Overexpression of the polynucleotide phosphorylase gene (pnp) of Streptomyces antibioticus affects mRNA stability and poly(A) tail length but not ppGpp levels

Patricia Bralley and George H. Jones

Department of Biology, 1510 Clifton Rd, Emory University, Atlanta, GA 30322, USA

Correspondence
G. H. Jones
gjones{at}biology.emory.edu


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
The pnp gene, encoding the enzyme polynucleotide phosphorylase (PNPase), was overexpressed in the actinomycin producer Streptomyces antibioticus. Integration of pIJ8600, bearing the thiostrepton-inducible tipA promoter, and its derivatives containing pnp into the S. antibioticus chromosome dramatically increased the growth rate of the resulting strains as compared with the parent strain. Thiostrepton induction of a strain containing pJSE340, bearing pnp with a 5'-flanking region containing an endogenous promoter, led to a 2·5–3 fold increase in PNPase activity levels, compared with controls. Induction of a strain containing pJSE343, with only the pnp ORF and some 3'-flanking sequence, led to lower levels of PNPase activity and a different pattern of pnp expression compared with pJSE340. Induction of pnp from pJSE340 resulted in a decrease in the chemical half-life of bulk mRNA and a decrease in poly(A) tail length as compared to RNAs from controls. Actinomycin production decreased in strains overexpressing pnp as compared with controls but it was not possible to attribute this decrease specifically to the increase in PNPase levels. Overexpression of pnp had no effect on ppGpp levels in the relevant strains. It was observed that the 3'-tails associated with RNAs from S. antibioticus are heteropolymeric. The authors argue that those tails are synthesized by PNPase rather than by a poly(A) polymerase similar to that found in Escherichia coli and that PNPase may be the sole RNA 3'-polynucleotide polymerase in streptomycetes.


Abbreviations: acm, actinomycin D; PNPase, polynucleotide phosphorylase


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
RNA degradation in bacteria, once thought to be a pathway whose only function was to scavenge nucleotides for the resynthesis of RNAs, is now known to be a complex, carefully regulated pathway that involves a number of key enzymes acting in concert. Some 20 ribonucleases (RNases) have been identified in Escherichia coli (Regnier & Arraiano, 2000), and several of these are known to be involved in the decay and processing of stable and messenger RNAs. The complement of RNases identified thus far in Bacillus is smaller than that of E. coli (Farr et al., 1999), but several of these enzymes function analogously to those of E. coli (Luttinger et al., 1996; Herskovitz & Bechhofer, 2000) and it is clear that in both species gene expression can be regulated at the level of RNA stability (Luttinger et al., 1996; Grunberg-Manago, 1999).

Far less is known about RNA degradation in the Gram-positive, antibiotic-producing genus Streptomyces. Three enzymes involved in RNA decay and processing have been identified in Streptomyces. An RNase E homologue has been found in Streptomyces coelicolor (Hagege & Cohen, 1997). RNase E is a single-strand-specific endoribonuclease which in E. coli is thought to be the key enzyme involved in the initiation of mRNA decay (Coburn & Mackie, 1999; Rauhut & Klug, 1999; Regnier & Arraiano, 2000). An ORF with significant sequence homology to the oligoribonuclease of E. coli was cloned from Streptomyces griseus (Ohnishi et al., 2000). Overexpression of the cloned gene in E. coli verified that the product did possess oligoribonuclease activity (Ohnishi et al., 2000). The oligoribonuclease is essential in E. coli (Ghosh & Deutscher, 1999). Disruption of the gene in S. griseus or S. coelicolor produced a slow-growth phenotype but the enzyme was not essential in either species (Ohnishi et al., 2000). A gene encoding a putative RNase III homologue was identified as the absB locus in S. coelicolor (Adamis & Champness, 1992; Price et al., 1999). Interestingly, mutations in absB abolish production of all four antibiotics normally formed by S. coelicolor (Adamis & Champness, 1992), suggesting a possible relationship between RNA degradation and the regulation of antibiotic production in Streptomyces. However, the product of the absB locus has yet to be characterized so it is not clear that the locus actually encodes a double-strand-specific endoribonuclease. It is noteworthy that Streptomyces spp. do not appear to contain homologues of RNase II or RNase R (Zuo & Deutscher, 2001), which are 3'-5'-exoribonucleases. In E. coli, RNase II constitutes the major 3'-5'-exonuclease activity (Deutscher & Reuven, 1991).

One enzyme that has been studied extensively in Streptomyces is polynucleotide phosphorylase (PNPase; EC 2.7.7.8). PNPase catalyses the 3'-5'-phosphorolysis of RNAs and can also polymerize nucleoside diphosphates to produce ribopolymers (Grunberg-Manago & Ochoa, 1955; Godefroy-Colburn & Grunberg-Manago, 1972; Littauer & Soreq, 1982). The PNPase of Streptomyces antibioticus was identified initially by virtue of its ability to catalyse a different reaction, the formation of guanosine pentaphosphate (pppGpp) from GTP and ATP (Jones, 1994a; Jones, 1994b). pppGpp is the precursor of ppGpp, the mediator of the stringent response in E. coli (Cashel et al., 1996), and ppGpp has been shown to be required for antibiotic synthesis in several streptomycetes (Chakraburtty et al., 1996; Chakraburtty & Bibb, 1997; Hoyt & Jones, 1999). Thus, this unusual activity of S. antibioticus PNPase suggested a possible relationship between RNA degradation, ppGpp formation and antibiotic synthesis. The crystal structure of the S. antibioticus PNPase has recently been determined and that structure has been used to examine structure–function relationships for the enzyme (Symmons et al., 2000; Jarrige et al., 2002).

PNPase may play another unusual role in Streptomyces. As in E. coli, the 3'-tails of RNA species are modified by polyadenylation in Streptomyces (Bralley & Jones, 2001). The major enzyme responsible for RNA polyadenylation in E. coli is a poly(A) polymerase (Cao & Sarkar, 1992). When that enzyme is absent, RNAs with 3'-tails are still produced but those tails contain significant levels of G, C and U in addition to A residues (Mohanty & Kushner, 2000b). Mohanty & Kushner (2000b) have argued that the enzyme responsible for the synthesis of those heteropolymeric tails in mutants lacking poly(A) polymerase is none other than PNPase. We have shown recently that the 3'-tails associated with RNAs from wild-type S. coelicolor are also heteropolymeric (Bralley & Jones, 2002). Thus we hypothesize that PNPase may be responsible for the 3'-modification of RNAs in Streptomyces and may further be the major if not the sole 3'-polyribonucleotide polymerase in those organisms. It is noteworthy in this regard that PNPase appears to be the sole 3'-polyribonucleotide polymerase in spinach chloroplasts and that the 3'-tails associated with chloroplast RNAs are heteropolymeric (Lisitsky et al., 1996; Yehudai-Resheff et al., 2001).

Given the dearth of information available on the mechanism and regulation of RNA decay in Streptomyces there is little indication at this point of the degree to which those mechanisms mimic or differ from those observed in E. coli and Bacillus. Moreover, and as indicated above, it appears that PNPase may play novel roles in Streptomyces related to the production of antibiotics and to the modification of RNA 3'-ends. We have attempted to disrupt the PNPase gene (pnp) in Streptomyces antibioticus and S. coelicolor as one approach to unravel the complexities of RNA degradation in those species. To date, although three different disruption methods have been used by workers in two different laboratories, it has not been possible to isolate pnp disruptants. Thus, and again unlike the situations in E. coli and Bacillus, it is possible that pnp is an essential gene and PNPase an essential enzyme in Streptomyces.

To learn more about RNA degradation and the role of PNPase in that process in Streptomyces we have overexpressed the S. antibioticus PNPase gene in an otherwise wild-type background. We report here the effects of that overexpression on growth of the resulting strains, PNPase activity, RNA stability, poly(A) tail length, ppGpp levels and antibiotic production.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Growth of organisms and biochemical assays.
The actinomycin-producing strain, S. antibioticus IMRU 3720, and strains containing derivatives of pIJ8600 were grown on NZ-amine and galactose-glutamic acid (GGA) media as previously described (Gallo & Katz, 1972; Jones & Bibb, 1996). Strains containing pIJ8600 derivatives were grown on NZ-amine medium in the presence of 50 µg apramycin ml-1. No apramycin was added to GGA cultures. Growth of strains was measured as mycelial dry weight and actinomycin production was assayed as previously described (Jones, 2000). ppGpp levels were measured by HPLC as described previously (Hoyt & Jones, 1999). PNPase activity was measured via the polymerization reaction using [3H]ADP as a substrate (Jones & Bibb, 1996) and expressed as nmol ADP incorporated into an acid-insoluble form in a 15 min incubation.

Construction of pJSE340 and pJSE343.
Plasmids for overexpression of S. antibioticus pnp were constructed using the streptomycete expression vector pIJ8600 (Sun et al., 1999; Kieser et al., 2000). For the construction of pJSE340 a 3·2 kb PCR fragment was prepared, extending from the beginning of the intergenic region between a putative integral membrane protein and the rpsO gene upstream of pnp, to a BamHI site in a putative protease gene downstream of pnp (see Fig. 1). This fragment was cloned into the BamHI site of pIJ8600. pJSE343 contains an approximately 2·5 kb PCR fragment, with flanking XbaI and BamHI sites, extending from a few bases upstream of the pnp translation start site to the BamHI site in the putative protease gene. Plasmid pJSE303 (Jones & Bibb, 1996) was used as the template for PCR for both constructs. Restriction digestion was utilized to identify E. coli transformants containing the constructs with the cloned inserts in the desired orientation relative to the tipA promoter (Fig. 1). The plasmids were transferred to S. antibioticus by conjugation from E. coli as described previously (Jones et al., 1997). A control strain was prepared by transferring pIJ8600 to the S. antibioticus chromosome by conjugation from E. coli. Integration of the plasmids into the S. antibioticus chromosome was confirmed by Southern blotting.



View larger version (13K):
[in this window]
[in a new window]
 
Fig. 1. Schematic representation of the inserts of plasmids pJSE340 and pJSE343. The figure is not drawn to scale.

 
PNPase production was induced from pJSE340 and 343 by transferring portions of NZ-amine cultures, following 36–48 h of growth, to GGA cultures (200–800 ml, depending on the experiment). Following growth in GGA for 12 h, thiostrepton was added as indicated in Results and in the legends to relevant figures. Samples were removed at various times after inoculation and assayed for PNPase as described above.

Determination of chemical half-lives and poly(A) tail lengths.
To measure chemical half-lives of bulk mRNA, 20 ml portions of GGA cultures of the wild-type strain or strains containing pIJ8600 or pJSE340 were removed following 24 h of incubation, 12 h after addition of thiostrepton (5 µg ml-1). One hundred microcuries of [3H]uridine (Dupont NEN; 31·9 Ci mmol-1; 1180 GBq mmol-1) was added to the 20 ml portions and incubation with shaking was continued at 30 °C for 30 min. Rifampicin was then added to each incubation to give a final concentration of 500 µg ml-1 and duplicate 1 ml portions were removed immediately to 1 ml 20 % trichloroacetic acid (TCA). Incubation was continued and duplicate 1 ml samples were removed from each culture every 5 min for 30 min. A final duplicate sample was removed at 45 min. TCA samples were collected on glassfibre filters and analysed by liquid scintillation counting. The unstable RNA (mRNA) fraction was defined as the 3H c.p.m. at any time point after rifampicin addition minus the c.p.m. in stable RNA (the c.p.m. remaining after 45 min of exposure to rifampicin). Pilot studies demonstrated that the 3H c.p.m. remained essentially constant after 45 min of incubation with rifampicin. Chemical half-lives were determined by regression analysis of the decay data.

Poly(A) tail lengths were measured by end-labelling of total RNA (Hsieh & Jones, 1995) and digestion with RNases as described previously (Bralley & Jones, 2001). The presence of heteropolymeric tails at the 3'-ends of S. antibioticus RNAs was demonstrated by cDNA cloning as described previously for S. coelicolor (Bralley & Jones, 2002).

Reproducibility.
The data shown in Figs 2–4 were obtained from duplicate samples isolated from single cultures. Duplicates did not differ by more than 15 % in the assays for mycelial dry weight (Fig. 2), PNPase activity (Fig. 3) or chemical half-lives (Fig. 4). Moreover, samples from independent duplicate cultures did not differ by more than 12–18 % in any of these assays.



View larger version (29K):
[in this window]
[in a new window]
 
Fig. 2. Effects of pnp overexpression on growth of S. antibioticus strains. Thiostrepton ({blacksquare}, 5 µg ml-1; {blacktriangleup}, 15 µg ml-1) was added to 200 ml cultures of the plasmid-bearing strains 12 h after inoculation of GGA medium ({bullet}, 0 µg thiostrepton ml-1 as control). No thiostrepton was added to the culture containing IMRU 3720 alone ({circ}). Duplicate 10 ml samples were removed for the determination of dry weights. The data presented represent means of duplicates from cultures of each strain.

 


View larger version (25K):
[in this window]
[in a new window]
 
Fig. 3. Effects of pnp overexpression on PNPase levels in S. antibioticus. PNPase activity was measured as described in Methods. Results are the means of duplicate assays of extracts prepared from cultures of each strain and are expressed as nmol ADP incorporated (mg protein)-1 in a 15 min incubation at 37 °C. {bullet}, No thiostrepton; {blacksquare}, 5 µg thiostrepton ml-1; {blacktriangleup}, 15 µg thiostrepton ml-1; {circ}, IMRU 3720 alone (no thiostrepton).

 


View larger version (18K):
[in this window]
[in a new window]
 
Fig. 4. Decay kinetics of total RNA in S. antibioticus strains. Labelling, rifampicin inhibition and sampling were performed as described in Methods. Data represent the means of duplicate samples which differed by no more than 15 %. For each analysis the 3H c.p.m. for time points between 0 and 30 min following rifampicin addition were corrected by subtracting the value observed 45 min following the addition of rifampicin, which value represents the level of stable RNA. C.p.m. values were converted to percentages based on the zero time level of uridine incorporation and lines were fitted to the data points by linear regression. {blacktriangleup}, IMRU 3720 (half-life 24 min); {square}, 3720/pIJ8600 (half-life 20 min); {circ}, 3720/pJSE340 (half-life 9·5 min).

 

   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Construction of pJSE340 and pJSE343
pIJ8600 is a streptomycete expression vector based on the thiostrepton-inducible promoter tipA (Holmes et al., 1993; Sun et al., 1999). The vector also contains a ribosome-binding site downstream of tipA, the attachment site for the streptomycete phage {phi}C31 and genes conferring resistance to apramycin and thiostrepton (Kieser et al., 2000). To construct plasmids for overexpression of S. antibioticus pnp, PCR fragments were cloned into the multiple cloning site of pIJ8600, downstream of tipA. The desired orientation of the cloned fragments was verified by restriction digestion and plasmids designated pJSE340 and pJSE343 were obtained by this procedure. pJSE340 contains an approximately 3·2 kb insert that includes sequences flanking both the 5'- and 3'-ends of pnp (Fig. 1). The 5'-flanking sequence contains at least one endogenous promoter, Ppnp, present in the intergenic region between rpsO and pnp, and the pnp ribosome-binding site (Jones & Bibb, 1996). We suspect that the cloned fragment also contains a promoter either upstream of or within rpsO, but we have not yet verified its presence. Ppnp has been shown to function in S. antibioticus (P. Bralley & G. H. Jones, unpublished). pJSE343 contains an approximately 2·5 kb insert lacking Ppnp and the pnp ribosome-binding site but containing the pnp ORF and the same 3'-flanking sequence as pJSE340 (Fig. 1). Thus, pnp expression from pJSE343 was expected to be largely if not totally dependent on thiostrepton induction. Previous studies from our laboratory demonstrated that the tipA promoter would function in S. antibioticus. Integration of pJSE340 and pJSE343 and the control vector pIJ8600 into the host chromosome was verified by Southern blotting (data not shown). Restriction-digested DNA isolated from S. antibioticus strains containing pIJ8600, pJSE340 and pJSE343 revealed a single hybridizing band when pIJ8600 was used as a probe and two bands (one of which represented the endogenous PNPase gene) when pnp was used as a probe.

Effects of pIJ8600, pJSE340 and pJSE343 on growth and PNPase activity
S. antibioticus cultures were grown initially on NZ-amine medium, a rich, complex medium which supports vigorous mycelial growth but not actinomycin production (Gallo & Katz, 1972). After 36–48 h of growth in NZ-amine medium, mycelium was washed and transferred to GGA medium (Gallo & Katz, 1972). Cultures were grown for 12 h and thiostrepton was then added to strains containing pIJ8600 and its derivatives. Growth was measured as mycelial dry weight with the results shown in Fig. 2. To our surprise, and unlike results previously reported for derivatives of pIJ8600 in other Streptomyces species (Hesketh et al., 2001), there was a striking effect of the integration of pIJ8600 derivatives into the S. antibioticus chromosome. As Fig. 2 demonstrates, even when the starting inocula were similar in terms of their dry weights, the strains bearing pIJ8600, pJSE340 and pJSE343 grew much more vigorously than the parent strain, IMRU 3720. This increase in growth rate did not require the presence of thiostrepton and indeed, the antibiotic had little effect on the increase in mycelial dry weight observed for strains containing pIJ8600 derivatives. Because of the nature of the mycelial growth of S. antibioticus, it was not possible to accurately estimate generation times for the cultures analysed in the experiments of Fig. 2. However, it was possible to estimate the time required for a doubling of the mycelial dry weight between 0 and 24 h post-inoculation of GGA media for all the cultures. Thus, about 14 h was required during this period to double the mycelial mass for the parent strain, while the values for strains bearing pIJ8600, pJSE340 and pJSE343 were 7 h, 9 h and 8 h, respectively. The mycelial dry weight in all strains bearing pIJ8600 derivatives decreased between 72 and 96 h post-inoculation, even in cultures lacking thiostrepton (Fig. 2). In contrast, the mycelial dry weight of the parent strain continued to increase until at least 96 h post-inoculation of GGA medium. In addition, strains containing pIJ8600 derivatives did not exhibit the decrease in growth rate between 24 and 48 h post-inoculation observed for the parent strain (Fig. 2). This period, denoted the transition phase, has been frequently observed during the growth of other antibiotic-producing streptomycetes (Vohradsky et al., 2000; Huang et al., 2001) and represents the period between exponential growth and stationary phase when antibiotic synthesis is initiated (Takano et al., 1992). The effects of the pIJ8600 derivatives on growth were clearly not due to the presence of thiostrepton or to the nature of the insert cloned into pIJ8600. We speculate below on possible reasons for the increase in growth rate in the strains bearing pIJ8600 derivatives.

PNPase activity was measured using the polymerization assay with [3H]ADP as a substrate (Jones & Bibb, 1996). Results of the analysis of the thiostrepton-induction of PNPase levels are shown in Fig. 3. It is apparent that thiostrepton had no effect on PNPase levels in the strain bearing pIJ8600 alone. In contrast, thiostrepton at either 5 or 15 µg ml-1 significantly induced pnp expression in strains bearing pJSE340 and 343. At 12 h after the addition of thiostrepton (24 h of total incubation time in GGA medium), the level of PNPase activity in 3720/pJSE340 was nearly three times higher (in cultures containing either 5 or 15 µg thiostrepton ml-1) than that observed in the parent strain or the strain bearing pIJ8600 alone. The activity then decreased, dropping to levels approaching those of the control strains by 96 h post-inoculation. A different pattern of expression was observed for 3720/pJSE343. The thiostrepton-inducible PNPase activity peaked at about 72 h post-inoculation and was significantly higher than that observed in 3720/pJSE340 at 96 h (Fig. 3). The patterns shown in Fig. 3 were observed reproducibly and the data indicate that 5 µg thiostrepton ml-1 was saturating for induction in these systems.

The data of Fig. 3 also demonstrate that there was significant production of PNPase (about twice the levels of control strains) by 3720/pJSE340 and 3720/pJSE343 in the absence of thiostrepton. This observation almost certainly reflects the combination of several effects. In 3720/pJSE340 PNPase levels observed in the absence of thiostrepton probably result in part from the presence of at least one functional promoter, Ppnp, in addition to the tipA promoter in pJSE340. In both 3720/pJSE340 and pJSE343, the activity in the absence of thiostrepton probably also reflects either some level of thiostrepton-independent transcription from tipA, transcription of the cloned inserts from other plasmid-borne promoters, or both.

It is noteworthy that the PNPase levels in the GGA cultures at zero time, corresponding to the activity present in the mycelium harvested from NZ-amine cultures, was significantly elevated compared to the activity in the GGA cultures after 12 h of growth in the latter medium. This was true even for the parent strain and has been observed in previous studies of pnp expression in S. antibioticus (P. Bralley & G. H. Jones, unpublished). We observed reproducibly, as shown in Fig. 3, that the PNPase levels in NZ-amine grown cultures containing pIJ8600 and its derivatives were higher than those observed in cultures of the parent strain, without integrated plasmids. We have no definitive explanation for this observation but speculate that it may be related to the more vigorous growth of the relevant strains in NZ-amine medium as compared with GGA medium.

Overexpression of S. antibioticus pnp decreases chemical half-lives of bulk RNA and poly(A) tail lengths
We anticipated that the observed 2·5–3-fold increase in PNPase activity observed in Fig. 3 would affect both the stabilities and the level of 3'-end modification of S. antibioticus RNAs. To test this hypothesis, we examined those parameters in control strains and in 3720/pJSE340 using 5 µg thiostrepton ml-1 to induce PNPase activity. Thiostrepton was added 12 h post-inoculation of GGA medium and half-life measurements were conducted 12 h later, when the PNPase level was at its maximum in 3720/pJSE340 (Fig. 3). Relevant cultures were then incubated for 30 min with [3H]uridine. RNA synthesis was stopped by adding rifampicin. Pilot studies demonstrated that the concentration of rifampicin used was sufficient to block transcription. Results of the chemical half-life studies are shown in Fig. 4. It can be seen that the kinetics of decay of bulk mRNA were similar for IMRU 3720 and 3720/pIJ8600, yielding chemical half-lives of about 20 min. In contrast, the induction of PNPase in 3720/pJSE340 (a twofold increase as compared with the strain containing pIJ8600 only) led to a significant decrease in the half-life of bulk mRNA, to about 10 min (Fig. 4).

We confirmed by cDNA cloning that, as we reported for S. coelicolor (Bralley & Jones, 2002), the 3'-tails of RNAs from S. antibioticus are heteropolymeric. Thus, the 16S and 23S rRNAs possess tails with significant numbers of G, C and T (U) residues, as does the mRNA for PNPase (Fig. 5). These studies also detected 3'-tails that were significantly longer than those previously reported for S. coelicolor. Of the seven S. antibioticus clones shown in Fig. 5, two had lengths of 98 and 116 residues while of the fourteen S. coelicolor clones, the two longest were 52 and 48 residues in length. The tail associated with 16S rRNA clone 1373 contained a stretch of 34 A residues (Fig. 5). As with S. coelicolor though, the A stretches in the cDNA clones analysed thus far from S. antibioticus are generally no longer than 15–18 residues (see further below). The data shown in Fig. 5 for S. coelicolor have been reported previously (Bralley & Jones, 2002).



View larger version (25K):
[in this window]
[in a new window]
 
Fig. 5. Locations and compositions of mRNA and rRNA poly(A) tails sequenced from cDNA clones isolated by RT-PCR of total RNA from S. antibioticus and S. coelicolor. The position of each tail is reckoned from the translation start codon of mRNA ORFs or from the mature 5'-end of rRNA genes and is used to designate the identity of each clone. The position of the last nucleotide of the mature rRNA or translation stop codon of mRNA is indicated at the top right corner of each box. Genes are not drawn to scale. The data for S. coelicolor have been reported previously (Bralley & Jones, 2002).

 
Despite the observation that streptomycete poly(A) tails are heteropolymeric, we reasoned that we might be able to detect an effect of overexpression of PNPase on overall tail length by looking specifically at the stretches of poly(A). It seemed reasonable to expect that if PNPase overexpression occasioned a shortening of overall tail lengths, the stretches containing only A residues would also be shorter. To test this possibility, the lengths of poly(A) stretches were measured for RNAs isolated from IMRU 3720, 3720/pIJ8600 and 3720/pJSE340 as described previously (Hsieh & Jones, 1995; Bralley & Jones, 2001). The pIJ8600 derivatives were grown for 12 h, 5 µg thiostrepton ml-1 was added and growth was continued for 12 h. RNA was then isolated from these cultures and from a culture of IMRU 3720 that had been grown for 24 h. Results of these analyses are shown in Fig. 6. It should be noted that these analyses were performed on total RNA and not on RNA that had been previously passed through oligo dT cellulose. As was observed for S. coelicolor, there is a strong band in the autoradiogram corresponding to poly(A) lengths of about 4–6 residues. The maximum length observed for IMRU 3720 and 3720/pIJ8600 was about 18 residues, although the smearing above that length may well indicate the presence of lower levels of longer stretches of poly(A). The cDNA cloning data clearly indicate that such stretches exist in the heteropolymeric tails (Fig. 5). In contrast, RNA from 3720/pJSE340 showed discrete poly(A) stretches that were no longer than about 6 residues in length. No discrete bands representing oligo(A) or poly(A) were observed above this length in the RNA preparation from strain 3720/pJSE340 (Fig. 6).



View larger version (58K):
[in this window]
[in a new window]
 
Fig. 6. Measurement of the lengths of oligo and poly(A) stretches found in the 3'-tails of RNAs isolated from S. antibioticus strains. Total RNA was labelled, digested and the digestion products were fractionated on a 12 % polyacrylamide gel as previously described (Bralley & Jones, 2001). The oligo dT18 standard was end-labelled. Strains containing pIJ8600 and pJSE340 were grown in the presence of thiostrepton as described in Methods.

 
Actinomycin and ppGpp levels in S. antibioticus strains overexpressing pnp
Given the observations described above, indicating an effect of PNPase overexpression on RNA stability and poly(A) tail lengths, it seemed possible that there might be an effect of overexpression on specific mRNAs, viz. those involved in antibiotic production. We therefore tested the possibility that such an effect might be manifested in terms of the levels of actinomycin (acm) production in the strains bearing pJSE340 and pJSE343 as compared with the parental strain and that bearing pIJ8600 alone. Thiostrepton induction was performed as described above and actinomycin levels were measured 48–96 h post-inoculation of GGA medium. Results of this analysis are shown in Fig. 7. In all strains listed in Fig. 7, acm production began 36–48 h post-inoculation of GGA medium. Results are expressed as µg acm produced per mg of mycelial dry weight. When expressed in this manner, acm levels produced by the strain containing pIJ8600 were significantly depressed by 48 h post-inoculation compared to the wild-type system. By 72 h, the level of repression of acm production was still pronounced and the increment in acm production between 72 and 96 h was slightly reduced in the strain bearing pIJ8600 as compared with the wild-type strain (Fig. 7). Specifically, it was observed that integration of pIJ8600 into the S. antibioticus chromosome resulted in a 50–60 % decrease in the level of actinomycin production as compared with the parent strain by 96 h post-inoculation, even in the absence of thiostrepton. When 3720/pIJ8600 was grown in the presence of thiostrepton, the level of acm production decreased consistently by an additional 15–20 % as compared with IMRU 3720. In the strain bearing pJSE340 at all three time points examined and in the absence or presence of thiostrepton, acm levels were similar to those observed at the same time points in the strain bearing pIJ8600 grown in the presence of thiostrepton (Fig. 7). At 48 and 72 h post-inoculation acm levels were higher in strain 3720/pJSE343 than in 3720/pJSE340 and were not diminished by growth of 3720/JSE343 in the presence of thiostrepton. Acm levels in cultures of 3720/pJSE343 remained essentially constant between 72 and 96 h post-inoculation in the absence of thiostrepton and increased only slightly (about 15 %) in its presence (Fig. 7). Expression of acm production as µg ml-1 rather than as µg mg-1 mycelium changed the quantitative values for the acm levels produced by the four strains but did not affect the qualitative relationships depicted in Fig. 7.



View larger version (23K):
[in this window]
[in a new window]
 
Fig. 7. Actinomycin production by S. antibioticus strains. Acm levels were measured as the A443 of neutral ethyl acetate extracts of growth medium and converted to concentrations using the conversion factor 1 A443=50·2 µg ml-1 (Jones, 2000). The data represent the means of duplicate samples taken from each of two independent cultures for each strain and are expressed as µg acm produced (mg mycelial dry weight)-1±SEM.

 
ppGpp levels in the strain overexpressing PNPase were measured in 800 ml cultures to which thiostrepton (5 µg ml-1) was added at 12 h post-inoculation of GGA medium. Mycelial samples were removed 12, 24, 48 and 72 h post-inoculation, extracted and analysed by HPLC as described previously (Hoyt & Jones, 1999). The pattern of ppGpp accumulation in all three strains was essentially identical to that previously reported for IMRU 3720 (Hoyt & Jones, 1999), and no difference was observed in the levels of ppGpp or pppGpp in 3720/pIJ8600 or 3720/pJSE340 as compared with IMRU 3720, in the absence or presence of thiostrepton (data not shown). Again, the levels of PNPase at 24 h post-inoculation were about 2·5-fold higher in this experiment in the strain containing pJSE340 as compared with the parent strain and that containing pIJ8600.


   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
A striking observation from the studies reported here is that the integration of pIJ8600 and its derivatives into the S. antibioticus chromosome dramatically increased the growth rate of the resulting strains (Fig. 2). The rate of increase of mycelial dry weight for the strains containing pIJ8600 constructs was approximately twice that of the parent strain. It is a near certainty that the integration of the pIJ8600 derivatives involved the chromosomal attachment site for the {phi}C31 actinophage as that site has been used previously for the integration of plasmids into the S. antibioticus chromosome (Jones et al., 1997). In S. coelicolor, the {phi}C31 attB site is located in a gene whose function is unknown but which has significant sequence similarity to eukaryotic proteins involved in chromosome condensation (www.sanger.ac.uk/Projects/S_coelicolor; Combes et al., 2002). It is not known whether the organization of sequences around the attB site in S. antibioticus mimics that in S. coelicolor but our data clearly indicate that the site is situated in a region that affects mycelial growth in the former organism. This region also appears to affect actinomycin production as the integration of pIJ8600 results in an approximately 50 % decrease in actinomycin levels as compared with the parent strain 96 h post-inoculation of antibiotic production medium (Fig. 7). It is noteworthy that integration of pIJ8600 does not appear to affect growth or antibiotic production in S. coelicolor (Sun et al., 1999; Hesketh et al., 2001). Thus, the effects of integration may be species specific and our data suggest that experiments in which the {phi}C31 attachment site is used for integration should always involve an analysis of the effects of integration on the physiology of the host organism. It should be noted here that Combes et al. (2002) have recently demonstrated that pSET152, a vector which also bears the {phi}C31 attP site, integrates at multiple sites in the genomes of S. coelicolor and Streptomyces lividans. However, using pIJ8600 as the probe, we observed only a single band corresponding to pIJ8600 integration on Southern blots of chromosomal DNA obtained from S. antibioticus exconjugants bearing pIJ8600, pJSE340 or pJSE343 (data not shown). Thus, our data suggest that integration at a single site is responsible for the effects on growth and antibiotic production that we report here.

Thiostrepton induction of PNPase expression in pJSE340 and 343 led to a maximum level of PNPase that was 2–3-fold higher than the levels observed in control strains. The data of Fig. 3 indicate a different pattern of PNPase overexpression in 3720/pJSE340 as compared with 3720/pJSE343. One explanation for this difference is the possibility that, as in E. coli, pnp expression is subject to autoregulation in S. antibioticus. In addition to the pnp ORF, pJSE340 contains the rpsO–pnp intergenic region, a region that functions in the autoregulation of pnp expression in E. coli (Jarrige et al., 2001). If this region functions similarly in S. antibioticus, autoregulation could explain the decrease in PNPase activity to pre-induction levels by 72 h post-inoculation. We are currently examining the structure and function of the rpsO–pnp region of S. antibioticus more closely.

We have confirmed that, as is the case in S. coelicolor, the 3'-tails found on RNAs from S. antibioticus are heteropolymeric (Fig. 5). These tails were detected by cloning cDNAs produced using a primer containing an adapter sequence with 17 T residues at its 3'-end and gene specific 5'-primers for RT-PCR (Bralley & Jones, 2002). Because of potential artifacts that may be associated with this approach the following comments are appropriate. First, we recognize that the 3'-primer probably does not bind to the very end of the RNA templates we used. Thus, the actual lengths of the tails associated with those templates may be greater than those shown in Fig. 5. Our goal was to detect heteropolymeric tails, not to determine their precise lengths. Second, it should be noted that our procedure does detect short stretches of A residues despite the fact that our 3'-primer contains 17 T residues. Thus, we detect a stretch of 6 A residues associated with clone 1583, 8 A residues associated with clone 1981 and 9 A residues in the 3'-tail of clone 2014 (Fig. 5). Third, it is important to note that it was not the goal of our analyses to identify the polyadenylation site(s) at the 3'-ends of mature RNAs. While it is possible to identify such sites using anchored primers, our objective was to examine polyadenylation patterns. It is significant in this regard that most of the cDNA clones shown in Fig. 5 appear to represent decay intermediates. Finally, Mohanty & Kushner (2000b) observed tails in pcnB mutants that were generally homopolymeric at the 3'-end but heteropolymeric at the 5'-end. Our studies do not show such a pattern for the streptomycete 3-tails. Mohanty & Kushner (2000b) note that the pattern they observed reflects the method used to obtain the cDNAs, a method that ensures that the 3'-ends of those cDNAs will possess a homopolymeric stretch of A residues.

Studies in E. coli, and in spinach chloroplasts, have implicated PNPase rather than poly(A) polymerase as the enzyme responsible for the synthesis of heteropolymeric tails in those systems (Mohanty & Kushner, 2000b; Yehudai-Resheff et al., 2001). That we can detect heteropolymeric 3'-tails in wild-type Streptomyces strains strongly suggests that such tails are ubiquitous in streptomycetes and that PNPase is the enzyme responsible for their synthesis. Thus, we argue that PNPase is the major if not the sole 3'-RNA polynucleotide polymerase in Streptomyces.

In E. coli, Mohanty & Kushner (2000a) demonstrated that overexpression of PNPase decreased both the stability of specific mRNAs and the length of the poly(A) tails associated with those mRNAs. We have made similar observations in the studies reported here. PNPase overexpression led to a 50 % decrease in the chemical half-life of bulk mRNA (Fig. 4). The most likely explanation for this observation is the function of PNPase in the phosphorolysis of RNAs in S. antibioticus. While we have demonstrated that the S. antibioticus enzyme catalyses both polymerization and phosphorolysis in vitro (Jones & Bibb, 1996), this report provides the first suggestion that the enzyme also functions phosphorolytically in vivo. We also observed an effect of PNPase overexpression on the distribution of poly(A) stretches in total S. antibioticus RNA. As demonstrated in Fig. 6, discrete bands containing 6–18 A residues could be obtained using end-labelled RNA preparations isolated from strain IMRU 3720 or 3720/pIJ8600. However, strain 3720/pJSE340 produced no discrete bands above about 6 residues in length. Our data suggest that PNPase overexpression leads to a decrease in the average length of the 3'-tails associated with S. antibioticus RNAs and that this decrease is manifested as a decrease in the average length of the poly(A) stretches associated with these tails (Fig. 6).

The integration of pIJ8600 into the S. antibioticus chromosome decreased the ability of strain 3720/pIJ8600 to make actinomycin and this capacity was further decreased by growth of the strain in the presence of thiostrepton. The effects of integration of plasmids bearing the pnp insert on acm production are more complex. Thus, the effects of thiostrepton on strains bearing either pJSE340 or pJSE343 were less pronounced than in the strain bearing pIJ8600 alone and acm levels were somewhat higher in the strain bearing pJSE343 than in 3720/pJSE340 (Fig. 7). In both strains acm levels were lower than those observed in the parental strain. The effects observed in the strains bearing pIJ8600 and derivatives containing pnp may be due to one or all of the following. (1) The lack of a transition phase in the growth of the relevant strains. These strains may thus be less efficient than the wild-type in making the physiological and biochemical modifications required for the initiation of antibiotic production. In this regard it should be noted that the growth kinetics observed in the experiments reported in Fig. 7 were identical to those of Fig. 2. (2) Specific effects of PNPase overexpression on acm production. Fig. 7 shows that acm production was higher in 3720/pJSE343, at least at 48 and 72 h post-inoculation, than in 3720/pJSE340. This could conceivably reflect the different maximum levels of PNPase expression in the two strains (Fig. 3). (3) Effects of integration of plasmids with inserts as opposed to pIJ8600 alone into the S. antibioticus chromosome. It is conceivable that the effect of integration of pIJ8600 on acm production is magnified by the presence of a cloned insert in the plasmid. Further studies will be required to unravel the complex effects on acm production observed in strains overexpressing pnp.

As PNPase from S. antibioticus also functions, in vitro at least, as a pppGpp synthetase, it was of interest in the present study to examine the effects of pnp overexpression on ppGpp levels. Our data indicate no effect of increased levels of PNPase on the accumulation of ppGpp. Levels observed in overexpression strains were identical to those found in controls. In experiments not reported here, we have also found that overexpression of PNPase from pJSE340 in a relA- background failed to promote the accumulation of ppGpp by that strain. These observations strongly suggest that PNPase is not involved in the maintenance of ppGpp levels in vivo. What then is the explanation for the pppGpp synthetase activity of S. antibioticus PNPase? We now believe that activity is an in vitro artifact and that rather than reflecting a biochemical function involved in cellular economy it may reflect the existence of a (p)ppGpp binding site on PNPase that may be involved in regulating the activity of that enzyme. Studies are in progress to examine this hypothesis.

It should be noted here that Rott et al. (2003) have recently demonstrated the presence of heteropolymeric 3'-poly(A) tails associated with RNAs from a wild-type strain of the cyanobacterium Synechocystis PCC6803. They present evidence that this species does not contain a dedicated poly(A) polymerase and that the tails are synthesized by PNPase. As cyanobacteria are believed to be an ancient phylum, these observations are consistent with our hypothesis that PNPase is the ancestral polyadenylating enzyme in bacteria. We are examining this hypothesis by phylogenetic analysis of bacterial PNPases.


   ACKNOWLEDGEMENTS
 
This work was supported by grants from the National Science Foundation (MCB-0133520) and Emory University to G. H. J.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Adamis, T. & Champness, W. (1992). Genetic analysis of absB, a Streptomyces coelicolor locus involved in global antibiotic regulation. J Bacteriol 174, 4622–4628.[Abstract]

Bralley, P. & Jones, G. H. (2001). Poly(A) polymerase activity and RNA polyadenylation in Streptomyces coelicolor A3(2). Mol Microbiol 40, 1155–1164.[CrossRef][Medline]

Bralley, P. & Jones, G. H. (2002). cDNA cloning confirms the polyadenylation of RNA decay intermediates in Streptomyces coelicolor. Microbiology 148, 1421–1425.[Abstract/Free Full Text]

Cao, G. J. & Sarkar, N. (1992). Identification of the gene for an Escherichia coli poly(A) polymerase. Proc Natl Acad Sci U S A 89, 10380–10384.[Abstract]

Cashel, M., Gentry, D. R., Hernandez, V. J. & Vinella, D. (1996). The stringent response. In Escherichia coli and Salmonella: Cellular and Molecular Biology, pp. 1458–1496. Edited by F. C. Neidhardt and others. Washington, DC: American Society for Microbiology.

Chakraburtty, R. & Bibb, M. J. (1997). The ppGpp synthetase gene (relA) of Streptomyces coelicolor A3(2) plays a conditional role in antibiotic production and morphological differentiation. J Bacteriol 179, 5854–5864.[Abstract]

Chakraburtty, R., White, J., Takano, E. & Bibb, M. J. (1996). Cloning and characterization and disruption of a (p)ppGpp synthetase gene (relA) of Streptomyces coelicolor A3(2). Mol Microbiol 19, 357–368.[Medline]

Coburn, G. A. & Mackie, G. A. (1999). Degradation of mRNA in Escherichia coli: an old problem with some new twists. Prog Nucleic Acids Res Mol Biol 62, 55–105.[Medline]

Combes, P., Till, R., Bee, S. & Smith, M. C. M. (2002). The Streptomyces genome contains multiple pseudo-attB sites for the {phi}C-31-encoded site-specific recombination system. J Bacteriol 184, 5746–5752.[Abstract/Free Full Text]

Deutscher, M. & Reuven, N. B. (1991). Enzymatic basis for hydrolytic versus phosphorolytic mRNA degradation in Escherichia coli and Bacillus subtilis. Proc Natl Acad Sci U S A 88, 3277–3280.[Abstract]

Farr, G. A., Oussenko, I. A. & Bechhofer, D. H. (1999). Protection against 3'-to-5'-RNA decay in Bacillus subtilis. J Bacteriol 181, 7323–7330.[Abstract/Free Full Text]

Gallo, M. & Katz, E. (1972). Regulation of secondary metabolite biosynthesis: catabolite repression of phenoxazinone synthase and actinomycin production by glucose. J Bacteriol 109, 659–667.[Medline]

Ghosh, S. & Deutscher, M. (1999). Oligoribonuclease is an essential component of the mRNA degradation pathway. Proc Natl Acad Sci U S A 96, 4372–4377.[Abstract/Free Full Text]

Godefroy-Colburn, T. & Grunberg-Manago, M. (1972). Polynucleotide phosphorylase. In The Enzymes vol. 7, pp. 533–574. Edited by H. D. Boyer. New York: Academic Press.

Grunberg-Manago, M. (1999). Messenger RNA stability and its role in control of gene expression in bacteria and phages. Annu Rev Genet 33, 193–227.[CrossRef][Medline]

Grunberg-Manago, M. & Ochoa, S. (1955). Enzymatic synthesis and breakdown of polynucleotides: polynucleotide phosphorylase. J Am Chem Soc 77, 3165–3166.

Hagege, J. M. & Cohen, S. N. (1997). A developmentally regulated Streptomyces endoribonuclease resembles ribonuclease of Escherichia coli. Mol Microbiol 25, 1077–1090.[Medline]

Herskovitz, M. A. & Bechhofer, D. H. (2000). Endoribonuclease RNase III is essential in Bacillus subtilis. Mol Microbiol 38, 1027–1033.[CrossRef][Medline]

Hesketh, A., Sun, J. & Bibb, M. J. (2001). Induction of ppGpp synthesis in Streptomyces coelicolor A3(2) grown under conditions of nutritional sufficiency elicits actII-orf4 transcription and actinorhodin biosynthesis. Mol Microbiol 39, 136–141.[CrossRef][Medline]

Holmes, D. J., Caso, J. & Thompson, C. (1993). Autogenous transcriptional activation of a thiostrepton-induced gene in Streptomyces lividans. EMBO J 12, 3183–3191.[Abstract]

Hoyt, S. & Jones, G. H. (1999). relA is required for actinomycin production in Streptomyces antibioticus. J Bacteriol 181, 3824–3829.[Abstract/Free Full Text]

Hsieh, C.-J. & Jones, G. H. (1995). Nucleotide sequence, transcriptional analysis and glucose regulation of the phenoxazinone synthase gene from Streptomyces antibioticus. J Bacteriol 177, 5740–5747.[Abstract]

Huang, J., Lih, C.-J., Pan, K.-H. & Cohen, S. N. (2001). Global analysis of growth phase responsive gene expression and regulation of antibiotic biosynthetic pathways in Streptomyces coelicolor using DNA microarrays. Genes Dev 15, 3183–3192.[Abstract/Free Full Text]

Jarrige, A.-C., Mathy, N. & Portier, C. (2001). PNPase autocontrols its expression by degrading a double-stranded structure in the pnp mRNA leader. EMBO J 20, 6845–6855.[Abstract/Free Full Text]

Jarrige, A.-C., Bréchemier-Baey, D., Mathy, N., Duché, O. & Portier, C. (2002). Mutational analysis of polynucleotide phosphorylase from Escherichia coli. J Mol Biol 312, 397–409.[CrossRef]

Jones, G. H. (1994a). Purification and properties of ATP:GTP 3'-pyrophosphotransferase (guanosine pentaphosphate synthetase) from Streptomyces antibioticus. J Bacteriol 176, 1475–1481.[Abstract]

Jones, G. H. (1994b). Activation of ATP-GTP 3'-pyrophosphotransferase (guanosine pentaphosphate synthetase) in Streptomyces antibioticus. J Bacteriol 176, 1482–1487.[Abstract]

Jones, G. H. (2000). Actinomycin synthesis persists in a strain of Streptomyces antibioticus lacking phenoxazinone synthase. Antimicrob Agents Chemother 44, 1322–1327.[Abstract/Free Full Text]

Jones, G. H. & Bibb, M. J. (1996). Guanosine pentaphosphate synthetase from Streptomyces antibioticus is also a polynucleotide phosphorylase. J Bacteriol 178, 4281–4288.[Abstract]

Jones, G. H., Paget, M. S. B., Chamberlin, L. & Buttner, M. J. (1997). Sigma-E is required for the production of the antibiotic actinomycin in Streptomyces antibioticus. Mol Microbiol 23, 169–178.[Medline]

Kieser, Y., Bibb, M. J., Buttner, M. J., Chater, K. F. & Hopwood, D. A. (2000). Practical Streptomyces Genetics. Norwich, UK: John Innes Foundation.

Lisitsky, I., Klaff, P. & Schuster, G. (1996). Addition of destabilizing poly(A) rich sequences to endonuclease cleavage sites during the degradation of chloroplast RNA. Proc Natl Acad Sci U S A 93, 13398–13403.[Abstract/Free Full Text]

Littauer, U. Z. & Soreq, H. (1982). Polynucleotide phosphorylase. In The Enzymes vol. 15, pp. 517–553. Edited by H. D. Boyer. New York: Academic Press.

Luttinger, A., Hahn, J. & Dubnau, D. (1996). Polynucleotide phosphorylase is necessary for competence development in Bacillus subtilis. Mol Microbiol 19, 343–356.[Medline]

Mohanty, B. K. & Kushner, S. R. (2000a). Polynucleotide phosphorylase, RNase II and RNase E play different roles in the in vivo modulation of polyadenylation in Escherichia coli. Mol Microbiol 36, 982–994.[CrossRef][Medline]

Mohanty, B. K. & Kushner, S. R. (2000b). Polynucleotide phosphorylase functions both as a 3'-5' exonuclease and a poly(A) polymerase in Escherichia coli. Proc Natl Acad Sci U S A 97, 11966–11971.[Abstract/Free Full Text]

Ohnishi, Y., Nishiyama, Y., Sato, R., Kameyama, S. & Horinouchi, S. (2000). An oligoribonuclease gene in Streptomyces griseus. J Bacteriol 182, 4647–4653.[Abstract/Free Full Text]

Price, B., Adamis, T. & Champness, W. (1999). A Streptomyces coelicolor antibiotic regulatory gene, absB, encodes an RNase III homolog. J Bacteriol 181, 6142–6151.[Abstract/Free Full Text]

Rauhut, R. & Klug, G. (1999). mRNA degradation in bacteria. FEMS Microbiol Rev 23, 353–370.[CrossRef][Medline]

Regnier, P. & Arraiano, C. M. (2000). Degradation of mRNA in bacteria: emergence of ubiquitous features. Bioessays 22, 235–244.[CrossRef][Medline]

Rott, R., Zipor, G., Portnoy, V., Liveanu, V. & Schuster, G. (2003). RNA polyadenylation and degradation in cyanobacteria are similar to the chloroplast but different from E. coli. J Biol Chem 278, 15771–15777.[Abstract/Free Full Text]

Sun, J., Kelemen, G. H., Fernandez-Abalos, J. M. & Bibb, M. J. (1999). Green fluorescent protein as a reporter for spatial and temporal gene expression in Streptomyces coelicolor A3(2). Microbiology 145, 2221–2227.[Abstract/Free Full Text]

Symmons, M., Jones, G. H. & Luisi, B. (2000). A duplicated fold is the structural basis for polynucleotide phosphorylase catalytic activity, processivity and regulation. Structure 8, 1215–1226.[CrossRef][Medline]

Takano, E., Gramajo, H. C., Strauch, E., Andres, N., White, J. & Bibb, M. J. (1992). Transcriptional regulation of the redD transcriptional activator gene accounts for growth-phase dependent production of the antibiotic undecylprodigiosin in Streptomyces coelicolor A3(2). Mol Microbiol 6, 2797–2804.[Medline]

Vohradsky, J., Li, X. M., Dale, G., Folcher, M., Nguyen, L., Viollier, P. H. & Thompson, C. J. (2000). Developmental control of stress stimulons in Streptomyces coelicolor revealed by statistical analysis of global gene expression patterns. J Bacteriol 182, 4979–4986.[Abstract/Free Full Text]

Yehudai-Resheff, S., Hirsh, M. & Schuster, G. (2001). Polynucleotide phosphorylase functions as both an exonuclease and a poly(A) polymerase in spinach chloroplasts. Mol Cell Biol 21, 5408–5416.[Abstract/Free Full Text]

Zuo, Y. & Deutscher, M. (2001). Exoribonuclease superfamilies: structural analysis and phylogenetic distribution. Nucleic Acids Res 29, 1017–1026.[Abstract/Free Full Text]

Received 7 March 2003; revised 16 April 2003; accepted 7 May 2003.