The cell wall pectic polymer rhamnogalacturonan-II is required for proper pollen tube elongation: implications of a putative sialyltransferase-like protein

Background and Aims

Rhamnogalacturonan-II (RG-II) is one of the pectin motifs found in the cell wall of all land plants. It contains sugars such as 2-keto-3-deoxy-d-lyxo-heptulosaric acid (Dha) and 2-keto-3-deoxy-d-manno-octulosonic acid (Kdo), and within the wall RG-II is mostly found as a dimer via a borate diester cross-link. To date, little is known regarding the biosynthesis of this motif. Here, after a brief review of our current knowledge on RG-II structure, biosynthesis and function in plants, this study explores the implications of the presence of a Golgi-localized sialyltransferase-like 2 (SIA2) protein that is possibly involved in the transfer of Dha or Kdo in the RG-II of Arabidopsis thaliana pollen tubes, a fast-growing cell type used as a model for the study of cell elongation.

Methods

Two heterozygous mutant lines of arabidopsis (sia2-1+/– and qrt1 × sia2-2+/–) were investigated. sia2-2+/– was in a quartet1 background and the inserted T-DNA contained the reporter gene β-glucuronidase (GUS) under the pollen-specific promoter LAT52. Pollen germination and pollen tube phenotype and growth were analysed both in vitro and in vivo by microscopy.

Key Results

Self-pollination of heterozygous lines produced no homozygous plants in the progeny, which may suggest that the mutation could be lethal. Heterozygous mutants displayed a much lower germination rate overall and exhibited a substantial delay in germination (20 h of delay to reach 30 % of pollen grain germination compared with the wild type). In both lines, mutant pollen grains that were able to produce a tube had tubes that were either bursting, abnormal (swollen or dichotomous branching tip) or much shorter compared with wild-type pollen tubes. In vivo, mutant pollen tubes were restricted to the style, whereas the wild-type pollen tubes were detected at the base of the ovary.

Conclusions

This study highlights that the mutation in arabidopsis SIA2 encoding a sialyltransferase-like protein that may transfer Dha or Kdo on the RG-II motif has a dramatic effect on the stability of the pollen tube cell wall.     

Characterization of Arabidopsis CTP:3-deoxy-D-manno-2-octulosonate cytidylyltransferase (CMP-KDO synthetase), the enzyme that activates KDO during rhamnogalacturonan II biosynthesis

In plant cells, boron (B) occurs predominantly as a borate ester associated with rhamnogalacturonan II (RG-II), but the function of this B-RG-II complex has yet to be investigated. 3-Deoxy-D-manno-2-octulosonic acid (KDO) is a specific component monosaccharide of RG-II. Mutant plants defective in KDO biosynthesis are expected to have altered RG-II structure, and would be useful for studying the physiological function of the B-RG-II complex. Here, we characterized Arabidopsis CTP:KDO cytidylyltransferase (CMP-KDO synthetase; CKS), the enzyme activating KDO as a nucleotide sugar prior to its incorporation into RG-II. Our analyses localized the Arabidopsis CKS protein to mitochondria. The Arabidopsis CKS gene occurs as a single-copy gene in the genome, and we could not obtain cks null mutants from T-DNA insertion lines. Analysis using +/cks heterozygotes in the quartet1 background demonstrated that the cks mutation rendered pollen infertile through the inhibition of pollen tube elongation. These results suggest that KDO is an indispensable component of RG-II, and that the complete B-RG-II complex is essential for the cell wall integrity of rapidly growing tissues.     

Regulatory Mutations of Inositol Biosynthesis in Yeast: Isolation of Inositol-Excreting Mutants

The enzyme inositol-1-phosphate synthase (I-1-P synthase), product of the INO1 locus, catalyzes the synthesis of inositol-1-phosphate from the substrate glucose-6-phosphate. The activity of this enzyme is dramatically repressed in the presence of inositol. By selecting for mutants which overproduce and excrete inositol, we have identified mutants constitutive for inositol-1-phosphate synthase as well as a mutation in phospholipid biosynthesis. Genetic analysis of the mutants indicates that at least three loci (designated OPI1, OPI2 and OPI4) direct inositol-mediated repression of I-1-P synthase. Mutants of these loci synthesize I-1-P synthase constitutively. Three loci are unlinked to each other and to INO1, the structural gene for the enzyme. A mutant of a fourth locus, OPI3, does not synthesize I-1-P synthase constitutively, despite its inositol excretion phenotype. This mutant is preliminarily identified as having a defect in phospholipid synthesis.     

The genus-specific lipopolysaccharide epitope of Chlamydia is assembled in C. psittaci and C. trachomatis by glycosyltransferases of low homology

Chlamydiae possess a genus-specific epitope that is located on the lipopolysaccharide (LPS) and is composed of a 3-deoxy-d -manno-octulosonic acid (Kdo) trisaccharide of the sequence αKdo-(2→8)–αKdo–(2→4)-αKdo. In Chlamydia trachomatis, this trisaccharide is biosynthetically generated through the action of a multi-functional Kdo-transferase encoded by the gene gseA. gseA of Chlamydia psittaci 6BC was cloned and expressed in a rough mutant (Re chemotype) of Escherichia coli (strain F515) that contains an LPS with only two α2→4-linked Kdo residues. Recombinant strains were able to add the immunodominant Kdo residue in a α2→8-linkage to the parental LPS, as determined by SDS–PAGE and Western blot analysis using a monoclonal antibody against the genus-specific epitope. The DNA sequence of gseA was determined and aligned to that published recently for C. trachomatis serovar L2. Most surprisingly, the two deduced amino acid sequences shared only an overall homology of 67%. Thus, gseA exhibits species specificity at the DNA level, whereas its gene product results in the synthesis of a carbohydrate antigen with genus specificity.     

Structural characterisation of the pectic polysaccharide rhamnogalacturonan II using an acidic fingerprinting methodology

Rhamnogalacturonan II (RG-II) is a structurally complex cell wall pectic polysaccharide. Despite its complexity, both the structure of RG-II and its ability to dimerise via a borate diester are conserved in vascular plants suggesting that RG-II has a fundamental role in primary cell wall organisation and function. The selection and analysis of new mutants affected in RG-II formation represents a promising strategy to unravel these functions and to identify genes encoding enzymes involved in RG-II biosynthesis. In this paper, a novel fingerprinting strategy is described for the screening of RG-II mutants based on the mild acid hydrolysis of RG-II coupled to the analysis of the resulting fragments by mass spectrometry. This methodology was developed using RG-II fractions isolated from citrus pectins and then validated for RG-II isolated from the Arabidopsis mur1 mutant and irx10 irx10-like double mutant.     

Relaxed sugar donor selectivity of a Sinorhizobium meliloti ortholog of the Rhizobium leguminosarum mannosyl transferase LpcC. Role of the lipopolysaccharide core in symbiosis of Rhizobiaceae with plants

The lpcC gene of Rhizobium leguminosarum and the lpsB gene of Sinorhizobium meliloti encode protein orthologs that are 58% identical over their entire lengths of about 350 amino acid residues. LpcC and LpsB are required for symbiosis with pea and Medicago plants, respectively. S. meliloti lpsB complements a mutant of R. leguminosarum defective in lpcC, but the converse does not occur. LpcC encodes a highly selective mannosyl transferase that utilizes GDP-mannose to glycosylate the inner 3-deoxy-D-manno-octulosonic acid (Kdo) residue of the lipopolysaccharide precursor Kdo(2)-lipid IV(A). We now demonstrate that LpsB can also efficiently mannosylate the same acceptor substrate as does LpcC. Unexpectedly, however, the sugar nucleotide selectivity of LpsB is greatly relaxed compared with that of LpcC. Membranes of the wild-type S. meliloti strain 2011 catalyze the glycosylation of Kdo(2)-[4'-(32)P]lipid IV(A) at comparable rates using a diverse set of sugar nucleotides, including GDP-mannose, ADP-mannose, UDP-glucose, and ADP-glucose. This complex pattern of glycosylation is due entirely to LpsB, since membranes of the S. meliloti lpsB mutant 6963 do not glycosylate Kdo(2)-[4'-(32)P]lipid IV(A) in the presence of any of these sugar nucleotides. Expression of lpsB in E. coli using a T7lac promoter-driven construct results in the appearance of similar multiple glycosyl transferase activities seen in S. meliloti 2011 membranes. Constructs expressing lpcC display only mannosyl transferase activity. We conclude that LpsB, despite its high degree of similarity to LpcC, is a much more versatile glycosyltransferase, probably accounting for the inability of lpcC to complement S. meliloti lpsB mutants. Our findings have important implications for the regulation of core glycosylation in S. meliloti and other bacteria containing LpcC orthologs.     

Identification of essential histidine residues in 3-deoxy-D-manno-octulosonic acid 8-phosphate synthase: analysis by chemical modification with diethyl pyrocarbonate and site-directed mutagenesis

The enzyme 3-deoxy-D-manno-octulosonic acid 8-phosphate (KDO 8-P) synthase from Escherichia coli that catalyzes the aldol-type condensation of D-arabinose 5-phosphate (A 5-P) and phosphoenolpyruvate (PEP) to give KDO 8-P and inorganic phosphate (P(i)) is inactivated by diethyl pyrocarbonate (DEPC). The inactivation is first-order in enzyme and DEPC. A second-order rate constant of 340 M(-1) min(-1) is obtained at pH 7.6 and 4 degrees C. The rate of inactivation is dependent on pH and the pH-inactivation rate data imply the involvement of an amino acid residue with a pK(a) value of 7.3. KDO 8-P synthase activity is not restored to the DEPC-inactivated enzyme following treatment with hydroxylamine. Complete loss of KDO 8-P synthase activity correlates with the ethoxyformylation of three histidine residues by DEPC. KDO 8-P synthase is protected against DEPC inactivation by PEP and partially protected against inactivation by A 5-P. To provide further evidence for the involvement or role of the histidine residues in the aldol-type condensation catalyzed by KDO 8-P synthase, all six histidines were individually mutated to either glycine or alanine. The kinetic constants for the three mutants H40A, H67G, and H246G were unaffected as compared to the wild type enzyme. In contrast, H241G demonstrates a >10-fold increase in K(M) for both PEP and A 5-P and a 4-fold reduction in k(cat), while H97G demonstrates an increase in K(M) for only A 5-P and a 2-fold reduction in k(cat). The activity of the H202G mutant was too low to be measured accurately but the data obtained indicated an approximate 400-fold reduction in k(cat). Circular dichroism measurements of the wild-type and mutant enzymes indicate modest structural changes in only the fully active H67G and H246G mutants. The H241G mutant is protected against DEPC inactivation by PEP and A 5-P to the same extent as the wild-type enzyme, suggesting that the functionally important H241 may not be located in the vicinity of the substrate binding sites. The H97G mutant is protected by PEP against DEPC inactivation to the same degree as the wild-type enzyme but is no longer protected by A 5-P. In the case of the H202G mutant, both A 5-P and PEP protect the mutant against DEPC inactivation but to different extents from those observed for the wild-type enzyme. The catalytic activity of the H97G mutant is partially restored (20% --> 60% of wild-type activity) in the presence of imidazole, while a minor amount of activity is restored to the H202G mutant (<1% --> 4% of wild-type activity) in the presence of imidazole.     

Sphingosine-1-phosphate plays a role in the suppression of lateral pseudopod formation during Dictyostelium discoideum cell migration and chemotaxis

Sphingosine-1-phosphate (S-1-P) is a bioactive lipid that plays a role in diverse biological processes. It functions both as an extracellular ligand through a family of high-affinity G-protein-coupled receptors, and intracellularly as a second messenger. A growing body of evidence has implicated S-1-P in controlling cell movement and chemotaxis in cultured mammalian cells. Mutant D. discoideum cells, in which the gene encoding the S-1-P lyase had been specifically disrupted by homologous recombination, previously were shown to be defective in pseudopod formation, suggesting that a resulting defect might exist in motility and/or chemotaxis. To test this prediction, we analyzed the behavior of mutant cells in buffer, and in both spatial and temporal gradients of the chemoattractant cAMP, using computer-assisted 2-D and 3-D motion analysis systems. Under all conditions, S-1-P lyase null mutants were unable to suppress lateral pseudopod formation like wild-type control cells. This resulted in a reduction in velocity in buffer and spatial gradients of cAMP. Mutant cells exhibited positive chemotaxis in spatial gradients of cAMP, but did so with lowered efficiency, again because of their inability to suppress lateral pseudopod formation. Mutant cells responded normally to simulated temporal waves of cAMP but mimicked the temporal dynamics of natural chemotactic waves. The effect must be intracellular since no homologs of the S-1-P receptors have been identified in the Dictyostelium genome. The defects in the S-1-P lyase null mutants were similar to those seen in mutants lacking the genes for myosin IA, myosin IB, and clathrin, indicating that S-1-P signaling may play a role in modulating the activity or organization of these cytoskeletal elements in the regulation of lateral pseudopod formation.     

Sphingosine kinase regulates the sensitivity of Dictyostelium discoideum cells to the anticancer drug cisplatin

The drug cisplatin is widely used to treat a number of tumor types. However, resistance to the drug, which remains poorly understood, limits its usefulness. Previous work using Dictyostelium discoideum as a model for studying drug resistance showed that mutants lacking sphingosine-1-phosphate (S-1-P) lyase, the enzyme that degrades S-1-P, had increased resistance to cisplatin, whereas mutants overexpressing the enzyme were more sensitive to the drug. S-1-P is synthesized from sphingosine and ATP by the enzyme sphingosine kinase. We have identified two sphingosine kinase genes in D. discoideum--sgkA and sgkB--that are homologous to those of other species. The biochemical properties of the SgkA and SgkB enzymes suggest that they are the equivalent of the human Sphk1 and Sphk2 enzymes, respectively. Disruption of the kinases by homologous recombination (both single and double mutants) or overexpression of the sgkA gene resulted in altered growth rates and altered response to cisplatin. The null mutants showed increased sensitivity to cisplatin, whereas mutants overexpressing the sphingosine kinase resulted in increased resistance compared to the parental cells. The results indicate that both the SgkA and the SgkB enzymes function in regulating cisplatin sensitivity. The increase in sensitivity of the sphingosine kinase-null mutants was reversed by the addition of S-1-P, and the increased resistance of the sphingosine kinase overexpressor mutant was reversed by the inhibitor N,N-dimethylsphingosine. Parallel changes in sensitivity of the null mutants are seen with the platinum-based drug carboplatin but not with doxorubicin, 5-fluorouracil, and etoposide. This pattern of specificity is similar to that observed with the S-1-P lyase mutants and should be useful in designing therapeutic schemes involving more than one drug. This study identifies the sphingosine kinases as new drug targets for modulating the sensitivity to platinum-based drugs.     

Occurence of antibodies against chlamydial lipopolysaccharide in human sera as measured by ELISA using an artificial glycoconjugate antigen

Abstract An artificial glycoconjugate containing, as a ligand, the deacylated carbohydrate backbone of a recombinant Chlamydia -specific lipopolysaccharide was used as a solid-phase antigen in ELISA to measure antibodies against chlamydial LPS. The specificity and reproducibility of the assay was shown by using a panel of prototype monoclonal antibodies representing the spectrum of antibodies also occuring in patient sera. These mAbs recognized Chlamydia -specific epitopes [ α 2→8-linked disaccharide of 3-deoxy- d - manno -octulosonic acid (Kdo) or the trisaccharide α Kdo-(2→8)-→Kdo] or those shared between chlamydial and Re-type LPS ( α Kdo, α →4-linked Kdo disacccharide). The assay was used to measure IgG, IgA and IgM antibodies against chlamydial LPS in patients with genital or respiratory tract infections. In comparison to the results obtained with sera from blood donors, it became evident that both types of infection result in significant changes in the profile of LPS antibodies.     

Engineering of a sialic acid synthesis pathway in transgenic plants by expression of bacterial Neu5Ac-synthesizing enzymes

Plants are a low-cost and contamination-free factory for the production of recombinant pharmaceutical proteins. However, plant-made pharmaceuticals differ from their mammalian homologues by the structure of their N -linked glycans. For instance, most mammalian glycoproteins harbour terminal sialic acids that control their half-life in the bloodstream. The absence of the whole sialylation machinery in plants is of major concern as non-sialylated plant-made pharmaceuticals may not perform at their full potential in humans, because of their removal from the circulation through the involvement of hepatic cell receptors. In this context, we have investigated the synthesis of N -acetylneuraminic acid (Neu5Ac) in the cytosol of plants by either the re-routing of the endogenous 3-deoxy- d - manno -2-octulosonic acid (Kdo) biosynthetic pathway or the expression of microbial Neu5Ac-synthesizing enzymes. In this paper, we demonstrate that the plant Kdo-8P synthase is not able to use N -acetyl d -mannosamine as a substrate, and thus re-routing of the Kdo pathway for the synthesis of Neu5Ac is not possible. Consequently, we expressed genes encoding Neu5Ac lyase from Escherichia coli and Neu5Ac synthase ( neuB2 ) from Campylobacter jejuni in plants. These resulted in the production of functional enzymes in the cytosol, which in turn can catalyse the synthesis of Neu5Ac in vitro . Experiments were carried out on two models, Bright Yellow 2 (BY2) tobacco cells and Medicago sativa (alfalfa), the perennial legume crop.     

Characterization of monoclonal antibodies recognizing three distinct, phosphorylated carbohydrate epitopes in the lipopolysaccharide of the deep rough mutant I-69 Rd−/b+ of Haemophilus influenzae

Monoclonal antibodies against the lipopolysaccharide (LPS) of the deep rough mutant I-69 Rd⁻/b⁺ of Haemophilus influenzae were obtained after immunization of mice with sheep erythrocytes which had been coated with de- O -acylated LPS. Characterization of antibodies was performed by enzyme immuno assay (EIA) using LPS or neoglycoconjugates containing partial structures of LPS as solid-phase antigens and by haemagglutination with sheep erythrocytes coated with de- O -acylated LPS. Binding data were confirmed by EIA inhibition experiments using deacylated LPS or synthetic partial structures thereof. Three antibodies were specific for 3-deoxy- d - manno -octulopyranosonic acid- (Kdo) 5-phosphate, one for Kdo-4-phosphate, and one required, in addition to a Kdo-phosphate, parts of the phosphorylated glucosamine backbone of lipid A. All antibodies also bound in (i) Western blots to bacterial whole-cell lysates or isolated LPS separated by SDS–PAGE, (ii) bacterial colony blots, and (iii) immunofluorescence with live bacteria. The latter result indicated that Kdo-4- and Kdo-5-phosphate are synthesized by the bacteria and are not the result of phosphate migration.     

Elimination of 2-keto-3-deoxy-D-glycero-D-galacto-nonulosonic acid 9-phosphate synthase activity from human N-acetylneuraminic acid 9-phosphate synthase by a single mutation

The most commonly occurring sialic acid Neu5Ac (N-acetylneuraminic acid) and its deaminated form, KDN (2-keto-3-deoxy-D-glycero-D-galacto-nonulosonic acid), participate in many biological functions. The human Neu5Ac-9-P (Neu5Ac 9-phosphate) synthase has the unique ability to catalyse the synthesis of not only Neu5Ac-9-P but also KDN-9-P (KDN 9-phosphate). Both reactions are catalysed by the mechanism of aldol condensation of PEP (phosphoenolpyruvate) with sugar substrates, ManNAc-6-P (N-acetylmannosamine 6-phosphate) or Man-6-P (mannose 6-phosphate). Mouse and putative rat Neu5Ac-9-P synthases, however, do not show KDN-9-P synthase activity, despite sharing high sequence identity (>95%) with the human enzyme. Here, we demonstrate that a single mutation, M42T, in human Neu5Ac-9-P synthase can abolish the KDN-9-P synthase activity completely without compromising the Neu5Ac-9-P synthase activity. Saturation mutagenesis of Met42 of the human Neu5Ac-9-P synthase showed that the substitution with all amino acids except leucine retains only the Neu5Ac-9-P synthase activity at levels comparable with the wild-type enzyme. The M42L mutant, like the wild-type enzyme, showed the additional KDN-9-P synthase activity. In the homology model of human Neu5Ac-9-P synthase, Met42 is located 22 A (1 A=0.1 nm) away from the substrate-binding site and the impact of this distant residue on the enzyme functions is discussed.     

CHS8-a fourth chitin synthase gene of Candida albicans contributes to in vitro chitin synthase activity, but is dispensable for growth

In silico analysis of the genome sequence of the human pathogenic fungus Candida albicans identified an open reading frame encoding a putative fourth member of the chitin synthase gene family. This gene, named CaCHS8, encodes an 1105 amino acid open reading frame with the conserved motifs characteristic of class I zymogenic chitin synthases with closest sequence similarity to the non-essential C. albicans class I CHS2 gene. Although the CaCHS8 gene was expressed in both yeast and hyphal cells, homozygous chs8 Delta null mutants had normal growth rates, cellular morphologies and chitin contents. The null mutant strains had a 25% reduction in chitin synthase activity and were hypersensitive to Calcofluor White. A chs2 Delta chs8 Delta double mutant had less than 3% of normal chitin synthase activity and had increased wall glucan and decreased mannan but was unaffected in growth or cell morphology. The C. albicans class I double mutant did not exhibit a bud-lysis phenotype as found in the class I chs1 Delta mutant of Saccharomyces cerevisiae. Therefore, C. albicans has four chitin synthases with two non-essential class I Chs isoenzymes that contribute collectively to more than 97% of the in vitro chitin synthase activity.     

Mutations of muscle glycogen synthase that disable activation by glucose 6-phosphate.

Glycogen synthase, an enzyme of historical importance in the field of reversible protein modification, is inactivated by phosphorylation and allosterically activated by glucose 6-phosphate (glucose-6-P). Previous analysis of yeast glycogen synthase had identified a conserved and highly basic 13-amino-acid segment in which mutation of Arg residues resulted in loss of activation by glucose-6-P. The equivalent mutations R578R579R581A (all three of the indicated Arg residues mutated to Ala) and R585R587R590A were introduced into rabbit muscle glycogen synthase. Whether expressed transiently in COS-1 cells or produced in and purified from Escherichia coli, both mutant enzymes were insensitive to activation by glucose-6-P. The effect of phosphorylation was studied in two ways. Purified, recombinant glycogen synthase was directly phosphorylated by casein kinase 2 and glycogen synthase kinase 3, under conditions that inactivate the wild-type enzyme. In addition, phosphorylation sites were converted to Ala by mutagenesis in wild-type and in the glucose-6-P desensitized mutants expressed in COS-1 cells. Phosphorylation inactivated the R578R579R581A mutant but had little effect on the R585R587R590A. This result was surprising since phosphorylation had the opposite effects on the corresponding yeast enzyme mutants. The results confirm that the region of glycogen synthase, Arg-578-Arg-590, is required for activation by glucose-6-P and suggest that it is part of a sensitive and critical switch involved in transitions between different conformational states. However, the role must differ subtly between the mammalian and the yeast enzymes.     

Evidence for trehalose-6-phosphate-dependent and -independent mechanisms in the control of sugar influx into yeast glycolysis

In the yeast Saccharomyces cerevisíae, trehalose-6-phosphate (tre-6-P) synthase encoded by GGS1/TPS1, is not only involved in the production of trehalose but also in restriction of sugar influx into glycolysis in an unknown fashion; it is therefore essential for growth on glucose or fructose. In this work, we have deleted the TPS2 gene encoding tre-6-P phosphatase in a strain which displays very low levels of Ggs1/Tps1, as a result of the presence of the byp1-3 allele of GGS1/TPS1. The byp1-3 tps2Δ double mutant showed elevated tre-6-P levels along with improved growth and ethanol production, although the estimated concentrations of glycolytic metabolites indicated excessive sugar influx. In the wild-type strain, the addition of glucose caused a rapid transient increase of tre-6-P. In tps2^Δ mutant cells, which showed a high tre-6-P level before glucose addition, sugar influx into glycolysis appeared to be diminished. Furthermore, we have confirmed that tre-6-P inhibits the hexokinases in vitro. These data are consistent with restriction of sugar influx into glycolysis through inhibition of the hexokinases by tre-6-P during the switch to fermentative metabolism. During logarithmic growth on glucose the tre-6-P level in wild-type cells was lower than that of the byp1-3 tps2Δ. mutant. However, the latter strain arrested growth and ethanol production on glucose after about four generations. Hence, other mechanisms, which also depend on Ggs1/Tps1, appear to control sugar influx during growth on glucose. In addition, we provide evidence that the requirement for Ggs1/Tps1 for sporulation may be unrelated to its involvement in trehalose metabolism or in the system controlling glycolysis.     

Yeast glycogen synthase kinase 3 is involved in protein degradation in cooperation with Bul1, Bul2, and Rsp5

The yeast Saccharomyces cerevisiae has four genes, MCK1, MDS1 (RIM11), MRK1, and YOL128c, that encode glycogen synthase kinase 3 (GSK-3) homologs. The gsk-3 null mutant, in which these four genes are disrupted, shows temperature sensitivity, which is suppressed by the expression of mammalian GSK-3beta and by an osmotic stabilizer. Suppression of temperature sensitivity by an osmotic stabilizer is also observed in the bul1 bul2 double null mutant, and the temperature sensitivity of the bul1 bul2 double null mutant is suppressed by multiple copies of MCK1. We have screened rog mutants (revertants of gsk-3) which suppress the temperature sensitivity of the mck1 mds1 double null mutant and found that two of them, rog1 and rog2, also suppress the temperature sensitivity of the bul1 bul2 double null mutant. Bul1 and Bul2 have been reported to bind to Rsp5, a hect (for homologous to E6-associated-protein carboxyl terminus)-type ubiquitin ligase, but involvement of Bul1 and Bul2 in protein degradation has not been demonstrated. We find that Rog1, but not Rog2, is stabilized in the gsk-3 null and the bul1 bul2 double null mutants. Rog1 binds directly to Rsp5, and their interaction is dependent on GSK-3. Furthermore, Rog1 is stabilized in the npi1 mutant, in which RSP5 expression levels are reduced. These results suggest that yeast GSK-3 regulates the stability of Rog1 in cooperation with Bul1, Bul2, and Rsp5.