Article

• The EMBO Journal (2005) 24, 1842 - 1851
• doi:10.1038/sj.emboj.7600666

Published online: 28 April 2005

• Subject Categories:

  • RNA | Proteins

A 'gain of function' mutation in a protein mediates production of novel modified nucleosides

Peng Chen¹, Pamela F Crain², S Joakim Näsvall¹, Steven C Pomerantz² and Glenn R Björk¹

1. Department of Molecular Biology, Umeå University, Umeå, Sweden
2. Department of Medicinal Chemistry, University of Utah, Salt Lake City, UT, USA

Correspondence to:

Glenn R Björk, Department of Molecular Biology, Umeå University, S-901 87 Umeå, Sweden. Tel.:+46 90 785 6756; Fax:+46 90 77 2630; E-mail: glenn.bjork@molbiol.umu.se

Received 15 November 2004; Accepted 8 April 2005

• Keywords:

  • evolution,
  • frameshift,
  • mnm⁵s²U,
  • modified nucleosides,
  • tRNA

Introduction

The translation apparatus has evolved to read long messages and make functional proteins required for life. Although this apparatus has the ability to decode genetic messages faithfully, errors occur with frequencies of 10^-3–10^-4 per codon (Kurland et al, 1996; Farabaugh and Björk, 1999). Most missense errors are not harmful, since many amino acids can be substituted without affecting the activity of the protein, whereas, in contrast, almost all frameshift errors are detrimental because a shift in the reading frame almost always results in a truncated, inactive peptide. Although the mechanism of how the ribosome maintains the reading frame is not known, we know that if the ribosome makes a pause by, for example, shortage of an aminoacyl-tRNA or a defective tRNA, the peptidyl-tRNA may slip and thereby cause a frameshift error (Farabaugh, 1996; Farabaugh and Björk, 1999). Structural changes of the tRNA, such as deficiency of modified nucleosides, which are derivatives of the four major nucleosides U, C, A and G, may affect the entry of the ternary complex into the ribosomal A site or the fit of the tRNA in the P site, and therefore influence frameshift frequency (Qian et al, 1998; Urbonavicius et al, 2001). (For details, see Supplementary Figure 1.)

More than 80 different modified nucleosides have, so far, been characterized in tRNA (Rozenski et al, 1999). Although they are present in many different positions of the molecule, position 34 (the wobble nucleoside) and position 37 (3' of and next to the anticodon) are frequently modified (Sprinzl and Vassilenko, 2003) . Moreover, at these two positions a great variety of modified nucleosides are found, suggesting that the modified nucleosides in these two positions play a pivotal role in the decoding process. Transfer RNA from all organisms contains modified nucleosides and some of them are found in the same position and in the same subset of tRNAs from all organisms, suggesting that at least some of them were present in the organisms before the emergence of the three phylogenetic domains (Björk, 1986; Cermakian and Cedergren, 1998).

According to some theories, the early tRNA might have been a small hairpin-RNA (Hopfield, 1978) with an amino acid bound to it (Di Giulio, 1998). A relic of this primordial tRNA may be the presence of some modified nucleosides, which are today formed from the major RNA nucleosides (A, C, G, U) by the addition of an intact amino acid (e.g. N⁶-threonyl-adenosine (t⁶A), 2-lysyl-cytidine (k²C), glutamate bound to Q of tRNA^Asp (GluQ)) or parts of amino acids (methylated and thiolated nucleosides and acp³U). Therefore, one way that modified nucleosides may have evolved is by altering the substrate recognition of an enzyme involved in intermediary metabolism to recognize tRNA as substrate. Alternatively, a protein already having an affinity for tRNA may evolve a new enzymatic activity. Such alterations are predicted to be dominant over their wild-type allele and thus result in a 'gain of function' phenotype. By performing a similar chemical reaction as earlier but now also using tRNA as substrate, a novel modified nucleoside may have evolved. If the presence of this novel modified nucleoside provides an evolutionary advantage, the alteration will be fixed and the enzyme may thereafter evolve into a sophisticated tRNA modifying enzyme. Results in this paper address this aspect of tRNA modification.

A +1 frameshift mutation results from addition of one base in a coding sequence and such mutations may be suppressed by various alterations in the translation apparatus (Farabaugh, 1996). The first suppressors characterized were dominant and they have gained the ability to suppress a +1 frameshift mutation. Such suppressors have an altered tRNA with an extra nucleoside in the anticodon loop (Roth, 1981). We have shown that a deficiency of some modified nucleosides in tRNA also induces the ability to suppress many different +1 frameshift mutations and, in fact, it seems to be a common function of many modified nucleosides to improve the reading frame maintenance (Urbonavicius et al, 2001). According to our frameshift model (Supplementary Figure 1) the aberrantly modified tRNA is either poorly aminoacylated or/and that the ternary complex containing such a defective tRNA per se enters the A-site inefficiently. Such events stall the ribosome and allow the peptidyl-tRNA to slip one nucleotide forward, which returns the ribosome to the zero frame. So far, mutations inducing aberrant tRNA modification are recessive and results in a deficiency of a modification ('loss of function mutations'). However, this paper describes, for the first time, a dominant 'gain of function' mutation that alters the amino-acid sequence of a protein, induces suppression of a frameshift mutation, and mediates an efficient synthesis of new modifications in the tRNA.

The sufY204 mutation suppresses the +1 frameshift mutation hisC3737 and causes an amino-acid substitution in a protein

A sufC10 mutant of Salmonella enterica serovar Typhimurium (Riddle and Roth, 1970) contains a mutation, sufY204, which suppresses the hisC3737 mutation (Sroga et al, 1992). Following localization of the sufY204 mutation on the chromosome of S. enterica, the DNA sequence of the sufY⁺ and sufY204 alleles revealed that the sufY204 mutation was a G to A base substitution resulting in a Gly67 Glu67 (G67E) change of the SufY protein sequence. (For details, see Supplementary data.) Two other sufY mutants (sufY205 and sufY206) isolated earlier (Riddle and Roth, 1970) were also sequenced and both had a substitution of the same amino acid of the SufY protein: sufY205 had the same amino-acid substitution as sufY204 (G67E), while sufY206 had a G67R substitution.

The mutation in hisC3737 is an insertion of C

We noticed that the sufY204 mutation suppressed only the hisC3737 and the hisC3734 mutations among many other frameshift mutations in the his operon available to us (Supplementary Table I and Björk et al, 1989). Sequencing revealed that mutations hisC3737 and hisC3734 are identical, both being C insertions that give a run of five Cs resulting in a proline codon CCC, a glutamine codon CAA, and a stop codon UAA in the zero frame (Supplementary Table I). Since the sufA6 and sufB2 frameshift suppressors, which encode an altered tRNA_CGG^Pro and tRNA_GGG^Pro, respectively, also suppress the hisC3737 and hisC3734 mutations and are known to induce frameshifts at runs of Cs (Supplementary Table I), we suspected that one of the two proline CCC codons present within the '+1 frameshifting window' in the zero frame might be the site of frameshifting.

The hisD3749 mutation is a frameshifting site containing a CCC codon followed by a stop codon in the zero frame (Supplementary Table I). The suppression of hisD3749 by the sufB2 mutation, which alters tRNA_GGG^Pro, is not mediated by this tRNA but by the near cognate tRNA^Pro_cmo⁵UGG (Qian et al, 1998). The mutated cognate tRNA_GGG^Pro is out-competed in the selection at the ribosomal A-site by the near-cognate tRNA^Pro_cmo⁵UGG at the CCC codon in the zero frame. Following a normal three-nucleotide translocation, the tRNA^Pro_cmo⁵UGG is in the P site. Provided that the ribosome pauses at the next codon, the peptidyl-tRNA^Pro_cmo⁵UGG slips forward one nucleotide causing the ribosome to return to the +1 frame (cf. sequence for hisD3749 in Supplementary Table I). This frameshifting is abolished in an aroD mutant, since the aroD mutation blocks the synthesis of cmo⁵U34 (Björk, 1980), which promotes tRNA^Pro_cmo⁵UGG to out-compete the mutated cognate sufB2 tRNA_GGG^Pro at a low frequency in reading the CCC codon. Since the sufY204-mediated suppression was inhibited by the presence of an aroD mutation (data not shown), the sufY204-induced suppression is dependent on a cmo⁵U34-containing tRNA. We therefore propose that at the two possible frameshifting sites, -CCC-GCG- or -CCC-CAA- present within the frameshifting window in a hisC3737 mutant (Supplementary Table I), the tRNA^Pro_cmo⁵UGG may slip in the P-site into the +1 frame provided that a slow entry of the next tRNA occurs, which should stall the ribosome. At the first possible site within the frameshifting window, CCC-GCG, the pausing should be induced by a slow entry of Ala-tRNA^Ala_cmo⁵UGC, which decodes GCG. At the second site, the slow entry should be caused by the Gln-tRNA^Gln_{cmnm⁵s²UUG}, which reads CAA and to a lesser extent CAG (Yokoyama et al, 1985). Which of these two sites is the frameshifting site is addressed in the next section.

Translational frameshifting occurs at the CCC-CAA-UAA sequence

To identify the frameshifting site in hisC3737, we used a series of plasmids containing a +1 frameshift mutation placed in the beginning of the lacZ gene (Figure 1). The frameshifting window is small, with only four codons between the upstream stop codon in the +1 frame and the downstream stop codon in the zero frame. The reporter lacZ gene is placed in the +1 frame and thus -galactosidase activity reflects the level of +1 frameshifting. When CCC-GCG was tested, we did not observe any change in the level of +1 frameshifting between sufY⁺ and sufY204 strains (data not shown). Frameshifting sites were constructed by placing a CCC codon in the beginning of the lacZ gene followed by any of the four CAN codons to test whether the second site CCC-CAA in the hisC3737 frameshifting window was the frameshifting site. We introduced these plasmids into strains GT5633 (sufY⁺) and GT5424 (sufY204). No increased frameshifting compared to the wild-type strain was observed when the CCC codon was followed by CAU, CAC or CAG codons. However, a four-fold increase in +1 frameshifting in the sufY204 strain compared to the sufY⁺ strain was observed, when plasmid (pTHF42), which contained the sequence CCC-CAA, was present in the cells (Figure 1). Although the surrounding codon context is different in plasmid pTHF42 compared to that around the hisC3737 mutation, frameshifting occurs in both assay systems, indicating that CCC-CAA is the minimal requirement for this +1 frameshifting to occur. According to our frameshifting model (Farabaugh and Björk, 1999; Urbonavicius et al, 2001) (Supplementary Figure 1), a defective tRNA^Gln_{cmnm⁵s²UUG}, which reads CAA codon and less well CAG, may cause a slow entry to the ribosomal A site, resulting in a pause that allows the P-site proline tRNA^Pro_cmo⁵UGG decoding the CCC codon to slip one nucleotide forward and thus shift frame. We therefore expect that the sufY204 mutation in some way alters the structure of the tRNA^Gln_{cmnm⁵s²UUG} resulting in poor coding capacity of tRNA^Gln_{cmnm⁵s²UUG} and/or poor charging of it, which both should stall the ribosome. The CAG codon is also decoded by tRNA_CUG^Gln, which only reads the CAG codon, consistent with our observation that no increased frameshifting occurred when the synonymous glutamine codon CAG was next to and 5' of CCC (Figure 1). To further support our suggestion that it was a proline tRNA that shifted frame, we determined the amino-acid sequence of the frameshift product. The frameshift sequence -CCC-CAA-UAA was inserted between the genes gst and malE encoding the glutathione-S-transferase (GST) and the maltose-binding protein, respectively. The malE gene is in the +1 frame relative to the gst gene (Herr et al, 2001; Hansen et al, 2003). The complete fusion protein was purified from a strain containing the sufY204 mutation and a plasmid (pUST290) harboring the above-mentioned 'GST-CCC-CAA-UAA-MalE' construct. To liberate the slippage junction fused to MalE, the frameshift product was treated with PreScission Protease. After digestion, the N-terminal sequence of the slippage junction was determined. The first 12 amino acids of the peptide were GPLGILNPKANN, where P was the last amino acid inserted in the zero frame, suggesting that the frameshifting tRNA at the frameshift site CCC-CAA-UAA was a proline tRNA (Supplementary data).

Figure 1.

(A) pTHF plasmid constructs used to measure +1 frameshifting (Hagervall et al, 1993). The RNA sequence is shown in triplets to indicate the reading frame. (B) +1 frameshifting (measured by -galactosidase activity) on pTHF40, pTHF41, pTHF42, pTHF43 plasmid in sufY⁺ (GT5633), sufY204 (GT5424) and sufYMudCam (GT5520) strains.

View full figure (55 KB)

Taken together, these results suggest that the frameshift occurs in the P-site by slippage of peptidyl-tRNA^Pro_cmo⁵UGG According to our frameshifting model (Supplementary Figure 1), such slippage requires that the ribosome stalls at the A-site, suggesting that the tRNA^Gln_{cmnm⁵s²UUG} is in some way defective in the sufY204 mutant causing such a ribosomal stalling.

Aminoacylation of tRNA^Gln_{cmnm⁵s²UUG} is reduced in the sufY204 mutant

A defective tRNA^Gln_{cmnm⁵s²UUG} may result in inefficient charging and thereby stalling the ribosome that would induce a slippage of the peptidyl-tRNA. We therefore prepared tRNA from sufY204 and sufY⁺ strains under acidic condition in order to preserve the amino acid charged on the tRNA molecules. By running an acidic denaturing polyacrylamide gel, Gln-tRNA^Gln_{cmnm⁵s²UUG} and deacylated tRNA^Gln_{cmnm⁵s²UUG} can be separated and detected by Northern blot (Figure 2). The charging level in the sufY⁺ strain was 86% and in the sufY204 mutant the charging level was reduced to 26%, which is more than a three-fold reduction. According to our frameshifting model, the low level of charged tRNA^Gln_{cmnm⁵s²UUG} would at least partly explain the suppression of the hisC3737 mutation.

Figure 2.

Charging level of tRNA_U^*UG^Gln in sufY204 or sufY⁺ strains. tRNA was prepared from strains GT5633 (sufY⁺) and GT5424 (sufY204) as described (Materials and methods). The charging level is calculated by dividing the amount of Gln-tRNA^Gln_{cmnm⁵s²UUG} by those of Gln-tRNA^Gln_{cmnm⁵s²UUG} and tRNA^Gln_{cmnm⁵s²UUG} Two clones were used for each strain, and a deacylated sample was loaded as a control. Lane 1: GT5424a (sufY204), 2: GT5424b (sufY204), 3: GT5424 (sufY204), deacylated tRNA, 4: GT5633a (sufY⁺), 5: GT5633b (sufY⁺), 6: GT5633 (sufY⁺), deacylated tRNA.

View full figure (25 KB)

The sufY204 mutation is dominant and results in a 'gain of function' phenotype

During the mapping of the sufY204 mutation, we obtained a MudCam insertion that abolished the ability to suppress the hisC3737 mutation (phenotypically His^-) as well as the +1 frameshifting monitored as -galactosidase activity (Figure 1). This sufY2522MudCam insertion was located at the end of sufY204 coding sequence resulting in a sufY gene having both the sufY204 mutation and the MudCam insertion (Supplementary data). Since the transposon insertion abolished the suppressor activity, the sufY204 mutation is a 'gain of function mutation' and consequently should be dominant over its wild-type allele. To test this hypothesis, we introduced a plasmid carrying the wild-type allele of sufY (sufY⁺) into strain GT6026 (hisC3737, sufY204) resulting in a strain having the sufY⁺ allele on the plasmid and the sufY204 allele on the chromosome. Such a strain is still His⁺ demonstrating that the sufY204 mutation is dominant over its wild-type allele (Table I). Introduction of a plasmid containing the sufY204 allele into a wild-type strain induced a His⁺ phenotype stronger than that induced by a single copy of the sufY204 mutation residing on the chromosome. Strong suppressor activity was observed irrespectively if the strain possessed the sufY⁺ or the sufY204 allele on the chromosome (Table I). Thus, the sufY204 mutation is dominant, consistent with the suggestion that it mediates a gain of function. We further noticed that multiple copies of the sufY204 allele resulted in a stronger suppression than that observed by the presence of only one copy (cf. strains GT6028, GT6027 and GT6030).

Table 1: Levels of (c)mnm⁵s²U, UK1 and UK2 and suppression of the +1 frameshift mutation hisC3737

View full table

The sufY204 mutation induces the appearance of novel modified nucleosides in bulk tRNA and in purified tRNA^Gln_{cmnm⁵s²UUG}

Possible explanations for the sufY204-mediated frameshifting might be a changed structure of the tRNA^Gln_{cmnm⁵s²UUG}per se or a reduced level of Gln-tRNA^Gln_{cmnm⁵s²UUG} (Figure 2) caused by the structural alteration. Transfer RNA from a sufY204 mutant strain possessed two prominent 'UnKnown' novel compounds (denoted UK1 and UK2, Figure 3). Note also that compounds UK1 and UK2 are present in the wild-type tRNA, although at very low levels. Since the sufY2522MudCam strain is His^-, we expected that these compounds would be absent, which was the case (Figure 3). Interestingly, if the sufY204 allele was present both on the chromosome and on a plasmid, the amount of these UKs increased, suggesting that their synthesis was sensitive to gene dosage. Moreover, the amounts of these two unknown compounds positively correlated with the efficiency of frameshift suppression (Table I).

Figure 3.

HPLC chromatograms of tRNA hydrolysates. The (A, C, E) panels show the chromatograms from 9.6 to 20 min and (B, D, F) panels show the chromatograms from 75.2 to 84.8 min. Note the different scales on the Y-axes between panels A, C and E versus panels B, D and F. The arrows indicate the elution positions of (c)mnm⁵s²U, UK1 or UK2 (relative retention time to ms²i⁶A: UK1=0.890, UK2=0.921). The (c)mnm⁵s²U denotes the combined areas of the cmnm⁵s²U and mnm⁵s²U, which migrates very close to each other. See quantification of various modified nucleosides in Table I. A, B: tRNA hydrolysate from strain GT6031 (pCL1921/hisO1242, hisC3737, sufY⁺, zbb2523Tn10); C, D: from strain GT6028 (pCL1921/hisO1242, hisC3737, sufY204, zbb2523Tn10); E, F: from strain GT5520 (hisO1242, hisC3737, sufY2522MudCam).

View full figure (67 KB)

Not only did we observe increased amounts of UKs by the action of the sufY204 allele, but also a decreased synthesis of (c)mnm⁵s²U34 (Table I). These modified nucleosides, which are present in the wobble position of tRNA^Gln_{cmnm⁵s²UUG}, tRNA^Lys_mnm⁵s²UUU, and tRNA^Lys_mnm⁵s²UUU were reduced to about 50% in the sufY204 strain compared to the levels found in a sufY⁺ strain (Table I, cf. strains GT6028 to GT6031). This reduction was more pronounced (33% of the wild-type level) in a strain containing the sufY204 allele both on the chromosome and on the plasmid. Thus, there is an inverse relationship between the amounts of UKs and (c)mnm⁵s²U, suggesting that the appearance of the novel nucleosides occurs at the expense of (c)mnm⁵s²U. Accordingly, when ³⁵S-labeled tRNA was analyzed, UK2 contained radioactivity consistent with the suggestion that it is a derivative of mnm⁵s²U (data not shown).

To verify the model of how sufY204 induced frameshift suppression, we purified tRNA^Gln_{cmnm⁵s²UUG} from wild-type and sufY204 mutant cells and analyzed the nucleoside modification pattern (Table II). An unknown modified nucleoside (UK in Table II) was found in purified tRNA^Gln_{cmnm⁵s²UUG} from the sufY204 mutant, but was absent in tRNA^Gln_{cmnm⁵s²UUG} from the sufY⁺ strain. The sample of tRNA^Gln_{cmnm⁵s²UUG} was pure, since other modified nucleosides, which should be present in tRNA^Gln_{cmnm⁵s²UUG} (, m⁵U, m²A, Gm, Um, s⁴U, cmnm⁵s²U) were observed, but not any other modified nucleosides that may have originated from contaminating tRNA species. The UK, which had a UV spectrum similar to UK1 and UK2 found in bulk tRNA, had a relative retention time that was not identical neither to UK1, nor to UK2, and nor to UK3 (see below). The reason that the UK found in purified tRNA^Gln_{cmnm⁵s²UUG} did not migrate as any of UK1, UK2 or UK3 may be caused by an alteration of the structure caused by the extensive purification of tRNA^Gln_{cmnm⁵s²UUG} Still, the UV spectrum is identical to that of UK1–UK3 and the level of cmnm⁵s²U is reduced in the tRNA^Gln_{cmnm⁵s²UUG} from the sufY204 mutant.

Table 2: Relative levels of modified nucleosides in tRNA^Gln_{cmnm⁵s²UUG} from sufY⁺ and sufY204 mutant

View full table

Structure studies of UK2 by electrospray ionization mass spectrometry (ESI-MS)

By LC/MS we determined the M_r of UK2 to be 745, which is greater than that of any modified nucleoside hitherto characterized (Rozenski et al, 1999). However, nuclease P1, which we routinely use to degrade the tRNA, is sensitive to bulky modifications (Gehrke and Kuo, 1989; Crain, 1990) and leaves as the end product a dinucleotide in which the hypermodified nucleoside is the 5' member. Given the U^*UG anticodon of tRNA^Gln (U^* denotes an uncharacterized derivative of uridine), we predict that UK2 may be a dinucleotide, U^*pU, whose degradation is resistant to nuclease P1. Indeed, when UK2 was degraded to uridine by snake venom diesterase, an endonuclease not as sensitive to bulky substituents as P1 (Crain, 1990), the 5' modified nucleoside component of UK2 could not be identified (data not shown). In view of the probable structure of this nucleoside (discussed below), it is likely that it was not eluted from the high-performance liquid chromatography (HPLC) column under the conditions used (Gehrke and Kuo, 1989).

If UK2 is a dinucleotide consisting of a normal U and the modified U^*, the indicated M_r of U^* is 439, which does not correspond to that of any known nucleoside (Rozenski et al, 1999). Therefore, UK2 was purified and its structural features were examined following collision-induced dissociation (CID) of the protonated molecular ion (MH⁺), m/z 746 (Figure 4A). From studies of CID of dinucleotides (Phillips and McCloskey, 1993), ions corresponding to protonated free base (BH₂⁺) ions resulting from cleavage of the glycosidic bond with proton transfer from ribose to the base (Pomerantz and McCloskey, 1990) can be expected for both U (B=111) and U^* (B=306) (Figure 4B); they occur at m/z 113 and 308, respectively. An unusual fragmentation pathway is manifested in the loss of 136 Da from both MH⁺ and BH₂⁺ ions of U^*, giving m/z 610 and m/z 172, respectively. The m/z 172 ion is consistent with the value expected for the BH₂⁺ ion of mnm⁵s²U, and m/z 610 is consistent with an assignment as the MH⁺ ion of the dinucleotide mnm⁵s²UpU. The difference in mass between UK2 (U^*pU) and mnm⁵s²UpU is 136, suggesting that UK2 may consist of a dinucleotide of mnm⁵s²UpU plus an additional 137 Da substituent (net addition of 136 Da) that is present on the mnm⁵s²U base moiety. Apparently, the presence of this additional 137 Da substituent is inhibiting the P1 nuclease activity, thereby generating the U^*pU dinucleotide of M_r 745.

Figure 4.

LC/ESI-MS analysis of UK2 isolated from a digest of total tRNA from strain GT6027 (psufY204/hisO1242, hisC3737, sufY204). (A) Product ion mass spectrum from collision-induced dissociation of the MH⁺ ion of UK2. (B) Partial structure of UK2 deduced from mass spectral data.

View full figure (22 KB)

The 137 Da fragment that has been added to mnm⁵s²U forms a stable ion, (m/z 137, Figure 4), and LC/MS analysis of UK2 using deuteriated solvents (Edmonds et al, 1988) was carried out to determine the number of exchangeable H atoms in this ion, and to confirm the other ion assignments proposed above. This measurement was conducted at high resolution to permit an accurate mass to be determined for m/z 137 at the same time. The results indicated that there are no exchangeable hydrogens in the 137 Da fragment, its composition is C₁₀H₁₇, and it contains two double bonds. The spectrum and its interpretation are presented in Supplementary Figure 2. The data do not permit determination of the nature of the C₁₀H₁₇ substituent, but its composition is compatible with that of a geranyl group (or an isomer of it).

The appearance of the novel modifications is related to the degree of modification of mnm⁵s²U

As stated above, the UKs are related to the wobble nucleoside mnm⁵s²U. If so, mutations that result in accumulation of intermediates in the synthesis of this modified nucleoside should influence the structure of the UKs in a way predicted from the structure of the intermediates. The proposed biosynthetic pathway of mnm⁵s²U is shown in Figure 5 (Björk and Hagervall, 2005). Mutations in mnmA result in a deficiency of the s²-group (Sullivan et al, 1985), whereas mutations in mnmE result in a deficiency of the mnm⁵-side chain (Elseviers et al, 1984). The mnmC1 mutation results in accumulation of cmnm⁵s²U, whereas mutation mnmC2 results in the accumulation of nm⁵s²U (Hagervall et al, 1987). In the mnmA mutant no detectable UKs were observed, and in the mnmE mutant the level of UK2 was reduced to about 1.1% of that found in the mnmE⁺ strain (Figure 5). Thus, the s²- and mnm⁵-groups are required for the synthesis of the UKs. In the mnmC1 mutant, which accumulates cmnm⁵s²U, UK1 becomes the major unknown, whereas in the mnmC2 mutant, which accumulates nm⁵s²U, UK3 becomes the major unknown compound (Figure 5). Note that UK3 was not induced by the sufY204 mutation in the wild-type background, in which only UK2 and UK1 were synthesized. Clearly, the various UKs are related to the mnm⁵s²U derivatives found in the mutants suggesting that UK1 is associated with cmnm⁵s²U, UK2 with mnm⁵s²U and UK3 with nm⁵s²U. Accordingly, the M_rs of UK1, UK2 and UK3 should be related to each other in the same manner as those of cmnm⁵s²U (M_r 347), mnm⁵s²U (M_r 303) and mnm⁵s²U (M_r 289), which indeed they are (Figure 5). Since the M_rs of the three UKs were related to the 5-side chain in the intermediates, the derivatization of these derivatives does not occur on this side chain but rather on some other part of the nucleoside. These results strengthen the suggestion that the altered SufY protein modifies the wobble nucleoside mnm⁵s²U34.

Figure 5.

Levels of UK1, UK2 and UK3 in E. coli mutants mnmA, mnmE, mnmC1,and mnmC2 containing a plasmid (pUST211, sufY204) harboring the Salmonella sufY204 allele. tRNAs were prepared and analyzed by HPLC. Amounts of UK1, UK2 and UK3 are presented relative to the amount of t⁶A. The retention times for UKs relative to ms²i⁶A are: UK1 (0.890), UK3 (0.917) and UK2 (0.921). ^*: Not detected (less than 0.003 relative to t⁶A).

View full figure (83 KB)

Selenation is not involved in the synthesis of UKs

Recently, the YbbB protein of Escherichia coli was shown to catalyze the selenation of tRNA and it was suggested that YbbB is the selenouridine synthase (Wolfe et al, 2004). This enzyme, which contains a rhodanese domain (Figure 6), exchanges the sulfur of mnm⁵s²U34 for selenium. The selenium donor is selenophosphate, which is produced by the selD gene product (Veres et al, 1994). Indeed, deletion of the selD gene results in inability to incorporate selenium into proteins and tRNAs (Kramer and Ames, 1988; Stadtman et al, 1989). It was further shown that Cys97, which is conserved in the sulfurtransferases with a rhodanese domain, is required for the incorporation of Se. Since the sufY gene is the Salmonella homologue to the E. coli ybbB gene, we wanted to know if selenation is involved in the synthesis of the UKs in S. enterica.

Figure 6.

Overview of the SufY protein in S. enterica (modified from Wolfe et al, 2004). The Rhodanese domain (Rhd: aa 3–130) and the catalytic Cys97 (star in the Rhd domain with arrow above) are indicated. The 36 amino acid insertion (aa 41–76) within the Rhd domain in nine YbbB orthologs according to Wolfe et al (2004) is shown in gray, the sequence of which is shown below. P-loop (ATP binding motif) is present between amino acids 149–156. IleS-like region (aa 324–356) is shown in dark gray according to (Wolfe et al, 2004). The MudCam transposon insertion is shown with a triangle at D342 in the S. enterica sequence. The amino–acid substitutions in sufY204 (G67E) or sufY206 (G67R) mutants are shown with arrows.

View full figure (39 KB)

Since selenation requires the selenium donor SePO₃, the formation of which is catalyzed by the SelD protein, we constructed a selD deletion mutant in Salmonella. Introduction of a sufY204 plasmid revealed that UK1 and UK2 were still made in the selD mutant, demonstrating that selenation is not required for the formation of these unknown nucleotides (Figure 7). Compared to a plasmid containing the sufY204 mutation, a plasmid containing the sufY204 mutation in conjunction with a C97A substitution reduced the level of UK1 and UK2 to 36 and 13%, respectively. Thus, the Cys97 is important but not absolutely required for the formation of UKs.

Figure 7.

Levels of UK1, UK3, and UK2 in sufYfrt or selDfrt mutants containing plasmid pUST211 (sufY204) or plasmid pUST273 (with C97A and G67E alterations in the SufY/YbbB). tRNAs were prepared and analyzed by HPLC (Materials and methods). Amounts of UK1, UK3 and UK2 are presented relative to the amount of t⁶A. ^*: not detected (less than 0.003 AU relative to t⁶A).

View full figure (17 KB)

In this paper we report, for the first time, a dominant mutation ('a gain of function mutation'), which alters a protein and influences the modification of tRNA. Apparently, the wild-type form of the protein has an intrinsic ability to modify tRNA, although at such a low efficiency that this modification hitherto has not been observed. Although the structure of the modification is not established, it is clear that it is a derivative of the wobble nucleoside (c)mnm⁵s²U34, which is present in tRNAs specific for Gln, Lys and Glu.

The sufY204 mutation was discovered as a suppressor of a +1 frameshift mutation. According to the model how tRNA modification influences reading frame maintenance (Urbonavicius et al, 2001), we predict that the sufY204 mutation should alter the tRNA^Gln_{cmnm⁵s²UUG} since it acted at the sequence CCC-CAA-UAA (Figure 1). If the unknown modification were located at or in the vicinity of the wobble nucleoside cmnm⁵s²U34 of tRNA^Gln_{cmnm⁵s²UUG} we expect that it might influence the aminoacylation of this tRNA, since the wobble nucleoside is part of the recognition determinant for the GlnRS (Ibba et al, 1996). Indeed, this was the case (Figure 2) and therefore the induced frameshifting can be explained as a low level of ternary complex that in turn causes the ribosome to stall and allow the peptidyl-tRNA^Pro_cmo⁵UGG to slip into the +1 frame. The C₁₀H₁₇ fragment, which we suggest is bound to the s²-group of mnm⁵s²U, can of course, directly influence the anticodon–codon interaction, provided that a ternary complex is formed containing the C₁₀H₁₇-modified tRNA. The s²-group of mnm⁵s²U34 negates the bifurcated hydrogen bonding to a G as third nucleoside in the codon. It further improves the stacking with U35 of the anticodon, which results in an efficient reading of the A-ending codon (Murphy et al, 2004). One can envision that a large bulky substituent, such as the C₁₀H₁₇-fragment, at the wobble nucleoside may reduce severely the entry of such a tRNA to the A-site. In fact, it is questionable if such a C₁₀H₁₇ containing tRNA ever enters the A-site. Alternatively, the C₁₀H₁₇-modified tRNA may never be charged resulting in a low level of aminoacylation, which causes stalling of the ribosome. Irrespective of what actually is the dominating cause of the induction of a +1 frameshift, both these alternatives are consistent with the model how altered modification of tRNA influences reading frame maintenance (Urbonavicius et al, 2001).

The SufY (YbbB) protein has in its C-terminal a helical region, which may contribute to binding of the anticodon of tRNA and a rhodanese domain in the N-terminal part (Figure 6). Moreover, a P loop (Walker motif), which is found in proteins that bind ATP or GTP, is present just downstream of the rhodanese domain. The YbbB homologs contain an insertion of about 36 amino acids that is present between the CH2A rhodanese motif and the active site of a rhodanese (Cys97) (Wolfe et al, 2004). This insertion is conserved among the YbbB homologs, and Gly67, which is substituted by glutamate in the sufY204 mutant inducing frameshift and the synthesis of two not characterized modified nucleosides, is conserved within this region in nine YbbB orthologs. Since the wild-type form of SufY can interact with tRNA (Wolfe et al, 2004) and it contains a possible tRNA binding domain (Figure 6), it suggests that the G67E alteration influences a potential enzymatic activity of SufY rather than the tRNA binding activity. The fact that SufY may contain a tRNA binding motif and uses tRNA as substrate in a tRNA modification reaction (Wolfe et al, 2004), suggests that SufY is a tRNA modifying enzyme rather than a regulatory protein influencing the synthesis of an uncharacterized tRNA modifying enzyme.

The emergence of various modifications has most likely occurred via different routes. One would be that an enzyme, which catalyzes the synthesis of an intermediate in central or intermediary metabolism, is altered to have an affinity for a tRNA. The enzyme will still catalyze the same or similar enzymatic reaction(s) but, in addition, will now also be able to use a base in the tRNA as substrate. Alternatively, the enzyme may already have an affinity for tRNA, which SufY apparently has (Wolfe et al, 2004), the alteration may change the efficiency of a low intrinsic catalytic activity. Interestingly, the wild-type form of SufY is able to catalyze the formation of UKs, although at a very low efficiency (Figure 3). Perhaps, the glycine at position 67 (or the entire 36 amino-acid insertion found in all YbbB orthologs, Figure 6) dampens an inherent enzymatic activity, which is catalyzing the formation of C₁₀H₁₇-mnm⁵s²U34. An appropriate amino-acid change at this position would relieve this negative effect, allowing the protein to regain its ability to form C₁₀H₁₇-mnm⁵s²U34. This suggests that the observed novel modification may have been present earlier in the evolution of S. enterica and E. coli. Perhaps the advantage of C₁₀H₁₇-mnm⁵s²U34 in tRNA at this early time was overriding the disadvantage of its ability to induce frameshifting as shown by us. However, if, in some way, this evolutionary advantage was lost, any negative element that abolished the formation of C₁₀H₁₇-mnm⁵s²U34, thereby reducing errors in reading frame maintenance to a level acceptable for the organism, would be selected.

Bacteria and growth conditions

All bacteria are derivatives of S. enterica serovar Typhimurium strain LT2 or E. coli (Supplementary Table II). Complex medium NAA was prepared from Difco nutrient broth (0.8%, Difco Laboratories, Detroit, MI) supplied with 0.5% NaCl, adenine, tryptophan, tyrosine, phenylalanine, and the three aromatic vitamins p-hydroxybenzoate, 2,3-dihydroxybenzoate and p-aminobenzoate. All supplements were added at concentrations recommended (Davis et al, 1980). For solid medium, TYS agar (10 g of Tryptone, 5 g of yeast extract, 15 g of NaCl and 15 g of agar per liter) was used. Minimal medium was prepared either from medium E (Vogel and Bonner, 1956) or MOPS (morpholinepropanesulfonic acid) medium (Neidhardt et al, 1974) supplemented with 0.4% carbon source. Rich MOPS medium was prepared as described before (Neidhardt et al, 1977). Antibiotics were used at the following concentrations: carbenicilin 100 g/ml, chloramphenicol 12.5 g/ml, tetracyclin 15 g/ml, spectinomycin 800 g/ml.

Genetic procedures

Transductions were performed using a derivative of phage P22 containing the mutations HT105/I (Schmieger, 1972) and int-201 (Scott et al, 1975). Transduction was performed as described (Davis et al, 1980).

Molecular cloning techniques

Standard procedures of molecular cloning were used. Chromosomal DNA prepared from strain GT5520 (hisO1242, hisC3737, sufY204MudCam, zbb-2523Tn10dTet) and from the vector plasmid pBR322 were digested with PstI and SalI and then ligated using T4 DNA ligase (Boehringer Mannheim). E. coli strain DH5 (supE44, lacU169 (80lacZM15) hsdR17, recA1, endA1, gyrA96, thi-1, relA1) was used in transformations. Plasmid preparation kit was from Qiagen, and to PCR amplify a gene of interest, puReTaq-Ready-to-go PCR beads (Amersham Pharmacia, Piscataway, NJ) were used. PCR fragments were purified using PCR Kleen Spin columns (Bio-Rad) for direct recovery of the PCR product. Alternatively, the Wizard DNA clean-up Resin (Promega, Madison, WI) was used for purifying PCR fragments from agarose gels. Automatic sequencing was performed using DYEnamic ET terminator cycle sequencing premix kit (Amersham Pharmacia Biotech Inc., Cleveland, OH).

Determination of -galactosidase activity

Strains were grown at 37°C in NAA medium supplemented with 50 g/ml carbenicilin overnight for selection of pTHF plasmids. Strains were subcultured to mid-log phase, -galactosidase activity was measured as described by Miller (1972). ONPG was from Sigma (St Louis, MO). For each strain to be tested, three independent cultures were grown.

HPLC analysis of modified nucleosides

Strains were grown in LB medium at 37°C from 5 Klett units to 100 Klett units (approx. 4 10⁸ cells/ml). Total tRNA was prepared by phenol-extraction in acidic conditions and purified through a Nucleobond column (Clontech) as described elsewhere (Urbonavicius et al, 2001; Björk and Nilsson, 2003). Total tRNA (100 g) was digested with nuclease P1 (Boehringer Mannheim) followed by alkaline phosphatase (Sigma) treatment as described (Gehrke et al, 1982). The hydrolysates (and authentic cmnm⁵s²U) were analyzed by HPLC using a Supelcosil LC18 column (Supelco) with a Waters HPLC system. The gradient used was as described by Gehrke and Kuo (1989).

Tandem mass spectrometry (MS/MS) of dinucleotide UK2

Electrospray ionization LC/MS/MS was used for structure studies of the isolated dinucleotide. The data shown in Figure 4 were acquired on a Quattro II (Micromass) triple quadrupole mass spectrometer interfaced with an HP 1090 liquid chromatograph. The HPLC gradient was essentially as described (Pomerantz and McCloskey, 1990) except that the ammonium acetate concentration and solvent flow rate were reduced to 5 mM and 200 l/min, respectively. The collision energy was 20 eV. The ion source and desolvation temperatures were 140 and 280°C, respectively.

Isolation of tRNA^Gln_{cmnm⁵s²UUG}

To isolate single tRNA species (tRNA^Gln_{cmnm⁵s²UUG}), Dynabead (Dynal, Oslo, Norway) was used. A biotinylated DNA oligo complementary to nucleotides 39–73 (35 nucleotides) of tRNA^Gln_{cmnm⁵s²UUG} was bound to the Dynabeads and tRNA^Gln_{cmnm⁵s²UUG} was hybridized to the Dynabeads in 2.4 M tetraethylammonium chloride (Tsurui et al, 1994). Purified tRNA^Gln_{cmnm⁵s²UUG} was desalted and concentrated using Centricon YM-3 cartridges (Millipore, Bedford, MA) three times with 2 ml MilliQ water. In all, 5–10 g specific tRNA was digested with nuclease P1 and bacterial alkaline phosphatase, and modified nucleosides were analyzed by HPLC as described above.

Tentative identification of the wobble nucleoside in tRNA^Gln_{cmnm⁵s²UUG} as cmnm⁵s²U

The wobble nucleoside in gln1 tRNA, which reads CAA and CAG, is a derivative of mnm⁵s²U34. To investigate whether it may be the cmnm⁵s²U derivative, we first determined the relative retention time of a synthetic cmnm⁵s²U to A (0.345) and to U (1.387) as compared to mnm⁵s²U, which migrated slightly different relative to A (0.369) and U (1.481). In purified gln1 tRNA, a compound with the same UV spectrum as mnm⁵s²U migrated similarly as synthetic cmnm⁵s²U relative to A and U. A minor peak (20 % of the compound suggested to be cmnm⁵s²U) was also observed with migration properties similar to mnm⁵s²U. Thus, we tentatively conclude that under the conditions used to purify gln1 tRNA, 80% of the wobble nucleoside in this tRNA is cmnm⁵s²U and about 20% is mnm⁵s²U. Although this analysis suggests that the wobble nucleoside is cmnm⁵s²U in gln1 tRNA, we did not spike the purified gln1 tRNA with the marker, nor have we determined its relative molecular mass (M_r). Therefore, it cannot be ruled out that the compound in gln1 tRNA is not cmnm⁵s²U but a compound with the same spectrum and the same relative retention time as cmnm⁵s²U. Nonetheless, we use the nomenclature tRNA^Gln_{cmnm⁵s²UUG} to indicate the gln1 tRNA.

Analysis of the aminoacylation level of tRNA^Gln_{cmnm⁵s²UUG}

Cells were grown at 37°C to a cell density of 2 10⁸ cells/ml in 100 ml LB medium. All subsequent steps were carried out in the cold. Total RNA was prepared under acidic conditions and electrophoresis of RNA was as described (Varshney et al, 1991) except that 8% denaturing polyacrylamide gel was used. The portion of the gel containing tRNAs of interest (which migrate at the same position as xylene cyanol under this condition) was electroblotted using a semidry electroblot apparatus (10 V for 30 min, 1 TBE as transfer buffer) onto a Zeta-Probe membrane (Bio-Rad). After transfer, the membrane was rinsed briefly in 2 SSC, and tRNA was crosslinked to the membrane using a UV Stratalinker (Stratagene). The oligonucleotide used to detect tRNA^Gln_{cmnm⁵s²UUG} was complementary to nucleotides 25–50 in the tRNA^Gln_{cmnm⁵s²UUG}. The oligonucleotide was labeled using adenosine 5'[³²P]-triphosphate (5000 Ci/mmol, Amersham) and polynucleotide kinase (Roche Molecular Biochemicals). The membrane was prehybridized in Church buffer (0.5 M Na₂HPO₄, 1 mM EDTA, 1% BSA, 7% SDS, pH 7.2) at 70°C for 30 min, hybridized with ³²P-labeled oligo at 70°C for 30 min, and cooled slowly to room temperature. The membrane was washed three times with 6 SSC, 0.1% SDS at 42°C for 5 min, and visualized and quantified by Phosphor-Imager analysis using ImageQuant software (Molecular Dynamics).

Deletion of the sufY and the selD genes by linear transformation

The method used has been described earlier (Datsenko and Wanner, 2000). Plasmid pKD4 was used as template plasmid. The 1.6 kb PCR fragment carrying the Kan^R cassette was electroporated into strain GT6315 (pKD46/LT2) and Kan^R clones were selected. The Kan^R cassette was then removed by recombination at FRT sites present on both sides of the cassette, thus the resulting strain carried a 'scar' replacing the sufY or selD genes. Such replacements are denoted sufYfrt and selDfrt, respectively. The primers for replacing sufY gene were designed to create a replacement of the sufY gene from 25nt upstream of the start codon to 2nt downstream of the stop codon. The primers for replacing selD were designed to create a replacement of the selD gene from 46nt upstream of the start codon to 23nt downstream of the stop codon (Supplementary data).

Site-directed mutagenesis of sufY

A QuickChange^TM Site-Directed Mutagenesis Kit (Stratagene) was used to change Cys97 to Ala97; plasmid pUST211 (carrying sufY204 allele) was used as template.

We thank Kerstin Jacobsson for HPLC analysis and for purification of UKs for mass spectrometry analyses. We are grateful for the generous gift of the strain containing the original sufC mutation from John Roth, University of California at Davis, California, USA and to Tord Hagervall for the plasmids pTHF40–43. Norma M Wills and John F Atkins are acknowledged for providing us with the vector pGHM57 and with the information on the assay system. The synthetic markers cmnm⁵s²U and mnm⁵s²U were kind gifts from Andrzej Malkiewicz, Technical University, Poland. We thank Tord Hagervall and Anders Byström (Umeå) and James McCloskey (Utah) for the critical reading of the manuscript. This work was supported by grants from the Swedish Cancer Foundation (Project 680) and Swedish Science Research council (Project BU-2930), and by NIH Grant GM29812.

Björk GR (1980) A novel link between the biosynthesis of aromatic amino acids and transfer RNA modification in Escherichia coli. J Mol Biol 140: 391–410 | PubMed | ISI |

Björk GR (1986) Transfer RNA modification in different organisms. Chemica Scripta 26B: 91–95

Björk GR, Hagervall TG (2005) Transfer RNA modification. In Escherichia coli and Salmonella Cellular and Molecular Biology, Curtiss III R, Ingraham JL, Kaper JB, Maloy S, Neidhardt FC, Riley MM, Squires CL, Wanner BL (eds). Washington, DC: American Society for Microbiology. http://www.ecosal.org/ecosal/module
s/index.jsp

Björk GR, Nilsson K (2003) 1-Methylguanosine-deficient tRNA of Salmonella enterica serovar Typhimurium affects thiamine metabolism. J Bacteriol 185: 750–759 | PubMed |

Björk GR, Wikström PM, Byström AS (1989) Prevention of translational frameshifting by the modified nucleoside 1-methylguanosine. Science 244: 986–989 | Article | PubMed | ISI

Cermakian N, Cedergren R (1998) Modified nucleosides always were: an evolutionary model. In Modification and Editing of RNA, Grosjean H, Benne R (eds), pp 535–541. Washington DC: ASM Press

Crain PF (1990) Preparation and enzymatic hydrolysis of DNA and RNA for mass spectrometry. Methods Enzymol 193: 782–790 | PubMed | ChemPort |

Datsenko KA, Wanner BL (2000) One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci USA 97: 6640–6645 | Article | PubMed | ChemPort |

Davis W, Botstein D, Roth JR (1980) A Manual for Genetic Engineering: Advanced Bacterial Genetics. New York: Cold Spring Harbor Laboratory

Di Giulio M (1998) Reflections on the origin of the genetic code: a hypothesis. J Theor Biol 191: 191–196 | Article | PubMed | ChemPort |

Edmonds CG, Pomerantz SC, Hsu FF, McCloskey JA (1988) Thermospray liquid chromatography/mass spectrometry in deuterium oxide. Anal Chem 60: 2314–2317 | Article | PubMed | ChemPort |

Elseviers D, Petrullo LA, Gallagher PJ (1984) Novel E. coli mutants deficient in biosynthesis of 5-methylaminomethyl-2-thiouridine. Nucleic Acids Res 12: 3521–3534 | PubMed | ISI | ChemPort |

Farabaugh PJ (1996) Programmed translational frameshifting [review]. Microbiol Rev 60: 103–134 | PubMed | ISI | ChemPort |

Farabaugh PJ, Björk GR (1999) How translational accuracy influences reading frame maintenance. EMBO J 18: 1427–1434 | Article | PubMed | ISI | ChemPort |

Gehrke CW, Kuo KC (1989) Ribonucleoside analysis by reversed-phase high-performance liquid chromatography. J Chromatogr 471: 3–36 | Article | PubMed | ChemPort |

Gehrke CW, Kuo KC, McCune RA, Gerhardt KO, Agris PF (1982) Quantitative enzymatic hydrolysis of tRNAs: reversed-phase high-performance liquid chromatography of tRNA nucleosides. J Chromatogr 230: 297–308 | Article | PubMed | ISI | ChemPort |

Hagervall TG, Edmonds CG, McCloskey JA, Björk GR (1987) Transfer RNA(5-methylaminomethyl-2-thiouridine)-methyltransferase from Escherichia coli K-12 has two enzymatic activities. J Biol Chem 262: 8488–8495 | PubMed | ISI | ChemPort |

Hagervall TG, Tuohy TM, Atkins JF, Björk GR (1993) Deficiency of 1-methylguanosine in tRNA from Salmonella typhimurium induces frameshifting by quadruplet translocation. J Mol Biol 232: 756–765 | Article | PubMed | ISI | ChemPort |

Hansen TM, Baranov PV, Ivanov IP, Gesteland RF, Atkins JF (2003) Maintenance of the correct open reading frame by the ribosome. EMBO Rep 4: 499–504 | Article

Herr AJ, Nelson CC, Wills NM, Gesteland RF, Atkins JF (2001) Analysis of the roles of tRNA structure, ribosomal protein L9, and the bacteriophage T4 gene 60 bypassing signals during ribosome slippage on mRNA. J Mol Biol 309: 1029–1048 | Article | PubMed | ISI | ChemPort |

Hopfield JJ (1978) Origin of the genetic code: a testable hypothesis based on tRNA structure, sequence, and kinetic proofreading. Proc Natl Acad Sci USA 75: 4334–4338 | PubMed | ChemPort |

Ibba M, Hong K-W, Soll D (1996) Glutaminyl-tRNA synthetase: from genetics to molecular recognition. Genes Cells 1: 421–427 | Article | PubMed | ChemPort |

Kramer GF, Ames BN (1988) Isolation and characterization of a selenium metabolism mutant of Salmonella typhimurium. J Bacteriol 170: 736–743 | PubMed | ChemPort |

Kurland CG, Hughes D, Ehrenberg M (1996) Limitations of translation accuracy. In Escherichia coli and Salmonella. Cellular and Molecular Biology, Neidhardt FC, Curtiss III R, Ingraham JL, Lin ECC, Brooks Low K, Magasanik B, Reznikoff WS, Riley M, Schaechter M, Umbarger HE (eds), pp 979–1004. Washington, DC: American Society for Microbiology

Miller JH (1972) Experiments in Molecular Genetics. New York: Cold Spring Harbor Laboratory

Murphy FV, Ramakrishnan V, Malkiewicz A, Agris PF (2004) The role of modifications in codon discrimination by tRNA(Lys)(UUU). Nat Struct Biol 11: 1186–1191 | Article | ChemPort |

Neidhardt FC, Bloch PL, Smith DF (1974) Culture medium for enterobacteria. J Bacteriol 119: 736–747 | PubMed | ISI | ChemPort |

Neidhardt FC, Bloch PL, Pedersen S, Reeh S (1977) Chemical measurement of steady-state levels of ten aminoacyl-transfer ribonucleic acid synthetases in Escherichia coli. J Bacteriol 129: 378–387 | PubMed | ChemPort |

Phillips DR, McCloskey JA (1993) A comprehensive study of the low energy collision-induced dissociation of dinucleoside monophosphates. Int J Mass Spectrom Ion Processes 128: 61–82 | Article | ChemPort |

Pomerantz SC, McCloskey JA (1990) Analysis of RNA hydrolyzates by liquid chromatography-mass spectrometry. Methods Enzymol 193: 796–824 | PubMed | ChemPort |

Qian Q, Li JN, Zhao H, Hagervall TG, Farabaugh PJ, Björk GR (1998) A new model for phenotypic suppression of frameshift mutations by mutant tRNAs. Mol Cell 1: 471–482 | Article | PubMed | ISI | ChemPort |

Riddle DL, Roth JR (1970) Suppressors of frameshift mutations in Salmonella typhimurium. J Mol Biol 54: 131–144 | Article | PubMed | ChemPort |

Roth JR (1981) Frameshift suppression. Cell 24: 601–602 | Article | PubMed | ISI | ChemPort |

Rozenski J, Crain PF, McCloskey JA (1999) The RNA modification database: 1999 update. Nucleic Acids Res 27: 196–197 | Article | PubMed | ISI | ChemPort |

Schmieger H (1972) Phage P22-mutants with increased or decreased transduction abilities. Mol Gen Genet 119: 75–88 | Article | PubMed | ISI | ChemPort |

Scott JF, Roth JR, Artz SW (1975) Regulation of histidine operon does not require hisG enzyme. Proc Natl Acad Sci USA 72: 5021–5025 | PubMed | ChemPort |

Sprinzl M, Vassilenko KS (2003) tRNA sequences and tRNA genes database http://www.uni-bayreuth.de/departme
nts/biochimie/sprinzl/trna/

Sroga GE, Nemoto F, Kuchino Y, Björk GR (1992) Insertion (sufB) in the anticodon loop or base substitution (sufC) in the anticodon stem of tRNA^Pro2 from Salmonella typhimurium induces suppression of frameshift mutations. Nucleic Acids Res 20: 3463–3469 | PubMed | ISI | ChemPort |

Stadtman TC, Davis JN, Zehelein E, Böck A (1989) Biochemical and genetic analysis of Salmonella typhimurium and Escherichia coli mutants defective in specific incorporation of selenium into formate dehydrogenase and tRNAs. Biofactors 2: 35–44 | PubMed | ChemPort |

Sullivan MA, Cannon JF, Webb FH, Bock RM (1985) Antisuppressor mutation in Escherichia coli defective in biosynthesis of 5-methylaminomethyl-2-thiouridine. J Bacteriol 161: 368–376 | PubMed | ISI | ChemPort |

Tsurui H, Kumazawa Y, Sanokawa R, Watanabe Y, Kuroda T, Wada A, Watanabe K, Shirai T (1994) Batchwise purification of specific tRNAs by a solid-phase DNA probe. Anal Biochem 221: 166–172 | Article | PubMed | ISI | ChemPort |

Urbonavicius J, Qian Q, Durand JM, Hagervall TG, Björk GR (2001) Improvement of reading frame maintenance is a common function for several tRNA modifications. EMBO J 20: 4863–4873 | Article | PubMed | ISI | ChemPort |

Varshney U, Lee CP, RajBhandary UL (1991) Direct analysis of aminoacylation levels of tRNAs in vivo. Application to studying recognition of Escherichia coli initiator tRNA mutants by glutaminyl-tRNA synthetase. J Biol Chem 266: 24712–24718 | PubMed | ISI | ChemPort |

Veres Z, Kim IY, Scholz TD, Stadtman TC (1994) Selenophosphate synthetase—enzyme properties and catalytic reaction. J Biol Chem 269: 10597–10603 | PubMed | ISI | ChemPort |

Vogel HJ, Bonner DM (1956) Acetylornithinase of Escherichia coli: partial purification and some properties. J Biol Chem 218: 97–106 | PubMed | ISI | ChemPort |

Wolfe MD, Ahmed F, Lacourciere GM, Lauhon CT, Stadtman TC, Larson TJ (2004) Functional diversity of the rhodanese homology domain: the Escherichia coli ybbB gene encodes a selenophosphate-dependent tRNA 2-selenouridine synthase. J Biol Chem 279: 1801–1809 | PubMed | ISI | ChemPort |

Yokoyama S, Watanabe T, Murao K, Ishikura H, Yamaizumi Z, Nishimura S, Miyazawa T (1985) Molecular mechanism of codon recognition by tRNA species with modified uridine in the first position of the anticodon. Proc Natl Acad Sci USA 82: 4905–4909 | Article | PubMed | ChemPort |

Main navigation

• Search EMBO Journal

  Search

• Advanced search

• The EMBO Journal

  • Home
  • Aims and scope
  • Current issue
  • Advance online publication
  • Web focuses
  • Archive:
  • Browse by issue
  • Browse by subject
  • Browse by category
  • Free online sample issue
  • Press releases

• Authors and Referees

  • Editorial process
  • Guide for authors
  • Submit an article
  • Guide for referees
  • Editors and Editorial Board
  • Contact editorial office

• Customer Service

  • Subscribe
  • Order sample copy
  • Purchase articles
  • Reprints and permissions
  • Contact NPG
  • Advertising