Institut für Mikro- und
Molekularbiologie der Justus-Liebig-Universität Giessen,
35392 Giessen, Germany
We provide experimental evidence for RNase III-dependent
processing in helix 9 of the 23S rRNA as a general feature of many species in the alpha subclass of Proteobacteria
(alpha-Proteobacteria). We investigated 12 Rhodobacter, Rhizobium,
Sinorhizobium, Rhodopseudomonas, and
Bartonella strains. The processed region is characterized by the presence of intervening sequences (IVSs). The 23S rDNA sequences
between positions 109 and 205 (Escherichia coli numbering) were determined, and potential secondary structures are proposed. Comparison of the IVSs indicates very different evolutionary rates in
some phylogenetic branches, lateral genetic transfer, and evolution by
insertion and/or deletion. We show that the IVS processing in
Rhodobacter capsulatus in vivo is RNase III-dependent
and that RNase III cleaves additional sites in vitro. While all
IVS-containing transcripts tested are processed in vitro by RNase
III from R. capsulatus, E. coli RNase
III recognizes only some of them as substrates and in these substrates
frequently cleaves at different scissile bonds. These results
demonstrate the different substrate specificities of the two enzymes.
Although RNase III plays an important role in the rRNA,
mRNA, and bacteriophage RNA maturation, its substrate specificity
is still not well understood. Comparison of the IVSs of helix 9 does
not hint at sequence motives involved in recognition but reveals that
the "antideterminant" model, which represents the most recent
attempt to explain the E. coli RNase III specificity in
vitro, cannot be applied to substrates derived from
alpha-Proteobacteria.
 |
INTRODUCTION |
rRNA and ribosomal DNA (rDNA)
sequences are widely used for bacterial classification, phylogenetic
studies, and identification purposes. Therefore, it is important to
know which sequences are removed from the primary rRNA transcript
during rRNA maturation. Ten years ago, it was believed that
fragmented 23S rRNAs in bacteria are an exception. Now it is known
that this phenomenon is widespread. Fragmented 23S rRNAs are found
in representatives of many species of the alpha, gamma, and epsilon
subclasses of Proteobacteria (alpha-, gamma-, and
epsilon-Proteobacteria, respectively) (4, 9, 12, 13,
16, 19, 21, 23, 28-31, 33, 35, 37) and in Spirochaeta
(26). In most cases, the processing sites are characterized
by the presence of internal transcribed spacers, or so-called
intervening sequences (IVSs), which are removed without splicing.
Instead, the resulting fragments are held together by the compact
structure of the ribosomes which remain functional.
In all bacteria with fragmented 23S rRNA investigated thus far,
with the exception of those belonging to
alpha-Proteobacteria, two possible processing sites were
found: at positions 540 (helix 25) and 1120 (helix 45)
(Escherichia coli numbering). The occurrence of IVSs at
these positions is sporadic: only some strains of a bacterial species
possess fragmented 23S rRNA, and often even in a given bacterial
strain not all rrn operons contain IVSs (4, 12, 21,
23). In alpha-Proteobacteria additional fragmentation sites were found. In Rhodobacter capsulatus and
Rhodobacter sphaeroides, the 23S rRNA is fragmented
at position 1200 (helix 46; E. coli numbering) due to IVS
processing (16, 27). In domain III of 23S rRNA of some
Rhizobium and Agrobacterium strains sporadic fragmentation without the involvement of IVSs has been described (28, 30). In all Rhizobium,
Agrobacterium, and Bradyrhizobium strains
investigated thus far, fragmentation near position 130 (E. coli numbering) of 23S rRNA was found (9, 28-30).
In E. coli, nucleotides 130 to 148 of the 23S rRNA form
the small helix 9, which is substantially extended due to extra
stem-loop structures (IVSs) in rhizobial and agrobacterial strains,
both showing fragmentation in this region (28, 30). The
occurrence of IVSs in helix 9 of these strains is not sporadic and may
be a common feature of a major part of the
alpha-Proteobacteria. We investigated the processing of
helix 9 of 23S rRNA in 12 Rhodobacter,
Rhodopseudomonas, Rhizobium,
Sinorhizobium, and Bartonella strains. Recently,
fragmentation in helix 9 of 23S rRNA in Rhodopseudomonas
palustris was confirmed and such fragmentation in
Rhodobacter species was predicted (37). 23S
rRNA fragmentation at this position in R. sphaeroides was predicted already 10 years ago (7),
but it was not confirmed until now. The processing mechanism of the
known helix 9 fragmentation in Rhizobiaceae was also not
investigated until now. On the basis of sequence comparisons (EMBL
database), we also predicted fragmentation in helix 9 of 23S rRNA
in Bartonella bacilliformis (22). To confirm the
occurrence of such fragmentation in Bartonella, we investigated the processing of helix 9 of Bartonella
henselae ATCC 49882 in vitro.
The IVS processing in helices 25, 45, and 46 of eubacterial 23S
rRNA is catalyzed by RNase III (4, 27). This
endoribonuclease cleaves rRNA precursors during maturation of
rRNA and is also involved in the maturation of some mRNA and
bacteriophage RNA species (8, 14). The enzyme is not
essential for viability; nevertheless, its primary sequence is highly
conserved (27). RNase III recognizes and precisely
cleaves double-helical RNA structures with a 20-bp minimal length.
These double helices contain some unpaired residues and do not exhibit
any consensus sequence (24). The substrate specificity of
the enzyme is poorly understood. An attempt to explain the E. coli RNase III substrate specificity was made using the
concept of "antideterminants": a cleavage site is defined by the
absence of disfavored sequence motifs in its vicinity (38).
The RNases III from E. coli and R. capsulatus (RNase IIIEc and RNase
IIIRc, respectively) are very similar in their
amino acid sequences, but it was shown that they show significant
differences in binding and cleavage of certain substrates in vitro
(5). These differences between both enzymes are not well
understood. It is also not known which properties make an RNA a
"good" substrate for the RNase IIIRc. One way to address these questions is to compare different
natural RNase IIIRc specific substrates and
their interaction in vitro with both purified enzymes.
The aim of this work was to investigate the occurrence of IVS-dependent
fragmentation in helix 9 of 23S rRNA in different phototrophic,
symbiotic, and pathogenic alpha-proteobacterial species. Further,
we analyzed the mechanism of 23S rRNA processing in this region.
The study of the IVSs found in helix 9 of alpha-proteobacterial 23S
rRNA and their RNase III-dependent cleavage can help us to better understand the mode of action of this enzyme as well as the
evolution of the rRNA genes.
| |
MATERIALS AND METHODS |
Bacterial strains and culture conditions.
The R. capsulatus strains used in this study were the wild-type strains
37b4 (DSM 938) and B10 (18) and the mutant strain Fm65
(16). R. sphaeroides wild-type strains
WS8 (32) and 17023 were used. R. palustris
5D and R. sphaeroides 17023 were obtained from G. Drews, Freiburg, Germany. All Rhodobacter strains as
well as the strains R. palustris 5D and
Rhodospirillum rubrum DSM 107 were grown in a
minimal malate salt medium (6). During growth of the
R. capsulatus Fm65 strain, carrying plasmid pRK2fm1
(16), tetracycline was added at a final concentration of 1 µg ml
1.
The rhizobial strains Rhizobium giardinii H152,
Rhizobium gallicum R602 (1), and
Sinorhizobium fredii MSDJ 1536 were obtained from N. Amarger
(INRA-CMSE, Dijon, France). Rhizobium etli strains CFN42
(USDA 9032) and Viking I (USDA 2743) and Rhizobium
leguminosarum ATCC 10004 (USDA 2370) were obtained from D. K. Jones (U.S. Department of Agriculture/Agricultural Research Service,
Beltsville Rhizobium Germplasm Collection). They were grown on tryptic
yeast media (3).
E. coli JM109 (Stratagene) was grown on standard I medium (Difco).
Oligonucleotides (primers).
The oligonucleotide used for
hybridizations was 5'-GGGTTTCCCCATTCGGAAA (23Sup130, 19 nucleotides [nt], complementary to the highly conserved positions 112 to 130 of the 23S rRNA; E. coli numbering). For PCR
amplification of the regions between positions 109 and 205 of rDNA
(E. coli numbering), the following primers were used:
5'-GGGGGGAA TTCTAATACGACTCACTATAG(G/A)AT(T/G)TCCGAATGGGGAAACCC- 3'
(23S-IVS-sense primer, 50 nt; the EcoRI site is
underlined; the T7 promoter region for transcription initiation is in
boldface), corresponding to the rDNA positions 109 to 130, and
5'-GGGGGAAGCTTCTTAG(T/A)(A/T)GTTTC(T/A)GTTCC-3' (23S-IVS-antisense primer, 30 nt; the HindIII site
is underlined), corresponding to the rDNA positions 185 to 205 (E. coli numbering). All oligonucleotides were synthesized
on a 380B DNA Synthesizer (Applied Biosystems).
Isolation, amplification, and analysis of nucleic acids.
Total DNA from alpha-Proteobacteria was isolated according
to the method of Ausubel et al. (2). Hot phenol RNA
isolation, Northern blotting, and hybridization were performed
with standard methods (10).
Aliquots of 10 pmol of primer were labeled for 30 to 60 min at 37°C
with 20 µCi of [
-32P]ATP using polynucleotide kinase
and subsequently purified with a NucTrap push column (Stratagene).
Labeled oligonucleotides were used for hybridization and primer
extension analysis.
PCR was carried out in a final volume of 50 µl with 200 ng of total
DNA as a template, using 0.8 U of Taq DNA polymerase
(Promega) at an annealing temperature of 42°C (45 s), followed by
extension at 72°C (30 s). Cycles were repeated 35 times. The
resulting PCR products were purified from 3.5% small DNA
low-melting-point agarose gels (FMC-Biozym).
The purified PCR products were used directly for in vitro transcription
or for direct cycle sequencing using the ABI Prism Dye Terminator Cycle
Sequencing kit (Perkin-Elmer). The sequencing reaction was done
according to the protocol of the manufacturer. The program used for
cycle sequencing was as follows: initial denaturation for 50 s at
96°C, with 25 cycles of 20 s at 96°C, 20 s at 55°C, and
2 min at 60°C. The resulting products were ethanol precipitated and
loaded on a 373 DNA Sequencer (Perkin-Elmer).
The agarose gel-purified PCR products encompassing positions 109 to 205 of 23S rDNA (E. coli numbering) were cut with
HindIII and EcoRI and cloned into pUC18
vectors. The resulting constructs were purified with Qiagen-tip 100 and
used for manual sequencing with the T7Sequencing Kit
(Pharmacia Biotech) and [
-35S]dATP.
In vitro transcription of RNAs and the RNase III assays.
In vitro transcriptions using T7 RNA polymerase and purification of the
internally labeled transcript on 10% polyacrylamide gel were performed
as previously described (5).
Each assay was performed in a 10-µl reaction volume. The cleavage
buffer consisted of 30 mM Tris-HCl (pH 7.5), 10 mM MgCl2, 130 or 250 mM KCl, and 5% glycerol. The concentration of RNase III
was approximately 3 nM (homodimer), and that of the substrate was
20 or 40 nM. The assays were performed for 3 min (at a 20 nM substrate
concentration) or for 5 min (at a 40 nM substrate concentration) at
35°C and stopped by adding 7 µl of formamide-containing dye.
Reaction products were denatured at 65°C for 5 min, placed on ice,
and analyzed by autoradiography after separation on a 10%
polyacrylamide-7 M urea gel. Quantification of the amount of the
cleaved transcripts was performed with a Bioimager (Fuji BAS 1000) and
TINA software (Raytest).
Mapping of RNA 5' ends by primer extension.
To determine the
exact RNase III cleavage position at the 3'-processing site in
helix 9 of 23S rRNA, we used primer extension analysis. The primer
for the extension reaction was the 23S-IVS-antisense primer,
corresponding to the 23S rDNA positions 185 to 205 (E. coli numbering). Unlabeled in vitro transcripts were incubated with purified R. capsulatus His6-RNase
III or E. coli RNase III in cleavage buffer containing
130 mM KCl at 35°C for 15 min. After phenol extraction and ethanol
precipitation, the processed RNA substrate was treated as previously
described (11). Primer extension reactions were also
performed using 2 µg of total RNA to determine the 5' ends of the
large rRNA fragment originating of in vivo processing of helix 9 in
the 23S rRNA. Radioactively labeled sequencing reactions of the
cloned DNA template were loaded onto the same gel to map the position
of the cleavage site for RNase III.
Sequence analysis.
Alignments were performed manually and
online using the CLUSTAL W computer program
(http://www2.ebi.ac.uk/clustalw/). Putative rRNA secondary
structure models were obtained online using the folding program MFOLD
(20, 39). For this analysis rDNA sequences obtained in this
work, as well as previously published sequences of some other 23S rDNAs
taken from the EMBL database, were used.
Accession numbers of nucleotide sequences.
The nucleotide
sequences determined in this work were deposited in the EMBL databank
under EMBL accession nos. AJ251255 to AJ251267.
| |
RESULTS |
Fragmentation in helix 9 of 23S rRNA in some
alpha-Proteobacteria.
Fragmentation in helix 9 creates a
small 5' segment of 23S rRNA, which corresponds to the first
approximately 135 nt of the intact 23S rRNA. This small rRNA
migrates in 1.2% agarose formaldehyde gels in one spot, together with
the 5S rRNA, and therefore cannot be visualized easily by ethidium
bromide staining. Northern hybridization helps to detect this 5'
segment of 23S rRNA (9, 28). Examples of fragmented and
intact 23S rRNAs observed in different bacterial strains are shown
in Fig. 1. As expected, the radioactive
probe complementary to 23S rRNA nt 112 to 130 (E. coli
numbering) hybridized to the intact 23S rRNA of E. coli,
which is 2.9 kb long (Fig. 1, lane 1). In contrast, the 23S rRNA of
R. capsulatus 37b4 is fragmented in helix 9 and the
small 5' segment of the 23S rRNA was detected with this probe (Fig.
1, lane 2). The observed helix 9 fragmentation in R. capsulatus 37b4 was found to be RNase III dependent in vivo.
In the RNase III-deficient strain R. capsulatus Fm65, fragmentation is not observed and the probe hybridizes with the
unprocessed 2.9-kb 23S rRNA. After complementation of the RNase III-deficient mutant with the plasmid pRK2fm1,
which contains the gene encoding RNase
IIIRc, the fragmentation of the 23S rRNA was
restored to the wild-type R. capsulatus 37b4 pattern (Fig. 1, lanes 2, 3, and 4). The helix 9 fragmentation was observed in
all investigated wild-type strains of alpha-Proteobacteria except R. rubrum DSM 107 (Fig. 1, lane 7).
View larger version (83K):
[in this window]
[in a new window]
|
FIG. 1.
Presence of short rRNA corresponding to
approximately the first 130 nt of the 23S rRNA in some
alpha-Proteobacteria as shown by Northern hybridization of
total RNA separated on a 1.2% agarose formaldehyde gel with the
radioactively labeled oligonucleotide 23Sup130. Lanes: 1, E. coli JM109; 2, R. capsulatus 37b4; 3, R. capsulatus Fm65; 4, R. capsulatus
Fm65 (pRK2fm1); 5, R. sphaeroides 17023; 6, R. palustris 5D; 7, R. rubrum DSM
107; 8, R. gallicum R602; 9, R. giardinii H152.
|
|
The lower intensity of the hybridization signals obtained in the lanes
containing RNA isolated from E. coli, R. sphaeroides, and R. rubrum (Fig. 1, lanes
1 and 5 to 7) is due to mismatches between the probe and the target sequences.
Primary and proposed secondary structures of helix 9 of 23S
rRNA.
The 23S rDNA region between positions 109 and 205 (E. coli numbering) was amplified and sequenced. IVSs were
found in helix 9 of all studied alpha-proteobacterial strains, with the
exception of R. rubrum DSM 107. The 23S rRNA
sequences were folded with the aid of the MFOLD computer program
(20, 39). The proposed secondary structures of the helices 9 are shown in Fig. 2.
Previously published helix 9 sequences from some phylogenetically
related bacterial strains were also included in the analysis (Fig. 2).
View larger version (63K):
[in this window]
[in a new window]
|
FIG. 2.
Models of potential secondary structure of helix 9 in
the 23S rRNA primary transcript of the following strains (asterisks
indicate sequences obtained from the data bank, the other sequences
were determined in our laboratory, and EMBL accession numbers are in
parentheses): A, R. rubrum DSM 107 (AJ251267); B,
R. sphaeroides WS8 (AJ251261); C, R. sphaeroides 17023 (AJ251260); D, R. capsulatus
B10 (AJ251256); E, R. capsulatus 37b4 (AJ251255); F,
R. capsulatus DSM 938* (reference
11); G, R. palustris 5D (AJ251262);
H, B. japonicum 110* (reference 17); I,
B. bacilliformis KC 584* (reference 22);
J, B. henselae ATCC 49882 (AJ251257); K, S. fredii MSDJ 1536 (AJ251258); L, R. giardinii H152
(AJ251263); M, R. etli CFN 42 (AJ251265); R. etli Viking I (AJ251266); O, R. gallicum R602
(AJ251259), R. leguminosarum ATCC 10004 (AJ251264). (B
to F) Rhodobacter group of helices. Boxes with
highly conservative base pair occupation, specific for this group, are
indicated. The differences between the sequences shown in panels E and
F are in boldface letters. (G to P) Rhizobium-Bradyrhizobium
group of helices. Boxes with highly conservative base pair occupation,
specific for this group, are indicated. In panels G and H, sequences of
high similarity around the putative deletion and/or insertion site are
underlined. Arrows indicate the approximate positions of the RNase
III processing sites as determined by RNA fragment length estimation
(Table 4). Arrows on the left side of the helices indicate
5'-processing sites; arrows on the right side of the helices indicate
3'-processing sites. Filled arrowheads indicate primary processing
sites; empty arrowheads indicate secondary processing sites.
|
|
When helix 9 sequences are compared, the first 4 to 6 bp are highly
conserved. Possibly, they correspond phylogenetically to the canonical
helix 9 of bacteria lacking an IVS in this region. A high degree of
conservation can also be observed in the next approximately 25 bp of
the helix, which are part of the IVS. The first 30 bp of the helices 9 were aligned, and the percentage of identically occupied positions was
calculated (Table 1). Values of 50 to
60% indicate difficulties in aligning the sequences; values of <50%
indicate that an alignment of the IVS sequences was impossible. The
analyzed sequences can be divided into two groups, which reflect well
the phylogenetic relationships between the corresponding bacteria.
View this table:
[in this window]
[in a new window]
|
TABLE 1.
Percentage of sequence identity found by comparison of
the first 30 bp of the helix 9 of 23S rRNA of various
bacterial strainsa
|
|
One of them is the Rhodobacter group (Fig. 2B to
F, Table 1). The identically occupied positions in the sequences of the R. sphaeroides strains WS8 and 17023 are
concentrated in four blocks (Fig. 2B and C). A few differences were
found in the sequences of the R. capsulatus strains
37b4 and DSM 938 (Fig. 2E and F). The sequence of R. capsulatus B10 is highly different. High divergence is also
observed when R. sphaeroides and R. capsulatus sequences are compared (Table 1). Thus, high
variability is observed in the Rhodobacter group
of sequences.
The second group includes sequences from representatives of five
bacterial genera, i.e., the Rhizobium-Bradyrhizobium group (Fig. 2G to P; Table 1). Frequently, the same high degrees of conservation (
70%) were observed between sequences belonging to
representatives of the same species compared to those between different
species and even genera (see boldfaced values in Table 1). In the
secondary structure proposals, we found blocks with very conservative
base-pair occupation in nearly identical positions (Fig. 2G to P). We
conclude that the investigated helix 9 region of the
Rhizobium-Bradyrhizobium group is more conservative than that of the Rhodobacter group.
The apical part of helices 9 is highly variable. Obviously, the IVSs of
the Bradyrhizobium and Rhodopseudomonas strains
evolved from each other by deletion and/or insertion events. The
first 30 bp of helices 9 of both strains exhibit 97% sequence identity (Table 1), but the B. japonicum helices 9 are approximately
50 nt longer than those of the R. palustris
strains. The additional nucleotide stretch in the B. japonicum helix 9 is inserted in a region with an almost identical
sequence in both the R. palustris and the B. japonicum helices (these sequences are underlined in Fig. 2G and H).
We determined the GC contents of the helix 9 sequences and compared
them with each other as well as with the GC contents of already-sequenced rrn operons or 23S rRNA sequences of
phylogenetically closely related bacterial strains (Table
2). Generally, the GC content of the
helices 9 is lower than that of the overall rRNA sequences.
Both values are lower than the GC content of genomes of the
Rhodobacter or Rhizobium species,
which is >65%. Noteworthy is the large difference between the GC
content of the Bartonella helix 9 sequences compared to that
of the respective 23S rRNA (Table 2).
View this table:
[in this window]
[in a new window]
|
TABLE 2.
GC content of helix 9 of 23S rRNA in comparison with
the overall GC content of complete rrn operons or
23S rRNA
|
|
The helices 9 of R. capsulatus 37b4 and R. gallicum R602 have the highest overall GC content (Table 2).
Separate analysis of the first 30 bp of these helices revealed that
these regions have a GC content of 40 to 45%, which is typical for
most of the other helices 9 studied here. In contrast, the GC content
of the apical helix 9 regions of these two strains is very high
(>70%). This remarkable difference between the GC content of the two
helix 9 regions was also found in R. leguminosarum ATCC
10004 (Table 2), suggesting that they are of different origin.
In vitro processing of helix 9 of 23S rRNA by RNase
III.
The purified PCR amplificates were used as templates for in
vitro transcription. The transcripts were assayed with RNases III
from R. capsulatus and E. coli, purified in
our laboratory. Both enzymes differ in their preferences for
monovalent cations during catalysis. For RNase
IIIEc, the standard assay buffer contains
250 mM KCl; for RNase IIIRc, the optimal KCl
concentration is 130 mM. Both enzymes exhibit lower activity at higher
KCl concentrations. It is known that RNase
IIIEc can cleave suboptimal substrates at
monovalent cation concentrations of <250 mM (38). On the
other hand, at 250 mM KCl RNase IIIRc does
not process in vitro its natural in vivo substrate, an IVS in helix 46 of 23S rRNA (5). We performed activity assays for both enzymes under both KCl concentrations.
In each experiment the amount of radioactivity in distinct product
bands was determined. The amount of cleaved substrate was calculated,
comparing this value with the decrease in uncleaved substrate. The
results are shown in Table 3.
View this table:
[in this window]
[in a new window]
|
TABLE 3.
Percentage of in vitro-cleaved transcripts containing
helix 9 of 23S rRNA using E. coli and R. capsulatus RNases III at 130 mM and 250 mM KCla
|
|
All transcripts were almost completely cleaved by RNase
IIIRc at 130 mM KCl. Interestingly, at 250 mM
KCl the transcripts from the Rhizobium group are much better
substrates for RNase IIIRc than the
Rhodobacter transcripts.
The transcripts derived from the Rhodobacter
strains are not cleaved by RNase IIIEc at
both KCl concentrations. All other transcripts can be cleaved by
RNase IIIEc at 130 mM KCl (with the
exception of the S. fredii MSDJ 1536 transcript, where only
2% is processed), but most of them cannot be used as efficient
substrates for this enzyme under its specific assay conditions of 250 mM KCl. Only two exceptions were found: 70 to 75% of the transcripts
from B. henselae ATCC 49882 and R. palustris 5D are processed at a high monovalent ion concentration.
Even transcripts containing helix 9 with very similar primary and
secondary structures differ markedly in their interaction with
RNase IIIEc. For example, the transcripts
derived from the strains S. fredii MSDJ 1536 and
R. giardinii H152 have 80% identity in their overall
helix 9 sequences. Nevertheless, the first transcript is less reactive
with RNase IIIEc than the latter transcript
(Table 3). Moreover, the transcripts derived from the R. etli strains CFN 42 and Viking I also show 81% sequence identity
in the first 30 bp of their helices 9 (Table 1). This is the region
recognized and cleaved by RNase III (see below). Despite their very
similar sequences, the reactivities of these substrates with RNase
IIIEc are markedly different (Table 3).
The RNase III cleavage patterns of the different substrates used
are shown in Fig. 3. Despite helix 9 variability, they all share bands of similar length because the
transcripts also include 22 nt upstream from helix 9 (containing helix
8 sequences) and approximately 70 nt downstream from helix 9 (containing helix 10, helix 11, and downstream primer sequences).
Usually, RNase III cleaves in both strands of a duplex, and the 5'-
and 3'-processing sites are separated by only 2 bp (24). The
scissile bonds in helices 9 studied here are localized after the first
18 bp of the helix (Fig. 2). After complete cleavage at the 5'- and
3'-processing sites, a 5' fragment of approximately 40 nt, a 3'
fragment of approximately 90 nt, and additional fragments corresponding
to internal parts of the processed IVS arise (Fig. 2 and 3). The length
of the transcripts and the corresponding fragments are summarized
in Table 4.
View larger version (72K):
[in this window]
[in a new window]
|
FIG. 3.
In vitro processing of transcripts containing helix 9 of
23S rRNA by R. capsulatus (Rc) and E. coli (Ec) RNase III at low and high monovalent ion
concentrations. L, 130 mM KCl; H, 250 mM KCl; C, uncleaved substrate.
Rp, R. palustris; Rc,
R. capsulatus; Bh, B. henselae;
Rhe, R. etli; Rs, R. sphaeroides; Rhgi, R. giardinii;
Sf, S. fredi; Rhga, R. gallicum; Rhl, R. leguminosarum.
|
|
View this table:
[in this window]
[in a new window]
|
TABLE 4.
Summary of the approximate lengths of the helix 9 containing transcripts from the strains studied here and the
approximate lengths of the fragments arising after in vitro cleavage of
these transcripts by RNases III from R. capsulatus
and E. colia
|
|
In R. capsulatus 37b4 (Fig. 3, lanes 6 to 10) and B10
(Fig. 3, lanes 46 to 50) transcripts and in the R. sphaeroides 17023 (Fig. 3, lanes 56 to 58) transcript the
3'-processing site is the primary cleavage site used by RNase
IIIRc, and only a small amount of these
substrates is also cleaved at the 5'-processing site (indicated also in
Fig. 2 and Table 4). In the R. sphaeroides WS8
(Fig. 3, lanes 21 to 25) transcript, as well as in all other transcripts used here, both sites are processed by RNase
IIIRc at 130 mM KCl (Fig. 2 and 3; Table 4).
Both RNases III create different cleavage patterns when they
process some of the transcripts (marked by asterisks in Table 4;
compare with Fig. 3). This may be due to the processing of different scissile bonds. Apparently, multiple scissile bonds are
cleaved at the processing sites in most of these transcripts. In their
cleavage patterns multiple 5'-end, 3'-end, and internal fragments occur
which differ in their lengths by a few nucleotides (marked by two
asterisks in Table 4; compare with Fig. 3). A typical example for these
phenomena is the processing pattern of the B. henselae ATCC
49882 transcript (Fig. 3, lanes 11 to 15). We performed primer
extension to analyze the 3'-processing sites of this transcript
in vitro. In addition, we determined the exact in vitro and in vivo
RNase III 3'-processing sites in helix 9 of the strains
R. capsulatus 37b4 and R. palustris 5D.
Primer extension determines the exact RNase III 3' cleavage
sites.
The results of primer extension analyses are shown in Fig.
4 and are schematically summarized in
Fig. 5.
View larger version (64K):
[in this window]
[in a new window]
|
FIG. 4.
Primer extension analysis determines the rRNA 5'
ends obtained during in vitro cleavage of the transcripts from
R. capsulatus 37b4 (A), R. palustris 5D
(B), and B. henselae ATCC 49882 (C) with R. capsulatus (lanes Rc) and E. coli (lanes Ec) RNases
III at the 3'-processing site in helix 9. The corresponding 5' ends in
vivo were detected by primer extension analysis using total RNA (lanes
R) isolated from these strains. Lanes G, A, T, and C each refer to the
corresponding nucleotide of the DNA template (cloned 23S rDNA region),
as determined by sequencing. Parts of the in vitro transcripts and of
the pre-rRNA sequences are indicated on the right side of each
panel. The detected 5' ends are marked by labeled arrows as follows:
 1, 5' end of the 23S rRNA unprocessed in helix 9; 2, 5' end
of the unprocessed in vitro transcript; Ec, 5' end after RNase
IIIEc cleavage in vitro; Rc, 5' end after
RNase IIIRc cleavage in vitro; R, 5' end
corresponding to the RNase III processing site in helix 9, detected
in vivo; M, 5' end arising from further maturation of rRNA in
vivo.
|
|
View larger version (19K):
[in this window]
[in a new window]
|
FIG. 5.
Schematic representation of the RNase III
3'-processing sites in vitro and the in vivo 5' ends found in helix 9 of 23S rRNA with primer extension analysis shown in Fig. 4. (A)
R. capsulatus 37b. (B) R. palustris 5D.
(C) B. henselae ATCC 49882. Labeled arrows: Ec, 5' end after
RNase IIIEc cleavage in vitro; Rc, 5' end
after RNase IIIRc cleavage in vitro; R, 5'
end corresponding to the RNase III processing site in helix 9, detected in vivo; PB and DB, proximal and distal boxes, respectively,
found in the vicinity of the corresponding scissile bonds (compare with
references 24 and 38).
|
|
At the 3'-processing site in helix 9 of the R. capsulatus 37b4 transcript two scissile bonds were cleaved by
RNase IIIRc in vitro, but only the upstream
one was detected in vivo (Fig. 4A and Fig. 5A). The primer extension
analysis also confirmed the existence of multiple scissile bonds in the
B. henselae ATCC 49882 transcript.
In the R. palustris 5D and B. henselae ATCC
transcripts, both RNases III can cleave different scissile bonds in
vitro. In vivo, cleavage sites identical to the RNase
IIIRc processing sites were detected, probably
because the endogenous RNases III of these bacteria are more
similar to the RNase IIIRc than to
RNase IIIEc (Fig. 4B and C and Fig. 5B and C).
The intensity of the signals which represent 5' ends created by
RNase III cleavage in vivo is similar to the intensity of the
signals corresponding to the unprocessed 23S rRNA (Fig. 4). Additional 5' ends downstream of the in vivo processed helix 9 of
R. palustris 5D were detected (Fig. 4B), probably
arising due to further maturation of the 5' end of the large rRNA
fragment. Much stronger signals, different from those corresponding to
the RNase III cleavage sites, were obtained by primer extension
analysis of total RNA of all three strains. They correspond to
structural obstacles for cDNA synthesis at the GC-rich stem of the
helix 11, as well as to 5' ends at the stem of helix 10 (data not
shown; see also reference 37). They were not
detectable in primer extension reactions when in vitro assays were used.
| |
DISCUSSION |
We found that in Rhodobacter,
Rhodopseudomonas, Rhizobium,
Sinorhizobium, and Bartonella strains 23S
rRNA is fragmented near position 130 (E. coli numbering,
helix 9) due to RNase III-dependent processing of IVSs. These IVSs
are present in all operons and in all strains of a species (this
work and references 28 and 29).
They do not occur sporadically only, in contrast to all other IVSs
found in other rRNA regions until now (4, 9, 12, 23).
This suggests that these IVSs are of ancient origin and already occur
in the last common ancestor of modern alpha-Proteobacteria. The ubiquitous distribution of IVSs in the helices 9 of 23S rRNA of
certain alpha-proteobacterial species probably reflects an evolutionary
pressure for fragmentation in this region. It was recently reported
that the 5' end at the base of the helix 10 stem in R. palustris represents the real 5' end of the mature large rRNA
segment. The excision of the IVS from helix 9 is followed by complete
removal of the helix 9 and 10 sequences from the 23S pre-rRNA
(37). The authors of that study also discuss the
consequences of such extensive processing for the ribosome structure
(37). We were able to detect the in vivo 3'-processing sites
of RNase III in helices 9 of three different bacterial species by
primer extension analysis (Fig. 4). Comparison of the intensity of the obtained signals leads to the conclusion that the amount of the large
rRNA segments with a 5' end corresponding to the RNase III cleavage site is comparable to the amount of the unprocessed 23S pre-rRNA (Fig. 4). In contrast, much stronger signals were obtained which correspond to downstream 5' ends at the stem of helix 10 (not
shown). These results are in accordance with observations of others
(37) and suggest that the ubiquitous distribution of IVS in
helix 9 of certain bacterial species may reflect the need for an
initial processing signal leading to further downstream 23S rRNA
processing necessary to create functionally optimal ribosomes.
The obviously different primary structures of many of the studied IVSs,
the fact that RNase IIIRc processes the
Rhodobacter transcripts less efficiently than
those derived from the Rhizobium-Bradyrhizobium group (Table
3), and the differences between the GC content of the IVSs and overall
23S rRNA sequences (Table 2) can be explained by lateral genetic
transfer events. On the other hand, AU-rich duplexes are preferred
substrates for RNase III (24).
It is interesting that the proximal 30 bp of helix 9 in representatives
of five different genera are very similar (Table 1). We showed that the
RNase III cleavage sites are positioned approximately in the middle
of this region (Fig. 2 and 5). The RNase III binding sites are
located in the vicinity of the cleavage sites (24). Obviously, these facts contribute to the relative sequence conservation of this part of helix 9. The slower divergence of these sequences may
also reflect stronger constraints in the recognition of the cleavage
sites by the Rhizobium-type RNases III in comparison to
the Rhodobacter-type RNases III.
The mosaic structure of the IVS found in the helices 9 of the strains
R. capsulatus 37b4, R. gallicum R602,
and R. leguminosarum ATCC 10004 (Table 2), together
with the above-described evolution of the B. japonicum and
R. palustris helix 9 sequences from each other and the
presence of blocks with very conservative base pair occupation, show
that IVSs often evolve via insertion and/or deletion events.
The helix 9 secondary structures shown in Fig. 2 are based only on
computer-assisted folding (MFOLD [20, 39]). When only the sequences including the 23S rRNA helices 8, 9, and 10 were folded, we always obtained correct secondary structures for helices 8 and 10 compared with the universal secondary structure models for the
23S rRNAs of E. coli, R. capsulatus,
R. sphaeroides, and R. palustris
(23S rRNA Comparative Structure Database,
http://www.rna.icmb.utexas.edu/). Alternative secondary structures for
the relatively short helices 9 shown in Fig. 2A to L were not obtained.
The differences in the alternative structures obtained for the very
long rhizobial helices 9 (Fig. 2M to P) were localized apical to their
first 20 to 30 bp, these first 20 to 30 bp remaining invariable. For the overall length of the transcripts, we obtained many different secondary structures, but the differences were localized mainly outside
of helix 9. Only the first 6 to 8 bp of the short helices 9 presented
in Fig. 2B to L were involved in alternative structures. The 20-bp
region, which harbors the RNase III cleavage sites, remained
invariable. We therefore suggest that our secondary structure models
can be used for analysis of enzyme-substrate recognition.
Based on the secondary-structure proposals, we analyzed whether our
RNase III substrates, shown in Fig. 5, fit into the pattern proposed by the "antideterminant" model (24, 38). This
model explains recognition and cleavage of perfect double-helical
substrates by RNase IIIEc in vitro.
According to this model the positions 
4 to 6 relative to the
scissile bond must not have the sequence G(G,C)G (proximal box) and the
positions 11 and 12 must not have the sequence UC (distal box). The
presence of only one antideterminant base at positions 4 or 5 in
the proximal box reduces the cleavage of the substrate by the RNase
IIIEc to 7 to 15%; the presence of only two
antideterminant bases at both positions reduces it to 3%. The use of
lower salt concentrations in the reaction mix (<160 mM KCl) can
promote cleavage of less-reactive substrates (38).
In the case of the R. capsulatus 37b4 transcript, the
putative proximal boxes (determined in relation to both RNase
IIIRc cleavage sites) are involved in, or are
positioned immediately next to, predicted helix distortions (Fig. 5A).
The irregular double-helical structure of this transcript could
explain why it is not processed by RNase
IIIEc even at 130 mM KCl. It is more difficult
to understand the ability of the RNase IIIEc to cleave the R. palustris 5D transcript at 250 mM KCl,
despite the presence of two strong antideterminant bases (GG at
positions 5 and 6 relative to the scissile bond) in the proximal
box (Fig. 5B). Using the antideterminant model, it is also not
possible to explain why RNase IIIEc cannot
use the RNase IIIRc unique processing site
in the B. henselae ATCC 49882 transcript (Fig. 5C).
The antideterminant model may be limited to a subset of
all possible RNase IIIEc-specific
substrates. In addition to the presence or absence of certain
nucleotides near the cleavage site, the overall secondary and tertiary
structure of this substrate may be responsible for recognition by the
enzyme. It is possible that transcripts with alternative structures
were present in our in vitro assays, which differ in their cleavage
sites and efficiency of the cleavage. This could explain the detection
of multiple scissile bonds at the RNase III processing sites in
some of the studied transcripts (Table 3). We also cannot exclude the
possibility of sequence differences between the various rrn
operons of a bacterial strain. The observation that RNase III
cleaves an additional site in vitro in R. capsulatus
37b4 transcript (Fig. 4A) could be explained by the existence of
transcripts with alternative secondary structures in the assay, as well
as by differences in RNase III specificities in vitro and in vivo.
We describe here important differences in the substrate specificities
and cleavage sites of the RNases III from R. capsulatus and E. coli. Nothing is known about the way
RNase IIIRc recognizes its substrates. Until
now, only one R. capsulatus RNase III-specific
substrate was available (5). Our set of natural substrates
provides new perspectives for studying this enzyme. The exact
determination of RNase IIIRc cleavage sites in many different substrates should at least allow us to find the
position-exclusion of base pairs near the processing site for this
enzyme. This should make possible the construction of an
uncleavable IVS which can be used to replace wild-type IVSs in a
model alpha-proteobacterial strain. The availability of such mutant
strains with intact 23S rRNA and otherwise-unchanged genetic backgrounds will for the first time allow study of the consequences of
the 23S rRNA fragmentation in bacteria.
We thank C. Conrad for providing purified RNases III and many
buffers. We are grateful to N. Amarger (INRA-CMSE, Dijon, France), and
D. K. Jones (USDA/ARS Beltsville Rhizobium Germplasm
Collection), for sending us bacterial strains. We thank O. Fuhrmann (Charite, Berlin) for providing DNA from B. henselae ATCC 49882 and R. Rauhut for reading the manuscript
and for help with the computing programs.
This work was supported by Deutsche Forschungsgemeinschaft (Kl
563/11-1) and the Fonds der Chemischen Industrie.
| 1.
|
Amarger, N.,
V. Macheret, and G. Laguerre.
1997.
Rhizobium gallicum sp. nov. and Rhizobium giardinii sp. nov., from Phaseolus vulgaris nodules.
Int. J. Syst. Bacteriol.
47:996-1006[CrossRef][Medline].
|
| 2.
|
Ausubel, F. M.,
R. Brent,
R. E. Kingston,
D. D. Moore,
J. G. Seidman,
J. A. Smith, and K. Struhl.
1989.
Current protocols in molecular biology, vol. I.
Greene Publishing Associates and Wiley-Interscience, New York, N.Y.
|
| 3.
|
Beringer, J. E.
1974.
R-factor transfer in Rhizobium leguminosarum.
J. Gen. Microbiol.
84:188-198[Medline].
|
| 4.
|
Burgin, A. B.,
K. Parodos,
D. J. Lane, and N. R. Pace.
1990.
The excision of intervening sequences from Salmonella 23S ribosomal RNA.
Cell
60:405-414[CrossRef][Medline].
|
| 5.
|
Conrad, C.,
R. Rauhut, and G. Klug.
1998.
Different cleavage specificities of RNases III from Rhodobacter capsulatus and Escherichia coli.
Nucleic Acids Res.
26:4446-4453[Abstract/Free Full Text].
|
| 6.
|
Drews, G.
1976.
Mikrobiologisches Praktikum, 3rd ed.
Springer-Verlag, Berlin, Germany.
|
| 7.
|
Dryden, S. C., and S. Kaplan.
1990.
Localization and structural analysis of the ribosomal RNA operons of Rhodobacter sphaeroides.
Nucleic Acids Res.
18:7267-7277[Abstract/Free Full Text].
|
| 8.
|
Dunn, J. J.
1982.
Ribonuclease III, p. 485-499.
In
P. D. Boyer (ed.), The enzymes, vol. XV. Nucleic acids, part B. Academic Press, Inc., New York, N.Y.
|
| 9.
|
Evguenieva-Hackenberg, E., and S. Selenska-Pobell.
1995.
Variability of the 5'-end of the large subunit rDNA and the presence of a new short class of rRNA in Rhizobiaceae.
Lett. Appl. Microbiol.
21:402-405[Medline].
|
| 10.
|
Heck, C.,
R. Rothfuchs,
A. Jäger,
R. Rauhut, and G. Klug.
1996.
Effect of the pufQ-pufB intercistronic region on puf mRNA stability in Rhodobacter capsulatus.
Mol. Microbiol.
20:1165-1178[CrossRef][Medline].
|
| 11.
|
Höpfl, P.,
W. Ludwig, and K. H. Schleifer.
1988.
Complete nucleotide sequence of a 23S ribosomal RNA gene from Rhodobacter capsulatus.
Nucleic Acids Res.
16:2343[Free Full Text].
|
| 12.
|
Hsu, D.,
Y. Z. Zee,
J. Ingreham, and L.-M. Shih.
1992.
Diversity of cleavage patterns of Salmonella 23S rRNA.
J. Gen. Microbiol.
138:199-203[Medline].
|
| 13.
|
Hurtado, A.,
J. P. Clewley,
D. Linton,
R. J. Owen, and J. Stanley.
1997.
Sequence similarities between large subunit ribosomal RNA gene intervening sequences from different Helicobacter species.
Gene
194:69-75[CrossRef][Medline].
|
| 14.
|
Jemiolo, D. K.
1996.
Processing of procaryotic ribosomal RNA, p. 453-468.
In
R. Zimmermann, and A. Dahlberg (ed.), Ribosomal RNA. CRC Press, Boca Raton, Fla.
|
| 15.
|
Klug, G., and G. Drews.
1984.
Construction of a gene bank of Rhodopseudomonas capsulata using a broad host range cloning system.
Arch. Microbiol.
139:319-325[CrossRef][Medline].
|
| 16.
|
Kordes, E.,
S. Jock,
J. Fritsch,
F. Bosch, and G. Klug.
1994.
Cloning of a gene involved in rRNA precursor processing and 23S rRNA cleavage in Rhodobacter capsulatus.
J. Bacteriol.
176:1121-1127[Abstract/Free Full Text].
|
| 17.
|
Kundig, C.,
C. Beck,
H. Hennecke, and M. Gottfert.
1995.
A single rRNA gene region in Bradyrhizobium japonicum.
J. Bacteriol.
177:5151-5154[Abstract/Free Full Text].
|
| 18.
|
Marrs, B.
1981.
Mobilization of the genes for photosynthesis from Rhodopseudomonas capsulata by a promiscuous plasmid.
J. Bacteriol.
146:1003-1012[Abstract/Free Full Text].
|
| 19.
|
Marrs, B. L., and S. Kaplan.
1970.
23S precursor ribosomal RNA of Rhodopseudomonas sphaeroides.
J. Mol. Biol.
49:297-317[CrossRef][Medline].
|
| 20.
|
Mathews, D. H.,
J. Sabina,
M. Zuker, and D. H. Turner.
1999.
Expanded sequence dependence of thermodynamic parameters provides robust prediction of RNA secondary structure.
J. Mol. Biol.
288:911-940[CrossRef][Medline].
|
| 21.
|
Mattatall, N. R., and K. E. Sanderson.
1996.
Salmonella typhimurium LT2 possesses three distinct 23S rRNA intervening sequences.
J. Bacteriol.
178:2272-2278[Abstract/Free Full Text].
|
| 22.
|
Minnick, M. F.,
S. J. Mitchell,
S. J. McAllister, and J. M. Battisti.
1995.
Nucleotide sequence analysis of the 23S ribosomal RNA-encoding gene of Bartonella bacilliformis.
Gene
162:75-79[CrossRef][Medline].
|
| 23.
|
Miller, W. L.,
K. Pabbaraju, and K. Sanderson.
2000.
Fragmentation of 23S rRNA in strains of Proteus and Providencia results from intervening sequences in the rrn (rRNA) genes.
J. Bacteriol.
182:1109-1117[Abstract/Free Full Text].
|
| 24.
|
Nicolson, A. W.
1999.
Function, mechanism and regulation of bacterial ribonucleases.
FEMS Microbiol. Rev.
23:371-390[CrossRef][Medline].
|
| 25.
|
Otten, L.,
P. De Ruffray,
P. de Lajudie, and B. Michot.
1996.
Sequence and characterization of a ribosomal RNA operon from Agrobacterium vitis.
Mol. Gen. Genet.
251:99-107[Medline].
|
| 26.
|
Ralph, D., and M. McClelland.
1993.
Intervening sequence with conserved open reading frame in eubacterial 23S rRNA genes.
Proc. Natl. Acad. Sci. USA
90:6864-6868[Abstract/Free Full Text].
|
| 27.
|
Rauhut, R.,
A. Jäger,
C. Conrad, and G. Klug.
1996.
Identification and analysis of the rnc gene for RNase III in Rhodobacter capsulatus.
Nucleic Acids Res.
24:1246-1251[Abstract/Free Full Text].
|
| 28.
|
Selenska-Pobell, S., and E. Evguenieva-Hackenberg.
1995.
Fragmentations of the large-subunit rRNA in the family Rhizobiaceae.
J. Bacteriol.
177:6993-6998[Abstract/Free Full Text].
|
| 29.
|
Selenska-Pobell, S.,
H. Döring, and E. Evguenieva-Hackenberg.
1996.
Unusual organization of the 23S rRNA genes in Rhizobiaceae.
Soil Biol. Biochem.
29:905-909[CrossRef].
|
| 30.
|
Selenska-Pobell, S., and H. Döring.
1998.
Sequences around the fragmentation sites of the large subunit ribosomal RRNA in the family Rhizobiaceae.
Antonie Leeuwenhoek
73:55-67[CrossRef][Medline].
|
| 31.
|
Skurnik, M., and P. Toivanen.
1991.
Intervening sequences (IVSs) in the 23S ribosomal RNA genes of pathogenic Yersinia enterocolitica strains. The IVSs in Y. enterocolitica and Salmonella typhimurium have a common origin.
Mol. Microbiol.
5:585-593[CrossRef][Medline].
|
| 32.
|
Sistrom, W. R.
1977.
Transfer of chromosomal genes mediated by plasmid R68.
J. Bacteriol.
131:526-532[Abstract/Free Full Text].
|
| 33.
|
Song, X. M.,
A. Forsgren, and H. Janson.
1999.
Fragmentation heterogeneity of 23S ribosomal RNA in Haemophilus species.
Gene
230:287-293[CrossRef][Medline].
|
| 34.
|
Springer, N.,
W. Ludwig, and G. Hardarson.
1993.
A 23S rRNA targeted specific hybridization probe for Bradyrhizobium japonicum.
Syst. Appl. Microbiol.
16:468-470.
|
| 35.
|
Trust, T. J.,
S. M. Logan,
C. E. Gustafson,
P. J. Romaniuk,
N. W. Kim,
V. L. Chan,
M. A. Ragan,
P. Guerry, and R. R. Gutell.
1994.
Phylogenetic and molecular characterization of a 23S rRNA gene positions the genus Campylobacter in the epsilon subdivision of the Proteobacteria and shows that the presence of transcribed spacers is common in Campylobacter spp.
J. Bacteriol.
176:4597-4609[Abstract/Free Full Text].
|
| 36.
|
von Gabain, A.,
J. G. Belasco,
J. L. Schottel,
A. C. Y. Chang, and S. N. Cohen.
1983.
Decay of mRNA in Escherichia coli: investigation of the fate of specific segments of transcripts.
Proc. Natl. Acad. Sci. USA
80:653-657[Abstract/Free Full Text].
|
| 37.
|
Zahn, K.,
M. Inui, and H. Yukawa.
1999.
Characterization of a separate small domain derived from the 5' end of 23S rRNA of an -proteobacterium.
Nucleic Acids Res.
27:4241-4250[Abstract/Free Full Text].
|
| 38.
|
Zhang, K., and A. Nicolson.
1997.
Regulation of ribonuclease III processing by double-helical sequence antideterminants.
Proc. Natl. Acad. Sci. USA
94:13437-13441[Abstract/Free Full Text].
|
| 39.
|
Zuker, M.,
D. H. Mathews, and D. H. Turner.
1999.
Algorithms and thermodynamics for RNA secondary structure prediction: a practical guide, p. 11-43.
In
J. Barciszewski, and B. F. C. Clark (ed.), RNA biochemistry and biotechnology. NATO ASI series. Kluwer Academic Publishers, Dordrecht, The Netherlands.
|
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Top
Abstract
Introduction
