Functional equivalence of hairpins in the RNA subunits of RNase MRP and RNase P in Saccharomyces cerevisiae.

RNase MRP and RNase P are both ribonucleoprotein enzymes performing endonucleolytic cleavage of RNA. RNase MRP cleaves at a specific site in the precursor-rRNA transcript to initiate processing of the 5.8S rRNA. RNase P cleaves precursor tRNAs to create the 5' end of the mature tRNAs. In spite of their different specificities, the two RNases have significant structural similarities. For example, the two enzymes in Saccharomyces cerevisiae share eight protein subunits; only one protein is unique to each enzyme. The RNA components of the two nucleases also show striking secondary-structure similarity. To begin to characterize the role of the RNA subunits in enzyme function and substrate specificity, we swapped two hairpin structures (MRP3 and P3) between RNase MRP RNA and RNase P RNA of S. cerevisiae. The hairpins in the two enzymes could be exchanged without loss of function or specificity. On the other hand, when the MRP3 hairpin in RNase MRP of S. cerevisiae was replaced with the corresponding hairpin from the RNA of Schizosaccharomyces pombe or human RNase MRP, no functional enzyme was assembled. We propose that the MRP3 and P3 hairpins in S. cerevisiae perform similar functions and have coevolved to maintain common features that are different from those of MRP3 and P3 hairpins in other species.

Phylogenetic analysis of L4-mediated autogenous control of the S10 ribosomal protein operon.

We investigated the regulation of the S10 ribosomal protein (r-protein) operon among members of the gamma subdivision of the proteobacteria, which includes Escherichia coli. In E. coli, this 11-gene operon is autogenously controlled by r-protein L4. This regulation requires specific determinants within the untranslated leader of the mRNA. Secondary structure analysis of the S10 leaders of five enterobacteria (Salmonella typhimurium, Citrobacter freundii, Yersinia enterocolitica, Serratia marcescens, and Morganella morganii) and two nonenteric members of the gamma subdivision (Haemophilus influenzae and Vibrio cholerae) shows that these foreign leaders share significant structural homology with the E. coli leader, particularly in the region which is critical for L4-mediated autogenous control in E. coli. Moreover, these heterologous leaders produce a regulatory response to L4 oversynthesis in E. coli. Our results suggest that an E. coli-like L4-mediated regulatory mechanism may operate in all of these species. However, the mechanism is not universally conserved among the gamma subdivision members, since at least one, Pseudomonas aeruginosa, does not contain the required S10 leader features, and its leader cannot provide the signals for regulation by L4 in E. coli. We speculate that L4-mediated autogenous control developed during the evolution of the gamma branch of proteobacteria.

Analysis of the Bacillus subtilis S10 ribosomal protein gene cluster identifies two promoters that may be responsible for transcription of the entire 15-kilobase S10-spc-alpha cluster.

We have sequenced a previously uncharacterized region of the Bacillus subtilis S10 ribosomal protein gene cluster. The new segment includes genes for S10, L3, L4, L23, L2, S19, L22, S3, and part of L16. These B. subtilis genes map in the same order as the genes in the Escherichia coli S10 ribosomal protein operon. Two potential promoter sequences were identified, one approximately 200 bases and the other approximately 140 bases upstream of the S10 gene. The activities of the two promoters were demonstrated by primer extension analysis, in vitro transcription experiments, and in vivo promoter fusion plasmid studies. In agreement with previous reports, our Northern analysis of exponentially growing cells failed to identify terminators or other active promoters within the S10-spc-alpha region. Our observations suggest that the two S10 promoters reported here are responsible for transcribing a 15-kb-long transcript for all of the genes in the B. subtilis S10, spc, and alpha clusters.

A novel protein shared by RNase MRP and RNase P.

We have isolated suppressors of the temperature-sensitive rRNA processing mutation rrp2-2 in Saccharomyces cerevisiae. A class of extragenic suppressors was mapped to the YBR257w reading frame in the right arm of Chromosome II. Characterization of this gene, renamed POP4, shows that the gene product is necessary both for normal 5.8S rRNA processing and for processing of tRNA. Immunoprecipitation studies indicate that Pop4p is associated with both RNase MRP and RNase P. The protein is also required for accumulation of RNA from each of the two ribonucleoprotein particles.

A hairpin structure upstream of the terminator hairpin required for ribosomal protein L4-mediated attenuation control of the S10 operon of Escherichia coli.

Ribosomal protein L4 of Escherichia coli regulates transcription of the 11-gene S1O operon by promoting premature termination of transcription (attenuation) at a specific site within the 172-base untranslated leader. We have analyzed the roles of various domains of the leader RNA in this transcription control. Our results indicate that the first 60 bases of the leader, forming the three proximal hairpin structures, are not essential for in vivo L4-mediated attenuation control. However, a deletion removing the fourth hairpin, which is immediately upstream of the terminator hairpin, eliminates L4's effect on transcription. Base changes disrupting complementarity in the 6-bp stem of this hairpin also abolish L4 control, but compensatory base changes that restore complementarity also restore L4's effect. In vitro transcription studies confirm that this hairpin structure is necessary for L4's role in stimulating transcription termination by RNA polymerase.

Ribosomal protein L4 from Escherichia coli utilizes nonidentical determinants for its structural and regulatory functions.

Escherichia coli ribosomal protein L4 has two functions: it is a structural component of the 50S ribosomal sub-unit and it is a repressor of both transcription and translation of its own transcription unit, the 11-gene S10 operon. Genetic and biochemical studies have suggested that L4 can interact with 23S rRNA as well as with both RNA interactions. However, no significant similarities between its two RNA targets can be found at the primary or secondary structure level. To test if identical determinants of L4 are involved in both ribosome assembly and autogenous control, we have isolated L4 mutants defective in either of these functions and asked if a mutant protein divested of one function is also deficient in the other. Several mutations eliminated autogenous control, but still allowed assembly of the mutant L4 protein into functional ribosomes. Conversely, several mutant L4 proteins that could not be detected in 50S subunits nevertheless could regulate expression of the S10 operon. These results indicate that the L4 determinants required for autogenous regulation and ribosome incorporation are not congruent.

Regulation of the Escherichia coli S10 ribosomal protein operon by heterologous L4 ribosomal proteins.

We have cloned the L4 ribosomal protein genes from Morganella morganii and Haemophilus influenza. The sequences of these genes were compared with published sequences for Escherichia coli, Yersinia pseudotuberculosis, and Bacillus stearothermophilus. All five of these L4 genes were expressed in E. coli and shown to function as repressors of both transcription and translation of the E. coli S10 operon. Possible implications for regulation of r-protein synthesis in species other E. coli are discussed.

RNA determinants required for L4-mediated attenuation control of the S10 r-protein operon of Escherichia coli.

We have probed regions of the S10 leader RNA to determine their role in L4-mediated, NusA-dependent attenuation control of the S10 ribosomal protein operon. Using genetic and "antisense" oligonucleotide competition approaches, we were able to distinguish between the determinants necessary for intrinsic (NusA-independent) pausing by RNA polymerase at the S10 attenuation site, for NusA-dependent enhancement of pausing, and for L4 stabilization of the paused ternary complex. The upper stem-loop structure in the attenuator hairpin is the major determinant for the NusA-dependent pause, while the sequence at the site of pausing is important for RNA polymerase to pause in the absence of NusA. The determinants for L4 stabilization of the paused complex include the hairpin immediately upstream of the attenuator hairpin as well as the ascending side of the attenuator structure. In conclusion, our results suggest that there are three distinct pausing activities by RNA polymerase during its transcription of the S10 leader, with three corresponding signals in the S10 leader.

Role of NusA in L4-mediated attenuation control of the S10 r-protein operon of Escherichia coli.

The transcription of the 11 gene S10 operon of Escherichia coli is autogenously regulated by one of the operon's products, ribosomal protein L4. This protein stimulates termination of transcription in vivo at a specific site within the S10 leader. The in vivo effect can be reproduced in a purified transcription system but requires an additional factor, NusA. Our earlier in vitro studies showed that NusA is required for RNA polymerase pausing at the termination site; such paused complexes are further stabilized by L4, which presumably accounts for L4's stimulation of termination in vivo. Here we show that NusA is not absolutely required for RNA polymerase to recognize the attenuation site: at low (5 microM) UTP concentration, RNA polymerase pauses at the site, although the paused transcription complex formed in the absence of NusA can be further stabilized by subsequent addition of the protein. Furthermore, RNA polymerase pausing at the attenuation site is not sufficient for the L4 effect, since L4 cannot stabilize a transcription complex paused at the attenuation site in the absence of NusA. We have been able to isolate paused complexes formed without NusA and/or L4; such complexes are active upon re-addition of NTPs, and respond as expected to the addition of L4 or NusA. Our experiments are consistent with the notion that L4 is a stable component of a paused transcription complex.

RNase MRP and rRNA processing.

RNase MRP is a ribonucleoprotein enzyme with a structure similar to RNase P. It is required for normal processing of precursor rRNA, cleaving it in the Internal Transcribed Spacer 1.

Alternate pathways for processing in the internal transcribed spacer 1 in pre-rRNA of Saccharomyces cerevisiae.

We have extended the system of Nogi et al. (Proc. Natl. Acad. Sci. USA 88, 1991, 3962-3966) for transcription of rRNA from an RNA polymerase II promoter in strains lacking functional RNA polymerase I. In our strains two differentially marked rRNA transcription units can be expressed alternately. Using this system we have shown that the A2 processing site in the internal transcribed spacer 1 (ITS1) of the pre-rRNA is dispensable. According to the accepted processing scheme, the A2 site serves to separate the parts of the primary rRNA transcript that are destined for incorporation into the two ribosomal subunits. However, we have found that, when A2 is impaired, separation of the small and large subunit rRNAs occurs at a processing site further downstream in ITS1, indicating that alternate pathways for ITS1 processing exist. Short deletions in the A2 region still allow residual processing at the A2 site. Mapping of the cleavage sites in such deletion transcripts suggests that sequences downstream of the A2 site are used for determining the position of the cleavage.

Correlation of translation efficiency with the decay of lacZ mRNA in Escherichia coli.

We isolated mutations in the leader of a ribosomal protein (r-protein)/lacZ fusion gene in Escherichia coli that caused the mRNA to be translated at efficiencies between < 1% and 62% of the rate of wild-type message. Using a subset of these mutants with translation efficiencies between 5% and 62%, we studied both physical and functional decay of the mRNA after rifampicin inhibition of transcription initiation. The decay of physically intact transcript was analyzed by gel electrophoresis of hybrid-selected messenger RNA segments. The output from the message was analyzed by measuring the synthesis rate of r-protein/lacZ fusion protein. Decay of physically intact message after rifampicin addition correlated with the translation efficiency, with the more active messengers being more stable. Different segments of the r-protein/lacZ fusion mRNA decayed with the same rate, indicating that there is no hyper-labile region in the transcript. The decay rate was also independent of the length of the segment probed, suggesting that the mRNA is not degraded by random attacks along the entire length of the molecule. Our results are consistent with an overall 5' to 3' degradation pathway. Surprisingly, the rate of fusion protein synthesis did not decrease immediately after rifampicin addition. Rather, a lag preceded the exponential decay phase; the length of this delay correlated with the translation efficiency, such that the lag increased with increasing efficiency of translation. We suggest that these lags indicate that mRNAs are normally competing for ribosomes during exponential growth and, after rifampicin addition, RNA molecules with longer physical half-lives are translated by ribosomes released from fast decaying messengers.

The RNA of RNase MRP is required for normal processing of ribosomal RNA.

We have isolated clones which complement the temperature sensitivity and abnormal rRNA processing pattern of the rrp2-2 mutant of Saccharomyces cerevisiae we previously described. DNA sequencing and restriction analysis demonstrated that all clones contain the NME1 gene encoding the RNA of the ribonucleprotein particle RNase MRP. Deletion analysis showed that the NME1 gene is responsible for the complementation of the rrp2-2 phenotype. A single base change was identified in the nme1 gene in the rrp2 mutant, confirming that the RRP2 and NME1 genes are identical. Our experiments therefore indicate that RNase MRP, in addition to its previously reported role in formation of RNA primers for mitochondrial DNA replication [Clayton, D. A. (1991) Trends Biochem. Sci. 16, 107-111], is involved in rRNA processing.

Domain I of 23S rRNA competes with a paused transcription complex for ribosomal protein L4 of Escherichia coli.

Ribosomal protein L4 of Escherichia coli regulates expression of its own eleven gene S10 operon both by inhibiting translation and by stimulating premature termination of transcription. Both regulatory processes presumably involve L4 recognition of the S10 leader RNA. To help define L4's regulatory target, we have investigated the protein's cognate target on 23S rRNA. Binding of L4 to various fragments of the 23S rRNA was monitored by determining their ability to sequester L4 in an in vitro transcription system and thereby eliminate the protein's effect on transcription. Using this approach we identified a region of about 110 bases within domain I of 23S rRNA which binds L4. A two base deletion within this region, close to the base to which L4 has been cross-linked in intact 50S subunits, eliminates L4 binding. These results also confirm the prediction of the autogenous control model, that L4 bound to its target on rRNA is not active in regulating transcription of the S10 operon.

Ribosomal protein L4 and transcription factor NusA have separable roles in mediating terminating of transcription within the leader of the S10 operon of Escherichia coli.

Ribosomal protein L4 of Escherichia coli autogenously regulates both transcription and translation of the 11-gene S10 operon. Transcription regulation occurs by L4-stimulated premature termination at an attenuator hairpin in the S10 leader. This effect can be reproduced in vitro but depends on the addition of transcription factor NusA. We show that NusA is required to promote RNA polymerase pausing at the termination site; such paused transcription complexes are then stabilized further by r-protein L4. The L4 effect is observed even if the protein is added after the NusA-modified RNA polymerase has already reached the pause site. Genetically separable regions of the S10 leader are required for NusA and L4 action: The attenuator hairpin is sufficient for NusA-dependent pausing, but upstream elements are necessary for L4 to prolong the pause.

A new rRNA processing mutant of Saccharomyces cerevisiae.

We have identified from a collection of temperature sensitive yeast mutants strains which fail to process rRNA normally. Characterization of one such mutant is reported here. This strain accumulates increased amounts of the 35S primary transcript, '24S' molecules extending from the transcription start site to the 5.8S region, and two classes of 5.8S rRNA with 5' extensions of 7 and 149 bases, respectively. We show that this pleiotropic change in the rRNA processing pattern is due to a single mutation. Possible models for the function of the mutated gene are discussed.

Ribosomal protein L4 of Escherichia coli: in vitro analysis of L4-mediated attenuation control.

Ribosomal protein L4 of Escherichia coli functions not only as a component of the ribosome but also as a regulatory factor inhibiting both transcription and translation of its own operon, the 11 gene S10 operon. L4-mediated transcription control results in premature termination of transcription within the 172 base S10 operon leader. This attenuation control can be reproduced in a purified transcription system containing RNA polymerase, but depends on the addition of transcription factor NusA. The NusA stimulation saturates at about 2-4 copies per RNA polymerase. The L4 effect plateaus at about 4 copies per RNA polymerase. The specific recognition sites on 23S rRNA and in the S10 leader for L4 binding are not yet known. However, we can demonstrate that a fragment of 23S rRNA containing the proximal 840 bases can eliminate in vitro L4-stimulated attenuation, and hence, contains the information sufficient for L4 binding to 23S rRNA.

Intermediates in the degradation of mRNA from the lactose operon of Escherichia coli.

We have analyzed the processing of mRNA from the lac operon in an Escherichia coli strain carrying the lac on a multicopy plasmid. Messenger RNA was analyzed by hybridization and nuclease protection of pulse-labeled RNA and precursor-product relationships were determined by quantitating radioactivity in primary and processed transcripts at various times after induction of the lac promoter or inhibition of transcription with rifampicin. Our results support the existence of two types of processed transcripts with endpoints in the lacZ-lacY intercistronic region. One of these carries lacZ sequences and has a 3' endpoint about 30 bases downstream of this gene. The other carries lacY sequences and has a 5' end in the translation termination region of the lacZ gene. Finally, we have found evidence that transcription is continued at least 268 bases beyond the last gene (lacA) and that this 3' non-translated region is shortened by post-transcriptional processing.