Dom34 Links Translation to Protein O-mannosylation

Abstract

In eukaryotes, Dom34 upregulates translation by securing levels of activatable ribosomal subunits. We found that in the yeast Saccharomyces cerevisiae and the human fungal pathogen Candida albicans, Dom34 interacts genetically with Pmt1, a major isoform of protein O-mannosyltransferase. In C. albicans, lack of Dom34 exacerbated defective phenotypes of pmt1 mutants, while they were ameliorated by Dom34 overproduction that enhanced Pmt1 protein but not PMT1 transcript levels. Translational effects of Dom34 required the 5'-UTR of the PMT1 transcript, which bound recombinant Dom34 directly at a CA/AC-rich sequence and regulated in vitro translation. Polysomal profiling revealed that Dom34 stimulates general translation moderately, but that it is especially required for translation of transcripts encoding Pmt isoforms 1, 4 and 6. Because defective protein N- or O-glycosylation upregulates transcription of PMT genes, it appears that Dom34-mediated specific translational upregulation of the PMT transcripts optimizes cellular responses to glycostress. Its translational function as an RNA binding protein acting at the 5'-UTR of specific transcripts adds another facet to the known ribosome-releasing functions of Dom34 at the 3'-UTR of transcripts.

Fig 1. Genetic interactions of dom34 and pmt1 mutations.

(A) Killer phenotypes of S. cerevisiae mutants. Strains were grown as streaks and replica-printed onto plates containing the killer K1-secreting strain RC130. Following incubation at 18°C for 4–7 d, the appearance of streaks was blue (dead cells stained by methylene blue appear dark on figure), or white (live cells). Strains tested were the parental strain YE449 (WT) and its mutant M577 and pmt1 and/or pmt2 derivatives; in addition, mutant strains W21 (pmt1 dom34), W12 (pmt1 yil001w) and transformants of strain W21 carrying pSW20 (PMT1), p577/20 (DOM34) and pSW577/2 (YIL001w) were tested. (B) Inserts in genomic clones complementing the killer K1-resistance of strain M577 pmt1. Regions of chromosomes XIV and IX are shown, along with genomic inserts in YEp13 (numbered bars). (C) Hygromycin B sensitivity. Strains were serially diluted and spotted on YPD plates without or with hygromycin B (50 μg/ml). Growth was for 2 d at 30°C. S. cerevisiae strains tested were YE449 (WT), YE449 pmt1, YE449 dom34 and W21 (pmt1 dom34). C. albicans strains tested were CAF2-1 (WT), SPCa2 (pmt1/pmt1), JH47-1 (dom34/dom34), JH24-4 (pmt1/pmt1 dom34/dom34), SPCa10 (pmt5/pmt5) and JH5-3-1 (dom34/dom34 pmt5/pmt5).

Fig 2. DOM34 overexpression suppresses pmt1 phenotypes.

(A) Relative DOM34 transcript level (RTL). Total RNA of strains CAF2-1 (+/+), SPCa2 (pmt1/pmt1), CAP1-3121[pSP38] (pmt1/pmt1[empty vector] and CAP1-3121[pSK2] (pmt1/pmt1[DOM34]) was isolated and amounts of the DOM34 transcript were determined by qPCR using ACT1 as the reference transcript. Two independent biological replicates of each strain were assayed. (B) Hypha formation of representative colonies grown on Spider medium for 2–5 d at 37°C. Strain designations as in A. (C) Effects of DOM34 overexpression on pmt mutant phenotypes. Serially diluted cultures of strains were spotted and grown on YPD agar at 30°C without or with 200 μg hygromycin B, or at 42°C without additions. pmt1 single mutant host strains were transformed with expression vector pSK2 encoding wild-type Dom34 protein (a) or with pSK2mut encoding a E21A variant of Dom34 (Dom34*) (d). In addition, transformants of double mutant strains pmt1 pmt5 (b) and pmt1 pmt6 (c) were tested. Strains P15-274 (pmt1/pmt1 pmt5/pmt5), P15-274-1[pSP38] (pmt1/pmt1 pmt5/pmt5[empty vector], P15-274-1[pSK2] (pmt1/pmt1 pmt5/pmt5[DOM34]) were compared (b); in addition, strains CPP1121[pSP38] (pmt1/pmt1 pmt6/pmt6[empty vector]) and CPP1121[pSK2] (pmt1/pmt1 pmt6/pmt6[DOM34] were compared (c). Single mutant strains SPCa2 (pmt1/pmt1), SPCa10 (pmt5/pmt5), SPCa8 (pmtD6/pmt6) and the wild-type strain CAF2-1 (+/+) were used as reference strains.

Fig 3. Dom34 overexpression upregulates Pmt1 amounts and depends on 5′-UTR of PMT1.

(A) Pmt1 amounts. Two independent strains of CIS23[pSP38] (PMT1/PMT1^HA [empty vector]) were compared with three independent strains of CIS23[pSK2] (PMT1/PMT1^HA [DOM34]); strain CAF2-1 (+/+) was used as negative reference strain. 5 μg of crude extract protein were separated by SDS-PAGE (10% acrylamide) and immunoblots were probed by rat anti-HA antibody (1:1000) and mouse anti-actin antibody (1:1000) followed by reaction with POD-coupled anti-rat and anti-mouse antibodies (1:20000). The signals of tagged Pmt1^HA and actin are indicated. (B) Importance of 5′-UTR on PMT1 regulation. The 5′-start of the PMT1 transcript at position -218 is indicated by the kinked arrow. Plasmids containing the PMT1 promoter either including the 5′-end of the UTR from position -218 to -167 containing one of three 11-mer repeats (pPdC2-HIS) (under-/overlined sequences) or lacking this 5′-UTR sequence (pPdC3-HIS) in fusion to the RLUC reporter gene were integrated into the PMT1 promoter of C. albicans strain RM1000 by transformation. These strains were additionally transformed with vectors pSK2 for DOM34 overexpression or with the pSP38 empty control vector. Protein extracts of three independent double transformants were tested for luciferase activity, which was calculated as relative light units (RLU) per μg of protein.

Fig 4. Transcript fractionation on polysome gradients.

(A) Cellular extracts of strain CAF2-1 (wild-type) and dom34 mutant JH47-2 were centrifuged in a 10–50% sucrose gradient, which was subsequently fractionated. Nucleic acids in gradient fractions were detected by absorbance (A₂₆₀). Note that pre-polysome fractions contain 40S, 60S and 80S ribosomal RNA. (B) Occurrence of ACT1 and PMT1 transcripts in gradient fractions. Transcripts were detected by qPCR after adding a known amount of an in vitro generated transcript of CaCBGluc as calibrator. Each bar represents the normalized mean ACT1 or PMT1 transcript level of two independent experiments including the standard error of the mean. (C) The Kolmogorov-Smirnov test was used to determine the distance “D-value” between the two distribution functions of the wild-type strain CAF2-1 and dom34 mutant for polysomal and pre-polysomal fractions of PMT1, PMT4, PMT6 and ACT1, respectively. Statistical relevance is indicated by the calculated p-value.

Fig 5. Dom34 inhibits translation when PMT1 5‘-UTR is present.

(A) Identical sizes of CBGluc proteins produced in an in vitro rabbit reticulocyte translational system using RNA templates containing or not containing the 5‘-UTR of PMT1. (a) Scheme of RNA templates, (b) Protein products derived from CBGluc RNA without (CBGluc) or with (UTR-CBGluc) the 5′-UTR. Equal amounts of in vitro transcribed RNA were translated and labeled by incorporation of biotinylated lysine in the rabbit reticulocyte lysate; proteins were separated by SDS-PAGE and detected by using HRP-conjugated streptavidin. Further lanes contain protein products of control RNAs (Promega) showing 61 kDa Coleoptera luciferase (Colluc) and 42 kDa mouse ß-actin (Ambion) (ACT) and a control without RNA template (control). (B) Dom34 regulates in vitro translation. (a) CBGluc production as in (A) using no or 2.5 μM final concentration of Dom34, in the presence or absence of the 5′-UTR, as indicated. (b) Time course of luminescence emitted by the in vitro translated CBGluc, (c) comparisons of CBGluc protein amounts and CBGluc enzyme activities (luminescence peaks at 1332 sec) in the presence of Dom34. Three independent experiments were analyzed; the activity of samples without added Dom34 was set to 1. (C) Concentration dependence of in vitro translation of CBGluc transcript containing the 5‘-UTR by increasing amounts of Dom34 protein.

Fig 6. Dom34 is able to bind the PMT1 5‘-UTR and cleave at distinct sites.

Radioactive 3′ end-labeled RNA was incubated with increasing amounts of recombinantly produced Dom34 or Dom34^E21A proteins and after complex formation samples were split and analysed either by 6% native PAGE (A, C) or by 10% denaturing PAGE (B). Dom34 or Dom34 E21A were present in final concentrations of 0.1/0.15/0.25/0.3/0.6/1/2.5 μM, while BSA as a specificity control was added also at 2.5 μM final concentration. For comparison of running behaviour under native and denaturing conditions 6S RNA from E. coli was loaded on the gel (M). Two UTR-Dom34 complexes (I, II) were observed under native separation conditions (A, C), which also revealed potential RNA degradation fragments (unbound or bound to Dom34) (asterisks). Under denaturing conditions Dom34 but not the Dom34 E21A variant generated specific UTR degradation fragments (B).

Fig 7. Dom34 binding to the 5′-UTR of the PMT1 transcript.

(Top) Sequence of oligonucleotide representing the 5′-end of the 5′-UTR of the PMT1 transcript. (Bottom) EMSA of biotinylated 5′-UTR oligonucleotide in the absence or presence of E. coli-produced Dom34, its E21A variant and its N317A variant (protein/oligonucleotide molar ratio = 20). A specificity control reaction contains 2.5 μM BSA instead of Dom34. Unlabeled oligonucleotide in 100-fold excess was added as competitor in the indicated samples. The migration of unbound oligonucleotide (f) and two retarded complexes (I, II) was assayed by blotting of RNA separated by agarose gel electrophoresis onto a nylon membrane, which was developed by a chemiluminescent substrate to detected biotin. Binding reactions marked by an asterisk (*) were performed in the presence of 0.05% lauryl sarcosinate.