A proline tRNA(CGG) gene encompassing the attachment site of temperate phage 16-3 is functional and convertible to suppressor tRNA


  • Béla Blaha,

    1. Institute of Genetics, Agricultural Biotechnology Center, Gödöllõ, Szent-Györgyi A. 4., H-2100, Hungary.
    2. Department of Genetics, Faculty of Science, Eötvös Loránd University and Research Group for Molecular Genetics of Hungarian Academy of Sciences, Budapest, Pázmány P. 1/C., H-1117, Hungary.
    Search for more papers by this author
  • Szabolcs Semsey,

    1. Institute of Genetics, Agricultural Biotechnology Center, Gödöllõ, Szent-Györgyi A. 4., H-2100, Hungary.
    2. Laboratory of Molecular Biology, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892-4264, USA.
    Search for more papers by this author
  • Szilamér Ferenczi,

    1. Institute of Genetics, Agricultural Biotechnology Center, Gödöllõ, Szent-Györgyi A. 4., H-2100, Hungary.
    Search for more papers by this author
  • Zsolt Csiszovszki,

    1. Institute of Genetics, Agricultural Biotechnology Center, Gödöllõ, Szent-Györgyi A. 4., H-2100, Hungary.
    2. Department of Genetics, Faculty of Science, Eötvös Loránd University and Research Group for Molecular Genetics of Hungarian Academy of Sciences, Budapest, Pázmány P. 1/C., H-1117, Hungary.
    Search for more papers by this author
  • Péter P. Papp,

    1. Institute of Genetics, Agricultural Biotechnology Center, Gödöllõ, Szent-Györgyi A. 4., H-2100, Hungary.
    Search for more papers by this author
  • László Orosz

    Corresponding author
    1. Institute of Genetics, Agricultural Biotechnology Center, Gödöllõ, Szent-Györgyi A. 4., H-2100, Hungary.
    2. Department of Genetics, Faculty of Science, Eötvös Loránd University and Research Group for Molecular Genetics of Hungarian Academy of Sciences, Budapest, Pázmány P. 1/C., H-1117, Hungary.
    • E-mail orosz@abc.hu; Tel. (+36) 1 209 0555/8686; Fax (+36) 1 209 0555/1841.

    Search for more papers by this author


Several temperate bacteriophage utilize chromosomal sequences encoding putative tRNA genes for phage attachment. However, whether these sequences belong to genes which are functional as tRNA is generally not known. In this article, we demonstrate that the attachment site of temperate phage 16-3 (attB) nests within an active proline tRNA gene in Rhizobium meliloti 41. A loss-of-function mutation in this tRNA gene leads to significant delay in switching from lag to exponential growth phase. We converted the putative Rhizobium gene to an active amber suppressor gene which suppressed amber mutant alleles of genes of 16-3 phage and of Escherichia coli origin in R. meliloti 41 and in Agrobacterium tumefaciens GV2260. Upon lysogenization of R. meliloti by phage 16-3, the proline tRNA gene retained its structural and functional integrity. Aspects of the co-evolution of a temperate phage and its bacterium host is discussed. The side product of this work, i.e. construction of amber suppressor tRNA genes in Rhizobium and Agrobacterium, for the first time widens the options of genetic study.


Site-specific recombination is a widespread phenomenon in living organisms (Craig, 1988). A number of temperate phages and plasmids utilize this pathway for integration into the genome of the cognate host bacteria (Landy, 1989). The host chromosomal target sites for integration, called attachment sites (attB), are specific and often overlap with putative tRNA gene sequences (Reiter et al., 1989). Examples can be found in both eubacteria and archebacteria. Attachment sites of Lactobacillus phage mv4 (Dupont et al., 1995), Streptomyces phage RP3 (Gabriel et al., 1995), Rhizobium phage phiU (Uchiumi et al., 1998) and virus-like plasmid SSV1 of archebacterium Sulfolobus (Reiter and Palm, 1990) overlap with putative arginine tRNA genes, whereas putative leucine tRNA genes serve as attB sites for integration of Haemophilus phage HP1 (Hauser and Scocca, 1992) and Escherichia coli phage P4 (Pierson and Kahn, 1987). Furthermore, because of the conserved sequence of the tRNA genes, attB sites for phage integration might be present even in unrelated bacterial species, as demonstrated by the case of phage A2 (Alvarez et al., 1998). This phage can integrate its genome into a putative leucine tRNA gene both in Gram-positive (Lactobacillus) and Gram-negative (E. coli) species. However, in most cases it is not known whether the tRNA genes into which the phage can be inserted are functional. Only one study has been reported so far about the activity of a tRNA gene with an internal attachment site. The chromosomal integration site of Streptomyces plasmid SLP1 was shown to be a functional tyrosin tRNA gene, essential for cell viability (Vogtli and Cohen, 1992).

Phage 16-3 is a temperate phage which can efficiently lysogenize the symbiotic nitrogen fixing bacterium Rhizobium meliloti 41 (Orosz et al., 1973; Orosz, 1980; Dallmann et al., 1987; Papp et al., 2002). The attB site has been mapped between the cys46 and met5 genes of the bacterium, close to cys46 as was shown by specialized transduction using the 16-3 phage and R′ plasmids (Svab et al., 1978; Kiss et al., 1980). The genetic system for the 16-3 integration function has been thoroughly documented (Olasz et al., 1985; Dorgai et al., 1993). Upon integration, the recombination between attB and the phage-borne attP elements results in the hybrid attL and attR sequences flanking the integrated prophage (Papp et al., 1993a). This reaction is catalysed by the phage-encoded integrase (Int) protein (Semsey et al., 1999) expressed under the control of the 16-3 immC regulatory region (Olasz et al., 1985; Csiszovszki et al., 2003). The reversal of integration, i.e. the 16-3 prophage excision via the attL × attR recombination, is governed by the combined activity of 16-3 genes int and xis (Semsey et al., 1999).

The nucleotide sequence of the attB and attP sites share a 51 bp homology region containing the 23 bp minimal attB sequence (Semsey et al., 2002). Furthermore, the sequence of the 51 bp region overlaps the 3′ end of a putative proline tRNA(CGG) coding region (Fig. 1; Papp et al., 1993a). Because of this, the putative proline tRNA(CGG) gene is not altered upon 16-3 integration (Papp et al., 1993a) and becomes part of the attL site. Orthologues of this putative gene were identified in a number of related bacterial species: Rhizobia, Bradyrhizobia, Azorhizobia and Agrobacteria. The putative tRNA gene sequences in the latter organisms also served as efficient and specific attB sites for 16-3 integration reaction in the cognate bacteria (Semsey et al., 2002). However, no experimental data are available as to whether these genes are silent or functional.

Figure 1.

Sequence alignments of attB, attP, attL and attR. Schematic representation of the secondary structure of inline image is shown on top. The 51 bp homology of the sequences is highlighted. Uppercase letters indicate the bases encoding the putative proline tRNA(CGG).

Here we report that the putative proline tRNA gene, serving as a site for 16-3 integration in R. meliloti 41 chromosome, is convertible to an active suppressor tRNA allele which is functional as suppressor of an amber mutation in genes of 16-3 phage and of E. coli origin. The suppressor tRNA also retains its activity upon phage integration. We also provide experimental evidence that this tRNA gene, although not essential, significantly increases the lag phase of bacterial growth. The amber suppressor strains, which are side products of this study, might be useful tools for genetic studies in R. meliloti 41 and in Agrobacterium tumefaciens, an important plant pathogen and a major tool in the construction of transgenic plants.


A putative proline tRNA gene in the attB region of R. meliloti 41 chromosome

The attB site was localized on a 1.8 kb DNA fragment (Papp et al., 1993a) in the vicinity of the cys46 gene of R. meliloti 41 chromosome (Svab et al., 1978; Kiss et al., 1980; Dorgai et al., 1981) by genetic and physical mapping. Sequence analysis of a 978 bp region (Accession No. AJ698943) of this chromosome fragment (Fig. 2A) revealed a putative proline tRNA(CGG) gene (nt 1–77), preceded by three putative promoter sequences (P1 with score value of 0.99, sequence from nt −62 to −13; P2 with score 0.69, from nt −49 to +1 and P3 with score 0.58, from nt −114 to −65; P scores were determined by the scoring matrix method; Reese et al., 1996), and followed by a putative transcription terminator region (T  B with P value of 5.09, sequence from nt 511–561). The sequence analysis of the attL site also revealed a putative transcription terminator region (TL with P value of 4.08, sequence from nt 654–704) which is about 600 bp downstream of the putative tRNA gene (Fig. 2B).

Figure 2.

A. The anatomy of AvaI[-319]–HincII[659] chromosomal segment of R. meliloti 41 containing the overlapping attB site (white box) and the proline tRNA (CGG) gene (black bar). Different segments were inserted between the EcoRI and KpnI sites of the pSEM211 promoter probing plasmid (Ferenczi et al., 2004), upstream of the promoterless lacZ reporter gene. Plasmids containing the cloned segments are indicated along with the cognate β-galactosidase activities. Light grey bars indicate deletions. Dotted bars indicate putative promoters identified by sequence analysis (P1, P2 and P3). Only one of these promoters, P2, was found to be active (see pBLA247 versus pBLA238 and Fig. 2C). Striped bars indicate putative transcription terminators identified by sequence analysis (T B and T L in Fig. 2B). TB is an active transcriptional terminator (see pBLA246 versus pBLA217). IR1 and IR2 are inverted repeats with partial terminator activity (see pBLA239 versus pBLA217). Figure not drawn to scale.
B. The anatomy of AvaI[-319]–KpnI[710] chromosomal segment of lysogenic strain R. meliloti 41(16-3) containing the overlapping attL site (white box) and the proline tRNA(CGG) gene (black bar). Other designations are identical to that in (A). Unlike in attB, transcription termination takes place in a complex terminator region downstream of the attL site (see pBLA261 and pBLA262 versus pBLA217); however, the terminator searching program of GCG has predicted only TL (sequence from 654 to 704).
C. Identification of the transcriptional start point (arrow) in the promoter of the proline tRNA gene (P2) by primer extension assay.

For experimental verifications, the intact and truncated derivatives of the 978 bp DNA fragment were inserted into the pSEM211 plasmid (Ferenczi et al., 2004) upstream of the promoterless reporter gene lacZ (Fig. 2A). β-Galactosidase assays were performed in R. meliloti 41 strains carrying these plasmids. Cells carrying the plasmid containing the intact 978 bp fragment (pBLA248) showed similar β-galactosidase activity to those carrying the vector (pSEM211) alone. Deletion of the sequence downstream from the attB site (pBLA217) increased β-galactosidase synthesis about 150-fold, indicating the presence of an active promoter within the upstream region. More deletions (pBLA219, pBLA247 and pBLA238) were tested to further map the region with promoter function upstream of the putative tRNA coding sequence. Results showed that the putative promoters P1 and P3 could be eliminated without any loss of promoter activity. To determine the transcription start site of the active promoter, primer extension was performed as described in Experimental procedures. The transcription start site calculated from the result of the primer extension is in the predicted P2 promoter region at nt −10 (Fig. 2C). The fragment between nt 430 and 659 contains an active transcription terminator (see pBLA246 versus pBLA217), consistent with the predictions of sequence analysis. An equally efficient, however, more complex terminator region, consisting of more than one element, was detected downstream of the attL site (see pBLA261 and pBLA262 versus pBLA217). The predicted T L sequence could be its most distal element, the only one recognized by the sequence analysis program (Fig. 2B). This terminator region may have a dual role, by terminating the transcription of the tRNA gene and by preventing the transcription of phage genes from the host promoters.

To obtain direct proof for the function of the putative tRNA gene, we converted its predicted CGG anti-codon to an amber anti-codon, and probed its function as a suppressor tRNA of an amber mutation as described below.

Amber suppressor genes of attB and attL origins

Two versions of amber suppressor proline tRNA alleles, attB su and inline image, were created by modifying the wild-type gene containing attB. The inline image gene was a modified version of the attB su containing an additional substitution at the nucleotide position 37 (G to A) in the anti-codon loop of the tRNA gene. Based on the comparison of suppressor mutants with enhanced activities reported previously (Kleina et al., 1990), this substitution was predicted to improve the efficiency of suppression. To study the effect of 16-3 integration on the activity of the tRNA gene, a third version of the amber suppressor allele, attLsu, was created by modifying the putative tRNA gene overlapping the attL site. These alleles were separately cloned into the plasmid pBBR1MCS-3 resulting in plasmids pBLA20, pBLA41 and pBLA34 respectively (for details, see Experimental procedures and Table 1).

Table 1. Plasmids used in this study.a
PlasmidRelevant genotype and useSource of reference
  • a

    . Plasmid maps and their construction schemes will be provided upon request.

pBBR1MCS-3Cloning vector Kovach et al. (1995)
pSEM167Clonings, pCU1ori, carries both attP and int Semsey et al. (1999)
pIP72Derivative of pEMBL18, carrying wild-type proline tRNA(CGG) gene containing attB, on a  1.8 kb DNA fragment of R. meliloti 41 chromosome Papp et al. (1993a)
pSEM289Clonings, pSC101ori, carries both attP and int Ferenczi et al. (2004)
pSEM211Promoter probing, pCU1ori, carries lacZ gene without promoter Ferenczi et al. (2004)
pBLA19Derivative of pSEM167, carrying amber mutation in int gene at codon position 90 (int P90*)This work
pBLA37Derivative of pBBR1MCS-3, carrying proline tRNA(CGG) gene overlapping with attBThis work
pBLA20Derivative of pBBR1MCS-3, carrying amber suppressor gene attB suThis work
pBLA34Derivative of pBBR1MCS-3, carrying amber suppressor gene attLsuThis work
pBLA41Derivative of pBBR1MCS-3, carrying modified attB su gene (inline image) made by changing G to A  at nucleotide position 37This work
pBLA127Derivative of pSEM289, carrying attP site, wild-type int and lacZThis work
pBLA128Derivative of pBLA127, carrying amber mutation in lacZ gene at codon position 31 (lacZ P31*)This work
pBLA48Derivative of pBBR1MCS-3, carrying wild-type lacZ and proline tRNA(CGG) genesThis work
pBLA49Derivative of pBBR1MCS-3, carrying lacZ P31* and proline tRNA(CGG) genesThis work
pBLA53Derivative of pBBR1MCS-3, carrying lacZ P31* and attB su genesThis work
pBLA54Derivative of pBBR1MCS-3, carrying lacZ P31* and attLsu genesThis work
pBLA55Derivative of pBBR1MCS-3, carrying lacZ P31* and inline image genesThis work
pBLA198Derivative of pSEM289, carrying mutant attP site (attPm)This work

Converting a proline codon to an amber codon in genes of 16-3 phage (int) and of E. coli (lacZ)

In order to demonstrate the suppression activity of plasmids carrying the modified putative tRNA gene sequences, we constructed amber mutant alleles of the 16-3 int and of the E. coli lacZ genes by converting the proline codons (CCG) at position 90 in the 16-3 int gene and at position 31 in the E. coli lacZ gene into amber codons (for details, see Experimental procedures). The resulting alleles were named int P90* and lacZ P31*. The plasmids carrying the two mutant genes were pBLA19 and pBLA128 respectively (Table 1).

Suppression of the activity of the 16-3 int amber mutant

The attB-related and attL-related wild-type and suppressor alleles (on plasmids pBLA37, pBLA20, pBLA34 and pBLA41) were transferred by triparental matings into R. meliloti 41 already containing a compatible plasmid with the wild-type (pSEM167) or amber mutant (pBLA19) allele of the 16-3 int gene. Plasmids pSEM167 and pBLA19 also carried the attP site to allow chromosomal integration when a functional Int protein is present (we note that integration reaction could also take place between plasmids pBLA37 and pSEM167; however, this does not change the conclusions of the experiment).

Suppression of the mutant phenotype of the int P90* allele by different alleles of the putative proline tRNA gene was tested by following the Int-dependent attP × attBattL × attR reaction. Figure 3A–C shows the different combinations of the suppressor tRNA and int alleles used in this experiment and the predicted outcome of the attB × attP reaction in each case. Figure 3D shows the results of polymerase chain reactions (PCRs) used to detect the hybrid attR fragments indicating chromosomal integration of plasmids carrying attP. The rationale for using a PCR-based method was that it is highly sensitive and allows us to detect even very low levels of suppression. The attP-specific PCR was used as a control to confirm the presence of the integrative plasmids. In the control set-up, the wild-type Int protein readily catalysed integration (Fig. 3A and D, lanes 7 and 8). However, the amber mutant allele of the int gene did not produce any Int activity (Fig. 3B and D, lanes 1 and 2). Successful amplification of the attR region clearly showed that both the attB-related (pBLA20) and attL-related (pBLA34) suppressor alleles restored the integrase activity of the int P90* allele (Fig. 3C and D, lanes 3–6).

Figure 3.

Schematic illustration of the experimental set-up for the assay to determine whether the attB-related proline tRNA(CGG) gene is functional.
A. Elements of the natural reaction. The recombination reaction between attP and attB sites is catalysed by the active integrase enzyme expressed from the pSEM167 integrative plasmid. Integration of the plasmid into the chromosome results in sites attL and attR.
B. In the presence of the plasmid pBLA19 carrying amber mutant integrase allele int P90*, no integration is expected.
C. When the plasmid pBLA20 (carrying the amber suppressor attB su allele) is added, integration of pBLA19 is expected again. Dotted lines represent the chromosome of R. meliloti 41.
D. Integration events as detected by attR-specific PCR amplification. The attP-specific PCR was used as a control. The total DNA extract of R. meliloti 41 strains containing combinations of plasmids carrying the different int and tRNA alleles served as templates. Columns 1, 3, 5, 7: PCRs for attP (255 bp); 2, 4, 6, 8 for attR (185 bp).

Quantitative measurement of the suppression activity

Because β-galactosidase activity was practically absent in R. meliloti 41 (Papp et al., 2002), we chose the lacZ reporter system to determine the efficiency of suppression by the suppressors described above. To create a single-copy chromosomal reporter system, the wild-type and amber mutant alleles of lacZ were inserted into the integrative vector pSEM289, which is unable to replicate in R. meliloti (Ferenczi et al., 2004). The resulting plasmids, pBLA127 (lacZ +) and pBLA128 (lacZP31*), were integrated into the R. meliloti 41 chromosome at the attB site.

To test the suppression activity of the different tRNA alleles, plasmids pBLA20 (attB su), pBLA34 (attLsu) and pBLA41 (inline image) were transferred into the R. meliloti strains, carrying a chromosomal copy of lacZ + or lacZ P31*, by triparental matings. The suppressor efficiency, if any, was determined by measuring β-galactosidase activity. As shown in Table 2A, both suppressors attB su and attLsu increased the β-galactosidase activity more than 2.5-fold (pBLA20 and pBLA34 versus pBLA37). The efficiency of suppression by these alleles was comparable to the efficiency of proline suppressors in E. coli (Kleina et al., 1990). However, significantly higher level of β-galactosidase activity was observed in the presence of the suppressor allele inline image (more than 12-fold, pBLA41 versus pBLA37), indicating that the attempt to enhance suppressor activity with a single point mutation was successful. In order to improve the fidelity of the measurements, we also measured suppression efficiency with the reporter gene present in multicopy (Table 2B). In this multicopy set-up, the reporter lacZ gene and the tRNA genes were in cis. As shown in Table 2B, the attB su and the attLsu increased the β-galactosidase activity by more than sevenfold (pBLA53 versus pBLA49) and more than sixfold (pBLA54 versus pBLA49), respectively, whereas the mutant suppressor allele inline image increased activity about 50-fold (pBLA55 versus pBLA49).

Table 2. β-Galactosidase assay of suppression in the single- (A) and multicopy (B) systems in R. meliloti 41 and in A. tumefaciens.
Plasmid pro-tRNA allelea lacZ gene on chromosomeβ-Galactosidase activityb% Wild-type activityc
R. meliloti A. tumefaciens R. meliloti A. tumefaciens
  • a

    . Proline tRNA alleles were delivered by plasmids. wt indicates the wild-type allele.

  • b


  • β

    β-Galactosidase activity in Miller units. Each value is the mean of six independent experiments. SD was less than 6.7% of the mean value in each case.

  • c


  • β

    β-Galactosidase activity expressed as a percentage of wild-type levels.

(A) Singly-copy system
pBLA37 wt wt 41383586100100
pBLA37 wt lacZP31*   1.9   2.9  0.05  0.08
pBLA20 attB  su lacZP31*   5.5  45  0.13  1.25
pBLA34 attL su lacZ  P31*   5.3  45  0.13  1.25
pBLA41 inline image lacZ  P31*  25  61  0.60  1.70
(B) Multicopy system  lacZ gene on plasmid    
pBLA48 wt wt 46583208100100
pBLA49 wt lacZ  P31*   2.1   3.6  0.04  0.11
pBLA53 attB  su lacZ  P31*  13  88  0.28  2.74
pBLA54 attL su lacZ  P31*  12  87  0.26  2.71
pBLA55 inline image lacZ  P31*  91 193  1.96  6.02

Amber suppression by attBsu and attLsu in Agrobacterium

We extended both single- and multicopy suppression studies to A. tumefaciens, which is an optimal assay organism for several reasons. A. tumefaciens is a competent recipient for triparental mating, carries the attB site in its chromosome (Semsey et al., 2002) and lacks β-galactosidase activity. As shown in Table 2A and B, the Rhizobium and Agrobacterium data were comparable. Interestingly, the suppression of lacZP31* by the attBsu and the attLsu alleles was more than ninefold higher in Agrobacterium than in Rhizobium. The suppression efficiency of inline image was more than 6% in the multicopy set-up in Agrobacterium.

Inactivation of the chromosomal proline tRNA(CGG) gene

A selective pressure can explain that the proline tRNA(CGG) gene remains intact and functional both in attB and, after prophage integration, in attL. To test this hypothesis, we studied the role of this tRNA gene in bacterial growth. The proline tRNA(CGG) gene was inactivated as follows (Fig. 4A). In the attP site of the integrative plasmid pSEM289, the GAGTGTT sequence in the ΤφC arm of the tRNA was substituted by the TGTGTGG sequence (attP m). The resulting plasmid, pBLA198, was introduced into R. meliloti 41. The attP m mutant site of pBLA198 and the chromosomal attB site recombined readily via the attP m × attB pathway to help the non-replicating pBLA198 plasmid integrate into the R. meliloti 41 genome (resulting in attLm and attR). Integration of the plasmid was confirmed by the presence of attR and the absence of attB in R. meliloti(pBLA198) cells (Fig. 4B). In this strain both the attL and the attR sites were PCR amplified and the presence of the heptanucleotide substitution was confirmed by DNA sequencing. The attL region carried the heptanucleotide substitution, whereas attR did not (i.e. no gene conversion was left behind). The heptanucleotide substitution in the stem and the conserved sites of the ΤφC arm (nucleotides 49–55) in the proline tRNA gene should have made the tRNA non-functional.

Figure 4.

Inactivation of wild-type gene proline tRNA(CGG) of R. meliloti 41.
A. Schematic illustration of the experimental set-up. The attP m contains mutations (i.e. a heptanucleotide substitution) which do not influence the att/int function, but eliminate the inline image function upon integration by transferring the mutations to the attL site. Note: if inactivation of proline tRNA(CGG) is lethal, no cell carrying attLm and attR will be viable. Dotted lines represent the chromosome of R. meliloti 41.
B. Integration of pBLA198 as detected by attR-specific PCR (lane 1). The attB-specific PCR reaction (lane 2) was used as a control. Template DNAs were prepared from R. meliloti(pBLA198).

We compared the growth of three strains: (i) R. meliloti 41, wild-type with intact proline tRNA(CGG) gene, (ii) R. meliloti(pBLA198), with destroyed proline tRNA(CGG) gene and (iii) R. meliloti(pBLA289), with proline tRNA gene overlapping the attL site. As shown in Fig. 5, the loss of the function of the proline tRNA(CGG) gene led to a significant delay in switching from the lag to the exponential growth phase in rich medium. Smaller but significant delay was observed in minimal medium (data not shown). Furthermore, the growth curve of strain R. meliloti(pBLA289) containing the attL-related allele of the proline tRNA gene was indistinguishable from the growth curve of the wild-type R. meliloti 41 strain. However, no detectable differences were noted in growth rates in exponential phase (neither in complete nor in minimal medium) between R. meliloti 41 wild-type, R. meliloti(pBLA198) and R. meliloti(pSEM289). The doubling time at 28°C was 360 min in rich medium, and 700 min in minimal medium for all strains in the logarithmic phase.

Figure 5.

Growth curves of isogenic R. meliloti 41 strains in rich medium (YTB) carrying alleles of proline tRNA(CGG) gene: ×, wild-type R. meliloti 41 (i.e. carrying wild-type allele in attB region); □, strain R. meliloti(pSEM289), carrying wild-type allele in the attL chromosomal region; ▴, R. meliloti(pBLA198), carrying mutant allele of the proline tRNA(CCG) gene in the attLm chromosomal region. OD for start was adjusted to 0.04 (corresponds to 4 × 107 cells ml−1) for each strain. Note that the mutant strain delays in switching from the lag to the exponential phase.


In current views, tRNA genes are regarded as favourable for evolving to attB and attP sites for temperate phage insertion systems for several reasons. These sequences are rich in inverted repeats which are often tandemly arranged. The conserved nature of the tRNA genes may help to widen the host range and spreading of the phage. They may also direct the co-evolution with phage genes for recombination enzymes (Reiter et al., 1989; Papp et al., 1993a; Cheetham and Katz, 1995). Once a prophage is inserted into the bacterial genome, new doors open for ‘evolution in quantum leaps’ (Groisman and Ochman, 1996) because the prophage DNA provides a chromosomal port for inserting further foreign DNAs which may then evolve to islands of foreign DNAs (e.g. pathogenicity islands) (Hacker et al., 1997; Hacker and Kaper, 2000; Campbell, 2003). However, the effect of these processes on the function of the tRNA gene is not known. In this article, we intend to call attention to some new strokes for this evolutionary scenario.

We provided unambiguous proof that the attachment site of temperate phage 16-3 nests in the active gene for proline tRNA(CGG) in R. meliloti 41. The gene does not loose its functional and structural integrity upon prophage insertion. For the functional test, we took the advantage of amber suppression. The proline tRNA(CGG) gene was successfully converted to an active suppressor allele by mutating its anti-codon. The resulting suppressor allele ensured full-length translation and restored activity of an amber mutant in the int gene of 16-3 and LacZ gene of E. coli. We identified the promoter and transcription start site of the tRNA gene, and localized its transcription terminator both in attB and, upon prophage integration, in attL.

Our experiments show that the loss of the proline tRNA(CGG) function is not lethal for R. meliloti 41, although it can lead to a severe delay, 180 min, in switching from lag to exponential phase of growth (Fig. 5). The mutant phenotype sheds light on the co-evolution of the att site of the phage and of the tRNA gene of the bacterium, i.e. why the selective pressure kept the tRNA gene upon prophage insertion. It also explains the reason of the observed extended homology of the attachment sites. The attB and attP sites share 51 bp homology, whereas a 23 bp homologous region is sufficient for phage integration (Semsey et al., 2002).

Analysis of the usage of the four proline codons in R. meliloti, CCG (59%), CCC (26%), CCU (8.6%) and CCA (6.3%), shows that CCG is the major proline codon (http://sequence.toulouse.inra.fr/meliloti.html) (Capela et al., 2001). Besides the proline tRNA(CGG) gene, identified in this study, there are two more putative proline tRNA genes in the R. meliloti genome, proline tRNA(TGG) and proline tRNA(GGG) (http://www.ncbi.nlm.nih.gov/genomes/rnatab.cgi?gi=188anddb=Genome). The proline tRNA(TGG) gene might be also active, because knocking out the proline tRNA(CGG) gene is not lethal; the inline image must be sufficient for translating both CCA and CCG codons by wobble (Söll and RajBhandary, 1995). Surprisingly, however, we found that inactivating the proline tRNA(CGG) gene did not influence the growth rate of R. meliloti 41 in the exponential phase (Fig. 5). It seems that the proline tRNA corresponding to the least frequent proline codon (CCA) was able to take over the role of the major tRNA.

Table 3 lists the subsets of Rhizobium genes with their codon usages for proline. Starting with the premise that the E. coli and the R. meliloti species are related closely enough in their basic gene function, we selected two groups of (orthologous) genes: (i) active in the stationary growth phase (Ishihama, 1997; Tani et al., 2002) and (ii) active in the exponential growth phase. The genes for exponential growth phase were chosen according to Tables 2 and 3 in a previous study (Tao et al., 1999), at a threshold value for normalized activities above 0.4 and below −0.4. Surprisingly, the high usage of the CCG codon for incorporation of proline was the characteristic of certain genes (e.g. elongation factors) prominently active in the exponential phase in rich medium (Table 3A, lines 1–6) (Tao et al., 1999).

Table 3. Proline usage of selected genes of pronounced activities for exponential growth (A) and stationary growth (B).
(A) Genes for exponential growth phase(B) Genes for stationary growth phase
orthologous genea
(No. of aa)
% of the
CCG codonsb
E. coli
Function in E. coli Rhizobium
orthologous genea
(No. of aa)
% of the
CCG codonsb
E. coli
Function in E. coli
tsf (307) 83 (5/6) tsf Elongation factor Ts SMa2389 (177)30 (3/10) osmC Osmotically induced protein
tufB (391) 91 (20/22) tufB Elongation factor Tu katC (705)55 (25/45) katE Catalase
tufA (391) 90 (19/21) tufA Elongation factor Tu katA (494)81 (30/37) katE Catalase
infB (889) 71 (20/28) infB Initiator factor 2 fic (302)50 (11/22) fic Induced in stationer phase
fusA (699) 83 (25/30) fusA Elongation factor G otsA (463)58 (14/24) otsA Trehalose 6-P synthase
efp (189)100 (8/8) efp Elongation factor P SMa1153 (224)60 (3/5) osmY Osmotically induced prot
trmU (398) 75 (15/20) trmU tRNA methyltransferase SMc00698 (93)60 (3/5) bolA Regulator of murein genes
rplJ (172) 60 (3/5) rplJ L10 ribosomal protein aidB (550)64 (14/22) aidB Putative oxidoreductase
rplX (103)100 (3/3) rplX L24 ribosomal protein SMc02689 (502)87 (21/24) aldB Aldehyde dehydrogenase B
rpsE (189) 66 (4/6) rpsE S5 ribosomal protein sodC (172)60 (3/5) sodC Superoxide dismutase
ilvC (339) 62 (5/8) ilvC Amino acid biosynthesis wrbA1 (199)50 (4/8) wrbA Trp repressor binding prot
leuD (201) 30 (3/10) leuD Amino acid biosynthesis wrbA2 (202) 0 (0/6) wrbA Trp repressor binding prot
tyrA (111) 80 (4/5) tyrA Amino acid biosynthesis ihfB (103)40 (2/5) ihfB IHF β-subunit
gltD (378) 63 (17/27) gltD Amino acid biosynthesis SMc01698 (251)75 (6/8) yohF Putative oxidoreductase
aroF (457) 73 (16/22) aroF Amino acid biosynthesis SMb21187 (276)78 (7/9) mlrA Transcriptional regulator
pyrB (313) 80 (8/10) pyrB Nucleotide biosynthesis ggt (581)59 (20/34) ggt γ-Glutamyltranspeptidase
upp (209) 80 (8/10) upp Nucleotide biosynthesis SMa0682 (731)49 (18/37) ldcC Lysine decarboxylase 2
grxC (88) 33 (1/3) nrdH Vitamins and cofactors SMa0116 (318)42 (8/19) cbpA Chaperone
hemC (309) 78 (14/18) hemC Vitamins and cofactors SMb20095 (605)74 (28/38) poxB Pyruvate oxidase
hemD (309) 45 (9/20) hemC Vitamins and cofactors cfa1 (413)69 (9/13) cfa Fatty acyl P-lipid synthase
sigA (684) 70 (16/23) rpoD Factor σ70 cfa2 (418)27 (4/15) cfa Fatty acyl P-lipid synthase
llpB (512) 70 (21/30) nlpD Lipoprotein dac (384)42 (8/19) dacC Carboxypeptidase

We speculate that in the absence of the proline tRNA(CGG) gene, the establishment of a certain kind of proteins needed for fast growth suffers a delay, but once a critical amount is accumulated to support exponential growth, the activity of the proline tRNA(TGG) gene can maintain a steady state for these protein factors. Such factors can be part of the translational machinery (Table 3A, lines 1–7). In stationary phase the amount of such requirements may be low (Bjork, 1977). Hence, in the cells lacking the major proline tRNA gene, synthesis of these proteins and tRNAs (particularly inline image) in sufficient amount may require longer time and result in longer lag phase.

The side products of this study are amber suppressor strains of R. meliloti and of A. tumefaciens which may be useful in genetic studies of these organisms. The efficiency of the proline inserting amber suppressor allele (attB su), although low, in agreement with the general observations concerning inline image suppressors (Kleina et al., 1990), was improved significantly by a single substitution in the anti-codon loop.

Experimental procedures

Bacterial strains and growth conditions

Escherichia coli strain DH5α (Hanahan, 1983) was used in all cloning experiments and served as a host for donor plasmids used to conjugate into recipient bacterial strains, including Rhizobium meliloti 41 (Szende and Ördögh, 1960), the native host of phage 16-3 and Agrobacterium tumefaciens GV2260. R. meliloti 41 and A. tumefaciens GV2260 served as host for study of genetic suppression. E. coli was grown in Luria broth at 37°C (Sambrook et al., 1989), and all the other bacteria were grown in yeast-tryptone broth (YTB) 1% tryptone, 0.1% yeast extract, 0.5% NaCl, 1 mM MgCl2 and 1 mM CaCl2 per 1000 ml, pH 7.0 at 28°C. For plating bacteria, both Luria broth and YTB were supplemented with 1.5% agar, resulting in Luria agar and YTA respectively. GTS was used for minimal media (Kiss et al., 1980). Tetracycline was used at a concentration of 15 µg ml−1, whereas the concentration of kanamycin was 30 µg ml−1 in the case of E. coli, 100 µg ml−1 in the case of A. tumefaciens and 400 µg ml−1 in the case of R. meliloti 41. Plasmids used in this study are listed in Table 1.

Triparental mating

Triparental mating to transfer different plasmids, resident in E. coli, into the recipient bacteria was as described earlier (Semsey et al., 2002). Helper bacteria were E. coli harbouring pRK2013 (Figurski and Helinski, 1979) or pCU101 (Thatte et al., 1985), depending on the origin of mob region carried by the given plasmid in the donor.

DNA procedures

Basic DNA manipulations and molecular techniques were used according to methods described previously (Sambrook et al., 1989). Total bacterial DNA was prepared by a method described earlier (Berman et al., 1982). Primers used in PCR amplifications are listed in Table 4. PCR-mediated DNA amplifications were carried out using Pfu polymerase (Promega) to generate DNA fragments for mutagenesis and cloning. PCR mutagenesis was performed according to method described previously (Landt et al., 1990). DNA fragments were digested with the appropriate restriction enzyme(s) and purified by extraction from agarose gels (QIAEX II Gel Extraction Kit, Qiagen) before cloning. DNA sequences of fragments resulting from PCR amplifications were verified in plasmids after cloning. Nucleotide sequence determination was performed by the dideoxy chain termination method (Sanger et al., 1977) using the TaqTrack Sequencing kit or the fmol DNA Cycle Sequencing System (Promega).

Table 4. Primers used in this study.
PrimerSequenceant positionbUse
  • a

    . Restriction sites used in cloning procedures (EcoRI, XbaI, PvuII, HindIII, NspV and KpnI) are boldfaced. Lowercase letters represent bases that are not present in the original sequence.

  • b

    . Nucleotide positions refer to the base positions of the 3

  • bases of the primers. (+) and (–) indicate upper strand and lower strand respectively.

  • c

    . Primers annealed to the AvaI[-319]–HincII[659] chromosomal segment of R. meliloti 41.

  • d

    . Primers annealed to phage 16-3 sequences (Semsey et al., 1999).

 1CGGAGTGTAGCGCAGTCTGGT  21 (+)c attB and attL detection
 2AATAATCGGGGGATTGTTTCAGC  78 (–)c attB and attR detection, construction of plasmids  pBLA219 and pBLA247
 3GGTCGTTTTTATTGCCGTGG 494 (+)d attP and attR detection and construction of attLsu
 4ATCTAGAAGTAAATCATTGCCGTAAT 707 (–)d attP and attL detection, construction of lacZ P31*
 5ATgaattCTTGTGAAATCAAAAATGCT −25 (+)cConstruction of plasmid pBLA247
 6GCGaatTCTTTTTTGCGATAGGCGCT −43 (+)cConstruction of plasmid pBLA219
 7GGCGCGCCCCCGGGTACC Primer extension assay
 8ttGAATTCTTTctaTCGCCTTGCTGTTG1584 (+)dConstruction of int P90*
 9TAtcTAGAtaAGAaGgAAGGAAAaTGCCTCGTAAA1826 (–)dConstruction of int P90*
10CTATgAattCAGCTGGCGAActaGGGATGT Construction of lacZ P31*
11TAGTCtaGAACGTGGTGCGCTACC  19 (–)cConstruction of attBsu
12GCAAGCTTGCGATAGGCGCTTGTGAAATC −34 (+)cConstruction of attB su
13AGTTCtaGACGTGGGGGTCGAGTGTT  56 (+)cConstruction of attB su
14TCACACAGGAAACAGCTATGAC Construction of attB su
15AAACTGCGGGGTATCTCGGAATGAC 824 (–)dConstruction of attLsu and pBLA263
16GTGATTCGAACACTCGACCCCCACGTttaGAACG  31 (–)cConstruction of inline image
17GTAAAACGACGGCCAGT Construction of inline image
18CGTGGCAgGTacCGCGGAGCTC1264 (–)dConstruction of plasmid pBLA263

Construction of the promoter and terminator probing plasmids

Different parts of the 978 bp chromosome region (Fig. 2A) were inserted between the EcoRI and KpnI sites of the pSEM211 promoter probing plasmid (Ferenczi et al., 2004). The pBLA248 plasmid contained the whole 978 bp fragment. The pBLA246 was created by deleting the PvuII[76]–NruI[430] fragment of pBLA248. The pBLA239, pBLA217 and pBLA238 plasmids contained the AvaI[-319]–NruI[154], AvaI[-319]–PvuII[76] and NarI[-14]–PvuII[76] fragments respectively. The pBLA219 and pBLA247 contained the fragments −62 to 78 and −44 to 78 made by PCR amplification with primers 6 and 5 opposed to primer 2 (Table 4), respectively, then digested with PvuII.

Plasmid pBLA263.

 The attL containing 1029 bp fragment from lysogenic strain R. meliloti 41(16-3) was amplified by PCR with primers 18 and 15. After digestion withNspV and KpnI, the PCR product was inserted into NspV–KpnI sites of pBLA217 resulting in pBLA263. The pBLA260 was made by deleting the AccIII[220]–KpnI[710] fragment of pBLA263. The pBLA261 and pBLA262 were made by cloning the HindIII[531]–KpnI[710] and AccIII[220]–HindIII[531] fragments of pBLA263 into the KpnI site (made blunt by end filling) of pBLA217 respectively.

attB, attP and attR diagnostic PCR assays

Polymerase chain reactions were performed by Taq polymerase on total bacterial DNA preparations from R. meliloti 41. Each 50 µl of reaction mixture contained 100 ng of the template and 30 pmol of each primer. After 30 cycles of 30 s at 95°C, 30 s at 53°C and 30 s at 72°C, the PCR products were analysed by electrophoresis through a 2% agarose gel. Primer 1 and primer 2 (Table 4) were used to detect specifically attB, which was indicated by the appearance of a 100-bp-long PCR product. Primer 3 and primer 4 were used to amplify a 255 bp DNA fragment containing attP. Use of Primer 2 and primer 3 resulted in a PCR product of 185 bp, indicating specifically attR.

β-Galactosidase assay

β-Galactosidase assay of R. meliloti 41 and A. tumefaciens was performed according to the method described previously (Miller, 1972).

Primer extension assay

Primer 7, which corresponds to the sequence located within the 5′ end of the lacZ gene of plasmid pBLA217, was end-labelled with [γ-32P]-ATP. The reaction mixture (20 µl) contained 2 pmol of end-labelled primer, 20 µg total RNA from R. meliloti cells (containing plasmid pBLA217), 200 µM of each dNTP, 2 M betaine, 1 mM MgCl2, 10 mM Tris-HCl (pH 8.3), 90 mM KCl and 5 units Tth DNA Polymerase (Promega). The reaction was performed at 74°C for 30 min and then terminated by EGTA (final concentration of 20 mM). The extended products were resolved on a denaturing 6% polyacrylamide gel. Sequencing reaction for size control was performed on plasmid pBLA217 with primer 7.

Construction of amber suppressor alleles

Construction of the plasmid pBLA20.

The amber (CTA) suppressor allele of the putative proline tRNA(CGG) gene containing the attB site (attBsu) was made by the overlap extension method (Landt et al., 1990). The first PCR was governed by primers 11 and 12 on template pIP72 (Papp et al., 1993a). The second PCR was made by primers 13 and 14 also on pIP72 template. The third PCR was made by the outside primers (12 and 13) using the first and second PCR products for template. The product of the third reaction was digested, and then cloned into NarI–BamHI sites of pIP72 resulting in plasmid pBLA11. Finally the 1.8 kb PstI–XhoI fragment containing the allel attB su of pBLA11 was inserted into PstI–XhoI sites of the pBBR1MCS-3 (Kovach et al., 1995) resulting in pBLA20.

Construction of plasmid pBLA34.

The amber suppressor allele of the putative proline tRNA(CGG) gene overlapping the attL site (attLsu) was synthesized by PCR with primers 3 and 15, using a total DNA extract of a R. meliloti 41(16-3) lysogenic strain as a template. After digestion withNspV and AccIII, the PCR product was inserted into NspV–XmaI sites of the pBLA11 resulting in pBLA31. Finally the PstI–KpnI fragment of the pBLA31 was inserted into PstI–KpnI sites of the pBBR1MCS-3 resulting in pBLA34.

Construction of the plasmid pBLA41.

 Polymerase chain reaction was performed on template pBLA11 by primers 16 and 17 to synthesize a mutant derivative of the attBsu allele containing a G to A substitution at position 37 (inline image). After digestion by NspV and HindIII, the PCR product was inserted into NspV–HindIII sites of the pBLA11 resulting in pBLA40. Finally the PstI–XhoI fragment of the pBLA40 was inserted into PstI–XhoI sites of the pBBR1MCS-3 resulting in pBLA41.

Construction of the Plasmid pBLA37.

 Plasmid pBLA37 carrying the wild-type proline tRNA(CGG) gene was created by cloning the PstI–XhoI fragment of the pIP72 into PstI–XhoI sites of the pBBR1MCS-3.

Construction of the amber mutant alleles

To create the amber mutant allele of the 16-3 int gene (intP90*), a PCR reaction was performed using primer 8 and primer 9 on template pSEM167. After digestion with XbaI and EcoRI, the amplified fragment was inserted between the XbaI and EcoRI sites of pSEM167, resulting in pBLA19. Plasmid pBLA127 was created by cloning the SmaI–DraI fragment (wild-type lacZ gene) of the pMLB1109 (Papp et al., 1993b) into the XbaI site (made blunt by end filling) of pSEM289 (Ferenczi et al., 2004). The amber mutant lacZ gene (lacZ P31*) was created by PCR mutagenesis using primer 4 and primer 10 on template pBLA127. After digestion with SalI and PvuII, the PCR fragment was inserted between the partially digested SalI and PvuII sites of pBLA127, resulting in pBLA128.

Construction of plasmids to test suppression when the reporter gene is present in multicopy

Plasmids pBLA53, pBLA54 and pBLA55 were created by inserting the SalI–SpeI (previously made blunt) fragment of pBLA128, carrying the lacZ P31* allele, at the SmaI site of plasmids containing one of the suppressor tRNA alleles. Plasmid pBLA49 was created by inserting the SalI–SpeI (previously made blunt) fragment of pBLA128, carrying the lacZ P31* allele, at the SmaI site of plasmid pBLA37. Plasmid pBLA48 was created by inserting the SalI–SpeI (previously made blunt) fragment of pBLA127, carrying the lacZ+ allele, at the SmaI site of plasmid pBLA37 (Table 1).

Sequence analysis

Putative promoter sites and putative transcription terminator sites were predicted by using a program available at http://www.fruitfly.org/seq_tools/promoter.html (Reese, 1996) and by the Wisconsin package (Genetics Computer Group). The R. meliloti 41 orthologous genes were identified by the microbial genomic blast program of NCBI (http://www.ncbi.nlm.nih.gov/sutils/genom_table.cgi) (Pearson and Lipman, 1988). Further details are available upon request.

Nucleotide sequence accession number

The 978-bp-long nucleotide sequence of the attB region of R. meliloti 41 has been submitted to the EMBL database under Accession No. AJ698943.


We thank Sankar Adhya and Andrei Trostel for critical reading, discussion and helpful comments on the manuscript; Kornélia Szóráth Gál, Magdolna Tóth Péli and Csilla Sánta Török for excellent technical assistance. This work was supported by grants from the Hungarian Scientific Research Fund (OTKA) (T 032205 and T 032255); from the National Research and Development Project (NKFP) (OM 0028/2001 and OM 278/2001); and from the Hungarian Academy of Sciences (MTA/TKI/AKT-F 1999-2001 and MTA/TKI/AKT-F 2003-2006).