Sit4p Protein Phosphatase Is Required for Sensitivity of Saccharomyces cerevisiae to Kluyveromyces lactis Zymocin
Daniel Jablonowski, Andrew R. Butler, Lars Fichtner, Donald Gardiner, Raffael Schaffrath, Michael J. R. Stark


We have identified two Saccharomyces cerevisiae genes that, in high copy, confer resistance to Kluyveromyces lactis zymocin, an inhibitor that blocks cells in the G1 phase of the cell cycle prior to budding and DNA replication. One gene (GRX3) encodes a glutaredoxin and is likely to act at the level of zymocin entry into sensitive cells, while the other encodes Sap155p, one of a family of four related proteins that function positively and interdependently with the Sit4p protein phosphatase. Increased SAP155 dosage protects cells by influencing the sensitivity of the intracellular target and is unique among the four SAP genes in conferring zymocin resistance in high copy, but is antagonized by high-copy SAP185 or SAP190. Since cells lacking SIT4 or deleted for both SAP185 and SAP190 are also zymocin resistant, our data support a model whereby high-copy SAP155 promotes resistance by competition with the endogenous levels of SAP185 and SAP190 expression. Zymocin sensitivity therefore requires a Sap185p/Sap190p-dependent function of Sit4p protein phosphatase. Mutations affecting the RNA polymerase II Elongator complex also confer K. lactis zymocin resistance. Since sit4Δ and SAP-deficient strains share in common several other phenotypes associated with Elongator mutants, Elongator function may be a Sit4p-dependent process.

KILLER strains of the yeast Kluyveromyces lactis secrete a protein toxin or zymocin that inhibits the proliferation of several different yeasts including Saccharomyces cerevisiae (Starket al. 1990; Schaffrath and Breunig 2000). In the presence of K. lactis zymocin, sensitive S. cerevisiae cells become blocked in the G1 phase of the cell division cycle prior to budding and with unreplicated DNA (Whiteet al. 1989; Butleret al. 1991c), suggesting that the zymocin acts to inhibit one or more of the events normally triggered at Start and that are required for G1 exit. Although the zymocin is a heterotrimeric protein, intracellular expression of just the γ-subunit is sufficient to promote the G1 arrest phenotype (Tokunagaet al. 1989; Butleret al. 1991b). The two larger subunits are probably involved in entry of the zymocin into the cell, a process that involves an interaction between the α-subunit and cell wall chitin. Thus chitin-deficient mutants are zymocin resistant (Takita and Castilho-Valavicius 1993; Jablonowskiet al. 2001) and the α-subunit has a domain that shows in vitro chitinase activity and has sequence similarity to other chitinases and chitin-binding proteins (Butleret al. 1991a). Recent work has demonstrated that mutations affecting at least five genes encoding components of the RNA polymerase II Elongator complex lead to zymocin resistance (Frohloffet al. 2001). Elongator is a multisubunit complex that binds to the elongating form of RNA polymerase II (RNAPII: Oteroet al. 1999; Wittschiebenet al. 1999; Fellowset al. 2000) and loss of Elongator function causes a range of phenotypes including delayed activation of genes under changing growth conditions, hypersensitivity to 6-azauracil and caffeine, slow growth, and temperature sensitivity (Oteroet al. 1999; Frohloffet al. 2001). However, Elongator itself is dispensable; so although zymocin inhibition is an Elongator-dependent process, the zymocin cannot simply be acting to block Elongator function.

The execution of Start requires the Sit4p protein phosphatase and temperature-sensitive sit4 mutants to arrest in G1 prior to bud emergence and DNA replication (Suttonet al. 1991; Fernandez-Sarabiaet al. 1992). This requirement is at least in part because Sit4p is needed for proper expression of G1 cyclins such as CLN1, CLN2, and PCL1. However, while ectopic expression of CLN2 from a Sit4p-independent promoter can relieve the block to DNA replication in sit4-102 Ts cells, these cells still remain largely unbudded and thus blocked for bud emergence (Fernandez-Sarabiaet al. 1992). Thus, independently of its effect on CLN2 expression, Sit4p is also needed for other events required for G1 exit. The phenotype of sit4 mutants is complicated by its dependence on a second polymorphic gene, SSD1 (Suttonet al. 1991). In ssd1-d strains deletion of SIT4 is lethal, but in SSD1-v strains sit4 deletion is tolerated, although it leads to a G1 delay and a slow growth rate that is not improved by ectopic CLN2 expression (Fernandez-Sarabiaet al. 1992). Sit4p associates in a cell-cycle-dependent manner with the members of a protein family termed the SAPs (Suttonet al. 1991) and deletion of all four SAP genes (SAP4, SAP155, SAP185, and SAP190) confers the same phenotype as loss of SIT4 (Lukeet al. 1996). A variety of evidence shows that the SAPs and Sit4p function in an interdependent manner, leading to models whereby the SAPs are either positive activators of Sit4p phosphatase or Sit4p effectors (Lukeet al. 1996). Sit4p also binds Tap42p, an element in the TOR signaling pathway; on nutrient starvation or inhibition of the TOR kinases by rapamycin, Tap42p dissociates from Sit4p (Di Como and Arndt 1996). Nutrient starvation or treatment with rapamycin leads to a variety of responses including a G1 arrest, reduced translational initiation, reduced ribosome biosynthesis, and changes to the amino acid transporters expressed on the cell surface (see Schmidtet al. 1998; Cardenaset al. 1999). Dissociation of Sit4p from Tap42p has been proposed to trigger dephosphorylation of at least some of the proteins that act as effectors of the TOR pathway and that mediate some of these responses (Schmidtet al. 1998; Beck and Hall 1999).

We previously isolated two genes that confer K. lactis zymocin resistance when present specifically on high-copy plasmids (Butleret al. 1994). One of these (KTI12) had previously been defined by recessive mutations that also lead to zymocin resistance. Since either loss of KTI12 function or high-copy KTI12 conferred zymocin resistance, we hypothesized that Kti12p could be part of a protein complex required for zymocin action that was perturbed by excess Kti12p (Butleret al. 1994). Consistent with this idea, KTI12 has recently been shown to be allelic with TOT4, a component of the Elongator complex (Frohloffet al. 2001). Here we describe the isolation of GRX3 and SAP155 (encoding one of the Sit4p-associated SAPs) as additional genes that promote zymocin resistance when present in high copy. In addition, we show that cells lacking SIT4 are both refractory to K. lactis zymocin and share many of the other phenotypes of Elongator mutants, placing Sit4p within a pathway required for zymocin action and linking Sit4p to Elongator function.


Strains: All yeast strains used in this study are listed in Table 1. LL20-2 was generated from LL20-1 by transformation with pHO1.5 (a YEp vector carrying the HO and URA3 genes; M. Pocklington) to promote mating-type switching, followed by 5-fluoroorotic acid-induced plasmid loss and verification of mating type by crossing with tester strains (Herskowitz and Jensen 1991; Sikorski and Boeke 1991). CY4029α was generated from CY4029 in the same manner. LFY5 and LFY6 were constructed from AY925 by one-step disruption using the PCR primers described previously (Frohloffet al. 2001).

General methods: All yeast growth media and general yeast methods were as described by Kaiser et al. (1994). Yeast transformation was carried out according to Gietz et al. (1992). When direct selection for the leu2d marker gene was employed, the selective plates were supplemented with 0.0075% (w/v) Bacto-yeast extract. To test multiple strains for growth phenotypes on agar plates, strains were first grown on plates with selection for all markers and then colonies of similar size were resuspended in fresh medium at the same cell density (~106 cells/ml). Culture samples (~5 μl) together with three 10-fold serial dilutions were spotted onto plates using a multipronged inoculating manifold (Dan-Kar). General recombinant DNA procedures were carried out as described by Sambrook et al. (1989). DNA sequencing was performed manually using the Sequenase version 2.0 kit (United States Biochemical, Cleveland) and double-stranded plasmid DNA templates. A combination of specific subclones and synthetic oligonucleotide primers allowed determination of the sequence of SAP155 on both strands. To identify pMA3a clones encoding tRNA3glu sequences, plasmid DNA was digested with EcoRI and SalI and Southern blot analysis was performed using the 435-bp BsaAI-HindIII fragment from pYF1 (Butleret al. 1994). Similar analysis was performed using a 550-bp fragment of KTI12 excised from pJHW27 to identify KTI12-encoding clones. In each case, radiolabeled probes were generated using random hexamer priming (Feinberg and Vogelstein 1983). RNA isolation and RT-PCR were both carried out essentially as described by Frohloff et al. (2001) using the following primer pairs: TOT1 (5′-CTTGGTGTATGAAACTCGCG-3′ and 5′-TTCTTACCTCTGCCAGTACC-3′), TOT2 (5′-AACCTGATGAGACTTCAGGC-3′ and 5′-CAAACCTAACACAGGAACGG-3′), TOT3 (5′-TCAGTCCTTGTACGAAGACG-3′ and 5′-ATAAGCTCGACCTGATCTGG-3′), TOT4 (5′-TCCGGTATCAACTTCACTGC-3′ and 5′-CTTGTTCCGTTACTTACCCC-3′), and HHT1 (5′-AGCAAGAAAGTCCACTGGTG-3′ and 5′-GAATGGCAGCCAAGTTGGTA-3′).

Plasmids: Table 2 summarizes plasmids used in this study. The TAP42 fragment in pDJ14 was obtained by PCR amplification followed by in vivo gap repair (to recover the coding region from the genome) and complemented the lethality of atap42 deletion strain (not shown). The high-copy library used in this study comprised partial Sau3A fragments from yeast genomic DNA (6–10 kb) inserted into the vector pMA3a (leu2d 2μ) and has also been described previously (Crouzet and Tuite 1987). pARB19 was constructed in two stages: the large EcoRI fragment from the pARB106 insert was first ligated into the EcoRI site of YEplac181 and then the SalI fragment from this construct was replaced with the SalI fragment of pARB15 (Table 2) such that the bulk of the original insert of pARB106 was reconstructed. pDG8 was made by first cloning the 1.96-kb XbaI fragment of pARB100 (encoding the promoter and 5′ half of SAP155) into the XbaI site of YEplac181 such that the SAP155 open reading frame (ORF) read in the opposite direction to the lacZ α-fragment. The SalI-PstI interval from the insert was then removed and replaced with the SAP155 SalI-PstI fragment from pDG6, reinstating a complete SAP155 gene. To make a sap155::HIS3 deletion allele, pDG8 was digested with BclI to excise the bulk of the SAP155 coding region, which was replaced by the 1.3-kb HIS3 BamHI fragment of YDpH (Berbenet al. 1991). This resulted in the deletion of codons 85–999 in SAP155 and yielded pDG11 and pDG12, which differ in the orientation of the HIS3 insert.

K. lactis zymocin methods: Killer eclipse assays for zymocin sensitivity were performed as described previously (Kishidaet al. 1996), using K. lactis strains AWJ137 (zymocin producer) and NK40 (nonproducer). For most other assays, serial dilutions of strains (grown under selection for plasmids as appropriate) were spotted out onto YPD agar with or without culture supernatant from AWJ137 as a source of toxin [65% (v/v) unless otherwise stated]. Growth was compared after 2 days incubation at 26°. To test plasmids for conferring resistance to intracellular expression of the zymocin γ-subunit, they were introduced into strains LA and LY (Table 1) with selection on SD medium with appropriate supplements, and then transformants were tested for growth on S agar containing 2% raffinose and 2% galactose. Other strains were similarly tested by introduction of pLF16 (encoding a GAL promoter-zymocin γ-subunit fusion) or YCplac111 (as a control).

View this table:

Yeast strains


Isolation of sequences conferring resistance to K. lactis zymocin in high copy: We previously isolated two S. cerevisiae genes that conferred resistance to K. lactis zymocin when introduced in high copy into sensitive strains of S. cerevisiae (Butleret al. 1994). One of these encoded a tRNA3glu while the other encoded KTI12, a gene previously defined by a recessive, zymocin-resistant mutation and recently shown to encode a component of the RNAPII Elongator complex (Frohloffet al. 2001). Since the previous screen was far from saturated we examined a different high-copy library for additional sequences that could confer zymocin resistance and that might, like KTI12, also correspond to genes defined by our collection of zymocin-resistant mutations (Butleret al. 1994). To avoid obtaining spontaneous zymocin-resistant mutants among our transformants, we conducted the present screen in a homozygous diploid strain (LL20-3) transformed with a yeast genomic library in pMA3a, selecting Leu+ transformants directly. Since pMA3a carries the leu2d marker it is expected to be present at very high copy number when transformants are grown under selection for Leu+ (Erhart and Hollenberg 1983). Cells from ~12,500 independent transformants (~1.5 genome equivalents) were recovered from the selective plates, resuspended, and plated on a range of concentrations of crude K. lactis zymocin in YPD agar. From 5.5 × 106 cells thus screened, ~200 clones that could grow at concentrations of 0.25% crude zymocin or above were selected. From these, 50 plasmids were recovered that conferred resistance to 1.8% crude zymocin when reintroduced into LL20-3. Restriction mapping allowed these plasmids to be classified into several different categories.

View this table:


Since we had already shown that tRNA3glu genes and KTI12 could each confer zymocin resistance in high copy, clones from the current screen were examined by Southern blot analysis using tRNA3glu and KTI12 probes (not shown). This allowed identification of four KTI12 clones and three clones encoding two tRNA3glu loci [tE(UUC)GL2 and tE(UUC)ER3: see Hani and Feldmann (1998)], which were distinct from the two other tRNA3glu loci we had identified previously [tE(UUC)B and tE(UUC)E2: Butler et al. (1994)]. Representative clones from each of the two remaining categories of clone (pARB100 and pARB106) were therefore characterized further.

High-copy GRX3 confers K. lactis zymocin resistance: The entire insert of pARB100 was transferred from pMA3a into both YCplac111 and YEplac181 as an EcoRI-SalI fragment (see Figure 1A). Since the YEplac181 derivative (pARB23) conferred zymocin resistance when transformed into LL20, we concluded that the extra high copy number of the original isolate (pARB100) imposed by leu2d selection was not essential for the insert to function as a multicopy resistance determinant. The low-copy construct (pARB22) was introduced into a representative strain from each of the 13 complementation groups of K. lactis zymocin-resistant mutants described by Butler et al. (1994) and several transformants were tested for zymocin resistance in each case. Since complementation was not observed in any instance (not shown), we concluded that pARB100 does not correspond to any of the previously described KTI loci. pARB22 also failed to confer resistance on wild-type strains, demonstrating that the resistance determinant was required specifically in high copy.

Figure 1.

Characterization of clones that promote zymocin resistance in high copy. Maps indicate key restriction sites in both the insert (—) and the flanking pMA3a vector sequences (Graphic), together with the genes encoded by the insert (Graphic). Below the maps, open boxes (Graphic) indicate subclones that failed to confer zymocin resistance in high copy, while solid boxes (Graphic) indicate subclones that continued to confer zymocin resistance when inserted into YEplac181. (A) pARB100 subclones. The inset compares the zymocin resistance of CY4029 (W303 SSD1-v1) carrying pARB100, pDG14, and pDG15 by plating 10-fold serial dilutions of cells onto YPD agar with or without K. lactis zymocin with a high-copy Formula clone (pYF1) as a control. (B) pARB106 subclones.

DNA sequencing revealed that pARB100 carried an insert from chromosome IV that carried three genes (Figure 1A). Subclones from pARB100 were made in YEplac181 and tested for their ability to confer K. lactis zymocin resistance (Figure 1A). A YEplac181 subclone carrying GRX3 (pDG15) was an effective zymocin resistance determinant (Figure 1A, inset). Since this region encodes a tRNAgln in addition to GRX3 and because tRNA3glu genes confer zymocin resistance in high copy, we tested a YEplac181 subclone carrying the tRNAgln alone (pDG14). pDG14 was unable to confer zymocin resistance (Figure 1A, inset), thereby ruling out the tRNAgln as the critical factor.

GRX3 is one of three related yeast glutaredoxin genes that differ from classical glutaredoxins by having a single cysteine residue at the putative active site (Grant 2001). To determine the level at which GRX3 functions, pARB100 was tested for its ability to protect yeast cells from growth arrest by conditional intracellular expression of the zymocin γ-subunit from the GAL promoter. This test has previously been used to distinguish two classes of zymocin-resistant mutant, namely those in which the zymocin's intracellular target is altered to render it insensitive and others that are resistant to exogenous zymocin but still contain a sensitive intracellular target (Butleret al. 1994). Since presence of pARB100 did not protect cells from intracellular expression of the zymocin γ-subunit (not shown), we conclude that high-copy GRX3 is unlikely to function at the level of the zymocin's intracellular target but is more likely to operate by impairing entry of native zymocin. By comparison, high-copy SAP155 (see below), tRNA3glu , or KTI12 (Butleret al. 1994) can each confer protection from zymocin γ-subunit expression.

The zymocin resistance determinant on pARB106 is SAP155: DNA sequencing demonstrated that pARB106 encoded two complete genes on chromosome VI, YFR039c and SAP155. SAP155 is one of four related SAP genes (SAP4, SAP155, SAP185, and SAP190) that show a functional relationship with the Sit4p protein phosphatase (Lukeet al. 1996). pARB106 subclones were made in YEplac181 (Figure 1B) and tested for their ability to confer zymocin resistance. Neither the EcoRI-SalI nor the SalI-SalI fragments derived from the insert could confer zymocin resistance, although a reconstruction of these two fragments (pARB19) was completely functional in the assay. This demonstrated that the extra high copy number of pMA3a was not essential for conferring resistance and also ruled out YFR039c as the resistance determinant. In contrast, SAP155 alone was capable of conferring resistance when subcloned into YEplac181 (Figure 1B, pDG8), confirming it as the resistance determinant on pARB106. Consistent with this conclusion, pDG8 derivatives in which the SAP155 coding region was replaced by HIS3 or a construct that carried a frameshift mutation in SAP155 were unable to confer zymocin resistance (not shown). Unlike pARB100, pARB106 could protect cells from growth arrest by intracellular expression of the zymocin γ-subunit, allowing growth of strain LA on galactose-containing medium (not shown). Thus, unlike GRX3, high-copy SAP155 confers zymocin resistance at the level of its intracellular target. The pARB19 insert was also cloned into YCplac111 and introduced into the representative K. lactis zymocin-resistant mutant strains described above. This construct (pARB18) failed either to confer zymocin resistance on wild-type strains or to complement zymocin resistance in any of the mutants (not shown), indicating that SAP155 is needed in high copy and that it does not correspond to any of the previously described KTI genes.

In the process of identifying SAP155 as the zymocin resistance determinant, a large section of the pARB106 insert surrounding the SalI site was sequenced on both strands (EMBL accession no. AJ318331). The sequenced region (corresponding to chromosome VI bases 233710–237771) differs in six positions from the current sequence of chromosome VI in the Saccharomyces Genome Database. One of these differences (deletion of G234450 in our sequence) extends the SAP155 open reading frame by 97 codons, in agreement with Luke et al. (1996). The putative amino-terminal extension generated by this change shows strong similarity to the amino-terminal sequences of the other three SAPs (Sap4p, Sap185p, and Sap190p; Figure 2) and is therefore likely to be correct.

Figure 2.

Alignment of the predicted amino-terminal sequences of the four SAP proteins. The similarity of the amino-terminal extension to Sap155p predicted by this and other work (Lukeet al. 1996) to Sap4p, Sap185p, and Sap190p is indicated. • indicates the start of the Sap155p ORF predicted from current database entries. The arrow and circled amino acid denote the position of the frameshift (in the glycine codon) that extends the ORF as shown.

Only SAP155 and none of the other SAP genes function as a high-copy zymocin resistance determinant: SAP4, SAP155, SAP185, and SAP190 encode a family of related proteins that function interdependently with the protein phosphatase Sit4p, each of which can partially suppress the Ts growth defect of sit4-102 ssd1-d strains (Lukeet al. 1996). The four SAP proteins fall into two groups based on sequence similarity, with Sap4p and Sap155p in one group and Sap185p and Sap190p in the other. These two groups are functionally distinct since, for example, Sap185p and Sap190p (but not Sap4p and Sap155p) are together required for a Sit4p function that is essential in the absence of BEM2 (Lukeet al. 1996). Although Sap4p could not be found in Sit4p immune precipitates, Sap155p, Sap185, and Sap190p each form complexes separately with Sit4p and, by several criteria, cells lacking all four SAP genes behave very like SIT4-deficient cells (Lukeet al. 1996). We therefore next tested YEp24 clones of all four SAP genes for ability to confer zymocin resistance. Unlike SAP155, each of the other three SAP genes failed to protect cells from inhibition by zymocin (Figure 3A).

In addition to forming complexes with the SAPs, Sit4p also interacts with Tap42p in a TOR-dependent and rapamycin-sensitive manner (Di Como and Arndt 1996). However, high-copy TAP42 also failed to promote zymocin resistance (Figure 3B). We also tested whether other genes that improve the growth defect of sit4 mutants in high copy or when overexpressed could promote zymocin resistance. However, neither high-level expression of CLN2 or PCL1 (from the S. cerevisiae ADH promoter) nor high-copy SIS2 conferred zymocin resistance (not shown). Thus the effect of SAP155 in this assay is highly specific.

Since the phenotype of sit4 mutants is highly dependent on the SSD1 locus, we also compared isogenic ssd1-d and SSD1-v strains for zymocin sensitivity and for the effect of high-copy SAP155. Both ssd1-d and SSD1-v strains are inhibited by zymocin (Figure 4), although ssd1-d cells are slightly more zymocin sensitive (Figure 4). Furthermore, high-copy SAP155 could promote zymocin resistance in both genetic backgrounds (Figure 4), while TAP42 had no effect in either (Figure 3B).

Sit4 phosphatase and the SAP185/SAP190 family are required for zymocin sensitivity: We next tested the zymocin sensitivity of isogenic SSD1-v strains that were either wild type or deleted for SIT4 or various combinations of SAP genes. As shown in Figure 5, deletion of no single SAP gene conferred significant zymocin resistance. In comparison, loss of SIT4 rendered cells resistant to zymocin. High-copy SAP155 was unable to promote additional zymocin resistance in sit4Δ strains (Figure 3C), consistent with high-copy SAP155 acting through Sit4p rather than by some alternative route. Although loss of single SAP genes was without effect on zymocin sensitivity, the double sap185Δ sap190Δ strain was as resistant to zymocin as the sit4Δ strain, and all other multiple SAP deletion strains that also lacked both SAP185 and SAP190 were zymocin resistant (Figure 5). In comparison, cells lacking both SAP4 and SAP155 were fully zymocin sensitive. Like high-copy SAP155, knockout strains lacking SIT4 or both SAP185 and SAP190 promoted zymocin resistance at the level of the zymocin's intracellular target, since these mutant strains were also resistant to intracellular expression of the zymocin γ-subunit from the GAL promoter (Figure 6). Thus zymocin sensitivity requires a Sap185p/Sap190p-mediated Sit4p function.

Since the SAPs have been shown to compete with each other for binding to Sit4p (Lukeet al. 1996), our data suggested a simple model for how high-copy SAP155 might promote zymocin resistance: higher levels of Sap155p in strains with extra copies of SAP155 might out-compete Sap185p and Sap190p for binding to Sit4p, thereby opposing the Sap185/190p-dependent Sit4p function required for zymocin sensitivity. Such a model predicts that extra copies of SAP185 or SAP190 should therefore counteract the effect of SAP155. Figure 3D confirms this prediction, showing that extra copies of either SAP185 or SAP190 greatly reduce the ability of high-copy SAP155 to promote zymocin resistance.

Figure 3.

Effect of high-copy genes on zymocin sensitivity. In A, C, and D, 10-fold serial dilutions (right to left) of either CY4029 (W303 SSD1-v1) or CY3938 (sit4::HIS3) transformed with control URA3 vectors or 2μ-URA3 plasmids carrying SAP genes were replica plated onto YPD (left) or YPD containing zymocin (middle) or assessed using the eclipse assay (A and D, right). In the latter assay, the size of the growth inhibition halo around the K. lactis zymocin-producing colony, inoculated on the edge of a spot of S. cerevisiae cells, indicates the level of sensitivity/resistance shown by the latter. In B, the eclipse assay was used to compare sensitivity of ssd1-d2(AY925) and SSD1-v1(CY4029) strains carrying high-copy TAP42 with untransformed strains.

Loss of Sit4p and SAP function share common phenotypes with Elongator mutations: Since mutations in components of the RNAPII Elongator complex cause zymocin resistance, we next examined whether the sit4Δ or zymocin-resistant sapΔ strains share any of the other phenotypes shown by Elongator mutations. In addition to zymocin resistance, these phenotypes include slow growth, temperature sensitivity, and hypersensitivity to 6-azauracil. As a control for these experiments we used a strain deleted for TOT3/ELP3, which encodes a known component of Elongator (Wittschiebenet al. 1999; Frohloffet al. 2001). Either loss of SIT4 or deletion of multiple SAP genes has already been shown to confer a slow growth phenotype (Lukeet al. 1996) that we have confirmed in our work (not shown). Both sit4Δ and the quadruple sap deletion strain also conferred temperature sensitivity and 6-azauracil hypersensitivity (Figure 7). Loss of Sap185p and Sap190p also conferred both phenotypes, although the level of 6-azauracil hypersensitivity was lower and comparable to that shown by the tot3Δ control strain. Surprisingly, the sap4Δ sap155Δ double mutant was more hypersensitive to 6-azauracil despite failing to confer temperature sensitivity or zymocin resistance. Because SIT4 was first identified through mutations that affect RNAPII transcription of a variety of genes (Arndtet al. 1989), it is possible that Sit4p phosphatase is required to activate Elongator function. Such activation might involve effects on either the expression or the activity of the Elongator complex. However, when the mRNA levels of TOT1/ELP1, TOT2/ELP2, TOT3/ELP3, TOT4/KTI12, and TOT5 were examined by RT-PCR, essentially identical levels of mRNA for each of these Elongator components were found in wild-type and sit4Δ strains (not shown). Furthermore, comparison of the protein levels of Tot1p to Tot5p in strains with or without high-copy SAP155 showed essentially identical protein levels (not shown). Thus, if Sit4p does affect Elongator function, it is more likely to result from post-translational effects. If Sit4p and Elongator do function in a common process, then a double mutant would not be expected to show any additional growth defect. Conversely, if they act in distinct processes, then the slow growth of sit4Δ strains would be predicted to be additive with that of Elongator deletion mutations. We therefore crossed a sit4Δ strain with either TOT3/ELP3 or TOT4/KTI12 deletion strains in a common genetic background and compared the growth defect of the single and double knockout progeny. The absence of any additional growth defect when sit4Δ was combined with either tot3Δ or tot4Δ (Figure 8) is consistent with the notion that the phosphatase and the Elongator complex may function in a common pathway.

Figure 4.

High-copy SAP155 promotes zymocin resistance in ssd1-d2 and SSD1-v1 strains. Tenfold serial dilutions (left to right) of ssd1-d2 (AY925) or SSD1-v1(CY4029) strains with or without high-copy SAP155 (pARB19) were spotted onto YPD agar containing 65% (v/v) culture supernatant from the zymocin nonproducing K. lactis strain NK40 (control) and onto YPD agar containing the indicated concentration of AWJ137 supernatant.


GRX3 and zymocin resistance: Glutaredoxins and thioredoxins are small, heat-stable oxido-reductases containing two active-site cysteine residues that have a variety of proposed roles, including protein folding and repair of oxidatively damaged proteins (Holmgren 1989). Grx3p and Grx4p are two highly related yeast glutaredoxins that belong to a subfamily that also includes Grx5p. Unlike other glutaredoxins and thioredoxins (including yeast Grx1p and Grx2p), Grx3p, Grx4p, and Grx5p have only a single cysteine residue at the active site in place of the usual -C-X-X-C-motif (Rodriguez-Manzanequeet al. 1999; Grant 2001). Similar glutaredoxin-like proteins, apparently with a single, active-site cysteine, are also found in other eukaryotic organisms. Like protein disulfide isomerases (Freedmanet al. 1994), which also contain thioredoxin-like domains, glutaredoxins can catalyze the reduction of disulfide bonds using glutathione, although protein disulfide reduction is generally thought to involve a dithiol mechanism (Bushwelleret al. 1992). It is therefore unclear whether Grx3p and its paralogues can function in this way, although Grx3p and Grx4p do have an additional cysteine elsewhere in the polypeptide that might conceivably be involved in redox function. We have identified GRX3 as a sequence that protects cells from K. lactic zymocin in high copy, but even on the original pMA3a vector (which uses leu2d selection), GRX3 could not protect cells from induction of the zymocin γ-subunit from the GAL promoter. Our data therefore suggest that elevated GRX3 dosage does not act by protecting the intracellular target from zymocin and that the most likely mode of action therefore concerns entry of zymocin into sensitive cells. Zymocin uptake is a process that is poorly understood at present, although initial binding to sensitive cells requires the chitin binding domain and chitinase activity of the α-subunit (Butleret al. 1991a; Jablonowskiet al. 2001). Zymocin entry into the cell may involve the intensely hydrophobic β-subunit, to which the zymocin γ-subunit is known to be disulfide bonded (Starket al. 1990). If this disulfide bond is required for γ-subunit uptake, it is possible that higher-than-normal extracellular levels of Grx3p might promote reduction of this bond and thereby block toxin entry. Since it is most likely that Grx3p is not normally an extracellular protein, it could be released by lysis of some proportion of cells in a colony, thereby protecting the remaining viable cells from zymocin. Alternatively, excess Grx3p within the cell might alter the normal redox balance and interfere with release of the γ-subunit in that way. Finally, high-copy GRX3 might also function less directly in blocking zymocin entry, for example, by leading to reduced cell-wall chitin levels. A better understanding of the zymocin entry process will be required to determine how GRX3 might antagonize zymocin function.

Figure 5.

Sit4 phosphatase and either SAP185 or SAP190 are required for zymocin sensitivity. Tenfold serial dilutions (left to right) of strains were spotted onto YPD agar (left) or YPD agar containing zymocin (middle) or assessed using the eclipse assay (see Figure 3 legend). All strains were W303 SSD1-v1 and either wild type (CY4029) or lacking SIT4 (CY3938) or one or more SAP genes as indicated (CY5220, CY5224, CY917, CY4380, and DJY8-DJY17; see Table 1). Note that the resistance of sit4Δ and some multiple sapΔ strains is evident despite their weaker growth.

Figure 6.

Loss of SIT4 or combined loss of SAP185 and SAP190 is resistant to intracellular zymocin expression. The indicated strains were transformed with either YCplac111 (as a control) or pLF16, which expresses the zymocin γ-subunit from the GAL promoter. Tenfold serial dilutions of cells were plated out on YP medium containing either 2% glucose or 2% galactose. While growth of the sap4Δ sap155Δ strain was inhibited by expression of the zymocin γ-subunit on galactose, strains lacking either SIT4 or both SAP185 and SAP190 were able to grow despite induction of the GAL-zymocin γ-subunit construct. The wild-type control strain (CY4029) behaved identically to CY5220 (not shown).

The role of Sit4p phosphatase in zymocin sensitivity: Starting with the finding that high-copy SAP155 confers resistance to K. lactis zymocin, we have demonstrated in this work that cells need a functional Sit4p protein phosphatase for zymocin sensitivity and that either Sap185p or Sap190p is also required. Although (like zymocin-treated cells) ssd1-d2 strains deficient in Sit4p function arrest in the G1 phase of the cell cycle, Sit4p itself cannot be the intracellular target of the zymocin; SSD1-v1 strains in which Sit4p function is dispensable are sensitive to zymocin inhibition. Sap4p, Sap155p, Sap185p, and Sap190p are Sit4p-associated proteins that could be either regulators or effectors of Sit4p function and our data underscore the division of these four proteins into two families (Sap4p/155p and Sap185p/190p), each of which has at least some unique function. Thus, only combined loss of Sap185p and Sap190p leads to zymocin resistance, while combined loss of Sap4p and Sap155p or triple sap deletion strains still containing either Sap185p or Sap190p are zymocin sensitive. This is consistent with previous work showing that SAP genes from one family are ineffective in high copy at rescuing the growth defect shown by loss of the other family and that sap185Δ sap190Δ (but not sap4Δ sap155Δ) is synthetically lethal with loss of BEM2, as is loss of SIT4.

Figure 7.

sit4Δ zymocin-resistant mutants share other phenotypes with RNAPII Elongator mutants. In A, W303 SSD1-v1 strains of the indicated genotype were tested for growth at 39°. Only the wild type and sap4Δ sap155Δ mutant show significant growth at this temperature, as indicated by the presence of individual colonies growing to the center of the plate. (B) Sensitivity of strains toward 6-azauracil. Tenfold serial dilutions of the indicated strains were plated on SD agar without uracil and either containing or lacking 50 μg/ml 6-azauracil. All strains contained YCplac33 (carrying URA3).

High-copy SAP155 confers zymocin resistance by a mechanism involving competition between the different SAP proteins, since its effect is antagonized by elevated dosage of either SAP185 or SAP190. Although we have not looked directly at the effect of SAP155 on the association of the different SAPs with Sit4p, it has previously been clearly demonstrated that high-copy SAP155 effectively dissociates Sap185p and Sap190p from Sit4p immunoprecipitates, while, conversely, high-copy SAP190 greatly reduced the level of Sap155p in Sit4p immune complexes (Lukeet al. 1996). Thus, given that the sap185Δ sap190Δ double mutant is zymocin resistant, all our data are fully consistent with high-copy SAP155 interfering with the formation of the Sap185p·Sit4p and Sap190p·Sit4p complexes. The alternative model, whereby either elevated dosage of SAP155 or loss of SAP185 and SAP190 promote zymocin resistance by increasing the amount of Sap155p·Sit4p in the cell, is inconsistent with the finding that loss of Sit4p itself also causes resistance.

Figure 8.

Loss of Elongator function fails to show an additive defect with loss of SIT4. W303 SSD1-v1 strains containing tot4/kti12Δ (LFY6; A) or tot3/elp3Δ (LFY5; B) mutations were crossed with CY4029 (sit4Δ) and sporulated, and tetrads were dissected. The growth of spores from a typical tetrad from each cross is shown, together with the doubling time (in minutes) subsequently determined in liquid YPD medium.

Another role of Sit4p phosphatase is in the TOR signaling pathway, where its interaction with Tap42p is important (Di Como and Arndt 1996). Tap42p is bound to Sit4p and to the catalytic subunit of yeast PP2A (PP2AC) under conditions of nutrient sufficiency, but on starvation or when the TOR kinases are inhibited by rapamycin, this interaction is lost. On release from Tap42p, Sit4p has been proposed to dephosphorylate at least some of the downstream effectors of the TOR pathway (Beck and Hall 1999). Since Tap42p binds to Sit4p in a complex distinct from the Sap·Sit4p complexes we also tested whether high-copy TAP42 could confer zymocin resistance. However, its failure to do so suggests that it is unable to compete effectively with Sap185p and Sap190p for binding to Sit4p, despite evidence that more Sit4p can associate with Tap42p in sap-deficient strains (Di Como and Arndt 1996). Since rapamycin also causes cells to arrest in G1 it is attractive to suppose that zymocin might mimic rapamycin in blocking growth-promoting signaling through the TOR pathway. Like rapamycin-treated cells, tap42-11 mutants arrest at their restrictive conditions because they induce a range of responses that is normally inhibited by TOR signaling; at their permissive temperature, however, they show dominant rapamycin resistance because the mutant protein cannot dissociate from Sit4p or PP2AC when TOR activity is inhibited (Di Como and Arndt 1996). However, since we have found that the tap42-11 mutant is fully sensitive to zymocin (not shown), this appears to rule out any effect of zymocin either on TOR itself or any upstream activators of TOR. While zymocin might conceivably block the TOR signaling at the level of Tap42p, the failure of high-copy TAP42 to confer even a low level of zymocin resistance does not support this notion.

Seven genes in addition to SIT4 and SAP185/SAP190 that confer zymocin resistance when deleted have now been identified and, of these, at least five encode components of Elongator, a multiprotein complex associated with the elongating form of RNAPII (Butleret al. 1994; Frohloffet al. 2001; our unpublished data). Elongator mutants share a number of phenotypes in common and we have found that mutants lacking SIT4 also exhibit these phenotypes. Thus an intriguing possibility is that Sit4p and Elongator are functionally linked and that, for example, Elongator function requires dephosphorylation by Sit4p. This is consistent with previous work showing that sit4 mutations can affect the transcription of many different genes (Arndtet al. 1989) and would provide a mechanism to link Sit4p to the transcriptional machinery. Further investigation will reveal if any of the components of the Elongator complex are phosphorylated and, if so, whether their phosphorylation state changes in strains lacking SIT4. Our failure to detect changes in either the mRNA or protein levels of several key components of Elongator would be consistent with such post-translational regulation. Like Sit4p, Elongator function is also dispensable and so it is at first sight difficult to understand how Elongator could be the target of zymocin. One possibility is that zymocin might block the recycling of RNAPII in an Elongator-dependent manner, i.e., by converting Elongator into a cellular poison. In support of this idea, it has been shown that several RNAPII-dependent genes are downregulated by zymocin treatment whereas RNAPI transcription is not affected (Frohloffet al. 2001). In addition, cells with reduced levels of Rpb1p are hypersensitive to zymocin (R. Schaffrath, unpublished data), consistent with the notion that RNAPII function becomes limiting in zymocin-treated cells.


We are extremely grateful to Kim Arndt for supplying many of the plasmids and strains used in this study. Thanks are also due to Mick Tuite for providing the pMA3a library, to Vera Martin for generating strain LL20-2, to Evelyn Tait for the TAP42 clone, to Doug Stirling for critical comments, and to the CRC Nucleic Acid Structure Research Group at Dundee for the synthesis of oligonucleotides. D.J. was supported by a Federation of European Biochemical Societies Fellowship and L.F. by a Federation of European Microbiological Societies Fellowship. This work was supported by project grant FG94/518 from the Agricultural and Food Research council to M.J.R.S. and by grants from Deutsche Forschungsgemeinschaft (Scha 750/2-1 and 2-2) to R.S.


  • Communicating editor: P. Russell

  • Received June 25, 2001.
  • Accepted September 18, 2001.


View Abstract