Regulation of E activity by the L4 ribosomal protein of

Dharam Singh, Ssu-Jean Chang1, Pei-Hsun Lin, Olga V. Averina2, Vladimir R. Kaberdin, and Sue Lin-Chao3

Institute of Molecular Biology, Academia Sinica, Taipei 11529, Taiwan

Edited by Stanley N. Cohen, Stanford University School of Medicine, Stanford, CA, and approved November 21, 2008 (received for review October 10, 2008) Whereas ribosomal proteins (r-proteins) are known primarily as interaction modulates RNase E activity, altering the steady-state components of the translational machinery, certain of these level and decay of affected regulatory and messenger . As r-proteins have been found to also have extraribosomal functions. the abundance of proteins encoded by some of these mRNAs is Here we report the novel ability of an r-protein, L4, to regulate RNA known to increase along with free r-proteins in response to degradation in Escherichia coli. We show by affinity purification, environmental stresses, our findings reveal a mechanism by immunoprecipitation analysis, and E. coli two-hybrid screening which L4 may regulate the production of stress-induced proteins that L4 interacts with a site outside of the catalytic domain of to enhance the survival of under adverse conditions. RNase E to regulate the endoribonucleolytic functions of the , thus inhibiting RNase E-specific cleavage in vitro, stabi- Results lizing mRNAs targeted by RNase E in vivo, and controlling plasmid L4 Directly Interacts with the C-Terminal Region of RNase E in Vivo and DNA replication by stabilizing an antisense regulatory RNA nor- in Vitro. To identify low-molecular-weight (Յ 30 kDa) proteins mally attacked by RNase E. Broader effects of the L4-RNase E that bind to RNase E, FLAG-tagged RNase E was overexpressed interaction on E. coli transcripts were shown by DNA microarray in E. coli and purified by affinity-chromatography as described analysis, which revealed changes in the abundance of 65 mRNAs previously (19). After electrophoretic analysis on 12% SDS gels encoding the stress response proteins HslO, Lon, CstA, YjiY, and followed by Coomassie Blue staining, the polypeptides co- YaeL, as well as proteins involved in carbohydrate and amino acid purifying with RNase E were identified by mass spectroscopy. metabolism and transport, transcription/translation, and DNA/ Several r-proteins, including L2, L3, L4, S3, and S4, were RNA synthesis. Analysis of mRNA stability showed that the half co-purified with the RNase E complex (the ) lives of stress-responsive transcripts were increased by ectopic (supporting information (SI) Table S1). We then used an E. coli expression of L4, which normally increases along with other two-hybrid system (27) to further investigate a possible interac- r-proteins in E. coli under stress conditions, and also by inactivation tion of each of these r-proteins with the major components of the of RNase E. Our finding that L4 can inhibit RNase E-dependent degradosome: RNase E, PNPase, RhlB helicase, or enolase (Fig. decay may account at least in part for the elevated production of S1). We observed that only L4 directly interacted with degra- stress-induced proteins during bacterial adaptation to adverse dosome proteins binding to the C-terminal half of RNase E and environments. also to PNPase (Fig. 1A and B). Co-immunoprecipitation ex- periments confirmed that L4 bound to RNase E and this posttranscriptional control ͉ RNA degradation ͉ stress responses ͉ interaction of L4 and RNase E was likewise dependent on the degradosome C-terminal half of the enzyme (Fig. 1C, compare lane 2 with lane 6), in particular on two regions (684–784 aa and 985-1061 aa, Fig. ver the past two decades, an understanding of mRNA decay 1B). Furthermore, by using micrococcal to digest RNA Opathways in Escherichia coli has advanced significantly (for co-purifying with the RNase E complex (19, 28), we found that reviews, see ref. 1–3), and RNase E has emerged as a key player the association of L4 with RNase E is independent of RNA (Fig. in mRNA turnover as well as in the processing and decay of 1D). In contrast, interaction between L4 and PNPase was weak noncoding RNAs (e.g., rRNAs [4, 5], tRNAs [6, 7], M1 RNA [8], (data not shown); we thus chose not to characterize it further. and 6S RNA [9]). RNase E is a multifunctional L4 is a structural protein of the 50S ribosomal subunit and also (10) known to preferentially cleave RNA within AU-rich single- a regulator of both transcription and translation of its own operon (24, 25). These functions require two independent stranded regions (11, 12) enriched in specific sequence deter- domains of L4 (26). To examine whether these domains are minants (13). The level of this enzyme in vivo is controlled via required also for interaction with RNase E, we separately autoregulation of its own synthesis (14–16). co-expressed FLAG-tagged RNase E with HA-tagged L4 (con- In addition to its N-terminal catalytic domain (N-RNase E), trol) or L4 mutants lacking either of these functional domains RNase E contains a C-terminal region (C-RNase E) that serves (Fig. S2A). Subsequently, by using affinity purification, we found as a scaffold (17, 18) for association with polynucleotide phos- phorylase (PNPase), RhlB RNA helicase, and the glycolytic

enzyme enolase to form the RNA-degrading complex known as Author contributions: D.S., S.-J.C., P.-H.L., O.V.A., V.R.K., and S.L.-C. designed research; D.S., the ‘‘degradosome’’ (19, 20). C-terminal truncation of RNase E, S.-J.C., and O.V.A. performed research; D.S. and S.L.-C. contributed new reagents/analytic which prevents degradosome assembly, leads to accumulation of tools; D.S., S.-J.C., V.R.K., and S.L.-C. analyzed data; and D.S., S.-J.C., P.-H.L., V.R.K., and RNase E-targeted mRNAs (21, 22), suggesting that degrado- S.L.-C. wrote the paper. some assembly and functional interactions of degradosome The authors declare no conflict of interest. components are necessary for normal mRNA turnover in E. coli. This article is a PNAS Direct Submission. Although ribosomal proteins (r-proteins) function primarily as Freely available online through the PNAS open access option. components of the translation machinery, some prokaryotic and 1Present address: Department of Life Sciences, National Science Council, Taipei 10622, eukaryotic r-proteins also have extraribosomal functions (23). Taiwan. For example, L4, an essential r-protein encoded by the S10 2Present address: P. P. Shirshov Institute of Oceanology, Kaliningrad 117997, Russia. operon in E. coli is a regulator of both transcription and 3To whom correspondence should be addressed. E-mail: [email protected]. translation of its own operon (24, 25). The regions within L4 This article contains supporting information online at www.pnas.org/cgi/content/full/ required for these distinct functions differ (26). Here we show 0810205106/DCSupplemental. that the E. coli L4 protein interacts with RNase E and that this © 2009 by The National Academy of Sciences of the USA

864–869 ͉ PNAS ͉ January 20, 2009 ͉ vol. 106 ͉ no. 3 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0810205106 Downloaded by guest on October 2, 2021 A ΔS10

RE14-L4

Interaction B with L4 RE1 1 597 Fig. 2. Testing the functionality of FLAG-L4 and its ability to inhibit RNase RE3 E-mediated cleavage of BR13 in vitro. (A) Testing the ability of FLAG-tagged 499 1061 RE12 L4 to inhibit its own synthesis by using an in vitro translation system. The 985 1061 RE14 plasmid pET29-⌬S10 (depicted on the left) encoding the S10 leader sequence 684 784 PNP and the first three genes (rpsJ, rplC, rplD) of the S10 operon was used as a 1 Eno 734 template for in vitro transcription-translation assays. Reactions were per- 1 432 35 RhlB formed in the presence of S-Met and increasing concentrations of FLAG-L4 1 421 and decreasing concentrations of BSA as indicated. The products of transla- tion were analyzed by electrophoresis on a 12% SDS polyacrylamide gel and C rne+ rne131 detected by autoradiography. (B) In vitro cleavage of oligonucleotide BR13 by T IP PI B T IP PI B the full-length RNase E (FL-RNase E) and its C-terminally truncated polypep- * L4 L4* tide (N-RNase E). Reactions were performed at 30 °C for 15 min in 20 ␮l with FL-RNase E N-RNase E equimolar amounts of RNase E and N-RNase E (500 ng of each protein) and PNPase PNPase increasing amounts of FLAG-L4 (25 to 100 ng). The bands corresponding to Western Blots 1 2 3 4 5 6 7 8 BR13 and its cleavage product BR13Ϫ5 are indicated by arrows.

Micrococcal Micrococcal D --+ nuclease M - + nuclease GENETICS T FR FR bp increasing amounts of FLAG-tagged L4 (FLAG-L4). Before testing cleavage ability, we used an in vitro translation system to RNase E confirm that FLAG-L4 was able to inhibit its own translation (Fig. 2A) and therefore functionally resembled the nontagged HA-L4 500 Western Blots 250 chromosomally encoded L4. Cleavage assays were carried out with oligonucleotide BR13 derived from RNAI, an antisense Fig. 1. L4 interacts with the E. coli degradosome in vivo and in vitro by RNA, whose cleavage by RNase E controls replication of binding to the C-terminal ‘‘scaffold’’ region of RNase E. (A) E. coli two-hybrid ColEI-type plasmids in vivo (29). As anticipated from previous assays demonstrating L4 interactions with RNase E and other major compo- studies (30), FLAG-RNase E efficiently cleaved this substrate; nents of the degradosome on MacConkey/maltose agar plates. The T25- and however, cleavage was reduced in the presence of increasing T18-based chimeric protein construct pairs (see SI Materials and Methods) amounts of L4 (Fig. 2B, lanes 2–4). In parallel control experi- screened for protein-protein interactions as well as positive [C(ϩ), red colo- nies] and negative [C(Ϫ), white colonies] controls (see supporting ref. 3) are ments, no inhibition of the ribonucleolytic activity of RNase A indicated for each sector of the plates shown. (B) The schematic representa- was observed (data not shown). In addition, no inhibition was tion of expressed polypeptides and results [positive (ϩ) and negative (Ϫ)] of observed when FLAG-RNase E was replaced with N-RNase E two-hybrid screening. (C) Analysis of the RNase E-L4 interaction by immuno- containing the catalytic domain alone (Fig. 2B, lanes 7–9). This precipitation of cell lysates prepared from the wild-type BL21 (DE3) strain result suggests that L4 inhibits RNase E activity by interacting (rneϩ) and its isogenic mutant (rne131) encoding C-terminally truncated with the enzyme’s C-terminal region. RNase E. Protein samples obtained by immunoprecipitation with L4-specific To learn whether L4 also can impair RNase E activity in vivo, antibodies (IP), preimmune serum (PI) as well as total cell lysate (T), and we tested the effect of L4 ectopic expression on RNase proteins that were bound to beads nonspecifically (B) were separated by electrophoresis on a 10% SDS gel and analyzed by Western blotting by using E-mediated decay of RNAI and its consequences on the copy antibodies specific for L4, full-length RNase E (FL-RNase E), N-terminal half of number of a ColE1-type plasmid in E. coli strains N3433 and RNase E (N-RNase E), and PNPase. The polypeptide nonspecifically interacting BZ453 (31) expressing the full-length and C-terminally trun- with anti-L4 antibody (in PI and IP) is indicated by an asterisk. (D) Co- cated RNase E polypeptides, respectively. Northern blot analysis immunopurification of L4 and RNase E is not dependent on E. coli RNA. revealed that elevation of L4 resulted in a prolongation of the FLAG-RNase E (FR) was affinity-purified on an M2 column (Sigma) from E. coli RNAI half-life from 3.4 min to 5.7 min (Fig. 3A) and a reduction cell extract (T) containing co-expressed FLAG-RNase E and HA-L4 before (Ϫ) in the copy number of a ColE1-type plasmid in strain N3433 ϩ and after ( ) treatment. The resulting samples of affin- strain (Fig. 3C), but that no detectable effect on RNAI half-life ity-purified FLAG-RNase E were either analyzed by Western blotting (on the left) by using antibodies specific for RNase E or HA-tag or were extracted with or plasmid copy number was seen in the BZ453 strain (Fig. 3C). phenol and analyzed by electrophoresis on a 2.0% agarose gel followed by These results suggest that L4 binding to the C-terminal region of ethidium bromide staining (on the right). M, 1 kb DNA ladder (Fermentas). RNase E inhibits RNase E activity and leads to prolongation of the RNAI half-life, decreasing copy number of ColE1-type plasmids. Consistent with the previous data, we also observed that both mutant L4 proteins interacted with RNase E (Fig. that in vivo production of L4 leads to an increased level of S2A), thus suggesting that neither domain is essential for L4 endogenous full-length RNase E (Fig. S3) and its mRNA (see binding to RNase E. microarray data in Fig. 4B) in N3433 but not the BZ453 strain. As RNase E is known to autoregulate its own level (16) by The L4-RNase E Interaction Inhibits RNase E Activity in Vitro and in cleaving its cognate mRNA, the observed increase in the full- Vivo. To determine whether binding of L4 to RNase E affects length RNase E and rne mRNA levels is consistent with the RNase E endonucleolytic activity, we performed cleavage assays observed inhibition of RNase E activity by L4. in vitro by using full-length RNase E (FLAG-RNase E) or its As the C-terminal half of RNase E is required for interaction N-terminal catalytic domain (N-RNase E) in the presence of of this endoribonuclease with L4 (present study) but is dispens-

Singh et al. PNAS ͉ January 20, 2009 ͉ vol. 106 ͉ no. 3 ͉ 865 Downloaded by guest on October 2, 2021 A N3433 (Rne) A C BZ453 N3433 Control L4 BZ453 or N3433 (N-Rne) (Rne) min yjiY 0 1.5 3 4.5 0 1.5 3 4.5 vector only vector + L4 after rif 5.0 lldP RNAI (control) (L4) srlB 4.0 RNAI-5 cDNA labeling hslO nanA 5S Alexa 555 Alexa 647 3.0 2.5 nanT t ~3.4 min t ~5.7 min 1/2 1/2 Hybridization 2.0 cspE DNA Microarray sdaC B BZ453 (N-Rne) 1.5 uxaC min Control L4 GeneSpring 1.2 cspC after rif 0 1.5 3 4.5 6 0 1.5 3 4.5 6 analysis cstA RNAI 1.0 ybeD BZ453 N3433 ybhT B (N-Rne) (Rne) 0.9

5S change Relative rpmE rpsJ 0.8 lon rplC t ~5.2 min 0.7 1/2 t1/2~5.2 min rplW sdhB rplB 0.6 tnaA C N3433 (Rne) BZ453 (N-Rne) D wt rnets rpsS 0.5 tnaC Control L4 Control L4 rplD 0.4 sdhA rne 0.3 srlD pre-M1 rpsT sdhC Chr M1 ompA 0.2 sdhD rpsO 0.1 yhjX P pre-6S 6S Fig. 4. Microarray identification of specific mRNA targets dependent on the 1 2 3 4 1 2 3 4 L4-RNase E interaction. (A) Shown is the experimental strategy used to per- form microarray analysis. The E. coli strains N3433 and BZ453 encoding wt rnets wt rnets E F full-length (Rne) and C-terminally truncated RNase E (N-Rne), respectively, and carrying the control plasmids pPW500flag (vector only) and plasmid pPWflagL4 (vector ϩ L4) encoding L4 were individually grown in LB medium 9S at 32 °C to an OD600 of Ϸ0.3 and then induced with IPTG (0.5 mM) for 30 min. Aliquots of total RNA extracted from each cell culture were used to synthesize cDNAs labeled with Alexa Fluor 647 or Alexa Fluor 555 as indicated on the pre-tRNA scheme. After hybridization with a DNA microarray, the resulting patterns 5S tRNAAsn were analyzed by using GeneSpring software GX 7.3.1 (Silicon Genetics) and 1 2 3 4 1 2 3 4 the results of this analysis were summarized in Table S3.(B) Microarray data Fig. 3. Effects of L4 ectopic expression on the RNase E-mediated decay (A–C) (shown in duplicate) are for selected transcripts known to be controlled either and processing (D–F) of noncoding RNAs. (A and B) Northern blot data by L4 or RNase E. As anticipated, several 5Ј-proximal species transcribed from demonstrating that the half-life of RNAI increases after L4 ectopic expression the S10 operon including rpsJ, rplC, rplW, rplB, and rpsS were down-regulated from pPWflagL4 (L4) in a strain encoding the full-length RNase E (N3433) but upon L4 (rplD) ectopic expression. In contrast, the level of rne mRNA (rne) was not in a strain (BZ453) expressing only the N-terminal domain (1–602 aa) of up-regulated, whereas the abundance of other well known RNase E substrates this protein. The half-life of RNAI in the absence of L4 ectopic expression was (rpsT, ompA, and rpsO) was not affected. A reference color bar showing also determined in the same strains harboring the plasmid pPW500flag (con- correlation between the observed color patterns and quantitative changes in trol) lacking the L4-coding sequence. (C) The observed increase in the half-life transcript levels is provided in panel C. (C) Microarray data (shown in dupli- of RNAI upon L4 ectopic expression (A and B) correlates with a decrease in the cate) are for 23 transcripts whose relative mRNA levels were significantly copy number of ColE1-type plasmids in N3433 (lanes 1 and 2) when compared increased in N3433 (i.e., in the presence of L4-RNase E interaction) but not in to that in BZ453 (lanes 3 and 4). Chromosomal (Chr) and plasmid (P) DNA are BZ453 (in which L4-RNase E interaction is impaired). The complete list of 65 indicated. The larger size of plasmid DNA (P) indicated in lane 2 and 4 is due transcripts that were reproducibly up-regulated upon L4 ectopic expression in to the presence of an extra DNA fragment encoding L4. (D–F) Equal amounts N3433 is shown in Dataset S1. The gradient color bar on the left indicates the of total RNA extracted from the E. coli wild-type N3433 (wt) strain carrying relative decrease or increase in the level of individual transcripts. The differ- control and L4 expressing plasmids after induction with IPTG (0.5 mM) for 30 ently colored circles (on the right) indicate functional classification of gene min as well as those prepared from the temperature sensitive N3431 (rnets) products (red, metabolism; green, transcriptional/translational regulation; strain at permissive (32 °C) and nonpermissive (44 °C) temperatures were blue, stress response; and black, hypothetical/uncharacterized). analyzed by Northern blotting. The 32P -labeled oligonucleotide probes were complementary to M1 RNA (D), 6S RNA (D), 5S rRNA (E), and tRNAAsn (F). None of the processing intermediates normally accumulating upon inactivation of Microarray Identification of Transcripts That Are Affected by L4 in an RNase E in N3431 (lane 4) were detected in the presence of increasing levels of RNase E-Dependent manner. To determine the breadth of the L4 in N3433 at 32 °C (compare lane 2 versus lane 1). effect of L4 inhibition of RNase E activity on the cellular abundance of E. coli transcripts in vivo, we used microarray analysis to compare the steady-state levels of 4290 E. coli transcripts in the presence (L4) or absence (control) of L4 able for processing of stable RNAs (22), we hypothesized that the ectopic expression by using the wild-type (N3433) and rne L4-RNase E interaction would most likely not affect stable RNA mutant BZ453 strains (Fig. 4A). As the RNase E polypeptide processing. Consistent with this notion, we found that the RNase produced in strain BZ453 is C-terminally truncated, and there- E-mediated processing of 5S rRNA (32), tRNA (6, 7), 6S RNA fore lacks the ability to interact with L4, transcripts whose (9), and M1 RNA (8), the catalytic RNA subunit of RNase P, in steady-state level is specifically dependent on the L4-RNase E vivo was similar in the presence (L4) or absence (control) of L4 interaction were expected to be detected as RNAs that are ectopic expression (Fig. 3D–F, compare lane 2 with lane 1). up-regulated during L4 ectopic expression in N3433, but not in This finding enables the use of stable RNAs as internal BZ453. We used these criteria to identify 65 transcripts whose standards for normalizing the amount of individual transcripts relative levels were increased at least 1.5 fold in N3433 (see Table during L4 ectopic expression. S2). Interestingly, these transcripts included products of five

866 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0810205106 Singh et al. Downloaded by guest on October 2, 2021 A wt E wt rnets 32°C °C 32°C °C min Control L4 44 44 after rif .25 1.5 3 4.5 6 8 .25 1.5 3 4.5 6 8 .25 1.5 3 4.5 6 8 .25 1.5 3 4.5 6 8 .25 1.5 3 6 12 .25 1.5 3 6 12 cstA

16S ± ± t ~3.3±0.2 min ND t ~3.4±0.1 min t ~9.5±0.7min t1/2~3.2 0.2 min t1/2~5.5 0.7 min 1/2 1/2 1/2

wt B F wt rnets min Control L4 32°C 44°C 32°C 44°C after rif .25 1.5 3 4.5 6 8 .25 1.5 3 4.5 6 8 .25 1.5 3 4.5 6 8 .25 1.5 3 4.5 6 8 .25 1.5 3 6 12 .25 1.5 3 6 12 yjiY

16S t ~3±0.1 min t ~6±0.1 min ± ± ± 1/2 1/2 t1/2 ~3.4 0.4 min ND t1/2~3.6 0.6 min t1/2~10.9 0.2 min

wt C G wt rnets min Control L4 32°C 44°C 32°C 44°C after rif .251.5 3 4.5 6 8 .25 1.5 3 4.5 6 8 .25 1.5 3 4.5 6 8 .25 1.5 3 4.5 6 8 .25 1.5 3 6 12 .25 1.5 3 6 12 lon

16S ± ± ± t1/2~3.4 0.2 min ± t ~3.4±.0.1 min ND t1/2~4.5 0.1 min t1/2~6.5 0.7 min t1/2~6.2 0.8 min 1/2

wt D H wt rnets min Control L4 32°C 44°C 32°C 44°C after rif .25 1.5 3 4.5 6 8 .25 1.5 3 4.5 6 8 .25 2 4 6 8 12 .25 2 4 6 8 12

.25 1.5 3 4.5 6 8 .25 1.5 3 4.5 6 8 GENETICS hslO

16S ± ± t1/2~2.7 0.6 min t1/2~5 0.2 min ± ND ± t1/2~2.6 0.22 min t1/2~3 0.1 min t1/2>12 min

Fig. 5. Northern blot analysis of transcripts whose rate of decay is controlled by RNase E in a L4-dependent manner. (A–D) Total RNA was isolated from the wild-type N3433 (wt) strain carrying plasmid pPW500flag (control) or plasmid pPWflagL4 encoding L4 (L4) after induction with IPTG (0.5 mM) for 30 min prior rifampicin treatment. Equal amounts of total RNA were analyzed by Northern blotting by using radioactively labeled RNA probes specific for the cstA, yjiY, lon, or hslO mRNAs (A–D, respectively) as described in Materials and Methods. The same procedure and probes were used to analyze RNA extracted from the rne ts wild-type N3433 (wt) and its isogenic mutant N3431 (rne ) at permissive (32 °C) and nonpermissive (44 °C) temperatures (see panel E–H). The half-life (t1/2)of each transcript and standard deviations were calculated based on the intensity of radioactive signals normalized to the amount 16S rRNA (internal loading control) shown at the bottom of each panel. Due to the low intensity of signals presumably caused by rapid turnover of these transcripts at 44 °C, it was not possible to calculate their half-lives in N3433 (ND).

genes (cstA, yjiY, lon, hslO, and yaeL) found previously to be several of the genes most prominently affected by the L4-RNase involved in bacterial stress responses, which are known to elevate E interaction (see Fig. 4C) was verified by quantitative real-time the level of free r-proteins in vivo (33). Genes that encode PCR (qRT-PCR). The qRT-PCR data obtained were in good remaining transcripts have been implicated in carbohydrate and agreement with the results of the microarray analysis (Table S3). amino acid metabolism and transport (membrane-associated proteins) (e.g., sdhA, sdhA, sdhC, and sdhD), transcription/ L4 Can Stabilize E. coli mRNA Species Encoding Stress-Induced Pro- translation (e.g., rimK, rpoC, rpoS, etc.), and DNA/RNA mod- teins. Previous findings that the intracellular concentration of ification (rmuC, topA, and rne). free r-proteins is elevated under stress conditions (33) raised the In addition, we found that among the 15% to 20% of genomic possibility that the observed L4-stimulated RNase E-dependent transcripts that were affected by the ectopic expression of L4 increase in abundance of stress-responsive transcripts (in par- (Table S3), Ϸ3.6% (154) of transcripts were down-regulated and ticular, cstA, yjiY, lon, hslO, and yaeL mRNA) (Fig. 4C, Table Ϸ2.3% (99) of transcripts were up-regulated in both the N3433 S2), results from inhibition of mRNA decay by the L4/RNase E and BZ453 strains (i.e., their abundance was decreased or interaction. To test this notion, we first determined whether the increased by ectopic L4 expression even in the absence of the half-lives of these transcripts are increased upon L4 ectopic RNase E segment required for interaction with L4), indicating expression by Northern blot analysis or ribonuclease protection the RNase E independence of this regulation. The down- assay. We found (Fig. 5A–D and Fig. S4A) that an increase in the regulated species include the polycistronic transcript detected level of L4 prolonged the half-life of each transcript, consistent with probes complementary to several ORFs (rpsJ, rplC, rplW, with the hypothesized role of this r-protein in inhibiting mRNA rplB, and rpsS, Fig. 4B) within the 5Ј-proximal portion of the S10 turnover. operon (Fig. 4B and Dataset S1), consistent with the previous The above result (Fig. 5 and Fig. S4A) together with microar- finding that L4 is capable of down-regulating transcription and ray data (Fig. 4C) argue that the observed stabilization of cstA, translation of its own operon (24, 25). Unlike its chromosomally yjiY, lon, hslO, and yaeL mRNAs upon L4 ectopic expression encoded counterpart, FLAG-L4 was ectopically expressed from results from inhibition of RNase E-mediated decay. Indeed, a high copy number plasmid and therefore the level of its own further analysis revealed that all five transcripts were stabilized mRNA was among the transcripts up-regulated in both strains upon inactivation of RNase E at nonpermissive temperature (Fig. 4B, and Dataset S1). (44 °C) in an RNase E temperature-sensitive mutant (rnets) but not To corroborate the microarray data, perturbed expression of in the wild-type strain (Fig. 5E–H and Fig. S4B; compare rnets with

Singh et al. PNAS ͉ January 20, 2009 ͉ vol. 106 ͉ no. 3 ͉ 867 Downloaded by guest on October 2, 2021 wt), demonstrating a role for RNase E in controlling the degrada- and RraB, proposed to impede RNA decay by remodeling the tion of these mRNA species (cstA, yjiY, lon, hslO, and yaeL). degradosome (36), L4 inhibits the RNase E-dependent RNA decay without altering the degradosome composition, which was Discussion found to be nearly the same in the absence or presence of L4 Previous work has shown that RNase E along with PNPase, RhlB ectopic expression (Fig. S2B). helicase, and enolase form a multienzyme complex termed as the Our microarray data not only suggest the existence of multiple ‘‘RNA degradosome.’’ Although a number of proteins are mRNA targets controlled by L4 in an RNase E-dependent known to be present in this complex as minor components, very manner but also disclose a large number of transcripts that are little is known about their specific roles in the regulation of up- or down-regulated by L4 independently of the L4/RNase E degradosome activity, assembly, or composition (for review, see interaction (Dataset S1). The group of down-regulated tran- ref. 3). Here, we affinity purified RNase E and determined the scripts includes species that are transcribed from the S10 operon nature of the co-purified proteins with molecular weights below known to be controlled by L4 at the transcriptional and trans- 30 kDa. Although several r-proteins were found to co-purify with lational levels (24, 25). It seems possible that, similar to its FLAG-RNase E (see Table S1), further analysis revealed that control of the S10-derived transcripts, L4 might down-regulate only one of them, the r-protein L4, could directly interact with the level of these mRNAs by impairing their translation, the degradosome by binding to the C-terminal half of RNase E. thereby increasing the level of ribosome-free mRNAs that are L4, an integral component of the 50S ribosomal subunit, is intrinsically more susceptible to degradation by the RNA known to regulate its own operon (i.e., the S10 operon) by decay machinery. repressing its transcription and translation (24, 25), which are actions involving two distinct nonoverlapping domains (26). In Materials and Methods the present study, L4 mutant variants lacking either of these Bacterial Strains and Plasmids. The E. coli strain BL21(DE3) (rne131) (22) functional domains were still able to interact with RNase E (Fig. encoding a truncated version of RNase E (amino acid 1–585), and its parental S2A), suggesting that L4 includes multiple domains that carry strain BL21(DE3) were used to prepare protein samples for immunoprecipi- out extraribosomal functions. tation, copurification of RNase E and L4 variants, and to overexpress FLAG- Cleavage assays revealed that RNase E cleavage of oligonu- tagged proteins. The strain DHP1, an adenylate cyclase-deficient derivative of Ј DH1 (38), was used in E. coli two-hybrid assays to study protein-protein cleotide BR13 representing the 5 -end single-stranded segment interactions in vivo (27) as described in the SI Materials and Methods. The E. of RNAI, an antisense RNA controlling the copy number of coli K-12 strains N3433 (39) and BZ453 (31) encoding the full-length RNase E ColE1-type plasmids (29), decreased on L4 binding. Moreover, and its truncated version (amino acid 1–602), respectively, were used for the L4-dependent inhibition of RNase E activity in vitro was not determining RNA stability, plasmid copy number, detection of endogenous observed for an RNase E variant protein containing only the RNase E, and microarray analysis. The pnpϪ strain YHC012 (40) was used to catalytic domain of the enzyme (Fig. 2). Consistent with the purify FLAG-tagged RNase E, N-RNase E (amino acid 1–597), C-RNase E (amino ability of L4 to inhibit RNase E cleavages in vitro, we found that acid 499–106), and FLAG-L4 proteins. ectopic expression of L4 in vivo increased the level and stability of RNAI in vivo, in turn resulting in a lower copy number of the Basic Biochemical and Molecular Biology Techniques. Protein purification, resident plasmid in cells synthesizing full length RNase E but not RNase E cleavage assays, Northern blot analysis, and plasmid copy number in cells producing an RNase E variant that lacks the C-terminal determination were performed as previously described (12, 19, 29) (see SI). region. Similarly, the L4-dependent decrease in RNase E activity Testing the Functionality of Affinity-Purified FLAG-L4. To confirm that, similar also increased the level of rne mRNA (see Fig. 4B), which is to its chromosomally encoded counterpart, the recombinant FLAG-L4 can known to be cleaved by this endoribonuclease during autoreg- autoregulate its own synthesis by binding to the S10 leader (41), we carried ulation of its own synthesis in vivo (15). Microarray analysis out in vitro translation by using pET29-⌬S10, a pET29-based plasmid encoding revealed that L4-mediated inhibition of RNase E activity was the S10 leader-rpsJ-rplC-rplD region, as a template and an in vitro translation associated with an increase in the steady-state level of numerous kit (Rapid Translation System RTS 100, E. coli HY kit, Roche). The resulting mRNAs, including multiple stress-responsive transcripts (cstA, 35S-labeled products of translation were further analyzed on 12% SDS poly- yjiY, lon, hslO, and yaeL). Moreover, Northern blot analysis acrylamide gels followed by exposure to x-ray films. revealed that the increased abundance of cstA, yjiY, lon, hslO, and yaeL mRNAs upon L4 ectopic expression resulted from Microarray and Quantitative RT-PCR. The relative mRNA abundance of 4290 E. inhibition of RNase E-dependent decay of these mRNAs. coli transcripts was analyzed in the presence (L4) or absence (control) of L4 ectopic expression by using the wild-type (N3433) and rne mutant BZ453 E. coli The level of free r-proteins is known to increase with strains carrying the plasmids pPW500flag (vector only) and pPWflagL4 (vector in response to stresses (e.g., temperature upshifts or amino acid plus L4). RNA was isolated from cells grown at 32 °C and induced with IPTG at starvation, a carbon source, nitrogen, or phosphate [33]) that an OD600 Ϸ 0.3 for 30 min. RNA isolation, synthesis of fluorescently labeled trigger the nucleolytic attack of ribosomal RNA, leading to cDNA, hybridization to microarrays, and data analysis were performed as ribosome disassembly (reviewed in ref. 34). The stress-induced described in SI Materials and Methods to generate Dataset S1. For quantita- disassembly of ribosome particles and accumulation of free tive RT-PCR, gene-specific primers (see Table S4) were designed with the r-proteins, including L4, can potentially inhibit the RNase Beacon Designer 5 software package (Premier Biosoft). Quantitative PCR and E-mediated decay of the above stress-responsive transcripts, data collection were performed in the real time PCR system (MiniOpticon, thereby facilitating bacterial adaptation to adverse environ- Bio-Rad). ments. Our finding that L4 can potentially increase the level of ACKNOWLEDGMENTS. We thank I. Moll (Vienna University, Vienna, Austria) Lon by inhibiting the degradation of lon mRNA suggests that the for kindly providing several antibodies specific for r-proteins S1, L2, L3, S3, S4, L4-mediated inhibition of RNase E activity might contribute to and L4; K.-F. Chak (National Yang-Ming University, Taipei, Taiwan) for specific the Lon-dependent degradation of r-proteins in response to antibodies for TolA and TolB; editors H. Wilson and M. Loney (Institute of amino acid starvation (stringent response) to provide the extra Molecular Biology, Academia Sinica) for help in editing the manuscript; S.-Y. Tung (IMB Microarray Core Facility) for excellent technical support; G.-G. Liou amino acids necessary for the synthesis of stress-related proteins (35). for generating preliminary data; Y.-G. Tsay (National Yang-Ming University) Interestingly, our data also suggest that L4 does not act as a for mass-spectroscopic analysis; and C.-S. Lin and H.-Y. Chen for technical general inhibitor of RNase E to stabilize every mRNA whose support. This work was supported by grants from the National Science Council, decay is mediated by this enzyme (e.g., ompA mRNA, see Fig. Taiwan (NSC 94/95-2311-B-001-034; NSC 97-2321-B-001-014) and by an intra- mural fund from Academia Sinica to S.L.-C. Authors D.S, S.-J.C., and O.V.A. 4C). This mode of L4 inhibitory action is similar to RraA and received Post Doctoral Fellowships from the National Science Council, Taiwan. RraB, which are also inhibitors of RNase E and affect only V.R.K. was supported by the Thematic Research Program, Academia Sinica (AS selective groups of transcripts (36, 37). However, unlike RraA 97-23-22).

868 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0810205106 Singh et al. Downloaded by guest on October 2, 2021 1. Carpousis AJ (2007) The RNA degradosome of Escherichia coli: An mRNA-degrading 21. Bernstein JA, Lin PH, Cohen SN, Lin-Chao S (2004) Global analysis of Escherichia coli RNA machine assembled on RNase E. Annu Rev Microbiol 61:71–87. degradosome function using DNA microarrays. Proc Natl Acad Sci USA 101:2758–2763. 2. Deutscher MP (2006) Degradation of RNA in bacteria: Comparison of mRNA and stable 22. Lopez PJ, Marchand I, Joyce SA, Dreyfus M (1999) The C-terminal half of RNase E, which RNA. Nucleic Acids Res 34:659–666. organizes the Escherichia coli degradosome, participates in mRNA degradation but not 3. Regnier P, Arraiano CM (2000) Degradation of mRNA in bacteria: Emergence of rRNA processing in vivo. Mol Microbiol 33:188–199. ubiquitous features. Bioessays 22:235–244. 23. Wool IG (1996) Extraribosomal functions of ribosomal proteins. Trends Biochem Sci 4. Li Z, Pandit S, Deutscher MP (1999) RNase G (CafA protein) and RNase E are both 21:164–165. required for the 5Ј maturation of 16S ribosomal RNA. EMBO J 18:2878–2885. 24. Lindahl L, Zengel JM (1979) Operon-specific regulation of ribosomal protein synthesis 5. Misra TK, Apirion D (1979) RNase E, an RNA processing enzyme from Escherichia coli. in Escherichia coli. Proc Natl Acad Sci USA 76:6542–6546. J Biol Chem 254:11154–11159. 25. Yates JL, Nomura M (1980) E. coli ribosomal protein L4 is a feedback regulatory protein. 6. Li Z, Deutscher MP (2002) RNase E plays an essential role in the maturation of Cell 21:517–522. Escherichia coli tRNA precursors. RNA 8:97–109. 26. Li X, Lindahl L, Zengel JM (1996) Ribosomal protein L4 from Escherichia coli utilizes 7. Ow MC, Kushner SR (2002) Initiation of tRNA maturation by RNase E is essential for cell nonidentical determinants for its structural and regulatory functions. RNA 2:24–37. viability in E. coli. Genes Dev 16:1102–1115. 27. Karimova G, Pidoux J, Ullmann A, Ladant D (1998) A bacterial two-hybrid system based 8. Lundberg U, Altman S (1995) Processing of the precursor to the catalytic RNA subunit on a reconstituted signal transduction pathway. Proc Natl Acad Sci USA 95:5752–5756. of RNase P from Escherichia coli. RNA 1:327–334. 28. Worrall JA, et al. (2008) Reconstitution and analysis of the multienzyme Escherichia coli 9. Kim KS, Lee Y (2004) Regulation of 6S RNA biogenesis by switching utilization of both RNA degradosome. J Mol Biol 382:870–883. sigma factors and . Nucleic Acids Res 32:6057–6068. 29. Lin-Chao S, Cohen SN (1991) The rate of processing and degradation of antisense RNAI 10. Cohen SN, McDowall KJ (1997) RNase E: still a wonderfully mysterious enzyme. Mol regulates the replication of ColE1-type plasmids in vivo. Cell 65:1233–1242. Microbiol 23:1099–1106. 30. McDowall KJ, Kaberdin VR, Wu SW, Cohen SN, Lin-Chao S (1995) Site-specific RNase E 11. Lin-Chao S, Wong TT, McDowall KJ, Cohen SN (1994) Effects of nucleotide sequence on cleavage of oligonucleotides and inhibition by stem-loops. Nature 374:287–290. the specificity of rne-dependent and RNase E-mediated cleavages of RNA I encoded by 31. Kido M, et al. (1996) RNase E polypeptides lacking a carboxyl-terminal half suppress a the pBR322 plasmid. J Biol Chem 269:10797–10803. mukB mutation in Escherichia coli. J Bacteriol 178:3917–3925. 12. McDowall KJ, Lin-Chao S, Cohen SN (1994) AϩU content rather than a particular 32. Ghora BK, Apirion D (1978) Structural analysis and in vitro processing to p5 rRNA of a nucleotide order determines the specificity of RNase E cleavage. J Biol Chem 9S RNA molecule isolated from an rne mutant of E. coli. Cell 15:1055–1066. 269:10790–10796. 33. Kaplan R, Apirion D (1975) Decay of ribosomal ribonucleic acid in Escherichia coli cells 13. Kaberdin VR (2003) Probing the substrate specificity of Escherichia coli RNase E using starved for various nutrients. J Biol Chem 250:3174–3178. a novel oligonucleotide-based assay. Nucleic Acids Res 31:4710–4716. 34. Deutscher MP (2003) Degradation of stable RNA in bacteria. J Biol Chem 278:45041– 14. Diwa A, Bricker AL, Jain C, Belasco JG (2000) An evolutionarily conserved RNA stem- loop functions as a sensor that directs feedback regulation of RNase E gene expression. 45044. Genes Dev 14:1249–1260. 35. Kuroda A, et al. (2001) Role of inorganic polyphosphate in promoting ribosomal 15. Jain C, Belasco JG (1995) RNase E autoregulates its synthesis by controlling the degra- protein degradation by the Lon protease in E. coli. Science 293:705–708. dation rate of its own mRNA in Escherichia coli: Unusual sensitivity of the rne transcript 36. Gao J, et al. (2006) Differential modulation of E. coli mRNA abundance by inhibitory

to RNase E activity. Genes Dev 9:84–96. proteins that alter the composition of the degradosome. Mol Microbiol 61:394–406. GENETICS 16. Mudd EA, Higgins CF (1993) Escherichia coli endoribonuclease RNase E: Autoregulation 37. Lee K, et al. (2003) RraA. a protein inhibitor of RNase E activity that globally modulates of expression and site-specific cleavage of mRNA. Mol Microbiol 9:557–568. RNA abundance in E. coli. Cell 114:623–634. 17. Kaberdin VR, et al. (1998) The endoribonucleolytic N-terminal half of Escherichia coli 38. Hanahan D (1983) Studies on transformation of Escherichia coli with plasmids. JMol RNase E is evolutionarily conserved in Synechocystis sp. and other bacteria but not the Biol 166:557–580. C-terminal half, which is sufficient for degradosome assembly. Proc Natl Acad Sci USA 39. Goldblum K, Apririon D (1981) Inactivation of the ribonucleic acid-processing enzyme 95:11637–11642. ribonuclease E blocks cell division. J Bacteriol 146:128–132. 18. Vanzo NF, et al. (1998) Ribonuclease E organizes the protein interactions in the 40. Kaberdin VR, Chao YH, Lin-Chao S (1996) RNase E cleaves at multiple sites in bubble Escherichia coli RNA degradosome. Genes Dev 12:2770–2781. regions of RNA I stem loops yielding products that dissociate differentially from the 19. Miczak A, Kaberdin VR, Wei CL, Lin-Chao S (1996) Proteins associated with RNase E in enzyme. J Biol Chem 271:13103–13109. a multicomponent ribonucleolytic complex. Proc Natl Acad Sci USA 93:3865–3869. 41. Freedman LP, Zengel JM, Archer RH, Lindahl L (1987) Autogenous control of the S10 20. Py B, Higgins CF, Krisch HM, Carpousis AJ (1996) A DEAD-box RNA helicase in the ribosomal protein operon of Escherichia coli: Genetic dissection of transcriptional and Escherichia coli RNA degradosome. Nature 381:169–172. posttranscriptional regulation. Proc Natl Acad Sci USA 84:6516–6520.

Singh et al. PNAS ͉ January 20, 2009 ͉ vol. 106 ͉ no. 3 ͉ 869 Downloaded by guest on October 2, 2021