑acetylation of Orc1. HML silencing was assayed in patch mating assays of MATa strains AEY3144 (sir3‑A2T), AEY3147 (orc1‑A2P sir3‑A2T), AEY3148 (orc1‑A2V sir3‑A2T) and AEY2912 (nat1Δ). Likewise, synthetic HMR silencing was tested in MATα strains AEY3145 (sir3‑A2T), AEY3149 (orc1‑A2P sir3‑A2T), AEY3151 (orc1‑A2V sir3‑A2T), and AEY2916 (nat1Δ).

Since the unacetylated forms of both Orc1 and Sir3 singly had a less pronounced silencing defect than nat1∆, we asked whether their combination would enhance the effect on silencing. However, orc1‑A2P sir3‑A2T and orc1‑A2V sir3‑A2T double mutants showed the same amount of HM derepression as sir3‑A2T alone, suggesting that NatA had other targets whose function was required for HM silencing (Fig. 3.15B).

 NatA activity was required to localize Sir3 to perinuclear foci

NatA is required for silencing of subtelomeric reporter genes (Fig. 3.1C) (Aparicio et al. 1991), and the lack of Nα‑acetylation of Orc1 and Sir3 resulted in derepression of subtelomeric URA3 (Fig. 3.5) (Stone et al. 2000). While the insertion of reporter genes generates truncated versions of these telomeres (Gottschling et al. 1990), Sir3 is also required for silencing in a native telomeric context (Vega-Palas et al. 1997; Venditti et al. 1999). Normally, Sir3 colocalizes with Rap1 and Sir4 in perinuclear foci (Grunstein 1998) whose structural integrity was proposed to be a prerequisite for telomeric silencing (Cockell et al. 1995).

In order to determine whether NatA played a role in chromatin organization of native telomeres, we investigated the localization of GFP‑tagged Sir3 in wild-type and nat1∆ strains. Interestingly, whereas GFP signals in wild-type cells showed the expected perinuclear foci, Sir3 became distributed throughout the nucleus in the absence of NAT1 (Fig. 3.16).


Fig. 3.16: The association of GFP-tagged Sir3 with telomeric foci was abrogated in nat1Δ cells. Strains AEY160 (WT) and AEY2786 (nat1Δ) transformed with pAE580 were examined by fluorescent microscopy using a FITC filter. Bar, 2 μm.

This suggested that the structure of native chromosomal ends depended on NatA activity. Since GFP was fused to the C-terminus of Sir3, it was probably still Nα-acetylated in the wild-type and unacetylated in the nat1Δ strain. Thus, it is conceivable, that the missing Nα‑acetylation caused Sir3 to detach from the perinuclear foci, rather then it was an indirect effect. However, this question was not answered in our experiment and requires further investigation.
A genetic screen for multicopy suppressors of the nat1Δ silencing defect Screening for restored silencing of HMR SS ΔI in a nat1Δ strain

The silencing phenotype of unacetylated orc1 sir3 double mutants suggested that the function of NatA in silencing comprises more than these two substrates. In order to identify more silencing components that require Nα‑acetylation by NatA, we performed a genetic screen for multi-copy suppressors of the nat1Δ silencing defect. This unbiased approach had the advantage that NatA substrates with so far unknown implication in silencing could be discovered.

Our experiment based on the assumption that the malfunction of a silencing component provoked by missing Nα-acetylation might be compensated for by its overexpression. It should therefore be possible to identify such a NatA substrate by screening for genes that, when overexpressed, are capable of restoring HM silencing in a nat1Δ strain. For the screen, we used a MATα strain with a synthetic HMR silencer (HMR SS ΔI), which was a complete non‑mater due to the deletion of NAT1 (Fig. 3.1A). In this background, multicopy suppressors of nat1Δ should be easily detectable by restored mating of the respective transformants. As a positive control, we expected to isolate NAT1, which should suppress its own deletion phenotype.

We transformed the strain with a 2μ‑based genomic library (YEp24) (Carlson and Botstein 1982) and tested the mating ability of 30.000 transformants by replica‑plating the colonies on MAT a tester plates. The 90 maters identified were verified by repeated patch-mating assays and only those with reproducible results were further tested. Also, candidates that were assigned to be identical according to the restriction pattern of their plasmid were rejected from further tests. Since many of the originally identified maters did not give reproducible results, only 15 candidates remained. Their plasmids were isolated and retransformed into the HMR SS ΔI nat1Δ strain to confirm their suppression potential. Furthermore, the mating ability of the candidate strains was tested after loss of the URA3 marked library plasmids on 5‑FOA medium. Interestingly, candidate strain number 23 displayed good mating in the absence of the plasmid (Tbl. 3.1), suggesting that its mating ability was reestablished by (an) additional chromosomal mutation(s) rather than the overexpression a suppressor gene. The retransformed plasmids of six out of the 15 candidates could induce mating. The inserts of these plasmids were sequenced and subsequently blasted using the Saccharomyces genome database (http://www.yeastgenome.org/) to determine the encoded chromosomal region. Among them were two different NAT1 containing clones (72 (Tbl. 3.1) and 89 (data not shown)), implicating that the tested number of transformants was sufficient to cover all open reading frames of the genome with the screen.

 Overexpression of SSF2 suppressed the nat1Δ mating defect

Each of the remaining four candidate clones encoded a gene whose function has been linked to mating in earlier studies. On candidate clone 22, RVS161 (reduced viability upon starvation), encodes a protein with a direct role in cell fusion during mating (Brizzio et al. 1998). RGA1 (rho-type GTPase-activating protein) (on candidate clone 80) was shown to act as a negative regulator of the pheromone response pathway by controlling the activity of Cdc42, a p21 GTPase required for polarity establishment and bud emergence (Stevenson et al. 1995). The overexpression of SSF2 (suppresor of Sterile Four) (candidate clone 83) was shown to increase the mating efficiency in an earlier study (Yu and Hirsch 1995), and acts directly in RNA processing (J. Hirsch; personal communication). NPL3 (nuclear protein localization) (candidate clone 84) is also involved in processing and nuclear-cytoplasmatic transport of RNA and is required for silencing of the mating-type loci (Loo et al. 1995b). However, as Npl3 does not act directly at HMR-E, this effect was proposed to be indirect.
Table 3.1: Positive candidates from a screen for multi-copy suppressors of the nat1Δ mating defect in the MATαHMR SS ΔI strain AEY1273.

In order to confirm the four genes as multicopy suppressors of nat1Δ, they were cloned in 2μ‑based pRS426 vectors and transformed individually into the HMR SS ΔI nat1Δ strain used above. Notably, only SSF2 restored mating, suggesting that RVS161, RGA1 and NPL3 were not responsible for the suppressing effect of the library plasmid they were derived from. However, further subcloning of these plasmids did not reveal any of the other encoded ORFs to be responsible for the suppressing effect. Thus, it appeared that for unknown reasons RVS161, RGA1 and NPL3 acted as suppressors of nat1Δ specifically in the context of the library vectors.

All four candidate suppressor genes were potential NatA substrates according to their penultimate amino acid, and therefore their dependence on Nα‑acetylation might have been suppressed by overexpression. However, they appeared to improve mating by processes distinct from HMR SS ΔI silencing, and therefore their suppressing effect on the nat1Δ phenotype was indirect.

In summary, our screen identified one gene, SSF2, as a multi-copy suppressor of the nat1Δ mating defect. Since the suppression phenotype of SSF2 appeared to be indirect, the screen failed to identify a NatA substrate directly involved in HM silencing.

 Overexpression of ORC1 did not suppress the mating defect caused by nat1Δ

We next tested whether overexpression of ORC1, which we had earlier identified as a NatA substrate, would suppress the mating defect of the nat1Δmutant. For this, ORC1 was placed in a 2μ‑based plasmid under control of the strong constitutive GPD‑promoter to ensure overexpression. This construct was biologically active, since it restored HMR silencing in the orc1Δ1‑51 mutant (Fig. 3.17A). Moreover, ORC1 was overexpressed efficiently in the MATαHMR SS ΔI nat1Δ strain, which was used before in the screen for multicopy suppressors (Fig. 3.17B). However, the mating defect of this strain was not suppressed by ORC1 overexpression (Fig. 3.17C).

Interestingly, overexpression of SIR3, which we had also identified as NatA substrate, also failed to suppress the nat1Δ mating defect in an earlier study (Stone et al. 1991).


Fig. 3.17: The mating defect of nat1Δ was not rescued by overexpressed ORC1.
(A) ORC1 expressed under control of the GPD-promoter on a 2μ-based plasmid rescued the silencing defect of orc1Δ1-51 at HMR::ADE2. Strain AEY2333 (orc1Δ1-51) was transformed with pAE866 (p2μGPDp-ORC1) and grown in serial dilutions on medium lacking adenine. Strain AEY743 (WT) was tested in parallel for comparison. (B) In strain AEY1273 (-2μGPDp-ORC1), more Orc1 protein was abundantupon transformation with the overexpressing construct pAE866 (+ 2μGPDp-ORC1). Equal amounts of protein from whole cell extracts were applied to a SDS gel and subsequently to Western blot analysis using antibodies against Orc1 and β tubuline (as loading control). (C) Overexpression of ORC1 did not increase silencing of the synthetic HMR silencer in a nat1Δ background. HMR silencing was determined by the mating ability of MATα-strain AEY1273 (HMR SSΔI nat1Δ) transformed with pRS316 (vector), pAE303 (NAT1) or pAE866 (2μGPDp-ORC1).