Mechanism of Regulation of Intrachromatid Recombination and Long-Range Chromosome Interactions in Saccharomyces cerevisiae

Abstract

The NAD-dependent histone deacetylase Sir2 controls ribosomal DNA (rDNA) silencing by inhibiting recombination and RNA polymerase II-catalyzed transcription in the rDNA of Saccharomyces cerevisiae. Sir2 is recruited to nontranscribed spacer 1 (NTS1) of the rDNA array by interaction between the RENT (regulation of nucleolar silencing and telophase exit) complex and the replication terminator protein Fob1. The latter binds to its cognate sites, called replication termini (Ter) or replication fork barriers (RFB), that are located in each copy of NTS1. This work provides new mechanistic insights into the regulation of rDNA silencing and intrachromatid recombination by showing that Sir2 recruitment is stringently regulated by Fob1 phosphorylation at specific sites in its C-terminal domain (C-Fob1), which also regulates long-range Ter-Ter interactions. We show further that long-range Fob1-mediated Ter-Ter interactions in trans are downregulated by Sir2. These regulatory mechanisms control intrachromatid recombination and the replicative life span (RLS).

INTRODUCTION

The ribosomal DNA (rDNA) of Saccharomyces cerevisiae is present as ∼200 tandem repeats in chromosome XII (1). In order to prevent physiologically unwarranted recombination provoked by such a large number of tandem repeats, and the consequent risk of rDNA copy number instability, yeast cells have evolved a mechanism that suppresses intrachromatid recombination and transcription by RNA polymerase II (Pol II) but not by RNA Pol I and Pol III (2). The suppression occurs by loading of the NAD-dependent histone deacetylase Sir2 onto each of the rDNA repeats at nontranscribed sequence 1 (NTS1) by interactions between Fob1 and the Net1 protein of the RENT (regulation of nucleolar silencing and telophase exit) complex (3). One of the two RENT loading sites in rDNA is located at or near the Ter sites in NTS1, and the second one is located at or near the promoter of the 35S precursor rRNA (4). The RENT complex consists of Net1 (a scaffold protein), Cdc14 (phosphatase), which is needed for escape from mitosis (5), and Sir2 (6, 7). An alternative pathway to RENT for loading of Sir2 comprises Tof2, Csm1, and Lrs4, which are components of the monopolin complex, and two inner nuclear membrane proteins called Heh1 and Nur1. Tof2 physically interacts with Fob1 (8, 9).

The Fob1 (fork blockage) protein specifically binds to the Ter sites and causes polar arrest of replication forks (10, 11). It has been suggested that Fob1-Ter interaction prevents collision between the rDNA transcription and replication forks approaching from the opposite direction (12).

The primary silencing activity of Sir2 is effected through repression of a bidirectional promoter called Epro, located in NTS1 (13). Transcription initiated at Epro is believed to displace cohesin from NTS1, thereby making the region recombinogenic (13).

Fob1 arrests replication forks at Ter sites and induces DNA bending, and the bent DNA presumably recruits topoisomerase I (Topo I) to generate DNA breaks that provoke recombination (14). However, maintenance of a certain level of topoisomerase activity prevents recombination at rDNA (15, 16). We have recently reported that Fob1 oligomerization is necessary to promote long-range Ter-Ter interactions in trans (also called “chromosome kissing”), which bring about synaptic interactions that initiate intrachromatid recombination (17). We have also shown that Fob1 remains in a conformation that is refractory to Fob1-Fob1 interaction. This is caused by intramolecular interactions between the N-terminal domain of Fob1 (N-Fob1) and the inhibitory C-terminal domain (C-Fob1). The inhibitory interaction is counteracted by phosphorylation of certain residues of C-Fob1, thereby promoting Fob1-Fob1 interaction and initiating intrachromatid recombination (17). It was previously shown that intrachromatid recombination reduces the replicative life span (RLS) of S. cerevisiae (18, 19).

This work was designed to address the following questions. First, how is the recruitment of RENT and Tof2 complexes to NTS1 regulated? Second, besides suppression of transcription from the Epro promoter, what other functions relevant to rDNA silencing are mediated by Sir2? We present evidence showing that phosphorylation of C-Fob1 is required for loading of both RENT and Tof2 complexes onto the NTS1 region of rDNA. We show further that Fob1 phosphorylation regulates long-range Ter-Ter interaction in trans as revealed by a modified circular chromosome conformation capture (4C) technique (17, 20, 21) and that Sir2 downregulates Fob1-mediated trans interaction between Ter sites. The data also show that the replicative life span (RLS) was significantly enhanced by 3 critical Ser-to-Ala substitutions at C-Fob1 and that phosphomimetic Asp substitutions at the same sites significantly reduced the RLS in comparison with the Ala-substituted triple mutant form and the wild-type (WT) Fob1. In summary, this work provides new insights into regulation of rDNA silencing, control of long-range Ter-Ter interaction, intrachromatid recombination, and RLS.

MATERIALS AND METHODS

Strains, plasmids, and primers.

The yeast strains, plasmids, and oligonucleotides used in this work are shown in Tables 1, 2, and 3.

TABLE 1.

Yeast strains and plasmids

Strain or plasmid	Genotype	Source or reference
S. cerevisiae strains
W303	MATa leu2-3,112 trp1-1 can1-100 ura3-1 ade2-1 his3-11,15	R. Rothstein
LPY11	W303a sir2Δ::HIS3	L. Pillus
Lfob1	LPY11 fob1Δ::G418	42
MC44	Lfob1 fob1S467A,S468A,S519A-Phleomycin	This study
MC55	Lfob1 fob1S467D,S468D,S519D-Phleomycin	This study
PJ69-4A	MATa trp1-Δ901 leu2-3,112 901 ura3-52 his3-Δ200 gal4Δ gal8Δ GAL2-ADE2 LYS2::GAL1-HIS3 met2::GAL7-lacZ	24
YSB348	MATα his3Δ200 leu2Δ1 ura3-167 RDN1(50L)::mURA3-HIS3	Jeff Smith
YSB348fob1Δ	YSB348 fob1Δ::G418	31
BY4742 fob1Δ	MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 fob1Δ::G418	Clontech
YPK	MATa ade2-101ochre his3-Δ200 leu2-Δ1 lys2-801amber trp1-Δ63 ura3-52	29
BJ5464	MATα ura3-52 trp1 leu2 Δ1 his3-Δ200 pep4::HIS3 prb1 Δ1.6R	Satya Prakash
DMY1690	NET1-TAP::K.l-TRP1	4
DMY2334	FOB1-TAP::K.l-TRP1	4
DMY1704	SF10 SIR2-TAP::K.l-TRP1	4
DMY1704W	DMY1704 SIR2-13myc-Phleomycin	This study
DMY1704A	DMY1704 SIR2-13myc-Phleomycin fob1 S467A,S468A,S519A-LEU2	This study
DMY1704D	DMY1704 SIR2-13myc-Phleomycin fob1 S467D,S468D,S519D-LEU2	This study
DMY1704fob1Δ	DMY1704 SIR2-13myc-Phleomycin fob1Δ::URA3	This study
Plasmids
pGAD424		Clontech
pGBT9		Clontech
pGAD424Fob1		This study
pGAD424fob1K89T		This study
pGAD424fob1M213L		This study
pGAD424fob1T322I		This study
pGAD424fob1E373V		This study
pGAD424fob1S467A,S468A,S519A		This study
pGAD424fob1S467D,S468D,S519D pGBT9Fob1		This study
pGBT9fob1 K89T		This study
pGBT9fob1 M213L		This study
pGBT9fob1T322I		This study
pGBT9fob1E373V		This study
pGBT9fob1S467A,S468A,S519A		This study
pGBT9fob1S467D,S468D,S519D		This study
pGBT9Net1		This study
pGBT9Sir2		This study
pGBT9sir2L159S		This study
pGBT9net1(1–341aa)		This study
pGBT9net1(566–801aa)		This study
pGBT9Ydr026		This study
pGAD424Tof2		This study
pMALnet11–341aaCKin		This study
pMALSir2Ckin		This study
pBJ842Fob1CKin		This study
pBJ842fob1T322ICKin		This study
pBBHyg2		This study

TABLE 2.

Missense mutants of Fob1 and their phenotypes

Fob1 mutant	Interaction			Silencing
	Fob1-Fob1	Fob1-Net1	Fob1-Sir2
pGAD	−	−	−	−
pGADFob1	+	+	+	+
S37L	+	+	+	+
S40L	+	+	+	+
S50L	+	+	+	+
L79F	−	−	−	−
K89T	−	+/−	+/−	+/−
V115L	+	+	+	+
D130E	−	−	−	+/−
D130G	−	−	−	+
D131E	+	+	+	+
R135K	+	+	+	+
G138R	+	+	+	+
K149N	+	+	+	+
A157P	−	−	−	−
M213L	−	+/−	+/−	+/−
L218I	+	+	+	+
K305I	+	+	+	+
T322I	+	−	−	−
R356L	+	+	+	+
E373V	−	+/−	+/−	+
P375T	+	+	+	+
K387W	+	+	+	+
K404R	+	+/−	+/−	+/−
I407T	+	+/−	+/−	+/−
L409 M	+	+/−	+/−	+/−
L417H	+	+/−	+/−	+/−
E420G	+	+	+	+
C425N	+	+	+	+
Q448H	+	+	+	+

TABLE 3.

Oligonucleotides

Purpose and primer	Sequence
Site-directed mutagenesis of Fob1
K89TF	5′GACAACTAACGCAGACAATATATGAACTAATAAAAAC3′
K89TR	5′GTTTTTATTAGTTCATATATTGTCTGCGTTAGTTGTC3′
M213LF	5′GGAATATAAACGTCCTGACTTGTACGATAAACTAC3′
M213LR	5′GTAGTTTATCGTACAAGTCAGGACGTTTATATTCC3′
T322IF	5′GTTCTGCGAGATTTCATATTGGGGGTATACTGTGC3′
T322IR	5′GCACAGTATACCCCCAATATGAAATCTCGCAGAAC3′
E373VF	5′TGCAAGTACTACTTAGTGTAGTTCCAGGTCACAATG3′
E373VR	5′GCATTGTGACCTGGAACTACACTAAGTAGTACTTGC3′
S467AS468AF	5′GATTTGGCACATGATGCTGCAGAGGGCGAATTTGAAC3′
S467AS468AR	5′GTTCAAATTCGCCCTCTGCAGCATCATGTGCCAAATC3′
S467DS468DF	5′GATTTGGCACATGATGATGATGAGGGCGAATTTGAAC3′
S467DS468DR	5′GTTCAAATTCGCCCTCATCATCATCATGTGCCAAATC3′
S519AF	5′AAAGTGACGGTGCAGCACAAGTAGATCAAAG3′
S519AR	5′CTTTGATCTACTTGTGCTGCACCGTCAC3′
S519DF	5′GAAAGTGACGGTGCAGACCAAGTAGATCAAAGTG3′
S519DR	5′GTTCAAATTCGCCCTCATCATCATCATGTGCCAAATC3′
Fob1 and Tof2 cloning in 2-hybrid vectors
FOBAMF	5′GCGGCGGATCCGGATGACGAAACCGCGTTACAATGACG3′
FOBSALR	5′GCGGTCGACCAATTCCATTGATGTGCCAAAGTCTCTTG3′
TOF2BAMF	5′GGCGGGGGATCCGGATGATAAAAATGTGGAGGTTACAG3′
TOF2SALR	5′CGGCGCGTCGACTTACTGGTCGTCTTCATCACTTTC3′
Fob1 cloning in expression vector
FOBBAMF	5′CACACGGATCCACATATGACGAAACCGCGTTACAATGACGTGTTG3′
FOBBAMR	5′CACACGGATCCTACAATTCATTGATGTGCCAAAGTCTCTTG3′
rDNA probe used in 2D gel
RDN1350	5′CTGAACATGTCTGGACCCTGCCCTC3′
RDN2800	5′AGGCGTCCTTGTGGCGTCGCTGAAC3′
Fob1 knock-in/knockout
FOB1F44 + 22	5′CAATTTAACGATTGTGTGAGTGTGAATTTGTGCTGAGGATAACAATGACGAAACCGCGTTACAATGACG3′
FOBCRELOXF	5′GAGATCAAACAAGAGACTTTGGCACATCAATGGAATTGTAACAGCTGAAGCTTCGTACGCTGCAG3′
FOB1CRELOXR	5′GCGTACGAAGCTTCAGCTGTTACAATTCCATTGATGTGTGCCAAAGTCT CTTGTTTGATCTC3′
FOB + 41TO1CRELOXR	5′CACCTATGGTGACTCCTCCTTTCATTCTATCCTACATATTAGCATAGGCC ACTAGTGGATCTGATAT3′
Detecting chromatin conformation
F1	5′GCACTGGCTATTCATCTTGCACTTTTCCTC3′
F2	5′GGAAAAGTGCAAGATGAATAGCCAGTG3′
F3	5′CGATGAGGATGATAGTGTGTAAAGAGTG3′
F4	5′GGTACACTCTTACACACTATCATCCTCATCG3′
Myc tagging of Sir2
Sir2mycF	5′AAGGGCGTGTATGTCGTTACATCAGATGAACATCCCAAAACCCTCCATGCCACTCTCGTCTTCGATGTGGAG3′
Sir2downstPhleoR	5′ATTGATATTAATTTGGCACTTTTAAATTATTAAATTGCCTTCTACGCATAGGCCACTAGTGGATCTG3′

Gene knockout.

Gene deletions were carried out by the one-step gene disruption method (22, 23). Phleomycin, G418, and nutritional markers were used for gene knockouts. The knockouts were confirmed by appropriate PCR amplification to obtain products diagnostic of a successful knockout.

Y2H interactions.

Yeast two-hybrid (Y2H) interactions were carried out using the yeast strain PJ69-4A as described previously (11, 24). Briefly, pGAD424 and pGBT9 vectors containing the appropriate gene were transformed in pairs into the yeast strain PJ694-A and selected on SD plates without Leu and Trp (SD/Leu⁻ Trp⁻ plates). Colonies from SD/Leu⁻ Trp⁻ plates were patched on SD/Leu⁻ Trp⁻ and SD/Leu⁻ Trp⁻Ade⁻ plates. Colonies growing on SD/Leu⁻ Trp⁻Ade⁻ plates were considered the initial signal for positive protein-protein interaction. The corresponding colonies from SD/Leu⁻ Trp⁻ plates were grown in liquid cultures and analyzed further by β-galactosidase assay as described in the Clontech manual.

Isolation of noninteracting mutants by yeast reverse 2-hybrid (YR2H) analysis.

The mutant isolation and selection strategy is schematically shown in Fig. 1G. Briefly, the rationale was as follows. Since yeast 2-hybrid (Y2H) analyses showed apparent interaction between Sir2 and Fob1 (Fob1 with Fob1 and Fob1 with Net1), any fob1 mutation that destroyed the interactions should not grow on Ade dropout indicator plates (expression of ADE was used as the indicator of positive Y2H interaction). Before DNA sequencing of the Fob1 open reading frame (ORF) of each putative mutant to eliminate nonsense and frameshift mutations and select the potentially useful missense alterations, we wished to further eliminate those that produced misfolded proteins. The rationale was as follows. If a putative mutant, besides having an apparent defect in protein-protein interaction with the prey, was also defective in DNA binding, its likelihood of being a globally defective one with possibly misfolded protein was deemed to be higher.

FIG 1.

Protein-protein interactions that recruit Net1 and Sir2 to NTS1 of rDNA. (A) Schematic representation of the tandem rDNA repeats. (B) Y2H analyses of Fob1-Net1 interaction (row 2), apparent Fob1-Sir2 interaction (row 4), and lack of Fob1-Sir2(L159S) interaction (row 5). (C) The Y2H data suggest 2 alternative models of Fob1-RENT interaction. (D) Quantification of the data shown in panel B by measurements of β-galactosidase activity of the LacZ reporter. (E) Biochemical analyses of protein-protein interactions shown in panels C and D, revealing that purified ³²P-Sir2 interacts with immobilized Net1 (Net1-TAP) but not with Fob1-TAP. (F) Flow chart from low-fidelity PCR mutagenesis of the Fob1 ORF, 2-hybrid test of interaction or lack of it with Sir2, and transfer of the 2-hybrid vector-based fob1 mutant pool into the HOT1 indicator strain. (G) A fob1 mutant that fails to bind to the Ter site present in the HOT1 element will fail to induce high-frequency recombination between the his4 alleles flanking the ADE5 reporter, producing solid red colonies, whereas high-frequency recombination by fob1 that binds to HOT1 produces red-white sectored colonies due to loss of ADE5.

The Fob1 ORF was mutagenized by low-fidelity PCR using Vent Exo⁻ DNA polymerase. Each mutagenesis reaction was carried out by using 4 reaction tubes containing 3 deoxynucleoside triphosphates (dNTPs) at 0.25 mM each and 1 dNTP at 0.06 mM. After PCR amplification, the 4 PCR products were mixed to generate a pool of putative fob1 mutants. Competent cells of yeast containing a resident pGBT9Sir2 plasmid were transformed with the mutagenized fob1 PCR products mixed with linearized pGAD424. The transformed cells were plated on SD/Leu⁻ Trp⁻ plates to select cells that contain pGBT9Sir2 and circularized pGAD424+fob1 PCR products. Colonies that grew on the SD/Leu⁻ Trp⁻ but not on the SD/Leu⁻ Trp⁻ Ade⁻ plates were expected to contain either a putative noninteracting mutant of Fob1 or a blank vector without an insert. Plasmids containing the putative fob1 mutants were isolated and checker by using PCR and Fob1-specific primers to confirm whether the pGAD424 plasmid contained the correct Fob1 ORF. These plasmids were further tested for retention of DNA binding as described below.

Color sectoring assay.

To weed out fob1 mutants that were likely to be globally misfolded, we wished to rapidly detect those that while being defective in protein-protein interaction were also defective in Ter binding. For a rapid genetic test for DNA binding, we used a Fob1-Ter binding-dependent recombination assay which used a color sectoring technique (25, 26) as follows (Fig. 1G). It has been reported that a yeast strain that is ade2 ade3 fob1Δ and contains a DNA fragment consisting of his4-1 and his4-2 repeated alleles flanking an ADE5 reporter at an ectopic location in chromosome III produces red colonies on plates with a low adenine concentration. A recombinogenic sequence called HOT1 that contains a promoter and enhancer of RNA Pol I transcription and a Fob1 binding Ter site was inserted into the left his4 repeat. Fob1 binding to the Ter site present in the HOT1 sequence causes high-frequency recombination between the flanking his4 alleles, leading to loss of the ADE5 marker. This loss is revealed by red-white sectoring of clones on low-adenine plates. As shown in Fig. 1G, without the presence of Fob1 or in a fob1 mutant, which fails to bind to Ter, the cells form solid red colonies. Thus, red-white sectoring promoted by Fob1 provides a rapid test of those mutant forms of the protein that retain Ter binding activity.

The putative mutant pool of fob1 present in the 2-hybrid plasmid vector were transformed into the HOT1 indicator strain, and colonies that generated red-white sectoring were retained, whereas those producing solid red colonies were discarded. Plasmid DNAs each carrying a putative fob1 mutant were prepared and subjected to DNA sequencing to identify and eliminate frameshift and nonsense mutations. All single and double missense mutations were saved for further analysis to confirm the apparent loss of protein-protein interaction with Sir2 (or with Fob1 or Net1, depending on the prey used) by Y2H analysis (Table 2) and biochemical assays using purified mutant proteins.

qChIP for Sir2 loading.

The quantitative chromatin immunoprecipitation (qChIP) for Sir2 loading at NTS1 was performed as described previously (27, 28). Yeast cultures (100 ml each) were grown to an optical density at 600 nm (OD₆₀₀) of 1.5 and cross-linked with formaldehyde to a final concentration of 1% (vol/vol) at room temperature (RT) for 15 min. Reactions were stopped by adding glycine to a final concentration of 375 mM, and the mixtures were incubated with shaking for 5 min at RT. Cells were washed twice with cold TBS (20 mM Tris-HCl [pH 7.6] and 150 mM NaCl), harvested, and stored at −80°C until further use. Frozen pellets were resuspended in 400 μl of lysis buffer (50 mM HEPES-KOH [pH 7.5], 500 mM NaCl, 1 mM EDTA [pH 8.0], 1% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS, and protease inhibitors) and lysed using glass beads four times for 30 s each at 4°C. Cell lysates were centrifuged at 12,000 × g for 15 min at 4°C, and the resulting chromatin pellet was resuspended in 400 μl of lysis buffer, sonicated 5 times for 20 s each on ice at 40% amplitude (Branson Digital Sonifier), and centrifuged at 12,000 × g for 30 min at 4°C. Prewashed IgG-agarose beads (30 μl) were added to the supernatant and mixed by rotation at 4°C for 2 h. Beads were washed thoroughly and divided into two equal parts, one for immunoprecipitation (IP) reaction (with 1 μg anti-c-Myc antibody [Sigma, USA]) and the other as a no-antibody control. Both IP and control tubes were kept on rotation overnight (O/N) at 4°C. IgG-agarose beads (preadsorbed O/N with bovine serum albumin [BSA] and sonicated salmon sperm DNA) were thoroughly washed, added to the IP and control tubes, and kept on rotation for another 4 h at 4°C. Antibody-bead complexes were washed twice with lysis buffer, twice with wash buffer (10 mM Tris-HCl [pH 8.0], 0.25 M LiCl, 0.5% NP-40, 0.5% sodium deoxycholate, and 1 mM EDTA), and twice with TE (10 mM Tris-HCl [pH 7.5], and 1 mM EDTA [pH 8.0]) at RT. Beads were incubated with 200 μl of elution buffer (50 mM Tris-HCl [pH 7.5], 10 mM EDTA [pH 8.0], and 1% SDS) at 65°C for 15 min, and DNA was eluted. Pronase (20 μg) was added to each IP and control tube and incubated further at 42°C for 2 h, followed by reversal of cross-linking at 65°C for 6 h. To the input DNA tube, an equal volume of elution buffer was added, followed by 20 μg of pronase, and incubated as described above. DNA was extracted, diluted to 1:50 and 1:100, and subjected to real-time PCR analysis using iQ SYBR green Supermix (Bio-Rad, Hercules, CA) according to the manufacturer's instructions on an Bio-Rad real-time PCR system. Primers will bind to the rDNA Ter site (enriched region in ChIP) where Fob1 binds. Quantification was performed using the ΔΔC_T method. Briefly, threshold cycle (C_T) values obtained by IP reactions were normalized by the no-antibody control for each cell type, which was further normalized by the Fob1Δ control and followed by linear conversion of values using the formula 2^(−ΔΔCT). The primer pairs RDN1F (5′AGGGCTTTCACAAAGCTTCC3′)-RDN1R (5′TCCCCACTGTTCACTGTTCA3′) was used for PCR.

RLS determination.

A knockout/knock-in method was employed to create yeast mutant strains for RLS determination. All fob1 mutants were made from S. cerevisiae strain YPK9 (MATa ade2-101^ochre his3-Δ200 leu2-Δ1 lys2-801^amber trp1-Δ63 ura3-52) (29). The FOB1 gene was knocked out with replacement by the Ura3 gene. The resulting fob1Δ strain was transformed with a PCR fragment amplified from a plasmid containing the desired fob1 mutants (fob1^{S467A, S468A, S519A} and fob1^{S467D, S468D, S519D}).

RLS determination was carried out on yeast-peptone-dextrose (YPD) agar plates, as described previously (30). Briefly, cells were spotted on the plates, and individual unbudded cells were micromanipulated to isolated spots. After the cell budded, the mother was removed and discarded. The RLS determination was initiated with the newborn daughter or virgin cell that had never budded before. RLSs of 40 cells of each strain were determined at 30°C. At each division, the daughter was removed and discarded, and the mother was counted another generation older. This was continued until the mother cell ceased budding completely and lost refractility. RLSs were determined for at least two separately isolated clones for each strain. The Mann-Whitney test was used to statistically evaluate any RLS differences in each experiment, with two-tailed P values reported (Table 4).

TABLE 4.

Statistical analysis of RLS data

Strain	Mean RLS (generations)	P value compared to:
		WT	fob1Δ mutant	fob1AAA mutant
WT	18.4
fob1Δ mutant	31.5	0.0001
fob1AAA mutant	32.1	0.0001	0.32
fob1DDD mutant	18.3	0.664	0.0001	0.0001

Silencing assay.

The gene silencing assay was carried out as described previously (31). The mURA3 cassette used for silencing was present at the end of the ribosomal DNA array in strain YSB348 (32). To study silencing by Fob1 and its mutant forms, plasmids containing these different ORFs were transformed into strain YSB348Fob1Δ and selected on SD/Leu⁻ plates. The transformants were grown in SD/Leu⁻ liquid medium overnight at 30°C. The cultures were adjusted to an A₆₀₀ of 2.0 and were washed with water. The test cultures with controls were serially diluted in 10-fold increments, spotted on SD/Leu⁻ and SD/Leu⁻ Ura⁻ plates, and incubated at 30°C for 2 days. All plates were scanned, and results in 3 replicates were recorded.

Protein-protein interaction in vitro.

The WT Fob1 and Net1 were expressed and purified as tandem affinity purification (TAP) tag fusion proteins from yeast. WT Fob1 was also expressed as a glutathione S-transferase (GST) fusion and purified from yeast. In addition, WT Fob1, its T322I mutant form, Sir2, and the N-terminal peptide from amino acid 1 to 341 of Net1 with a kinase tag were cloned into pMAL vector that included a tobacco etch virus (TEV) protease recognition site that enabled us to cleave off the kinase-tagged protein from the maltose binding protein (MBP) tag. The yeast and bacterial strains and plasmids used for expressing various genes and their derivatives are shown in Tables 1 and 2. Fob1 fusion proteins expressed in yeast and Escherichia coli were immobilized on GST-agarose, MBP-agarose, or calmodulin-agarose affinity columns, depending on the affinity tag, and purified. The immobilized proteins on the respective affinity beads were allowed to interact with [γ-³²P]ATP-labeled putative interacting proteins or peptides, eluted, and resolved in SDS-polyacrylamide gels. The images from the phosphorimager were quantified with the Image J program (NIH).

Brewer-Fangman 2D gels.

The DNA replication intermediates were digested with EcoRV and BglII and resolved in two-dimensional (2D) gels to follow replication fork arrest as described previously (11, 33). After blotting onto nitrocellulose membranes, Southern blots of replication intermediates were developed using an α-³²P-labeled rDNA probe, washed, and visualized with a phosphorimager.

4C analysis.

4C analyses were performed as described previously (17), with some additional modifications, to investigate the effect of phosphorylation of C-Fob1 and of Sir2 on long-range Ter-Ter interactions. The various strains used in the experiments were transformed with plasmid pBBHyg2 containing a modified NTS1 sequence that was constructed as follows. It was necessary to modify the plasmid-borne NTS1 sequence from the natural chromosomal NTS1 sequence, which is 525 bp long and flanked on either ends by AflIII sites (Fig. 2A, blue arrows) to enable us to distinguish it from chromosomal NTS1, as discussed below.

FIG 2.

Impact of Sir2 and phosphorylated Fob1 on chromosome kissing as determined by 4C analysis. (A) Schematic representations of Fob1-mediated trans interactions between a Ter site located in the chromosomal NTS1 (blue arrows) in the rDNA array and a plasmid-borne modified NTS1 (depicted as a green arrow with a red Ter site), which is extended by 150 bp of a non-rDNA tag (red line). (B) Fob1-mediated cis interaction caused by DNA looping between two chromosomal NTS1s and captured by the 4C procedure (blue circle with tandem arrows). (C) Capture by the chromosomal NTS1 (blue arrow) of the plasmid-based modified NTS1 (green arrow with red extension) (in trans). (D) Extension of the trans ligation product by primer pair 2-4 captured (in the presence of WT Fob1), as expected, a 675-bp PCR product. PCR by the primer pairs generated a 525-bp product (cis interaction) that also required Fob1 and ligation following chromosome capture. Fob1^AAA did not show any 675-bp product from either the Sir2 or Sir2Δ samples. The WT Fob1 (Sir2Δ) sample showed a majority 675-bp product, whereas it was minimized in the WT Fob1 (Sir2) sample. The Fob1^AAA sample from either Sir2 or Sir2Δ cells did not reveal a detectable 675-bp product. (E) Primer pair 1-3 generated a 525-bp product (by amplifying the bait sequence) that, as expected, did not require Fob1 or ligation following capture to form the circular template shown in panels B and C. (F) Quantification of the WT Fob1 products from 3 replicates amplified with primer pair 2-4 from the trans capture; note that the 675-bp product was in the minority in the WT Fob1 (Sir2), but the pattern was reversed in the WT Fob1 (Sir2Δ) template DNA. (G) Quantification of the PCR products from panel D. (H) Schematic of the plasmid integration reaction. (I) Southern blots showing that plasmid integration occurred only in WT FOB1 sir2Δ cells but not in their SIR2 counterparts.

The plasmid-borne copy of NTS1 (Fig. 2A, green arrow) was mutagenized to remove one of the two flanking AflIII sites and ligated at that end to a 150-bp-long marker sequence (shown in red in Fig. 2A). Therefore, while the chromosomal NTS1 was located in a 525-bp-long AflIII fragment (blue), the plasmid borne-NTS1 was 675 bp in length (green-red). The pHyg2-NTS1^modified plasmid was transformed into yeast cells of the designated genotypes as described above. Transformed yeast cells were grown overnight (O/N) in YPD containing hygromycin (200 μg/ml) at 30°C. The next day, cells were diluted to an A₆₀₀ of 0.2 in 500-ml cultures and incubated at 30°C until the A₆₀₀ reached 0.7. Fresh formaldehyde was added (1% final volume), left for 10 min with constant stirring, and quenched with 125 mM glycine for 15 min at room temperature. Control cells without formaldehyde treatment were also processed in parallel. The cells were harvested and resuspended in spheroplast buffer (0.4 M sorbitol, 0.4 M KCl, 40 mM potassium phosphate buffer [pH 7.2], and 0.5 mM MgCl₂) with 30 mM dithiothreitol (DTT) and centrifuged at 1,500 × g for 10 min. The pellet was again resuspended in spheroplast buffer with 1 mM DTT. Fixed cells were converted to spheroplasts with 10 mg/ml Zymolyase for 1 h at 30°C with very gentle rotation. Spheroplasts were collected by centrifugation and washed, and pellets were finally resuspended in 8 ml RE buffer (50 mM Tris-HCl [pH 7.9], 100 mM MgCl₂, 1 mM DTT, and 100 μg/ml BSA) plus 80 μl of 10% SDS and 440 μl of 20% Triton X-100 and divided into 17 tubes. The tubes were incubated at 65°C for 20 min followed by 37°C for 1 h with shaking.

Lysates were digested with EcoRI and AflIII enzymes (30 units/chromatin sample) O/N at 37°C with constant mixing. The next day, more enzymes (10 units per sample) were added for continuing digestion. DNA was sheared by sonication using a number of short pulses (10 s) with pauses (30 s) while controlling the temperature by keeping the suspension over ice. DNA was fragmented to <1,000 bp and was checked by gel electrophoresis. Enzymes were inactivated by adding 90 μl of 10% SDS–1% Triton X-100 and incubated at 80°C for 10 min, 65°C for 20 min, and 37°C for 1 h. Digested and sheared DNA was pooled, diluted into ligation buffer (10×; 500 mM Tris-HCl [pH 7.5], 100 mM MgCl₂, 100 mM DTT, 10× BSA [10 mg/ml], and 100 mM ATP), and incubated for 1 h at 37°C. T4 ligase (200 units/sample) was added, and ligation was carried out at 4°C for 72 h with continuous slow rotation. Every 12 h, fresh ATP was added to the ligation mix at 2 mM. After completion of ligation for reversal of cross-linking, proteinase K (0.5%, 20 mg/ml) was added, and the sample was incubated O/N at 65°C. The next day, additional proteinase K (0.5%, 20 mg/ml) was added, and incubation was continued for 2 h at 55°C. DNA was precipitated, washed, and resuspended in 500 μl of TE with RNase A (10 mg/ml) at 37°C O/N. DNA was extracted 3 times with phenol-chloroform and was alcohol precipitated in the presence of 10 μg/ml glycogen and washed. The DNA pellet was dissolved in 50 μl of TE. To detect ligated product, PCR was conducted with two pairs of primers, i.e., primers 1-3 and 2-4. PCR conditions were 95°C for 5 min and 18 cycles of 95°C for 10 s, 54°C for 10 s, and 72°C for 10 s). The PCR products resolved in ethidium bromide-stained agarose gels were analyzed with the Image J program (NIH) and quantified. The flow chart for the 4C manipulations is schematically shown in Fig. 2A.

The primer pair 1-3 was used to amplify the chromosomal NTS1 bait sequence of 525 bp (Fig. 2E). The PCR products of each sample were used to calculate the initial concentrations of the template sequence after determining the intensity of the bands using the program Image J. This information was used as an internal standard to normalize the PCR product generated by the primer pair 2-4 (Fig. 2D). It should be noted that generation of the 525-bp PCR product with primer pair 1-3 does not require Fob1-mediated chromosome capture.

In contrast, the primer pair 2-4 generated a 675-bp product for trans capture of the plasmid-borne prey sequence by a 525-bp-long chromosomal NTS1 bait (Fig. 2C). The bait, as expected, also captured another chromosomal 525-bp NTS1 because of Fob-mediated looping (cis) interaction in the rDNA array (Fig. 2B). It should be noted that the two types of products generated by the 2-4 primer pair were dependent on Fob1-mediated chromosome capture (Fig. 2D). Two sets of PCRs using approximately the same amount of starting DNA template for each sample with primer pair 2-4 and separately with primer pair 1-3 were carried out under identical conditions, with the latter PCR amplification serving as an internal control.

The data in Fig. 2D were generated from 3 independent replicates. The quantitative analysis of 4C capture by Fob1^AAA, Fob1^DDD, and the controls was derived from 5 replicates (Fig. 2D). The template DNA generated after trans capture (Fig. 2C) was also amplified with the primer pair 1-3 to generate a 525-bp PCR product from each sample and was used to estimate the quantity of starting template. These data were used to make sure that the PCRs of each sample contained approximately the same amount of starting template DNA. The latter PCR product, it should be noted, was independent of the presence of Fob1 and did not require chromosome capture because it was a measure of total NTS1 DNA present in the bait.

Plasmid integration assay.

The plasmid integration assay was done as previously described (17). It was totally dependent on the absence of Sir2 (Fig. 2H and I).

Mass spectrometry.

For details of enrichment of phosphopeptides and mass spectrometry, see the supplemental material.

RESULTS

Protein-protein interactions and loading of RENT complex at NTS1.

A schematic diagram of the rDNA repeat region of S. cerevisiae is shown in Fig. 1A. Although it has been reported that Net1, a scaffold protein, acts as an adapter in loading Sir2 onto NTS1 (4), our experiments to further analyze regulation of protein-protein interactions between Fob1 and the RENT complex by yeast two-hybrid (Y2H) analysis (24) yielded positive interaction signals for both Fob1-Net1 and Fob1-Sir2 (Fig. 1B, rows 2 and 4, and D). We wished to reconcile our data with the aforementioned published work that favored direct interaction between Net1 and Fob1 and not between the latter and Sir2 for rDNA silencing (4).

We considered two alternative models of loading of Sir2 at NTS1. Model 1 posits that Sir2 protein interacts with both Net1 and Fob1, whereas model 2 suggests that Sir2 directly interacts with Net1 but not with Fob1 (Fig. 1C). As a passenger on Net1, Sir2 gets loaded onto rDNA by Fob1-Net1 interaction. We wished to obtain further evidence to distinguish between the two models. First, using a yeast reverse 2-hybrid (YR2H) selection, we selected several mutants of Fob1 that appeared to disrupt its interaction with Sir2 (Fig. 1G; Table 2). The strategy and the rationale for the YR2H selection and authentication of the missense mutation and elimination of those that caused global misfolding of Fob1 by HOT1 assay are described in Materials and Methods. Briefly, binding of Fob1 to the Ter site, present in the recombinogenic HOT1 element, causes high-frequency recombination between the his4 alleles flanking the ADE5 marker, which causes its loss and is manifested by red-white sectoring. The Fob1 mutants which fail to bind to HOT1 generate solid red colonies (25, 26). Thus, transformation of the fob1 mutant pool into the HOT1 indicator strain helped us to weed out those fob1 mutants that failed to bind to DNA and were considered globally defective for having lost both DNA binding and protein-protein interaction (Fig. 1G).

First, every mutant form of Fob1 that appeared to be defective in interaction with Sir2 in Y2H screens was also found to be defective in interaction with Net1. Second, the Sir2 L159S mutant, which is known to be defective in interaction with Net1 (34), did not interact with Fob1 (Fig. 1B, row 5, and D; Table 2). Third, we purified calmodulin binding peptide-tagged Fob1 (present in a tandem affinity purification [TAP] module) and similarly tagged Net1, separately immobilized the fusion proteins on a calmodulin-Sepharose affinity matrix, and challenged these with an equal range of concentrations of ³²P-labeled Sir2. The data showed that Net1, but not Fob1, bound to Sir2 (Fig. 1E). Taken together, the 3 lines of evidence were consistent with model 2 (Fig. 1C). It should be noted that mutant forms of Fob1 that mostly retained the ability to interact with themselves in comparison with WT Fob1 were defective to various degrees in their ability to silence rDNA, thereby separating Fob1 oligomerization from rDNA silencing (Table 2).

The experimental characterizations of the phenotypes of the selected mutants are presented in Fig. 3 (also see Table 2). For example, Fob1^K89T showed a significant reduction in Fob1 self-interaction in comparison with that of the WT Fob1 protein but showed a modest reduction in interaction with Net1. In contrast, Fob1^T322I, while retaining a substantial level of self-interaction (>80% of the WT level), caused a significant reduction (to 16 to 17% of the WT levels) in Net1-Fob1 interaction (Fig. 3A and B and Table 2). It should be noted that none of the mutations caused global inactivation of Fob1, because these mutants retained their ability to arrest replication forks at Ter, as revealed by Brewer-Fangman 2D gels (Fig. 3C).

FIG 3.

Fob1-Fob1 and Net1 interaction with the WT and single point mutants of Fob1. (A) Y2H data (β-galactosidase activities) showing that while Fob1^K89T and Fob1^M213L mostly lost their ability to oligomerize, Fob1^T322I retained >80% of its ability to oligomerize in comparison with WT Fob1. (B) In contrast with the data shown in panel A, whereas K89T and M213L both retained ∼80% of the WT level of interaction with Net1, Y322I had a greatly reduced (just above background level) ability to interact with Net1. (C) Brewer-Fangman 2D gel data showing that all three mutant forms of Fob1 shown in panels A and B retained their ability to arrest replication forks, thereby showing that the mutations did not cause global inactivation of the mutant forms of the protein.

Fob1 specifically interacts with the N-terminal region of Net1.

We wished to localize the Fob1-interacting domain of Net1, which is a large, multifunctional protein, by investigation of its partially deleted forms for interaction with Fob1, by Y2H analysis, and by direct binding of purified proteins to each other. We discovered that the N-terminal 341 residues of Net1 contained the Fob1 interaction domain. It was detected using Y2H analysis on the basis of the expression of the stringent Ade reporter and was further confirmed and quantified by measurements of the activity of the lacZ reporter (Fig. 4A to C). The data were also confirmed by expressing a kinase-tagged N-terminal peptide of Net1 including residues 1 to 341 expressed in E. coli and labeling it with [γ-³²P]ATP and muscle kinase. The peptide tag has a recognition site for muscle kinase so that the fusion protein can be labeled with [γ-³²P]ATP, as described previously (11). Purified WT Fob1 and the mutant T322I form were expressed as glutathione S-transferase (GST) fusion proteins in yeast, purified, and immobilized on GST columns. The magnitude of binding of labeled N-terminal Net1 to the immobilized protein forms was measured. The background binding to GST was low and was subtracted from the data points. The results showed that the N-terminal peptide of Net1, as expected, readily bound to WT Fob1 and that the T322I mutant form, as expected, showed significantly reduced binding (Fig. 4D).

FIG 4.

The N-terminal domain of Net1 interacts with Fob1. (A) Schematic representation of the peptides of Net1 examined for their interaction with Fob1. (B) Y2H data showing that the peptide from residue 1 to 341, but not that from residue 566 to 801, of Net1 interacts with Fob1. (C) β-Galactosidase activities of the lacZ reporter, confirming the data shown in panel B. (D) Direct protein-protein interactions show that the N-terminal peptide from residue 1 to 341 of WT Net1 binds to WT Fob1, whereas Fob1 T322I, which is known to be defective in interaction with Net1 (see the text), shows reduced interaction.

Identification of additional phosphorylation sites in C-Fob1 by mass spectrometry.

Our previous work, using the phosphorylation data available online (Phosphogrid; www.phosphoGRID.org), has revealed that Ala substitutions at T504 and S519 of the C-terminal domain of Fob1 reduced oligomerization of Fob1 monomers, whereas phosphomimetic Asp substitutions at the same residues partially restored Fob1-Fob1 interaction. However, the contribution of T504 substitutions to the phenotypic differences was minimal, and therefore the double mutants (T504A/D S519A/D) were not robust enough to help investigate directly the effect of phosphorylation on protein-protein interactions and on Fob1-mediated trans interactions. Therefore, the S504 residue was not studied further. We wished to identify additional phosphorylation sites in the C-terminal domain of Fob1 (C-Fob1) in the hope that multiple substitutions at these residues might generate mutant forms with stronger phenotypes. Mass spectrometry of phosphorylated peptides identified two additional phosphorylated serine residues, S467 and S468, in C-Fob1 (see Fig. S1 in the supplemental material) as potential targets for mutagenesis. We substituted Ala and separately Asp for the Ser residues at positions 467, 468, and 519 to generate fob1^AAA and fob1^DDD triple mutants, respectively, by site-directed mutagenesis (Fig. 5A).

FIG 5.

Phosphorylation of C-terminal Ser residues of Fob1 controls its ability to interact with Net1 and reduce the replicative life span. (A) Schematic representation of the 3 critical phosphoserine residues of Fob1 (see Fig. S1 in the supplemental material) that regulate Fob1-Net1 interaction. (B) Left, Y2H data showing that Fob1^AAA reduced self-oligomerization to background levels, whereas Fob1^DDD restored it close to that of the WT. Middle, Fob1^AAA reduces interaction with Net1 down to background levels, whereas the Fob1^DDD form restores the interaction to close to the WT Fob1 level. Right, in contrast, phosphorylation of Fob1 has a very modest effect on its interaction with Ytt1. (C) 2D gel analyses show that neither the AAA nor the DDD mutant form of the protein has a detectable change in ability to arrest replication forks. (D) Western blots of Fob1 mutant forms. Actin was used as the loading control in each case, showing that the mutations did not detectably alter the intracellular levels of the protein. (E) RLS of strain YPK9 carrying a wild-type FOB1 and of the isogenic fob1Δ, fob1^{S467A,S468A,S519A}, and fob1^{S467D,S468D,S519D} mutant strains.

Phosphorylation of C-Fob1 is essential for recruitment of the Net1-Sir2 complex.

In order to investigate the possible effects of phosphorylation of Fob1 on Net1 recruitment (and by extension that of its passenger protein Sir2), we examined protein-protein interactions between WT Fob1, Fob1^{S467A,S468A,S519A} (here called Fob1^AAA), and Fob1^{S467D,S468D,S519D} (here called Fob1^DDD) by Y2H analyses (Fig. 5). The activities of the LacZ reporter, used for the quantifications of Y2H data, were derived from three independent sets of experiments, and each data point was repeated three times within each set. The data are shown with standard error bars in Fig. 5B. The left panel shows that in the Fob1^AAA protein, Fob1-Fob1 interaction was reduced to almost background levels. In contrast, in the Fob1^DDD protein, Fob1-Fob1 interaction was restored almost to the normal WT levels. The impact of the AAA and DDD mutations on interaction between Fob1 and Net1 is shown in the middle panel of Fig. 5B. The data show that whereas in the Fob1^AAA mutant form Fob1-Net1 interactions were reduced down to background levels, in the Fob1^DDD mutant the interaction was restored almost to the WT levels.

Does phosphorylation of C-Fob1 at the specified residues (Fig. 5A) impact protein-protein interactions with all of its known interacting partners? The answer to this question is in the negative because Ytt1 (yeast transcription terminator) protein, encoded in the Ydr026C open reading frame (ORF) (35) (Fig. 5B, right panel), showed only a small difference in its interaction with the two Fob1 triple mutant forms of the protein. YTT1 was initially discovered by us in a yeast monohybrid screen for genes that encoded proteins interacting with the Ter region of NTS (11). The protein was subsequently called NsiI (36). Since, there is already a restriction enzyme with the identical name, and in order to avoid possible confusion in nomenclature, we suggest the name YTT1 for this gene. It should be noted that the mutants of Fob1 described here, with the exception of the fob1Δ mutant, were able to arrest replication forks, which suggests that these were not globally misfolded (Fig. 5C). Furthermore, the differences in protein-protein interaction as determined by Y2H analysis could not be attributed to changes in the intracellular levels of expression of the WT and the mutant forms of the proteins because Western blots showed that the protein levels were approximately equivalent. Actin was used as a loading control (Fig. 5D).

Effect of phosphorylation of Fob1 on RLS.

In order to measure the physiological impact of the fob1^AAA and fob1^DDD mutations on silencing and recombination, etc., we used RLS as an additional biological readout. All RLS measurements were carried out using two independent clones of each mutant. WT FOB1, fob1Δ, fob1^AAA, and fob1^DDD strains were grown and micromanipulated to separate the daughter cells away from the mother cells until they stopped dividing in order to measure their RLS expressed as cell survival as a function of the number of cell generations or divisions. This is described in more detail in the preceding section (30). The results showed that whereas the WT cells (YPK9) and the isogenic fob1^DDD strain had an average RLS of ∼18 generations, both the fob1Δ and the fob1^AAA strains had an average RLS of ∼31.5 generations (Fig. 5E and Table 4). The relative differences in the RLSs of the AAA and DDD fob1 mutants were robust and clearly distinct from each other. The data showed that the DDD form had a significantly shorter RLS than the AAA form (Fig. 5E). The latter form was substantially deficient in recruiting Sir2 via the RENT pathway but had an extended RLS. We have suggested an explanation for this apparent conundrum in Discussion.

4C analysis for detection of long-range Ter-Ter interactions.

We wished to investigate the possible effects of Fob1 phosphorylation and the potential regulatory role of Sir2, if any, on long-range Fob1-mediated Ter-Ter interactions by using a modified circular chromosome conformation capture (4C) approach. This is described in detail in Materials and Methods and is schematically presented in detail in Fig. 2.

Briefly, we wished to detect relative cis and trans interactions mediated by WT Fob1 and its AAA and DDD mutant forms of the protein in the presence and absence of Sir2. As shown in Fig. 2A, a looping interaction between two chromosomal NTS1 sequences in cis (Fig. 2B), when challenged by primer pair 2-4, yielded a 525-bp PCR product that was Fob1 and ligation dependent. In contrast, such an interaction between chromosomal NTS1 bait (blue arrow) and the presumptive prey, the 675-bp-long plasmid-borne, modified NTS1 (green arrow) with a 150-bp-long foreign DNA sequence (red) flanked by AflIII sites, yielded a 675-bp PCR product diagnostic of a trans interaction. This product was observed when WT Fob1 in the presence and absence of Sir2 and Fob1^DDD only in Sir2Δ samples were used as templates (Fig. 2D, E, and G). It should be noted that while the 675-bp band was the minority product in WT Fob1 (Sir2) samples, it was the majority product in the Fob1 WT (Sir2Δ) sample and was not detectable in Fob1^AAA (Sir2 or Sir2Δ) samples when the primer pair 2-4 was used in PCR amplification (Fig. 2C and D).

Keeping in mind that the WT Fob1 is phosphorylated by an unidentified kinase (17) that acts in a direction opposite to that of Sir2 during recombination and loss of rDNA silencing (LRS) (37), we decided to eliminate the possible effect of this kinase from further consideration by focusing on the investigation of the regulatory impact of Sir2, if any, on long-range interactions by using the mutants Fob1Δ, Fob1^AAA, and Fob1^DDD, which function independently of this phosphorylation. A representative picture of an agarose gel of the PCR products generated by the primer pair 2-4 for each sample is shown in Fig. 2D, and the quantifications of the PCR products are displayed in Fig. 2G. The Fob1^AAA sample derived from either Sir2 or Sir2Δ cells generated no detectable 675-bp band diagnostic of plasmid NTS1-chromosomal NTS1 (trans interaction) but contained various amounts of the 525-bp product that were believed to have been generated by cis interactions between the bait and a chromosomal NTS1 caused by Fob1-mediated DNA looping. In contrast, the Fob1^DDD samples from Sir2Δ cells showed the 675-bp diagnostic bands only in the samples from Sir2Δ cells but not in those prepared from the Sir2 cells (Fig. 2D and G). We wish to point out that a trans interaction, as suggested diagrammatically in Fig. 2H, leads to plasmid integration only in sir2Δ but not in SIR2 cells, as confirmed by the Southern blots visualized with a labeled plasmid-specific probe (Fig. 2I).

In summary, the data supported the conclusion that Sir2 downregulated long-range Fob1-mediated interaction between the bait and the prey sequences occurring in trans. However, it allowed some cis interaction between chromosomal NTS1 sequences to occur. Since the 4C approach as employed here is mostly qualitative in nature, it is not regarded as suitable for finer quantitative measurements. Other methods have to be invented in the future for conducting finer measurements of the extent of cis (and trans) interactions. We conclude from the data that Sir2-modulated long-range Ter-Ter interactions occur in trans and that phosphorylation of the C-Fob1 also regulates the same interactions.

Regulation of rDNA silencing by Fob1 phosphorylation.

We wished to investigate the possible impacts of the fob1^AAA and fob1^DDD mutations on rDNA silencing, which was measured by the activity of the mURA3 reporter inserted immediately downstream from the Ter site located in the last, centromere-proximal NTS1 of the rDNA array in chromosome XII (4, 38) (Fig. 6C). First, we wished to measure recruitment of Sir2 to the rDNA at or near the Ter sites by quantitative chromatin immunoprecipitation (qChIP). For this purpose, we replaced chromosomal WT FOB1 with the DDD and AAA mutant forms. All of the strains harbored a 13-Myc epitope-tagged Sir2. We performed qChIP analyses, as described in detail in a preceding section, and measured the relative levels of Sir2 recruited to the region of rDNA at or near the Ter sites. The data showed that the relative amounts of Sir2 recruited were fob1^DDD > WT FOB1 > fob1^AAA (Fig. 6A).

FIG 6.

Silencing of rDNA. (A) qChIP analysis of Sir2 in the region of the Ter sites. (B) Left panel, spot test for growth of 1:20 dilutions for measuring silencing of the mURA3 reporter as revealed by growth on Ura dropout plates; right panel, effects of selected fob1 single point mutants. (C) A model showing how phosphorylation-dependent “opening” of Fob1 promotes its interaction with Net1, leading to rDNA silencing.

Measurements of the relative levels of silencing elicited by these fob1 mutants, as contrasted with that of WT FOB1, by colony spotting experiments on selective medium were not inconsistent with the ChIP data for Net1 recruitment and by extension that of Sir2 to NTS1. By monitoring the degrees of transcriptional activation of the mURA3 reporter (Fig. 6B and C), a comparative analysis of the silencing abilities of the mutants under study could be carried out. The appropriate strains were constructed by complementing a fob1Δ strain in vivo with WT FOB1 and its various mutant forms, carried in a LEU2 plasmid vector. These were grown to log phase, and serial dilutions of the concentrated cell suspensions were spotted onto both Leu dropout and Leu and Ura dropout plates and incubated for various periods of time at 30°C. Figure 6B shows spot tests for growth of cultures diluted 1/20, which displayed the clearest differences in silencing among the different genotypes under our experimental conditions. Each silencing experiment was done in 3 sets. The results showed, on the basis of the extent of growth on the Leu and Ura dropout plates, that WT FOB1, as expected, caused significant mURA3 silencing. In contrast, there was greatly reduced silencing of the mURA3 reporter in the fob1Δ and fob1^AAA cells. The fob1^DDD mutant showed significant restoration of silencing in comparison with the WT FOB1 and fob1^AAA cells. The results (Fig. 6B, top) clearly supported the conclusion that phosphorylation of C-Fob1 regulated rDNA silencing.

We also examined the K89T, M213L, T322I, and E373V single mutants of fob1 with the goal to achieve separation of the various functions of Fob1 from rDNA silencing (Table 2). We found that these were all partially defective in silencing. The E373V mutant, which is severely defective in Fob1-Fob1 interaction and in long-range trans interactions (17), was as proficient in silencing as WT FOB1. While the K89T and M213L mutants were defective in self-interaction, these retained full fork arrest activity in comparison with WT Fob1. T322I retained high levels of self-interaction while suffering significant loss of interaction with Net1 (Fig. 2C and 6B, bottom; Table 2). As shown before, the E373V mutant was normal in these activities in comparison with the WT but had lost both its self-interaction and the ability to promote plasmid integration. It also had a significantly reduced RLS (17).

A model of the regulation of Sir2 loading by Fob1 phosphorylation is shown in Fig. 6C. It posits that without phosphorylation of C-Fob1, it remains bound to N-Fob1 and allows it to interact with Net1 and by extension with its passenger protein Sir2 and loads the latter onto NTS1 of rDNA. We speculate that the Fob1^AAA form remains constitutively closed and the Fob1^DDD form remains constitutively open to interaction with Net1.

Interaction between Fob1 and Tof2 is regulated by Fob1 phosphorylation.

Although the Net1 pathway seems to be the major vehicle for Sir2 loading, it has been reported that there is an alternative pathway that depends on the interaction between Fob1 and Tof2. The latter form a complex with two proteins of the monopolin complex called Csm1 and Lrs4 and two inner membrane proteins called Huh1 and Nur1 (8, 9). Tof2, besides interaction with Fob1, also interacts with Sir2 and recruits it to NTS1, as shown by a pulldown assay (8). The latter pathway also loads condensin to rDNA (Fig. 7A).

FIG 7.

Fob1 phosphorylation is also required for loading of the Tof2-silencing complex at Ter sites. (A) Schematic diagram showing that interaction of Tof2 with Fob1 and with the monopolin complex and the two inner membrane proteins Heh1 and Nur1 loads Sir2 onto NTS1 and cohesin onto rDNA. (B) Two-hybrid analysis showing that WT Fob1 and Fob1^DDD, but not Fob1^AAA, interact with Tof2. (C) β-Galactosidase reporter activity, confirming the 2-hybrid data shown in panel B.

We wished to find out whether phosphorylation of the C-terminal domain of Fob1 regulated its interaction with Tof2. We carried out Y2H analyses using both the ADE and the lacZ reporters to measure the interaction of Tof2 with the WT and the two triple mutant forms of Fob1 (Fig. 7B). The data clearly and consistently showed that WT Fob1 interacted with Tof2. The data were quantified by measuring the activities of the lacZ reporter present in the 2-hybrid indicator strain (Fig. 7C). Whereas the AAA form significantly reduced the interaction, the phosphomimetic DDD mutant form restored the interaction to close to the level of the WT Fob1.

DISCUSSION

Previous work has established that Sir2 causes rDNA silencing in yeast by inhibiting transcription catalyzed by RNA polymerase II and intrachromatid recombination (32, 39) and that the two protein complexes called RENT and the monopolin complex interact with the replication terminator protein Fob1 to recruit Sir2 onto rDNA. However, the mechanism of regulation of Sir2 recruitment was unknown. Our previous work showed that the C-terminal domain of Fob1 is inhibitory and interacts with the N-terminal activity domain to inhibit Fob1-Fob1 interaction that suppresses intrachromatid recombination and enhances the RLS. Phosphorylation of its C-terminal domain inhibits the intramolecular interaction and promotes Ter-Ter interaction in trans (17). In this work, we wished to address further the mechanism of regulation of rDNA silencing, intrachromatid recombination, and RLS by investigating what regulates Fob1-mediated Sir2 recruitment and what are the factors that modulate long-range trans interactions between Ter sites. The data presented in this work show that phosphorylation of C-Fob1 is essential for both Net1-Fob1 and Tof2-Fob1 interactions. Because Sir2 is a passenger protein, on both Net1 and the Tof2 complexes, its recruitment, by extension, is also regulated by phosphorylation of C-Fob1. Therefore, we conclude that regulation of rDNA silencing, intrachromatid recombination, and RLS is controlled by C-Fob1 phosphorylation.

Sir2 is known to suppress transcription emanating from a bipolar promoter called Epro, which is located in the spacer region of rDNA (13). The transcription is believed to displace cohesin rings from the NTS1 of rDNA, thereby allowing the unconstrained chromatids to undergo intrachromatid recombination. This causes repeat expansion and contraction (13). Is this the only mode of action of Sir2 in rDNA silencing? The following work shows that Sir2 activity in the control of rDNA stability is multimodal.

It has been suggested that Sir2 deacetylates histones and that the resultant altered chromatin structure renders it inaccessible to the recombination machinery (39) that triggers repeat expansion and contraction and reduction of RLS. The work reported here also shows that Sir2 downregulates long-range trans interactions at Ter sites, thereby providing another mechanism for modulation of intrachromatid recombination. Long-range Ter-Ter interaction has been shown to be a key step in initiating intrachromatid recombination (17). These interactions, as drivers of 3-dimensional control of various DNA transactions, have emerged as a topic of considerable current interest.

Herman Muller was the first to point out that somatic pairing in Drosophila chromosomes indicated that a regulatory factor encoded in one chromosome could diffuse across to the corresponding site of its homolog to control gene activation in trans (40). Subsequently, this prediction was experimentally verified and named transvection by E. B. Lewis (41).

What might be the intermediate steps that lead to the long-range interactions on chromosomes, and what are the proteins that mediate these hypothetical steps? While it is clear that oligomerization of a protein that can simultaneously bind to two binding sites on a chromosome in trans can cause the trans interactions, it also seemed reasonable to postulate that accessibility of a binding site on a chromosome to the binding protein(s) should also be regulated by chromatin remodeling and that the latter therefore should also be a controlling factor.

In mammalian cells, a complex of CTCF (CCCTC binding factor) and cohesin promotes long-range interactions. Yeast does not encode CTCF, but the presence of paralogs is a possibility that remains open. Whereas Sir2, which modifies chromatin structure by covalent modification (deacetylation) of K16 of histone H4, acts as a regulatory factor in long-range interactions, it is tempting to suggest that other remodeling proteins, which do not leave covalent signatures on histones but use ATP hydrolysis for chromatin modifications, are also likely to regulate the interactions.

How does Sir2 regulate Fob1-mediated site-site interactions in trans and perhaps also in cis? We have observed that both Sir2 and sir2Δ strains show comparable magnitudes of fork arrest at Ter, which leads to the conclusion that these are probably equally accessible to Fob1 binding independently of Sir2 activity. However, the data presented in this work are not incompatible with the suggestion that Sir2-mediated changes in chromatin structure could significantly reduce the accessibility of a Fob1 monomer bound to a Ter site for interaction with another such complex in trans.

The effects of fob1^AAA and fob1^DDD on RLS deserve additional comments. The two forms of FOB1 were observed to impact RLS differently, as measured in congenic strains expressing or lacking SIR2. Reduction of Fob1 oligomerization caused by the AAA form significantly reduced long-range trans interactions and markedly extended the RLS to the same extent as in the fob1Δ strain. This was observed despite reduced recruitment of Sir2 due to greatly decreased Fob1-Net1 and probably Fob1-Tof2 interaction by the AAA form of FOB1. In contrast, in the DDD mutant, despite enhanced Sir2 recruitment and presumably greater downregulation of trans interactions by Sir2 at Ter, RLS was significantly reduced in comparison with that of the AAA mutant and was indistinguishable from that of the WT. In order to explain this apparent conundrum, we suggest that the regulation of RLS, when measured in our standard Sir2-plus strain and under the experimental conditions used, is mostly controlled by Fob1 self-interaction that is epistatic over that of the impact of Sir2. The epistasis is probably determined by the physiological state of the cells, such as the metabolic flux during a switch from poor growth to rapid growth, and perhaps for other reasons.

A better understanding of this important regulatory mechanism would probably require high-resolution structural information on open and closed forms of Fob1, with and without Ter DNA, and perhaps a ternary complex consisting of Fob1^DDD-Ter DNA and the N-terminal interacting peptide of Net1, which was identified in this work. With these goals in mind, constructions of suitable overproducer strains of yeast and purification of suitable quantities of the various forms of Fob1 are now in progress. Future work, using novel genetic selection schemes coupled with high-throughput screening, we hope will uncover additional candidate proteins that might regulate chromosome kissing. Furthermore, identification of the kinase and hypothetical phosphatase that cause the conformational alterations of Fob1 and their regulation also remains a significant goal for the future.

This article is dedicated to Wolfgang K. Joklik.