Mutations in Mtr4 Structural Domains Reveal Their Important Role in Regulating tRNAiMet Turnover in Saccharomyces cerevisiae and Mtr4p Enzymatic Activities In Vitro

RNA processing and turnover play important roles in the maturation, metabolism and quality control of a large variety of RNAs thereby contributing to gene expression and cellular health. The TRAMP complex, composed of Air2p, Trf4p and Mtr4p, stimulates nuclear exosome-dependent RNA processing and degradation in Saccharomyces cerevisiae. The Mtr4 protein structure is composed of a helicase core and a novel so-called arch domain, which protrudes from the core. The helicase core contains highly conserved helicase domains RecA-1 and 2, and two structural domains of unclear functions, winged helix domain (WH) and ratchet domain. How the structural domains (arch, WH and ratchet domain) coordinate with the helicase domains and what roles they are playing in regulating Mtr4p helicase activity are unknown. We created a library of Mtr4p structural domain mutants for the first time and screened for those defective in the turnover of TRAMP and exosome substrate, hypomodified tRNAiMet. We found these domains regulate Mtr4p enzymatic activities differently through characterizing the arch domain mutants K700N and P731S, WH mutant K904N, and ratchet domain mutant R1030G. Arch domain mutants greatly reduced Mtr4p RNA binding, which surprisingly did not lead to significant defects on either in vivo tRNAiMet turnover, or in vitro unwinding activities. WH mutant K904N and Ratchet domain mutant R1030G showed decreased tRNAiMet turnover in vivo, as well as reduced RNA binding, ATPase and unwinding activities of Mtr4p in vitro. Particularly, K904 was found to be very important for steady protein levels in vivo. Overall, we conclude that arch domain plays a role in RNA binding but is largely dispensable for Mtr4p enzymatic activities, however the structural domains in the helicase core significantly contribute to Mtr4p ATPase and unwinding activities.


Introduction
Degradation of RNA can modulate gene expression or help cells eliminate old and abnormally formed RNAs. RNA processing is essential to modify and generate varieties of mature and functional RNAs to perform their correct cellular functions. In human cells, defective RNAs without proper processing or degradation could result in cancer or other neurodegenerative diseases [1]. In Saccharomyces cerevisiae, nuclear RNA 3' to 5' degradation and processing are initiated by Trf4/5p-Air1/2p-Mtr4p polyadenylation complex (TRAMP) [2,3]. TRAMP is composed of a non-canonical poly (A) polymerase Trf4/5p, an RNA binding protein Air1/2p and a member of DExH RNA helicase superfamily, Mtr4p. All three subunits of the TRAMP complex are highly conserved in eukaryotes, notably, all human homologs have been identified [4][5][6][7]. Air1/2p association with RNA helps Trf4/5 to append a poly (A) tail to the 3' end of target RNAs [8][9][10][11]. Mtr4p presumably removes highly ordered RNA structures and proteins bound to RNA substrates that would inhibit exosome degradation, and it seems to be necessary for efficient recruitment of the nuclear exosome for complete degradation of the RNA [12,13]. The conserved eukaryotic exosome, which is composed of 9 non-catalytic subunits and a tenth nuclease Rrp44p, can completely degrade RNA through exonucleolytic and endonucleolytic activities with a 3'to 5' polarity [14][15][16][17].
The TRAMP subunit Mtr4p was first identified in Saccharomyces cerevisiae with a genetic screen used to identify proteins playing a role in mRNA nucleocytoplasmic transport [18]. Mtr4p is localized to both nucleoplasm and nucleolus [19] where it plays an essential role in nuclear RNA processing and turnover of a variety of RNAs synthesized by all three RNA polymerases [20]. It was reported that Mtr4p is involved in degradation of cryptic unstable transcripts (CUTs) and 5' externally transcribed spacer (5' ETS) of pre-rRNA, as well as 5.8S rRNA processing [12,21]. Previous studies from our lab demonstrated that Mtr4p is required for degradation of hypomodified tRNA i Met by the nuclear exosome, as a component of TRAMP complex [8,20,22]. Mtr4p function, inside and outside of TRAMP has been reported in nuclear RNA metabolism [12]. It was reported that Mtr4p requires ATP hydrolysis to remove RNA secondary structure in a 3' to 5' direction [20]. Mtr4p performs RNA unwinding only in the presence of hydrolysable ATP or dATP [12,20], and ATP hydrolysis is fully dependent on the presence of RNA [20,23]. Thus, RNA-binding, ATPase activity and unwinding activity are tightly coupled in the helicase function of Mtr4p. Mtr4p belongs to the helicase superfamily 2 and is a structural and functional homolog of Ski2p, which plays a similar role in exosome-mediated RNA degradation in the cytoplasm [24,25]. Two groups solved the three-dimensional structure of Mtr4p and each showed that Mtr4p contains a helicase core and a protruding arch domain [26,27]. The helicase core consists of two highly conserved RecA-like domains that exist in all superfamily II DNA and RNA helicases, together with the winged helix (WH) and ratchet domains. The auxiliary arch domain, which is unique to the exosome-linked helicases Mtr4 and Ski2, contains the arm (also called stalk) and the fist (also called KOW) [25][26][27]. Other than the RecA1 and RecA2 domains that play catalytic function, we refer to as the arch domain, winged helix and ratchet domains as structural domains in this study.
Hel308, also a superfamily 2 archaeal DNA helicase, contains helicase core that is highly conserved to Mtr4p. The Winged helix and ratchet domains of Hel308, are believed to impart some, but not all DNA binding activity to this helicase [28,29]. Whether winged helix and the ratchet domains of Mtr4p execute similar function and how they participate in enzymatic mechanism have not been fully characterized. Studies on the arch domain revealed that complete removal of the arch domain of Mtr4p (Mtr4p-archless) enfeebled its in vivo functions including processing of 5.8S rRNA and degradation of the 5' ETS [26]. It also has been shown that the arch domain itself is sufficient to bind tRNA i Met in vitro [27], and the KOW domain is required to activate TRAMP-mediated exosome degradation in vitro [30]. It was reported that the KOW domain was found in rRNA binding proteins [26,27] and Mtr4p crystal structure indicates that two long loops β 2~β 3 and β 3~β 4 of the KOW domain protrude towards the RNA that was co-crystalized with protein [26,27]

Plasmids and Oligonucleotides
All plasmids and oligonucleotides used in this study were listed separately in Tables 1 and 2.

Error-prone PCR and Dominant-negative Screen
Original yeast strains used in this study are listed in Table 3. Error-prone PCR [31]  Mtr4p identified in this screen will exhibit reduced function in tRNA i Met turnover (negative) and compete with wild-type Mtr4p (dominant) to assemble TRAMP complexes in vivo. The mutant plasmids resulting in trm6-504 phenotype suppression were rescued from yeast, the gap-repaired region of DNA sequenced, and then each was retested by trm6-504 transformations to confirm the original phenotype.

Creation of Mutant Integrated trm6-504 Strains
To create Mtr4 genomic mutant strains in trm6-504 background, we cloned a N-terminally truncated wild type MTR4 gene containing only 1472 bp to the stop codon plus 228 bp C-terminal of the stop codon into an integrative plasmid YIplac211. The mutant constructs were made using QuickChange by site-directed mutagenesis (Stratagene). YIplac211 plasmids harboring wild type or mutant truncated MTR4 gene were linearized by BspEI restriction enzyme (NEB) and used to transform trm6-504 to URA+. Stable transformants were initially tested for integration into the MTR4 locus by PCR and phenotype before being used in subsequent experiments. Genomic DNA of integrated strains was isolated and integration and mutations were confirmed by sequencing. This method will recreate a full length mtr4 gene that now harbors the mutation we created and an additional N-terminally truncated non-functional mtr4 downstream.
Yeast Total Cell Protein Isolation and Western Blot 20 ml of indicated yeast stains were grown in liquid YPD media at 30°C until OD 1. Cells were harvested at 4, 000 x g at 4°C for 10 min, and washed with 20 ml of 1x TBS buffer (150 mM NaCl, 2 mM KCl, 25 mM Tris pH 7.5). Cell pellets were weighed and resuspended in cold breaking buffer (1 X TBS, 1 mM DTT, cOmplete EDTA-free protease inhibitor cocktail tablet (Roche)) at 2ml/1g wet cells. 250 μl of acid-washed glass beads (Sigma) were added to the cells, followed by intermittent vortexing for 5min. After spinning at 4°C for 10 min at 14, 000 x g, supernatant was carefully taken out by U-100 1cc insulin syringe (28G, ½ inch) (Becton Dickinson), avoiding sucking up any glass beads, and transferred into a fresh cold microcentrifuge tube. Another spin was repeated to remove any insoluble part in the total cell protein. Supernatant was transferred into a new tube, and protein concentration was determined by Bradford assay. Protein aliquots were stored in laemmli buffer at -80°C. Protein samples were separated on 8% SDS-PAGE gel, followed by transferring onto a Nitrocellulose blotting membrane (Pall Corporation) at 25 Volts for 90 min in 1X transfer buffer (48 mM Tris, 39 mM glycine, 0.0375% SDS, 20% vol/vol methanol) at 4°C. Membrane was blocked in 5% milk in PBST for 1 h at room temperature, followed by primary antibody (anti-Mtr4p at 1:10000 [12], anti-Nab2p at 1:5000 [33], anti-his at 1:500 (Calbiochem)) incubation on a nutator overnight at 4°C. After wash, blots were incubated with appropriate secondary antibody (1:5000 dilution in 5% milk-PBST) at room temperature for 90 min with gently shaking. The blot was washed and developed with ECL reagents and visualized by Chemiluminescence System (UVP). ImageJ software (NIH) was used to quantify western blot results.

Yeast Total RNA isolation and Northern Blotting
Yeast total RNA was isolated as described [34]. 10 μg RNA sample was separated on 6% denaturing PAGE gel (8M Urea) in 0.5X TBE buffer, followed by transferring onto a Hybond-N + membrane (GE Healthcare) at 12 V for 5 h in 0.5× TBE at 4°C. RNA blots were probed with radiolabelled deoxyoligonucletides as described [22], detected by phosphorimager (Molecular Dynamics), and quantified by the ImageQuant 7.0 software.

Mtr4p Expression and Purification
Recombinant Mtr4p was expressed in BL21 (DE3) competent E. coli cells (NEB) by autoinduction [35]. Cells were collected and frozen at -80°C. Cell pellets were resuspended and disrupted by sonication in equilibration buffer (50 mM sodium phosphate pH 7.4, 10 mM β-ME, 10% Glycerol, with cOmplete EDTA-free protease inhibitor cocktail (Roche)). Cell lysate was clarified by centrifugation at 40, 000 x g for 30 min at 4°C and loaded onto a column containing His60 Ni superflow Resin (Clontech). The resins were washed with 10-column volumes of equilibration buffer with 20 mM imidazole and eluted with 5-column volumes of the same buffer containing 300 mM imidazole. Purified protein subsequently underwent gel filtration chromatography on HiPrep 16/60 Sephacryl S-200 column (GE Healthcare) in gel filtration buffer (20 mM HEPES (pH 7.5), 150 mM NaCl, and 1 mM DTT). Fractions with pure Mtr4p were collected and concentrated using Macrosep advance centrifugal devices (Pall Corporation). Protein aliquots were stored in a buffer containing 50 mM sodium phosphate pH 7.4, 10 mM β-ME and 20% Glycerol at -80°C.

Electrophoretic Mobility Shift Assays (EMSA)
Single-stranded RNAs were synthesized commercially (Thermo Scientific). RNA sequences are listed as below.
EMSA was modified from previously reported protocols [37]. Radiolabelled R 1-4 RNA duplex preps were used as RNA substrates in EMSA. Indicated concentrations of recombinant Mtr4p was incubated with 0.5 nM of radiolabeled RNA in a buffer containing 10 mM Tris (pH 7.5), 1 mM EDTA, 100 mM KCl, 0.1 mM DTT, 5% vol/vol glycerol, and 1% NP40 on ice for 30min. Reactions samples were then prepared in a buffer containing 1 mM Tris, 0.1 mM EDTA, 1% glycerol, 0.0001% wt/vol bromophenol Blue, 0.0001% wt/vol xylene cyanol FF. Prepared samples were loaded onto a 5% nondenaturing PAGE, and separated at 120 V for 100 min at 4°C. Gels were dried and visualized by phosphorimager (Molecular Dynamics).

ATPase Assay
ATP/NADH coupled assay were carried out and modified as described previously [23]. 83 nM of Mtr4p was pre-incubated with 9 μM E.coli total tRNA (Roche) in a cuvette containing 8 mM MgSO 4 , 1.5 mM phosphoenolpyruvate (PEP), 0.15 mM NADH, 2 units pyruvate kinase (PK) and 7 units L-Lactate dehydrogenase (LDH) for 5 min at 30°C. Next, ATP was added to initiate the reactions at indicated concentration (0.25 mM, 0.5 mM, 1 mM, 2 mM, 4 mM, 8 mM). Reactions were incubated in cuvettes at 30°C for 15 min, continuously monitored by Shimadzu UV-1800 spectrophotometer. Upon one molecule of ATP hydrolyzed by Mtr4p, PK converts PEP into pyruvate, which is subsequently converted into lactate by LDH, followed by one NADH molecule oxidized to NAD + . Reaction rate of NADH depletion over time was recorded at wavelength of 340 nm. P value was calculated by T test for unpaired data based on triplicates.

Unwinding Reactions
Unwinding reactions were performed as described [38]. 50 nM of Mtr4p (wild type or mutants) were used in each reaction to unwind 0.5 nM radiolabelled R 1-4 . Aliquots of reactions were stopped at 0 min (time of ATP-MgCl 2 added to reactions), 3 min, 8 min, 15 min, 25 min, 40 min, and run on a 15% nondenaturing PAGE gel to separate single-stranded RNA and double-stranded RNA. Gels were dried, visualized by phosphorimager (Molecular Dynamics), and quantified by the ImageQuant 7.0 software.

TRAMP Reconstitution
In vitro TRAMP reconstitution was modified from a previous description [39]. Recombinant His-tagged wild type or mutant Mtr4p was overexpressed in BL21 (DE3) competent E. coli cells (NEB) by auto-induction [35]. Recombinant Flag-tagged-Trf4p and His-tagged-Air2p was also co-overexpressed by auto-induction. Cells were collected and weighed. 100 mg of cells were suspended in 1 ml of equilibration buffer (50 mM NaH 2 PO 4 pH 7.0, 250 mM NaCl, 1 mM ZnC l2 , 10%glycerol, with cOmplete EDTA-free protease inhibitor cocktail (Roche)). This was followed by sonication disruption and centrifugation to clarify the extract. Supernatants were mixed in a ratio of 8:1 for Mtr4 extract (mutant or wild type) to Trf4-Air2 extract. 100 μl of Trf4-Air2-extract was mixed with 800 μl equilibration buffer as a negative control. Cell mixtures were mixed by nutator overnight, and went through Flag purification using EZ-view red anti-Flag M2 affinity gel (Sigma-Aldrich). Pull down of Mtr4p and Air2p were detected by western blotting using anti-his antibody (Calbiochem).

Results
A plasmid library of Mtr4p structural domain mutants that are defective in tRNA i Met turnover A schematic diagram of MTR4 gene (Fig 1A) shows that the RecA-like domains reside in the N-terminal half of MTR4, while the C-terminal region contains the arch, winged helix and ratchet domains. To detect how structural domains play a role in Mtr4p function, structural domain mutants were created by error-prone PCR ( Fig 1A) on a high-copy-number plasmid YEplac195 (HC). A dominant-negative screen was performed to identify mutants that are defective in tRNA i Met turnover, using yeast strain trm6-504 (Materials & Methods). In the presence of Mtr4p mutants, perturbed in tRNA i Met degradation, the trm6-504 strain will grow better than the HC wild-type Mtr4p control at temperatures above 30°C (Fig 1B). Compared to HC wild-type MTR4 transformants, 59 HC mtr4 mutant transformants (out of 78 tested transformants,~1000 untested transformants) exhibiting better growth at 33°C were considered as mutants defective in hypomodified tRNA i Met degradation (Fig 1B). Of the 59 defective Mtr4p structural domain mutants identified in this screen, 12 exhibiting the most severe phenotypes were chosen for DNA sequence analysis ( and β 3~β 4 loop (Fig 2A). K700N is a double mutant (Mtr4-23) with ratchet mutation R1030G, which we chose to study because both destroyed a positive charge that could potentially affect RNA binding (Fig 2A). Finally, winged helix mutation, K904N, was discovered as a single mutation (Mtr4-30). Three of the four mutations identified here are highly conserved in other eukaryotes including mammals. In particular, lysine 904 is conserved 100% in Mtr4 proteins from different eukaryotic species as well as Ski2p in S. cerevisiae (Fig 2B). Lysine 700 and arginine 1030 are 100% conserved only in Mtr4 proteins across eukaryotes examined where P731S was not found to be conserved ( Fig 2B).
Before further study, all four mutations were created as single mutations in HC plasmids and tested again in trm6-504 to determine if they reproduced the suppression phenotype. HC mtr4-K904N and HC mtr4-R1030G showed better growth than trm6-504 with wild-type HC MTR4. In contrast, arch domain mutants HC mtr4-K700N and HC mtr4-P731S exhibited similar growth at 36°or 30°C as wild-type HC MTR4 (S1 Fig). In conclusion, overexpression of winged helix mutant K904N, and ratchet domain mutant R1030G suppressed the growth defect of trm6-504, but not K700N or P731S, suggesting that two of the mutants identified affect hypomodified tRNA i Met turnover in yeast while two appear to exhibit modest or no defect in hypomodified tRNA i Met turnover.
Single mutants Mtr4p-K700N and Mtr4p-P731S degrade tRNA i Met like wild-type Mtr4p, while Mtr4p lacking the arch domain is defective in tRNA i Met turnover in vivo The dominant-negative screen was carried out with mtr4 mutants on a HC plasmid, which resulted in increased mutant Mtr4p expression over endogenous Mtr4p. In order to mimic endogenous MTR4 expression conditions and remove effects of overexpression, mtr4 single Consistent with what was observed for growth of yeast bearing K700N or P731S mutants, trm6-504/mtr4-700 and mtr4-731 had a similar level of tRNA i Met as trm6-504/MTR4 (Fig 3B).
To insure that these results are not a reflection of reduced Mtr4 protein levels (likely masked in high-copy number plasmids), western blots were done to detect Mtr4p, and integrated WT Mtr4p, Mtr4p-K700N and Mtr4p-P731S were expressed at similar levels ( Fig 3C). These data indicate that arch domain mutants K700N and P731S did not affect tRNA i Met turnover in vivo.
The reduced Mtr4p level in trm6-504/mtr4-904 raised the question whether defects in tRNA i Met turnover were caused by reduced protein levels, or by loss of Mtr4p function, or both. In order to test whether Mtr4p-K904N defects in tRNA i Met turnover exist under conditions where Mtr4p-K904N is expressed at comparable or higher levels than Mtr4p-WT, both were expressed from individual HC plasmids under control of their own promoter in trm6-504. HC Mtr4p-K904N partially suppressed the temperature-sensitivity of trm6-504 ( Fig  4A). Consistently, tRNA i Met degradation was also dramatically reduced in the presence of HC  (Fig 5B). The Michaelis constants (K m ) of Mtr4p-P731S for ATP-binding was comparable to the K m of wild-type (P > 0.1), indicating ATP-binding was not affected (Fig 5C), however Mtr4p-K700N showed significantly lower ATP affinity (p < 0.01) (Fig 5C). Consistently trending with defects in RNA binding, Mtr4p-K700N and Mtr4p-P731S displayed~50% and~70% RNA-dependent ATPase Mtr4 Structural Domain Mutants Affect Helicase Function activities (k cat /K m ) compared to wild-type Mtr4p. This suggests that some amino acids in the KOW domain affect Mtr4p ATPase activity, possibly through a failed synergy where reduced RNA-binding results in a reduction of ATP-binding and hydrolysis.
To provide a mechanistic view of ATPase activity coupled unwinding activity, we next tested Mtr4p wild-type and mutants in a timed unwinding assay using R 1-4 as substrate.
Mtr4p-K700N showed impaired unwinding (Fig 5D & 5E), and the observed unwinding rate constant (K obs =~0.071 min -1 ) was~25% lower than that of Mtr4-WT (K obs =~0.095 min -1 ) (Fig 5F). Mtr4p-P731S did not show defects in the unwinding assay (Fig 5D, 5E & 5F). Taken together with the wild-type-like phenotypes of arch domain mutants in vivo, it indicates that the KOW domain amino acids K700 and P731 had little effect on Mtr4p unwinding activity. We conclude that although arch domain mutants reduced RNA binding activity, which could induce lower ATPase activity, the reduction of RNA-binding was not sufficient enough to impair Mtr4p, or TRAMP function in vivo.
Mtr4p structural domain mutants in the helicase core reduced Mtr4p RNA binding, ATPase activity and unwinding activity in vitro The same series of in vitro biochemical studies were conducted on WH mutant Mtr4p-K904N and ratchet domain mutant Mtr4p-R1030G. The EMSA results with Mtr4p-R1030G showed it formed stable Mtr4p-RNA complexes at a significantly lower level than Mtr4p (Fig 6A). Mtr4p-K904N showed a similar level of reduced protein-RNA complex formation as Mtr4-R1030G, whereas a quantification of free duplex RNA revealed little differences between Mtr4p-K904N and Mtr4p (Fig 6A). Slightly higher levels of smearing between unbound and bound RNA duplex were consistently present in the K904N mutant protein lanes but not in lanes with wild-type Mtr4p, possibly indicating a lower binding affinity or higher dissociation rate of Mtr4p-K904N. We also cannot rule out the possibility of greater contaminating RNase activity consistently purifying with K904, which might lead to modest amounts of substrate degradation during the experiment, although this seems less likely given that several independent purifications of mtr4p-K904N gave the same or similar results.
Both mutants Mtr4p-K904N and Mtr4p-R1030G showed reduced ATPase activities compared to wild-type Mtr4p (Fig 6B). The Michaelis constant (K m ) values for ATP affinity of Mtr4p-K904N were statistically comparable to the K m of Mtr4p-WT, indicating ATP binding was not affected (P > 0.1) (Fig 6C). In contrast,~35% more ATP was required for Mtr4p-R1030G to achieve its maximum rate of hydrolysis (Fig 6C), indicating a lower ATP affinity (P < 0.05). The specificity constants (k cat /K m ) of both mutants were markedly lower than wild type Mtr4p (Fig 6C).
Both mutants exhibited different degrees of impaired unwinding activity (Fig 6D & 6E). The observed unwinding rate constant of Mtr4p-K904N (K obs =~0.060 min -1 ) and Mtr4p-1030p (K obs =~0.063 min -1 ) were only 60% of Mtr4p-WT (K obs =~0.095 min -1 ) (Fig 6F). The Mtr4p-K904N, which exhibited better RNA-binding than Mtr4p-R1030G, surprisingly performed the worst in the unwinding assay, suggesting that lysine 904 may play a significant role in stimulating the Mtr4p helicase core to unwind bound RNA. Likewise, Mtr4p-R1030G retained 60% unwinding activity, also indicating an important role of ratchet domain arginine 1030 in RNA unwinding. Taken into account the in vivo defects of these two mutants in tRNA i Met turnover, we conclude that single amino acids in the structural domains of the helicase core are important for Mtr4p to fully function in TRAMP as a helicase to degrade tRNA i Met in vivo.

Discussion
In this study, we created a library of Mtr4p structural domain mutants targeting its arch, ratchet and winged helix domains to begin to genetically probe the function of each domain. A dominant-negative screen was designed to identify mutants in those targeted domains that are defective in hypomodified tRNA i Met turnover. Four mutants from each structural domain have revealed aspects about the working structure of Mtr4p. The arch-KOW outside of the helicase core is more likely to impact RNA binding and the RNA-dependent ATPase activity, which surprisingly did not affect the efficiency of unwinding or in vivo function. The winged helix and ratchet domains in the core not only contribute to RNA and ATP binding, but also seem to be more intimately involved in unwinding activity, which resulted in reduced Mtr4p function in vivo. The winged helix domain was also found to play a role in protein stability, and the ratchet domain is especially important for RNA binding. Our study established that the structural domains outside (arch) and inside (WH & ratchet) of the helicase core confer absolute different functions due to their relative positions.
The dominant-negative screen provides an efficient way to identify important amino acids of Mtr4p Unfortunately, arch domain mutants were not identified that significantly affect degradation of hypomodified tRNA i Met in vivo, possibly because any mutation that phenotypically is like an [26] would not be selected during the screening process due to its own slow growth. An advantage of this screen was that it generated mutations randomly, was nonbiased, and it insured only Mtr4p mutants capable of assembling TRAMP complexes would be selected based on our screening criteria. The second advantage is that the screen ignored null MTR4 mutations, since these would have been inviable after selection against the wild-type MTR4 to reveal potential phenotypes. Thirdly, it could be utilized in other genetic backgrounds expressing RNAs subject to TRAMP surveillance.
The KOW domain mutants affect RNA binding, but marginally affect Mtr4p function as a helicase Our results showed that Mtr4p KOW domain mutants Mtr4p-K700N and Mtr4p-P731S exhibited significantly less RNA binding activity than wild-type (Fig 5A), indicating the RNA binding function of the long loops (β2~β3 and β3~β4) of Mtr4p arch domain. It was reported that the KOW domain itself binds the highly structured tRNA i Met in vitro [27] but not singlestranded RNA [25]. Our study began to assign RNA binding to the specific loops and amino acids. In particular, the K700N mutation is found in a disordered loop where it is clustered with 3 other positively charged amino acids (R701, R705 and K704) (S3A Fig). It will be interesting to determine if other positive charges in this loop of the KOW domain lying very near K700 contribute similarly to its predicted RNA-binding function. Given that Mtr4p has been characterized as a RNA-dependent ATPase, and higher RNA concentration induces greater ATP hydrolysis [23], we concluded that impaired RNA binding might contribute to the reduced ATP binding of Mtr4p-K700N, and ATP hydrolysis of both arch mutants. We also do not exclude the possibility that lysine700 may communicate with the helicase core, affecting ATP affinity via unknown conformational changes. enhance RNA-binding to assist Ski2p unwinding function [40]. The TRAMP complex exhibits similar regulation where it was noted that Mtr4p unwinds substrate RNAs more efficiently when it is in complex with Trf4-Air2 [38]. The diversity of substrates and binding sites on substrate RNAs in vivo is more complex than our defined in vitro system, which could account for differences in RNA-binding between defined model substrate in vitro and natural substrates in vivo. From this, we propose that there is a minimum threshold of Mtr4p RNA-binding activity by the arch domain required to support yeast growth [20], which implies that there is a range of Mtr4 RNA-binding activity that could result in similar phenotypes in vivo, but vastly different biochemical activities in vitro.
The winged helix domain contributes to enzymatic activity and is important for maintaining Mtr4p steady-state levels Consistent with the finding that Hel308 WH domain binds with DNA substrate [28,29], we found K904N mutation in the Mtr4p WH domain reduced RNA binding in vitro. Additionally, we experimentally demonstrated an essential role of the winged helix domain for Mtr4p function, not only in tRNA i Met degradation in vivo (Fig 4A and 4B), but also in Mtr4p enzymatic activities in vitro (Fig 6). The instability of Mtr4p-K904N isolated from bacteria (easily degraded during purification) and yeast ( Fig 3C)  The Ratchet domain of Mtr4p plays a role in RNA binding and enzymatic activity It was believed that ratchet domain serves as a scaffold providing an RNA binding surface that stabilizes protein-RNA interactions during each cycle of unwinding [42]. A recent publication targeted arginine 1030 as a potential RNA binding site through its position in 3D structure [43], while we discovered the same residue from our dominant-negative screen. The coincidence of both studies indicates an essential role of R1030 in Mtr4p function, and both studies showed mutation of arginine 1030 reduced Mtr4p RNA binding and unwinding activity. Our data of ATPase activity indicates that arginine 1030 possesses additional function in regulating ATP affinity and ATP hydrolysis. Upon further analysis of the Mtr4p 3D structure, we noticed two positively charged residues R1026 (5 Å to RNA) and K1029 (12 Å to RNA) are positioned close to R1030, indicating this might be a RNA binding surface (S3C Fig). A recent study of TRAMP complex assembly [44] showed that mutations in the Ratchet domain (K1015E and M1016E, or Y1020A and E1021R) disrupted TRAMP assembly. In vitro reconstitution was performed on each mutant in this study, and all mutants were able to assemble the TRAMP complex with Trf4p-Air2p (S4 Fig). This would suggest that the outside surface of the ratchet domain facing the solvent (K1015E and M1016E, or Y1020A and E1021R) participates in TRAMP assembly, while the inside surface of the ratchet domain facing the center of the helicase core (R1026, R1030 and K1029 helix) provides RNA-binding (S3D Fig). Since it seems that the R1030G region of the ratchet domain acts as a binding track, defects in RNA binding might result in greater protein dissociation, which would reduce its RNA dependent ATPase activities and unwinding efficiency. How this domain coordinates its complicated functions including RNA binding, ATP hydrolysis and unwinding activities needs further investigation.
Supporting Information