• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of rnaThe RNA SocietyeTOC AlertsSubscriptionsJournal HomeCSHL PressRNA
RNA. May 2005; 11(5): 567–577.
PMCID: PMC1370745

Co-evolution of tRNA 3′ trailer sequences with 3′ processing enzymes in bacteria


Maturation of the tRNA 3′ terminus is a complicated process in bacteria. Usually, it is initiated by an endonucleolytic cleavage carried out by RNase E and Z in different bacteria. In Escherichia coli, RNase E cleaves AU-rich sequences downstream of tRNA, producing processing intermediates with a few extra residues at the 3′ end; these are then removed by exoribonuclease trimming to generate the mature 3′ end. Here we show that essentially all E. coli tRNA precursors contain a potential RNase E cleavage site, the AU-rich sequence element (AUE), in the 3′ trailer. This suggests that RNase E cleavage and exonucleolytic trimming is a general pathway for tRNA maturation in this organism. Remarkably, the AUE immediately downstream of each tRNA is selectively conserved in bacteria having RNase E and tRNA-specific exoribonucleases, suggesting that this pathway for tRNA processing is also commonly used in these bacteria. Two types of RNase E-like proteins are identified in actinobacteria and the α-subdivision of proteobacteria. The tRNA 3′ proximal AUE is conserved in bacteria with only one type of E-like protein. Selective conservation of the AUE is usually not observed in bacteria without RNase E. These results demonstrate a novel example of co-evolution of RNA sequences with processing activities.

Keywords: tRNA, processing, RNase E, AU-rich, 3′trailer


Maturation of tRNA is a challenging problem in all living systems. In bacteria, the number of tRNA genes varies from ~30 to >100 in each genome (Lowe and Eddy 1997; Minagawa et al. 2004). tRNAs vary in their primary structures, but they conserve similar secondary and tertiary folds. It is common for bacterial tRNAs to be encoded by multicistronic operons. A multicistronic tRNA transcript may contain more than one tRNA, or rRNA and mRNA in addition to tRNA (Wawrousek and Hansen 1983; Komine et al. 1990; Sedlmeier et al. 1994). Some bacteria have genetically encoded 3′ terminal CCA for all tRNAs, whereas others only encode CCA for fractions of their total tRNAs (Deutscher 1984; Minagawa et al. 2004). In all cases, extra sequences present at both the 5′ and 3′ termini of a tRNA are removed during tRNA maturation (Deutscher 1984). Removal of the 5′ leader sequences from tRNA precursors is catalyzed by a single endonucleolytic reaction catalyzed by RNase P (Frank and Pace 1998; Altman and Kirsebom 1999), a ribozyme found in all bacteria with few exceptions (Willkomm et al. 2002; Li and Altman 2004).

Processing at the 3′ end is more complicated. Our knowledge of tRNA 3′ maturation is most advanced in Escherichia coli, an organism in which most, if not all, ribonuclease activities have been identified and studied in detail (Li and Deutscher 2004). A total of 86 tRNAs are identified in E. coli, with 3′ terminal CCA encoded for all of them. The current model suggests that the endoribonuclease RNase E carries out the first step of tRNA maturation. RNase E has been shown to cleave a number of primary tRNA transcripts at positions usually a few nucleotides downstream of the 3′ end of tRNAs (Li and Deutscher 2002; Ow and Kushner 2002). The intermediates with short 3′ ends are then cleaved efficiently by RNase P to generate the mature 5′ end. The few extra residues at the 3′ end are removed by exoribonucleases, mainly RNase T and PH (Reuven and Deutscher 1993; Li and Deutscher 1994 Li and Deutscher 1996), simultaneously with or after RNase P action (Li et al. 1998; Li and Deutscher 2002). Other exoribonucleases, including the tRNA-specific RNases D and BN, and the nonspecific RNase II, remove the 3′ extra residues less efficiently (Reuven and Deutscher 1993; Li and Deutscher 1994 Li and Deutscher 1996). In the absence of RNase T and PH activities, tRNA 3′ maturation is greatly impaired and cell growth slows down (Kelly and Deutscher 1992; Li and Deutscher 1994 Li and Deutscher 1996). In a few cases, RNase E cleavage occurs further downstream (Li and Deutscher 2002). The resulting longer 3′ trailer sequences are shortened by the nonspecific exoribonucleases, RNase II and polynucleotide phosphorylase (PNPase), prior to the actions of RNase P, T, and PH (Li and Deutscher 1994, 2002).

As described above, it is apparent that RNase E is a key enzyme for the maturation of tRNA precursors in E. coli. Its cleavage initiates maturation pathways. The position of cleavage determines the reactions to follow. RNase E has also been shown to cleave mRNA, antisense RNAs, and other noncoding small RNAs, and precursors of rRNA (for review, see Cohen and McDowall 1997; Kushner 2002; Li and Deutscher 2004). It has been shown that RNase E recognizes AU-rich sequences as cleavage sites for various substrates (for review, see Cohen and McDowall 1997). However, a sequence consensus has not been found (McDowall et al. 1994), and nearby secondary structures play controversial roles in determining cleavage sites (Bouvet and Belasco 1992; Mackie and Genereaux 1993; McDowall et al. 1995). These have prevented accurate predictions of RNase E cleavage sites in other E. coli tRNA precursors, which are important to decipher the global role of RNase E in tRNA maturation.

RNase E is an essential enzyme in E. coli (Ghora and Apirion 1978). Despite its importance in RNA metabolism, its distribution is limited to a relatively small group of bacteria, including actinobacteria and proteobacteria α, β, and γ subdivisions (Condon and Putzer 2002). Interestingly, a similar enzyme, RNase G, exists in E. coli (Li et al. 1999; Wachi et al. 1999). RNase G is smaller than RNase E and is homologous to the N-terminal catalytic domain of RNase E. RNase G is responsible for the maturation of the 5′ end of 16S rRNA and for the degradation of a number of mRNAs (for review, see Li and Deutscher 2004). RNase G, like RNase E, cleaves at AU-rich sites of RNA (Li et al. 1999; Jiang et al. 2000; Tock et al. 2000). However, the substrate spectrum of RNase G seems to be much narrower than that of RNase E (Lee et al. 2002). Preliminary studies suggest that RNase G may have a limited function in tRNA maturation in the absence of RNase E (J. Rubine and Z. Li, unpubl.). Overexpression of RNase G causes the formation of cytoplasmic axial filaments (Okada et al. 1994). Interruption of the rng gene encoding RNase G does not affect growth, but results in accumulation of a precursor to 16S rRNA (Li et al. 1999; Wachi et al. 1999). RNase G is present in a broader range of bacteria (Condon and Putzer 2002). Some bacteria have enzymes whose sequence features are similar to both RNase E and G. It has been shown that such an RNase E/G enzyme in Aquifex aeolicus cleaves tRNA precursors in vitro at AU-rich sequences (Willkomm et al. 2002).

Endonucleolytic activities that are responsible for eukaryotic tRNA 3′ processing (3′ tRNase or RNase Z) have been described in various organisms (Castano et al. 1985; Oommen et al. 1992; Nashimoto 1997; Kunzmann et al. 1998; Levinger et al. 1998). Recently, such activities were purified and cloned from plant and archaea, and are shown to be members of the ELAC1/2 family of proteins (Schiffer et al. 2002). RNase Z cleaves after the discriminator nucleotide of eukaryotic tRNA precursors that are generally CCA-less. The addition of CCA to CCA-less tRNAs is catalyzed by tRNA nucleotidyl transferase (Deutscher 1990; Schurer et al. 2001). RNase Z is widely present in bacteria (Condon and Putzer 2002). Recent work has demonstrated that RNase Z from Bacillus subtilis participates in the maturation of CCA-less tRNAs in the same way as its archaeal counterpart (Pellegrini et al. 2003). This reaction is less efficient if a long 5′ leader sequence has not been previously removed. The presence of a residue of C after the discriminator inhibits the activity of B. subtilis RNase Z, suggesting that this enzyme is probably not responsible for the maturation of tRNAs with encoded CCA in this organism. In contrast, RNase Z (3′ tRNase) in Thermotoga maritima works very differently from that in B. subtilis. The tRNase of T. maritima cleaves after the encoded CCA sequence, generating mature 3′ termini directly (Minagawa et al. 2004). In this case, the terminal CA residues are required for cleavage at the correct site. Interestingly, E. coli also harbors a RNase Z homolog encoded by the elaC gene. The E. coli RNase Z is nonessential (Schilling et al. 2004). Its role in the processing of E. coli tRNAs remains elusive, although the enzyme cleaves human pre-tRNA after the discriminator residue in vitro (Minagawa et al. 2004). Various mechanisms of tRNA processing are described in Figure 1 [triangle].

Maturation of tRNA 3′ termini in bacteria. Three models are presented for the 3′ maturation of (A) tRNA-CCA in E. coli (Li and Deutscher 1996, 2002), (B) CCA-less tRNA in B. subtilis (Pellegrini et al. 2003), and (C) tRNA-CCA in T. maritima ...

To better understand the mechanisms that are used by various bacteria for the 3′ processing of tRNA, we have analyzed completely sequenced bacterial genomes for tRNA 3′ trailer sequences and the presence of tRNA-processing enzymes. Here we show that RNase E is present in organisms that also have the processing exoribonucleases, and that the potential RNase E cleavage sites, AU-rich elements, are selectively enriched immediately downstream of tRNAs in only these organisms. The results suggest co-evolution of tRNA 3′ trailer sequences with 3′ processing activities.


RNase E cleaves at AU-rich sequences that are abundant in tRNA 3′ trailers in E. coli

It has been shown previously that in E. coli, RNase E cleaves several primary tRNA transcripts. This cleavage is the initial processing step for those tRNAs. As shown in Figure 2 [triangle], it is apparent that the sequences surrounding RNase E cleavage sites are rich in adenines and uridines. This observation is consistent with known RNase E cleavage sites on other substrates that are described above in the Introduction section (for review, see Cohen and McDowall 1997; Kushner 2002; Li and Deutscher 2004). In addition, RNase E cleaves primarily at positions a few nucleotides downstream of the 3′ terminus of tRNAs where AU-rich sequences are present. Cleavages at AU-rich sequences in more distal locations have been also observed. Therefore, if the 3′ trailer sequence of a tRNA precursor contains AU-rich sequence(s), it is probably processed by RNase E. This prompted us to study AU contents and AU-rich sequences in the 3′ trailers of other tRNAs in E. coli and tRNAs in other bacteria.

RNase E cleavage sites of E. coli tRNA transcripts. Cleavage sites on two tRNA transcripts were determined by Northern blotting and primer extension (Li and Deutscher 2002). Some well-defined cleavage sites are shown by “E” with vertical ...

In all, 86 tRNAs are encoded in E. coli K12. We extracted their 3′ trailer sequences as follows. The sequences start immediately after the encoded CCA terminus, and end at a terminator or the beginning of a known downstream RNA. If the trailers are very long, only the first 100 nt are recorded.

The frequencies of the four nucleotides were counted in the first 20 positions downstream of CCA, a region in which most known RNase E cleavage sites reside. As shown in Figure 3 [triangle], adenosines and uridines dominate in the first 10 positions. The percentage of uridine is very high at positions 2–6. Adenosine is high in the first two positions and from positions 7 to 10. The percentage of guanosine is low in the first 10 positions, with <10% in each of the first five positions. Cytosine is low at positions 2–4. The percentages of the four nucleotides become much closer between positions 11 and 20. Within the first 10 nucleotides after the tRNA, the GC percentage is only 29%, whereas that of the whole genome is 51%. This analysis clearly demonstrates a high AU content in the 3′ trailers immediately downstream of E. coli tRNAs. Note that within the 20 nucleotides in any of the sequences, no uridine was from a Rho-independent terminator that usually contains a run of U residues. Therefore, such uneven distribution of nucleotides is not the consequence of transcription termination. Rather, it is most likely that the distribution is conserved for the sake of tRNA processing. The high AU content within the first 10 nucleotides is consistent with the notion that RNase E may cleave most tRNA precursors in this region.

Nucleotide frequency of the 3′ trailers of E. coli tRNA. Nucleotide percentages at each of 20 positions were calculated by dividing the number of occurrences of a nucleotide by the total number counted at that position.

We further analyzed if the potential RNase E cleavages sites, AU-rich elements (AUEs), are present in the 3′ trailers of the majority of tRNAs in E. coli. Based on known RNase E cleavage sites in various RNA substrates, we arbitrarily defined the minimum AUE as a stretch of 5 nt containing a minimum of four A or U bases. The length of AUEs can be extended in both directions when more A/U is present, either continuously or separated by a single G/C. The presence of two G/Cs in any window of 3 nt will stop the extension. The results are summarized in Table 1 [triangle].

AU-rich elements in the 3′ trailer sequences of E. coli tRNAs

AUEs can be identified in the 3′ trailers of 82 out of 86 tRNAs. The four AUE-less pre-tRNAs contain only two to four nucleotides in the 3′ trailers, and their 3′ maturation presumably bypasses RNase E cleavage. Each of the four AUE-less pre-tRNAs is followed by another tRNA. Maturation of the 5′ termini of the downstream tRNAs by RNase P generates precursors of the AUE-less tRNAs with only two to four extra residues at the 3′ termini, which can be removed by solely exonucleolytic trimming reactions.

Most tRNA precursors (73 out of 82) contain at least one AUE within the first 10 nt of the 3′ trailers. It should be noted that the majority of AUEs are quite long and may harbor more than one cleavage site for RNase E. It is common that multiple AUEs are present in the long 3′ trailer sequences. Altogether, the results clearly demonstrate that potential RNase E cleavage sites are present in the 3′ trailer sequences of almost all tRNA precursors in E. coli, and are highly abundant in the region close to the 3′ end. This suggests that tRNA-processing initiated by RNase E cleavage is a common mechanism in this organism.

Distribution of AU-rich elements in tRNA 3′ trailers in 51 bacterial genomes

After observing the high occurrence of AUEs in the 3′ trailers of E. coli tRNAs, we set up a search of the AUEs in other bacteria. The purpose is to obtain information regarding AUE abundance in 3′ trailers and to understand the relationship, if any, between the presence of AUE and tRNA 3′ maturation in other bacteria.

To do this, we selected 51 genomes that are well annotated and have coordinates for all tRNA genes as well as downstream genes. Complete genome sequences and annotation data (GenBank format) for all bacteria were obtained from NCBI (http://www.ncbi.nlm.nih.gov/genomes/static/eub_g.html). Using procedures described above, tRNA 3′ trailer sequences were extracted and AUEs were searched within the 3′ trailers for all bacteria. Annotation errors present in some genomes were fixed manually. Perl scripts were developed to automate the search.

In order to compare the abundance of AUEs among all genomes, we have developed a numeric system to record the number of AUEs in different regions of the 3′ trailers. We divided the 100-nt 3′ trailer sequences into 10 regions, each containing 10 nt. The number of AUEs in each of the 10 regions is counted as described below. This enabled us to compare AUE abundance among different regions of the 3′ trailers, as well as across genomes.

The presence of an AUE in a region is defined as follows. (1) At least 3 nt of the AUE occupies this region; (2) one AUE can be counted in more than one region if it is long enough; and (3) one 3′ trailer sequence may contain 0–2 AUEs in a region. If taking the last encoded residue of a tRNA (the 3′ A or discriminator) as position 0, any AUE starting from -1 to 8 or stopping before 12 is counted in region 1–10; an AUE starting from 9 to 18 or stopping before 22 is counted in region 11–20; and so on.

Note that the lengths of 3′ trailers vary. It is necessary to calculate AUE abundance in each region according to the number of 3′ trailer sequences present in that region. The number of 3′ trailers in a particular region is determined by the presence of at least 3 nt of the trailer sequence in that region. For each bacterial genome, AUE abundance in each region (denoted as R1, R2, …, R10) is calculated as the number of AUEs divided by the number of 3′ trailers in that region. The results are presented in Table 2 [triangle].

AUE abundance in tRNA 3′ trailer sequences in 51 bacteria

Interestingly, AUE abundance varies greatly among genomes (Table 2 [triangle]). AUE abundance across the 10 regions is high in some genomes but low in some others. This may reflect low or high GC percentages in those genome sequences (see below). Many genomes have a higher AUE abundance in the first region than in other regions of the 3′ trailers. The AUEs, especially those selectively conserved close to the 3′ end of tRNAs, probably are related to tRNA 3′ processing. This point is further analyzed in the following sections.

Distribution of tRNA 3′-processing ribonucleases in 51 bacterial genomes

The differences in AUE abundance in the bacterial genomes may be associated with differences in tRNA 3′ processing activities. We carried out a search for the activities potentially involved in the 3′ processing of tRNA in all the genomes described above. In addition to RNase E/G and Z, we also searched for the exoribonucleases specific for tRNA 3′ maturation (namely, RNase T, PH, and D). RNase BN was not included in this search because of the uncertainty in the identification of the gene encoding this enzyme (M.P. Deutscher, pers. comm.). Searches for the nonspecific exoribonucleases RNase II and PNPase were carried out to compare with the distribution of tRNA-specific enzymes. RNase R, a nonspecific exoribonuclease whose function has not yet been related to tRNA maturation, is also included in this analysis because of its similarity to RNase II. In fact, RNase II, R, and proteins with similar sequences are grouped into the RNR family (Zuo and Deutscher 2001).

The presence of the aforementioned ribonucleases in each organism was determined primarily by genomic BLAST searches using the E. coli proteins as query sequences (http://www.ncbi.nlm.nih.gov/sutils/genom_table.cgi). Sequences of ElaC from E. coli and YqjK from B. subtilis were both used as queries in the RNase Z search. Membership of clusters of orthologous groups is also used in the search for RNase E/G (COG1530). We found that the BLAST result is inclusive of all COG members of RNase E/G, which indicates that genomic BLAST is a good tool for finding ribonucleases in bacteria. Since remote homologs of ribonucleases may or may not actually function as RNases, we did not include BLAST hits that have E-values >e-40. For the same reason, more sophisticated searches such as psiBLAST were not used. The RNases found in our search are essentially the ones that have already been annotated as the same enzymes in GenBank records. Ribonucleases found in various bacterial genomes are listed in Table 3 [triangle]. A similar search for most of the above enzymes was reported previously for some of the bacteria in our list (Minagawa et al. 2004). Our results agree well with the results of that search.

Ribonucleases present in the 51 bacteria

Based on the presence of RNase E or G, the bacteria can be divided into four groups (Table 3 [triangle]). Group I contains both RNase E and G. Interestingly, bacteria in this group also have the three tRNA-specific exoribonucleases RNase T, PH, and D, with a few exceptions in which one or two of the three enzymes exist. Note that RNase T is exclusively found in this group of organisms. Bacteria of this group except Caulobacter crescentus belong to proteobacteria β-and γ-subdivisions. C. crescentus is the only member from the α-subdivision. Bacteria in Group II each have one RNase E/G-like enzyme that is close to RNase E in size and contains the RNase E/G domain. This domain contains an S1 RNA-binding domain, confers catalytic activity, and encompasses the N-terminal half of the E. coli RNase E and almost the full length of the E. coli RNase G (Taraseviciene et al 1995; McDowall and Cohen 1996; Bycroft et al. 1997). Most of these bacteria also have RNase PH and D. This group of bacteria belongs to the proteobacteria α-subdivision and actinobacteria. Group III bacteria clearly have RNase G but not E, since the enzymes are much closer to RNase G in size and sequence. Group II bacteria are present in a wider range of eubacterial families, including actinobacteria, chlamydiae, clostridia, aquifecales, thermotogales, and the Chlorobium–Flavobacterium group. Group IV has neither RNase E nor G. Bacteria in Group IV are widely present in actinobacteria, the proteobacteria epsilon-subdivision, spirochaetales, firmicutes (bacillales mollicutes lactobacillales), and the Thermus–Deinococcus group. Very few bacteria in the third and fourth groups have RNase PH or D. In contrast to the tRNA-specific exoribonucleases, the nonspecific exoribonucleases RNase II, R, and PNPase are present indistinguishably among all groups. PNPase is present in all bacteria in this list except the two mycoplasmas. RNase R is present in most of the bacteria, and is much more frequently found than RNase II, in agreement with previous observations (Zuo and Deutscher 2001). RNase Z is present in most bacteria in Table 3 [triangle], and its presence does not seem to be different among the four groups.

AUE abundance in tRNA 3′ trailers among genomes with different percentages of GC

As described above, the abundance of AU-rich elements in tRNA 3′ trailers varies among bacterial genomes. Analysis of nucleotide frequencies in E. coli tRNA 3′ trailers suggested that adenine and uridine are most abundant within the first 10 nt downstream of tRNA. AUEs are present in the first 10 nt of the 3′ trailers in most E. coli tRNA precursors. Therefore, we focused our analysis on the comparison of AUE abundance in the first 10-nt region (R1) and the average AUE abundance of the other nine regions (Ř2–10) in the 3′ trailers.

Through random variation, AUE is expected to exist at a higher abundance in genomes with a lower percentage of G and C (GC%). We plotted AUE abundance (R1 and Ř2–10) in parallel to GC percentage in order to discriminate the effect of GC% and selective conservation on AUE abundance. In Figure 4 [triangle], bacteria are presented on the X-axis with increasing GC%. R1 and Ř2–10 were plotted separately. Ř2–10 decreases as GC% increases. Although the curve of Ř2–10 is not smooth, an inverse correlation is clearly present between Ř2–10 and GC percentage. The distribution of Ř 2–10 is consistent with the notion that AUEs in the region 11–100 nt in the 3′ trailers occur in a relatively random manner. This finding suggests that their relationship to tRNA maturation, if any, is weak.

The abundance of AU-rich elements (AUEs) in 3′ trailers of bacterial tRNA as an effect of GC percentage. AUE abundances in the first 10-nt region after tRNA (R1) and the average of abundance in the nine regions downstream (Ř 2–10 ...

In contrast, R1 shows a very interesting pattern when GC% increases. At low GC percentages, R1 is not distinguishable from Ř 2–10. However, the two curves separate as GC% increases. At higher GC%, R1 is lower in some genomes, but stays high in many others. In a few genomes with >60% GC, R1 is as high as that in genomes with <40% GC. This indicates that AUE in the first 10 nt is strongly conserved in tRNA precursors in certain high-GC bacteria. This behavior is not related to transcription termination since no U residues in the first 10 nt are from the terminator. Instead, the conserved high abundance of AUE immediately after tRNA is most likely related to tRNA processing.

AUEs are highly conserved immediately downstream of tRNA in organisms with RNase E and tRNA-maturation exoribonucleases

To further investigate the relationship between AUE abundance and tRNA processing, we have studied AUE abundance within the four groups of bacteria that have different RNase profiles (Table 3 [triangle]). AUE abundance in the first region (R1) and the average of the next nine regions (Ř 2–10) was plotted in the order of increasing GC percentages for each group of bacteria. As shown in Figure 5 [triangle], the difference of R1 and Ř 2–10 now is shown almost exclusively in Group I and II bacteria having RNase E or E-like enzymes.

tRNA 3′ AUE abundance in bacterial groups harboring different RNases. Bacteria are divided into four groups based on the presence or absence of different RNase E/G proteins, as demonstrated in Table 3 [triangle]. AUE abundances are plotted the ...

In Group I bacteria, R1 stays high across the group, whereas Ř 2–10 drops quickly as GC% increases. The high-GC bacterium C. crescentus CB15 is exceptional in that R1 is lower but is still much higher than Ř 2–10. The result indicates that AUEs are selectively conserved close to the 3′ end of tRNA in all members of this group. This is consistent with the idea that RNase E may play a major role in tRNA 3′ processing in these bacteria.

Group II bacteria show a mixed pattern. As GC% increases, Ř 2–10 drops in the same way as in Group I. R1 stays high in half of the bacteria. However, in other members of this group, R1 decreases and is not different from Ř 2–10. Therefore, it appears that only some members of Group II resemble Group I bacteria.

In order to understand the mixed pattern of R1 in Group II bacteria, we have compared RNase E and E-like enzymes in Group I and II by sequence alignment (Fig. 6 [triangle]). Interestingly, RNase E proteins in Group I all have similar features, i.e., close in sizes and very similar in sequences to the N-terminal half of E. coli RNase E. The RNase E-like enzymes in Group II, however, are divided into two types. Type I has an insert of varying lengths within the N-terminal catalytic domain, interrupting the S1 RNA-binding domain. Type II has a long N-terminal sequence in front of the continuous catalytic domain. Remarkably, all the Group II bacteria with high R1 have Type I enzymes, whereas the bacteria with low R1 have Type II enzymes. This strongly suggests that Type I enzymes function like RNase E in tRNA processing, depending on the selectively conserved AUE immediately downstream of tRNA. The role of Type II enzymes in tRNA 3′ maturation, if any, may not be related to AU-rich elements. It should be noted that all bacteria with Type I enzymes belong to the α-subdivision of proteobacteria, whereas those with Type II enzyme are members of the actinobacteria.

Alignment of RNase E and E-like proteins. The alignment was constructed based on BLAST pairwise alignments of each protein with the E. coli RNase E with E-values shown. Black bars represent highly homologous regions. Gray bars are regions showing no homology ...

Group III and IV bacteria show essentially the same pattern: both R1 and Ř 2–10 decrease with increasing GC%. This indicates that AUE is generally not selectively conserved in Group III and IV bacteria. A Group III bacterium, Aquifex aeolicus VF5, shows a much higher R1 than Ř 2–10, suggesting a possible selective pressure for 3′ proximal AUE in this organism. Alternatively, it is likely that its Ř 2–10 is unusually lower than those with similar GC% (Fig. 5 [triangle]) for unknown reasons. Interestingly, an activity has been demonstrated in A. aeolicus that cleaves AU-rich sequences in tRNA 3′ trailers in an in vitro processing reaction (Willkomm et al. 2002). It remains to be determined if RNase G confers this activity, and if it plays a role in tRNA 3′ maturation in this bacterium.

It should be noted that GC percentages also segregate among the bacterial groups. The majority of high GC bacteria fall into Group I and II. Only one organism in each of Groups I and II has a GC percentage <40%. In contrast, most low-GC bacteria belong to Groups II and IV. For the few high GC% organisms in these groups, both R1 and Ř 2–10 drop sharply as GC increases. Ř 2–10 drops steadily above ~45% GC (Fig. 5 [triangle]), indicating that selective pressure is required to keep high AUE abundance in bacteria with GC of >45%. R1 is high in all the low-GC bacteria. It remains to be determined if the high R1 in Groups III and IV is important for tRNA 3′ processing by enzyme(s) other than RNase E.


Our analysis has shown that RNase E and the Type I E-like proteins are present in the proteobacteria α-, β-, and γ-subdivisions. These bacteria also have an abundant AU-rich element (AUE) identified immediately downstream of their tRNA-coding sequences. The abundance of AUE in the distal regions of 3′ trailers becomes much lower when GC% increases. In contrast, AUE present within 10 nt downstream of tRNA is selectively conserved even in genomes of high GC%. Selective conservation of the proximal AUE is almost exclusively observed in the aforementioned bacteria. It has been previously shown that in E. coli, RNase E is responsible for the maturation of a number of tRNAs by cleaving at AU-rich sequences in the 3′ trailers (Li and Deutscher 2002; Ow and Kushner 2002). RNase E cleavages within the first 10 nt (R1) may be responsible for the processing of most tRNAs in E. coli since AUEs are highly present in this region. Cleavages also happen at downstream AUEs (Fig. 2 [triangle]), and can be important for tRNA processing if no cleavage occurs in R1 for a pre-tRNA. The general importance of AUEs in R1 for tRNA processing is further demonstrated by its selective conservation in bacteria having RNase E. Together, the results demonstrate that RNase E co-evolves with its potential cleavage sites, i.e., AUEs immediately downstream of the tRNA. This work strongly suggests that RNase E is generally important for tRNA 3′ processing in the above bacteria.

It should be noted that the definition of AUE is quite arbitrary since RNase E cleavage sites do not have a known sequence consensus. However, we find that the AUE designation is useful in finding sequence features for RNase E-dependent tRNA processing. A closer examination of the content of AUEs may be helpful to better decipher RNase E cleavage sites. Furthermore, neighboring structural features have been shown to be important for RNase E cleavage (for review, see Li and Deutscher 2004). It remains to be elucidated if tRNA itself and any other sequence or structures surrounding AUEs play roles for determination of RNase E cleavage sites in the 3′ trailers. The conservation of AUEs proximal to tRNA 3′ ends adds another example to the list of conserved sequence elements important for RNA processing. Conservation of nucleotides in the 3′ trailer that is important for tRNA processing was also described in mammalian systems (Nashimoto 1997).

Interestingly, as described above, the tRNA-specific exoribonucleases also co-evolve with RNase E. It is therefore likely that in bacteria having these enzymes, RNase E usually cleaves a few nucleotides downstream of tRNA, and that the extra residues are removed by exoribonucleases. Therefore, RNase E-containing bacteria may use a pathway for tRNA processing very similar to that in E. coli (Fig. 1A [triangle]).

We are prompted to ask if the sequences downstream of tRNA also affect exonucleolytic trimming reactions. RNase T, one of the major tRNA maturation exoribonucleases, has been shown to preferentially remove 3′ terminal residues in the order of A ≈ G > U >> C (Zuo and Deutscher 2002). It has also been demonstrated that in vivo, RNase PH is the most efficient enzyme for removal of C at the 3′ +2 position of pre-tRNA1Tyr (Li and Deutscher 1996). U is most abundant at positions +2 to +6 of E. coli tRNAs (Fig. 3 [triangle]), and can be removed efficiently by RNase T or PH under normal growth conditions (Li and Deutscher 1996). The base specificity of other tRNA-processing exoribonucleases has not yet been defined. It is likely that no individual exoribo-nuclease can trim all the extra residues efficiently enough to support optimal cell growth. On the other hand, RNase T’s ability to efficiently remove A residues results in the loss of the 3′ terminal A in a fraction of mature tRNAs, and the A is added back by tRNA nucleotidyltransferase (Deutscher 1990). Therefore, multiple exoribonucleases and tRNA nucleotidyltransferase act together for generating mature 3′ ends of tRNA in E. coli (Reuven and Deutscher 1993; Li and Deutscher 1994 Li and Deutscher 1996), and they may also be required in other bacteria using this pathway.

RNase E-like enzymes were identified in our search. These enzymes resemble RNase E more than G. However, sequence alignments show two different types of arrangement of protein segments among the E-like enzymes: Type I with inserts in the catalytic domain and Type II with a large domain at the N terminus (Fig. 6 [triangle]). Since AUEs are abundant immediately downstream of tRNA in bacteria having Type I but not Type II proteins, Type I proteins may function in tRNA processing similarly to RNase E. This finding also predicts that the two types of proteins may have different activities. Such an alignment provides a way of distinguishing RNase E/G-like proteins.

Despite the importance of RNase E in tRNA maturation in E. coli, this enzyme, and the Type I E-like proteins, are only present in three subdivisions of proteobacteria. RNase Z may function in tRNA processing in many of the other bacteria, such as B. subtilis and T. maritima (Fig. 1 [triangle]). However, the substrate specificity of RNase Z (3′ tRNase) in B. subtilis has been shown to be different from that in T. maritima (Pellegrini et al. 2003; Minagawa et al. 2004). In addition, RNase Z does not seem to be important for tRNA maturation in E. coli (Minagawa et al. 2004; Schilling et al. 2004). Therefore, it is difficult to predict the role of RNase Z in tRNA processing in other bacteria. RNase G and Type II RNase E-like proteins are found in some bacteria without RNase E. It remains to be determined if these enzymes play roles in tRNA processing in these organisms. Furthermore, the Group IV bacteria Mycoplasma pulmonis, Mycoplasma pneumoniae, and Helicobacter pylori do not have any of the endoribonucleases discussed above. These bacteria also lack the tRNA-specific exoribonucleases. It is likely that tRNA 3′ processing in these bacteria uses a mechanism yet different from any of the ones described in Figure 1 [triangle]. For instance, the 3′ trailers may be removed solely by the previously known nonspecific exoribonucleases (RNase R, II, or PNPase). Finally, it will be interesting to learn if there are sequence features important for tRNA 3′ processing by the other enzymes.


The authors are grateful to Keith Brew and Herbert Weissbach for helpful comments on the manuscript. This work was supported in part by a Florida Atlantic University New Project Development Award to Z.L.


Article published online ahead of print. Article and publication date are at http://www.rnajournal.org/cgi/doi/10.1261/rna.7287505.


  • Altman, S. and Kirsebom, L.A. 1999. Ribonuclease P. In The RNA world (eds. R.F. Gesteland et al.), pp. 351–380. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
  • Bouvet, P. and Belasco, J.G. 1992. Control of RNase E-mediated RNA degradation by 5′-terminal base pairing in E. coli. Nature 360: 488–491. [PubMed]
  • Bycroft, M., Hubbard, T.J., Proctor, M., Freund, S.M., and Murzin, A.G. 1997. The solution structure of the S1 RNA binding domain: A member of an ancient nucleic acid-binding fold. Cell 88: 235–242. [PubMed]
  • Castano, J.G., Tobian, J.A., and Zasloff, M. 1985. Purification and characterization of an endonuclease from Xenopus laevis ovaries which accurately processes the 3′ terminus of human pre-tRNA- Meti (3′ pre-tRNase). J. Biol. Chem. 260: 9002–9008. [PubMed]
  • Cohen, S.N. and McDowall, K.J. 1997. RNase E: Still a wonderfully mysterious enzyme. Mol. Microbiol. 23: 1099–1106. [PubMed]
  • Condon, C. and Putzer, H. 2002. The phylogenetic distribution of bacterial ribonucleases. Nucleic Acids Res. 30: 5339–5346. [PMC free article] [PubMed]
  • Deutscher, M.P. 1984. Processing of tRNA in prokaryotes and eukaryotes. CRC Crit. Rev. Biochem. 17: 45–71. [PubMed]
  • ———. 1990. Ribonucleases, tRNA nucleotidyltransferase, and the 3′ processing of tRNA. Prog. Nucleic Acid Res. Mol. Biol. 39: 209–240. [PubMed]
  • Diwa, A.A., Jiang, X., Schapira, M., and Belasco, J.G. 2002. Two distinct regions on the surface of an RNA-binding domain are crucial for RNase E function. Mol. Microbiol. 46: 959–969. [PubMed]
  • Frank, D.N. and Pace, N.R. 1998. Ribonuclease P: Unity and diversity in a tRNA processing ribozyme. Annu. Rev. Biochem. 67: 153–180. [PubMed]
  • Ghora, B.K. and Apirion, D. 1978. Structural analysis and in vitro processing to p5 rRNA of a 9S RNA molecule isolated from an rne mutant of E. coli. Cell 15: 1055–1066. [PubMed]
  • Jiang, X., Diwa, A., and Belasco, J.G. 2000. Regions of RNase E important for 5′-end-dependent RNA cleavage and autoregulated synthesis. J. Bacteriol. 182: 2468–2475. [PMC free article] [PubMed]
  • Kelly, K.O. and Deutscher, M.P. 1992. The presence of only one of five exoribonucleases is sufficient to support the growth of Escherichia coli. J. Bacteriol. 174: 6682–6684. [PMC free article] [PubMed]
  • Komine, Y., Adachi, T., Inokuchi, H., and Ozeki, H. 1990. Genomic organization and physical mapping of the transfer RNA genes in Escherichia coli K12. J. Mol. Biol. 212: 579–598. [PubMed]
  • Kunzmann, A., Brennicke, A., and Marchfelder, A. 1998. 5′ end maturation and RNA editing have to precede tRNA 3′ processing in plant mitochondria. Proc. Natl. Acad. Sci. 95: 108–113. [PMC free article] [PubMed]
  • Kushner, S.R. 2002. mRNA decay in Escherichia coli comes of age. J. Bacteriol. 184: 4658–4665. [PMC free article] [PubMed]
  • Lee, K., Bernstein, J.A., and Cohen, S.N. 2002. RNase G complementation of rne null mutation identifies functional interrelationships with RNase E in Escherichia coli. Mol. Microbiol. 43: 1445–1456. Erratum. Mol. Microbiol. 46: 295. [PubMed]
  • Levinger, L., Bourne, R., Kolla, S., Cylin, E., Russell, K., Wang, X., and Mohan, A. 1998. Matrices of paired substitutions show the effects of tRNA D/T loop sequence on Drosophila RNase P and 3′-tRNase processing. J. Biol. Chem. 273: 1015–1025. [PubMed]
  • Li, Y. and Altman, S. 2004. In search of RNase P RNA from microbial genomes. RNA 10: 1533–1540. [PMC free article] [PubMed]
  • Li, Z. and Deutscher, M.P. 1994. The role of individual exoribonucleases in processing at the 3′ end of Escherichia coli tRNA precursors. J. Biol. Chem. 269: 6064–6071. [PubMed]
  • ———. 1996. Maturation pathways for E. coli tRNA precursors: A random multienzyme process in vivo. Cell 86: 503–512. [PubMed]
  • ———. 2002. RNase E plays an essential role in the maturation of Escherichia coli tRNA precursors. RNA 8: 97–109. [PMC free article] [PubMed]
  • ———. 2004. Exoribonucleases and endoribonucleases. In Escherichia coli and Salmonella: Cellular and molecular biology, the Eco-Sal on-line edition (eds. R. Curtiss et al.), pp. 1–26. American Society for Microbiology, Washington, DC.
  • Li, Z., Pandit, S., and Deutscher, M.P. 1998. 3′ exoribonucleolytic trimming is a common feature of the maturation of small, stable RNAs in Escherichia coli. Proc. Natl. Acad. Sci. 95: 2856–2861. [PMC free article] [PubMed]
  • ———. 1999. RNase G (CafA protein) and RNase E are both required for the 5′ maturation of 16S ribosomal RNA. EMBO J. 18: 2878–2885. [PMC free article] [PubMed]
  • Lowe, T.M. and Eddy, S.R. 1997. tRNAscan-SE: A program for improved detection of transfer RNA genes in genomic sequence. Nucleic Acids Res. 25: 955–964. [PMC free article] [PubMed]
  • Mackie, G.A. and Genereaux, J.L. 1993. The role of RNA structure in determining RNase E-dependent cleavage sites in the mRNA for ribosomal protein S20 in vitro. J. Mol. Biol. 234: 998–1012. [PubMed]
  • McDowall, K.J. and Cohen, S.N. 1996. The N-terminal domain of the rne gene product has RNase E activity and is non-overlapping with the arginine-rich RNA-binding site. J. Mol. Biol. 255: 349–355. [PubMed]
  • McDowall, K.J., Lin-Chao, S., and Cohen, S.N. 1994. A+U content rather than a particular nucleotide order determines the specificity of RNase E cleavage. J. Biol. Chem. 269: 10790–10796. [PubMed]
  • McDowall, K.J., Kaberdin, V.R., Wu, S.W., Cohen, S.N., and Lin-Chao, S. 1995. Site-specific RNase E cleavage of oligonucleotides and inhibition by stem–loops. Nature 374: 287–290. [PubMed]
  • Minagawa, A., Takaku, H., Takagi, M., and Nashimoto, M. 2004. A novel endonucleolytic mechanism to generate the CCA 3′ termini of tRNA molecules in Thermotoga maritime. J. Biol. Chem. 279: 15688–15697. [PubMed]
  • Nashimoto, M. 1997. Distribution of both lengths and 5′ terminal nucleotides of mammalian pre-tRNA 3′ trailers reflects properties of 3′ processing endoribonuclease. Nucleic Acids Res. 25: 1148–1154. [PMC free article] [PubMed]
  • Okada, Y., Wachi, M., Hirata, A., Suzuki, K., Nagai, K., and Matsuhashi, M. 1994. Cytoplasmic axial filaments in Escherichia coli cells: Possible function in the mechanism of chromosome segregation and cell division. J. Bacteriol. 176: 917–922. [PMC free article] [PubMed]
  • Oommen, A., Li, X.Q., and Gegenheimer, P. 1992. Cleavage specificity of chloroplast and nuclear tRNA 3′-processing nucleases. Mol. Cell. Biol. 12: 865–875. [PMC free article] [PubMed]
  • Ow, M.C. and Kushner, S.R. 2002. Initiation of tRNA maturation by RNase E is essential for cell viability in E. coli. Genes & Dev. 16: 1102–1115. [PMC free article] [PubMed]
  • Pellegrini, O., Nezzar, J., Marchfelder, A., Putzer, H., and Condon, C. 2003. Endonucleolytic processing of CCA-less tRNA precursors by RNase Z in Bacillus subtilis. EMBO J. 22: 4534–4543. [PMC free article] [PubMed]
  • Reuven, N.B. and Deutscher, M.P. 1993. Multiple exoribonucleases are required for the 3′ processing of Escherichia coli tRNA precursors in vivo. FASEB J. 7: 143–148. [PubMed]
  • Schiffer, S., Rosch, S., and Marchfelder, A. 2002. Assigning a function to a conserved group of proteins: The tRNA 3′-processing enzymes. EMBO J. 21: 2769–2777. [PMC free article] [PubMed]
  • Schilling, O., Ruggeberg, S., Vogel, A., Rittner, N., Weichert, S., Schmidt, S., Doig, S., Franz, T., Benes, V., Andrews, S.C., et al. 2004. Characterization of an Escherichia coli elaC deletion mutant. Biochem. Biophys. Res. Commun. 320: 1365–1373. [PubMed]
  • Schurer, H., Schiffer, S., Marchfelder, A., and Morl, M. 2001. This is the end: Processing, editing and repair at the tRNA 3′-terminus. Biol. Chem. 382: 1147–1156. [PubMed]
  • Sedlmeier, R., Werner, T., Kieser, H.M., Hopwood, D.A., and Schmieger, H. 1994. tRNA genes of Streptomyces lividans: New sequences and comparison of structure and organization with those of other bacteria. J. Bacteriol. 176: 5550–5553. [PMC free article] [PubMed]
  • Taraseviciene, L., Björk, G.R., and Uhlin, B.E. 1995. Evidence for an RNA binding region in the Escherichia coli processing endoribonuclease RNase E. J. Biol. Chem. 270: 26391–26398. [PubMed]
  • Tock, M.R., Walsh, A.P., Carroll, G., and McDowall, K.J. 2000. The CafA protein required for the 5′-maturation of 16 S rRNA is a 5′-end-dependent ribonuclease that has context-dependent broad sequence specificity. J. Biol. Chem. 275: 8726–8732. [PubMed]
  • Wachi, M., Umitsuki, G., Shimizu, M., Takada, A., and Nagai, K. 1999. Escherichia coli cafA gene encodes a novel RNase, designated as RNase G, involved in processing of the 5′ end of 16S rRNA. Biochem. Biophys. Res. Commun. 259: 483–488. [PubMed]
  • Wawrousek, E.F. and Hansen, J.N. 1983. Structure and organization of a cluster of sic tRNA genes in the space between tandem ribosomal RNA gene sets in Bacillus subtilis. J. Biol. Chem. 258: 291–298. [PubMed]
  • Willkomm, D.K., Feltens, R., and Hartmann, R.K. 2002. tRNA maturation in Aquifex aeolicus. Biochimie 84: 713–722. [PubMed]
  • Zuo, Y. and Deutscher, M.P. 2001. Exoribonuclease superfamilies: Structural analysis and phylogenetic distribution. Nucleic Acids Res. 29: 1017–1026. [PMC free article] [PubMed]
  • ———. 2002. The physiological role of RNase T can be explained by its unusual substrate specificity. J. Biol. Chem. 277: 29654–29661. [PubMed]

Articles from RNA are provided here courtesy of The RNA Society
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...