A Small Regulatory RNA Generated from the malK 5′ Untranslated Region Targets Gluconeogenesis in Vibrio Species

ABSTRACT Vsr217 is a small RNA from Vibrio tasmaniensis LGP32, a pathogen associated with mortality events affecting juvenile oysters. The vsr217 gene is located within the 5′ untranslated region (UTR) of malK, encoding the ATPase component of the maltose importer, and is conserved within the genus Vibrio. In the presence of maltose, vsr217 is regulated by MalT, the positive regulator of the maltose regulon. vsr217 is required in cis for the full expression of malK. In addition, Vsr217 acts in trans to downregulate the expression of fbp encoding fructose-1,6-bisphosphatase, an enzyme involved in gluconeogenesis. Thus, in the presence of maltose, the induction of Vsr217 is expected to promote glycolysis by negatively regulating the expression of a key enzyme of gluconeogenesis. IMPORTANCE Juvenile pacific oysters have been subject in recent years to summer mortality episodes with deep economic consequences. The pathogen Vibrio tasmaniensis has been associated with such mortality events. For bacterial pathogens, survival within the host requires profound metabolic adaptations according to available resources. All kinds of regulatory elements, including noncoding RNAs, orchestrate this response. Oysters are rich in glycogen, a precursor of maltose, and we previously reported that V. tasmaniensis maltose-regulated genes are strongly induced during oyster infection. Here, we report the dual mechanism by which a small regulatory RNA, generated from the 5′ untranslated region of a gene belonging to the maltose regulon, acts both in cis and trans. In cis, it stimulates growth on maltose, and in trans, it downregulates the expression of a gene associated with gluconeogenesis, thus coordinating maltose utilization with central carbon metabolism.

IMPORTANCE Juvenile pacific oysters have been subject in recent years to summer mortality episodes with deep economic consequences. The pathogen Vibrio tasmaniensis has been associated with such mortality events. For bacterial pathogens, survival within the host requires profound metabolic adaptations according to available resources. All kinds of regulatory elements, including noncoding RNAs, orchestrate this response. Oysters are rich in glycogen, a precursor of maltose, and we previously reported that V. tasmaniensis maltose-regulated genes are strongly induced during oyster infection. Here, we report the dual mechanism by which a small regulatory RNA, generated from the 59 untranslated region of a gene belonging to the maltose regulon, acts both in cis and trans. In cis, it stimulates growth on maltose, and in trans, it downregulates the expression of a gene associated with gluconeogenesis, thus coordinating maltose utilization with central carbon metabolism. KEYWORDS Vibrio, sRNA, 59 UTR, maltose regulon, gluconeogenesis I n the last 20 years, the importance of noncoding RNAs (sRNAs) in bacterial regulatory networks has become widely recognized. The development of genome-wide transcriptomic studies has led to the exploration of transcriptomic landscapes of nonmodel bacterial organisms, either pathogens or those of environmental interest, leading to the discovery of many sRNAs (for reviews, see references 1-3). While classical bona fide sRNAs are expressed from intergenic regions, a newly recognized class of sRNAs includes sRNAs generated from 59 and 39 untranslated regions (UTRs) of coding genes that act in trans on genetically unlinked targets. A genome-wide search of small regulatory RNAs in Escherichia coli led to the identification of several sRNAs expressed from 59 or 39 UTRs (4), but only a few of them, mostly stemming from 39 UTRs, have been characterized. For instance, CpxQ is generated from the 39 UTR of cpxP mRNA, which encodes a periplasmic chaperone involved in the inner membrane stress response. CpxQ was shown to repress several inner membrane proteins, thus mitigating membrane stress (5). Another example is MicL, which is transcribed from a s E -dependent promoter internal to the upstream protein coding sequence, and then cleaved to a shorter product to act in cis to regulate the expression of the major outer membrane lipoprotein Lpp (6). A few examples are associated with Vibrio species. In Vibrio cholerae, MicX is generated by an internal promoter within the malG open reading frame. The transcript is processed, and the shorter version, more stable and corresponding largely to the malG 39 UTR, regulates an outer membrane protein, ChiP, annotated as a chitoporin, and an ABC transporter (7). Recently, two V. cholerae sRNAs, CarZ and OppZ, generated by processing of the 3'UTRs of carAB and oppABCDF mRNA, respectively, were shown to negatively regulate their respective operons (8).
trans-acting sRNAs can also be generated from 59 UTRs. In the human pathogen Listeria monocytogenes, PrfA is a virulence regulator whose expression is controlled at the translational level by a thermosensor located in the 59 UTR of the prfA transcript (9). However, two S-adenosylmethionine (SAM) riboswitches, SreA and SreB, not only regulate their downstream gene expression in cis, but also act in trans to target the 59 UTR of the prfA mRNA (10). Another recent example is the sRNA rnTrpL (11) in Sinorhizobium meliloti, which is derived from a transcriptional attenuator of one of the three tryptophan operons and regulate in trans a second tryptophan operon, which does not possess the leader peptide-encoding attenuator. Other examples include Teg49 in Staphylococcus aureus, derived from the 59 UTR of the virulence regulator gene sarA (12), or malS 59 UTR, a Salmonella enterica noncoding RNA generated from the 59 UTR of malS that was proposed to regulate in trans hisG, a gene involved in the histidine biosynthetic pathway (13).
Vibrio tasmaniensis LGP32 (previously V. splendidus LGP32) is a member of the V. splendidus clade (14) that has been associated with summer mortalities of juvenile Pacific oysters Crassostrea gigas (15). It is a facultative intracellular pathogen of oysters, invading the oyster hemocytes, the host immune cells (16)(17)(18). A genome-wide search for noncoding sRNAs in V. tasmaniensis LGP32 by transcriptome sequencing (RNA-seq) led us to identify 250 potential sRNAs, among which, several were confirmed by Northern blots and reverse transcription-PCRs (RT-PCRs) (19). More recently, a transcriptomic analysis of LGP32 within the host revealed that vsr217 was one of the most highly induced genes compared to that during growth under seawater conditions (18). Interestingly, vsr217 is located upstream of malK, a gene encoding the ATPase subunit of the maltose transporter, a member of the ABC transporter family (reviewed in reference 20). In addition, the maltose regulon was highly induced in oysters, a metabolic switch that can be related to the fact that oysters are rich in glycogen, a precursor of maltose (18).
Here, we show that vsr217 is a member of the maltose regulon and is positively controlled by the transcriptional regulator MalT. While, within the 59 UTR of malK mRNA, the Vsr217 sequence acts in cis to stabilize malK mRNA, it also generates an sRNA that acts in trans to repress the expression of a key gene involved in gluconeogenesis.

RESULTS
The genetic organization of vsr217-malK is conserved in different Vibrio species. Vsr217 is an sRNA from V. tasmaniensis expressed from the malK 59 UTR. We investigated its conservation within the Vibrio genus by sequence comparison of malK upstream sequences from species representative of different Vibrio clades (14) (Fig. 1). Five regions are especially conserved. Region I is immediately upstream of a predicted 235 promoter sequence (P1). Region II overlaps with the 210 sequence of the P1 promoter and includes a putative MalT operator. MalT is the positive regulator of the maltose regulon in E. coli and other Enterobacteriaceae (21) which binds to MalT consensus sequences (22). Region III corresponds to the 59 end of Vsr217 (see below) and includes the 32 first nucleotides of the sRNA. Region IV contains a conserved TTTTC motif present at the 39 end of vsr217. Region V is the largest conserved sequence and contains a second predicted s 70 promoter (P2) as well as the ribosome binding site for malK. No Rho-independent terminator was predicted between the 39 end of vsr217 and malK. This analysis highlights the conservation of a large putative regulatory region upstream of malK in Vibrio.
Vsr217 is induced by maltose in a MalT-dependent manner and is generated from the 59 UTR of malK. The conserved localization of vsr217 upstream of malK and the presence of a MalT box overlapping the putative P1 promoter of vsr217 ( Fig. 1) suggested that Vsr217 is a member of the maltose regulon controlled by MalT. The production of Vsr217 and malK mRNA was tested by Northern blots in the presence or absence of maltose ( Fig. 2A). Both Vsr217 and malK mRNA were strongly induced by  (14) were aligned using Muscle (39). For reason of space, a less-conserved sequence between regions III and IV is not shown, as indicated by two slashes. Conserved nucleotides are indicated by * under the alignment. V. tasmaniensis sequence is in red, with the vsr217 sequence in italic font. Important features are in boldface font, including Vsr217 59 and 39 ends; 235 and 210 regions of promoters predicted using BPROM (see Materials and Methods) are underlined. malK initiation codon is indicated in boldface red font.
Vsr217, a Novel Maltose Regulon sRNA in Vibrio maltose, whereas no expression was detected in the presence of glucose. Vsr217 and malK mRNA were not detected in the malT mutant, indicating that their expression is MalT dependent. Under the maltose condition, a high-molecular-weight band was detected with both vsr217 and malK probes, indicating the existence of a large transcript encompassing the two genes. A smaller band was detected with the malK probe but not with the vsr217 probe, indicating the existence of a transcript including malK RNA without Vsr217. The transcription of malK was also tested in a mutant harboring a complete deletion of the vsr217 sequence (see below and Fig. 2B). Using the malK probe, a single band was observed whose size was compatible with that of a malK transcript originating from the putative promoter P1, shortened by the absence of vsr217 (Fig. 1A). Of note, malK mRNA quantity was reduced in the Dvsr217 mutant compared to that in its parental strain ( Fig. 2A). The 59 and 39 ends of transcripts from the vsr217-malK locus were determined by 59 and 39 rapid amplification of cDNA ends (RACE) mappings (Fig. 2B). For Vsr217, all clones displayed the same extremities. The Vsr217 59 end was compatible with a transcription stemming from the P1 promoter, while the transcript terminated at the level of a conserved TTTTC motif. In contrast, for malK mRNA, if the 39 end was unique, due to the presence of a Rho-independent terminator, different malK 59 ends were identified. The majority of them (57.5%) corresponded to a transcription start site from the P2 promoter. Other 59 end sequences mapped upstream of P2, some of them within vsr217. Among them, 6% had a 59 end corresponding to a transcription initiation from P1, hence including the whole Vsr217 sequence (Fig. 2B). The existence of a vsr217-malK transcript was also confirmed by RT-PCR, whose amount increased in the presence of maltose (Fig. 2C).
We concluded that Vsr217 originates from the malK P1 promoter. The existence of transcripts encompassing Vsr217-malK mRNA suggested that Vsr217 is generated from malK 59 UTR and that processing of the 59 end of this larger transcript generates malK transcripts.
The 59 end of the malK mRNA in the Dvsr217 mutant was determined by 59 RACE. In contrast to the parental strain, the majority of the malK transcripts (64.2%) in the Dvsr217 mutant had a 59 end generated from the P1 promoter, whereas only 21.4% corresponded to transcripts generated from the P2 promoter (Fig. 3C).
Does the malK mRNA level reduction in the Dvsr217 mutant affect growth in the presence of maltose as the sole carbon source? Indeed, when grown in minimal medium supplemented with maltose, the Dvsr217 mutant had a lower growth rate than its parental strain. This difference was not observed in minimal medium supplemented with glucose (see Fig. S1 in the supplemental material). The maltose-dependent growth defect of the Dvsr217 mutant was corrected by the presence of a plasmid expressing malK only but not by a plasmid expressing vsr217 under the P1 promoter (Table 1). We concluded that the growth defect in the mutant is due to the cis effect of the deletion of vsr217 on malK mRNA stability, leading to the decrease of the message.
A U-rich terminal motif of Vsr217 is a stabilizing factor for Vsr217 but not for malK mRNA. Vsr217 contains a conserved U-rich motif at its 39 end (Fig. 1A). We hypothesized that it could play a role in generating Vsr217, either by promoting transcription termination, by being a processing site, or by contributing to its stability. To test this, we constructed a vsr217 mutant lacking its 13 last nucleotides (DpolyT) and checked by Northern blotting how the deletion would affect the amounts of Vsr217, malK mRNA, and their precursor (Fig. 4A).
The absence of the last 13 nucleotides (DpolyT mutant) led to a 100-fold reduction of Vsr217 compared to that in the WT strain, whereas the amount of malK mRNA was not affected in contrast to what was observed in the Dvsr217 mutant. Consistent with this result, we observed that malK mRNA stability did not decrease in the DpolyT mutant compared to that in the WT strain, contrary to Vsr217 stability (Fig. 4B). Surprisingly, the quantity of the precursor transcript vsr217-malK mRNA was also not affected in the DpolyT mutant as could have been expected if the deleted terminal motif was involved in a processing event generating Vsr217 and malK mRNA. Altogether, our results suggest that the U-rich terminal motif is involved in Vsr217 stability but not in malK mRNA stability.
In addition, no growth difference was observed between the DpolyT mutant and its parental strain in minimal medium supplemented with either glucose or maltose Vsr217, a Novel Maltose Regulon sRNA in Vibrio ( Fig. 4C), confirming that the growth defect observed in the Dvsr217 mutant was due to a cis effect on malK expression and not to the absence of Vsr217 production.
Vsr217 affects the expression of fbp, a gluconeogenic enzyme. The existence of a stable, abundant Vsr217 sRNA suggests a trans-acting function on targets possibly not directly related to maltose utilization but nevertheless regulated in response to maltose. Putative Vsr217 targets were searched by identifying proteins whose expression levels varied between the WT and Dvsr217 strains, especially when grown in maltose. Using a proteomic mass spectrometry approach, 2,104 proteins from 4,371 coding DNA sequences (CDSs) harbored by the V. tasmaniensis genome were identified. In glucose, the protein contents were the same between the two strains. However, in maltose, the amounts of 22 proteins were significantly altered (fold change of .2 or ,0.5, P value , 0.05) in the mutant, with 11 increasing and 11 decreasing (Table 2). Among them, MalK decreased in the Dvsr217 mutant, consistent with the reduced amount of malK mRNA (Fig. 2).
Among these 22 potential targets of Vsr217, to identify candidates for direct mRNA targets, we used the software CopraRNA, which predicts sRNA targets and their pairing regions with the sRNA by comparative genomics (23). From the list of proteins whose levels were affected in the Dvsr217 mutant, only one came out in the list of the first 100 best candidates for direct interaction: VS_2771, ranked 8th by CopraRNA, corresponds to Fbp, a class 1 fructose-1,6-bisphosphatase that catalyzes the conversion of fructose-1,6-bisphosphate to fructose-6-phosphate, a rate-limiting step of the gluconeogenesis pathway ( Table 2). Fbp expression increased in the Dvsr217 mutant, suggesting that Vsr217 could be a negative regulator of fbp.
Since posttranscriptional inhibition by sRNAs often leads to a stability reduction of target mRNAs (24), we determined by Northern blotting the amount of fbp mRNA in minimal medium supplemented by either glucose or maltose (Fig. 5). To circumvent the cis effect of vsr217 RNA on malK expression, we took advantage of the DpolyT mutant, which has a drastically reduced amount of Vsr217 but still retains a wild-type expression level of malK. fbp was repressed in maltose compared to that in glucose in the WT, a difference that was abolished in the DpolyT mutant, suggesting that Vsr217 downregulates fbp expression.
fbp mRNA is a direct target of Vsr217. A biocomputing analysis using IntaRNA identified a potential base-pairing region between Vsr217 and fbp mRNA, ending 29 nucleotides (nt) upstream of the fbp initiation codon (Fig. 6A). To confirm a putative direct interaction, the Vsr217 trans activity on fbp expression was tested in E. coli using a two-plasmid-based reporter system (25,26). The first plasmid carried the vsr217 gene cloned under a P LlacO promoter (pPLac-Vsr217). The second plasmid contained an inframe gene fusion between the predicted pairing region of fbp mRNA including fbp initiation codon and the second codon of the gfp CDS ( Table 1). The reporter fusion was cloned under the control of a P Ltet promoter, which is constitutive in the absence of a TetR repressor. pPtet-Fbp-GFP was introduced in E. coli together with pPlac-Vsr217, and the effect of Vsr217 production on the expression of fbp was monitored via fluorescence measurements (see Materials and Methods).
The presence of pPlac-Vsr217 plasmid led to a 20% reduction in the expression of the fbp-gfp fusion compared to that in the strain containing a control empty plasmid LGP32 carrying a deletion of vsr217 This study DmalT LGP32 carrying a deletion of malT This study DpolyT LGP32 carrying a deletion of polyT at the 39 end of vsr217 This study E. coli DH5a F 2 endA1 glnV44 thi-1 recA1 relA1 gyrA96 deoR nupG purB20 w80dlacZDM15 D(lacZYA-argF)U169  (Fig. 6B). Expression was further reduced to ;40% when isopropyl-b-D-thiogalactopyranoside (IPTG) was added to fully induce vsr217 expression. The RNA chaperone Hfq is often required for the interaction of sRNAs with their targets, including in Vibrio (27)(28)(29). To determine whether the potential interaction  Vsr217, a Novel Maltose Regulon sRNA in Vibrio between Vsr217 and fbp mRNA was dependent upon Hfq, we also transformed a MG1655 Dhfq mutant with the same set of plasmids.
The fbp-gfp fusion had a 2-fold reduction of expression in the Dhfq mutant compared to that in the WT strain, but the presence of pPlac-Vsr217 had no effect in the absence of Hfq, indicating that the interaction was Hfq dependent. Altogether, these results indicate that fbp mRNA is submitted to posttranscriptional inhibition by Vsr217 in an Hfq-dependent manner.

DISCUSSION
V. tasmaniensis LGP32 Vsr217, identified in a genome-wide search for sRNAs (19), is highly induced during oyster infection (18). In this study, we show that vsr217 is a conserved gene in Vibrio, located upstream of malK, which is inducible by maltose in a MalT-dependent manner. Induction of Vsr217 by maltose could account for its high expression in oysters. Indeed, the latter are rich in glycogen, a precursor of maltose, and the maltose regulon of V. tasmaniensis is induced upon oyster colonization (18).
An original feature of Vsr217 is that it is generated from the 59 UTR of malK mRNA as attested by the detection of a vsr217-malK RNA. malK is transcribed from two promoters, P1, upstream of vsr217, and P2, located in the intergenic region between vsr217 and malK. P2 is responsible for the transcription of more than half of malK transcripts. The other half was distributed between transcripts with several 59 ends that could correspond to the results of processing of transcripts stemming from the P1 promoter. Altogether, these results are in favor of Vsr217 being generated by the processing of the larger transcript stemming from the P1 promoter, although we cannot exclude that Vsr217, which is much more stable than the malK transcript(s), is generated by termination of transcription at its unique 39 end. However, no Rho-independent terminator was detected to support this hypothesis. In addition, we found that deleting the conserved 13 last nucleotides of vsr217 (called here polyT motif) led to a nearly complete absence of Vsr217 but had no effect on vsr217-malK or malK mRNA levels. These results suggest that the polyT motif is important for Vsr217 stability but not for vsr217-malK processing.
Vsr217 RNA has two functions. In cis, it affects the stability of the malK mRNA, which decreases in the Dvsr217 deletion strain, leading to a reduction of MalK and a decreased ability to use maltose as a carbon source, which is suppressed by the expression in trans of malK but not of vsr217. In addition, Vsr217 acts as a bona fide sRNA to control the expression of fbp, a gene involved in the neoglucogenesis/glycolytic pathway. A decrease of fbp mRNA was observed in the presence of maltose, when Vsr217 is strongly induced, whereas no change was observed in glucose. This decrease was abolished in the DpolyT mutant, confirming that it is Vsr217 dependent in the WT strain. Direct interaction between Vsr217 and fbp was further supported by Vsr217-dependent downregulation of a fbp-gfp reporter fusion in E. coli (Fig. 6). In addition, we found that the Vsr217-dependent posttranscriptional downregulation of the fbp-gfp fusion is indeed dependent upon Hfq in E. coli. V. tasmaniensis Hfq has 94% identity and 99% similarity with E. coli Hfq. Accordingly, it is likely that the Vsr217-fbp interaction is dependent upon Hfq in V. tasmaniensis as well.
Interestingly, fbp encodes one of the two enzymes in the glycolytic/gluconeogenesis pathway that catalyze the two irreversible reactions distinguishing neoglucogenesis from glycolysis, the other one being ppsA, encoding phosphoenolpyruvate synthase. PpsA was also identified in our proteomic approach as increasing in the Dvsr217 mutant, suggesting that it could also be negatively regulated by Vsr217. However, ppsA mRNA was not predicted as a target by CopraRNA. Further studies will be required to determine the potential involvement of Vsr217 in ppsA expression.
In E. coli, two major transcriptional regulators have been described to regulate the central carbon metabolism, CRP (catabolite repressor protein) and Cra (catabolite repressor activator, also known as FruR) (30)(31)(32). Whereas on neoglucogenic substrates such as acetate, CRP activates key enzymes of glycolysis, Cra downregulates the majority of enzymes in the glycolytic pathway and activates the two gluconeogenesis genes fbp and ppsA. The action of Cra is dominant over that of CRP (33). Both CRP and Cra have homologues in V. tasmaniensis, VS_2855 and VS_II1031, respectively. We propose that Vsr217 contributes to fine-tuning the production of a gluconeogenic enzyme, bringing an additional control to Vibrio metabolic regulations (Fig. 7).
In the presence of maltose and absence of glucose, the subunit EIIA of the glucose phosphotransferase system (PTS) transporter is phosphorylated and activates FIG 7 Role of Vsr217 in carbon metabolism regulation of V. tasmaniensis. The model is inferred from knowledge in enterobacteria and the existence of CRP (VS_2855) and Cra (VS_II1031) orthologs in V. tasmaniensis. In the presence of maltose or maltotriose as a carbon source instead of glucose, EIIA(glc) is phosphorylated and activates adenylate cyclase (not shown). cAMP accumulates and activates the cAMP receptor protein (CRP), which regulates positively glycolysis. On the other hand, the catabolite repressor activator Cra stimulates gluconeogenesis by activating fbp (and possibly ppsA). Maltose activates the MalT regulon including maltose transport genes such as malK, malE, malFG and maltose metabolic genes malPQ, which transform maltose into the first compound of glycolysis, glucose 6-phosphate. Vsr217 is highly expressed and represses fbp expression, inhibiting gluconeogenesis and favoring optimal maltose utilization. Glycolysis is in green, with only relevant steps for this study being shown. Gluconeogenesis, orange; maltose regulon, yellow; glucose PTS transporter, dark yellow; mRNAs, blue; Vsr217, red. adenylate cyclase: cAMP accumulates in the cell, activating CRP, which upregulates glycolysis. Maltose is transformed into glucose-6-phosphate and enters the glycolytic pathway. However, upregulation of glycolysis is counteracted by Cra, which drives the carbon flux toward gluconeogenesis. On the other hand, induction of the maltose regulon leads to a strong increase of Vsr217 that will inhibit the expression of the gluconeogenic enzyme Fbp, thus favoring glycolysis. Hence, in oysters, which are rich in the maltose precursor glycogen, Vsr217, a member of the growing family of trans-acting sRNAs derived from a 59 UTR, may contribute to the fitness of V. tasmaniensis by optimizing maltose utilization through inhibition of gluconeogenesis.

MATERIALS AND METHODS
Bacterial strains and plasmids. Bacterial strains and plasmids used in this study are listed in Table 1. V. tasmaniensis LGP32 mutants were constructed by allele replacement through homologous recombination as described previously (34). The presence of the desired mutations was confirmed by sequencing. Plasmids were constructed by one-step isothermal assembly (Gibson assembly) (35). Their main features are indicated in Table 1, and details of construction are presented in Fig. S2 in the supplemental material. We also constructed a derivative of the cloning vector pXG-10 (26), deleting the fragment coding for the alpha peptide present upstream of the gfp gene which led to a background expression of gfp. The resulting plasmid, pXG-10Dalpha was used as a control plasmid to determine the basal intrinsic cellular fluorescence level. Plasmids expressing either Vsr217 or MalK alone, under the control of their native promoters P1 and P2, were constructed by deleting either coding region from a parental plasmid carrying the complete locus P1-vsr217-P2-malK (Fig. S2). Primers used for DNA amplification are listed in Table 3.
To measure growth of LGP32, bacteria (3 replicate cultures in each case) were cultured overnight in minimal medium supplemented with 2 g/liter D-glucose or D-maltose. If needed, Cm was added. Overnight cultures were diluted to an optical density at 600 nm (OD 600 ) of ;0.05 in the same medium, and growth was monitored in a 96-well plate, by measuring cell density at OD 600 at regular time intervals in a CLARIOstar (BMG Labtech, France) microplate reader.
RT-PCR and Northern blotting. Overnight cultured bacteria were diluted to an OD 600 of ;0.05 in minimal medium, supplemented by glucose or maltose where indicated, except for the results presented in Fig. 2, where cells were grown in Zobell medium, since the malT mutant does not grow in minimal medium plus maltose. Cells were grown with continuous shaking at 180 rpm at 20°C. When the OD 600 reached ;0.4, bacteria were collected by centrifugation and resuspended in TRI reagent (Sigma) before total RNA extraction and purification, which was carried out using Direct-zol MiniPrep Plus (Zymo research). DNA was removed by TURBO DNase (Ambion) treatment. RNA was further purified by phenolchloroform extraction and RNA precipitation. Ten micrograms (5 mg in the case shown in Fig. 4B) purified RNA was used for Northern blotting as described previously (19). Oligonucleotide probes (Table 3) complementary to the gene to be assayed were labeled at their 39 ends using terminal transferase (Fermentas, USA) and [a-32 P]dCTP, according to the manufacturer's instructions. Transfer-messenger RNA (tmRNA) was used as a loading control. Membranes were scanned using an Amersham Typhoon biomolecular imager (GE Healthcare) in the phosphorimager mode. Transcript signals were quantified by ImageJ (36) and normalized to the signal of tmRNA.
For RT-PCR, 1 mg RNA was used as a template for the first-strand synthesis catalyzed by SuperScript III reverse transcriptase (Invitrogen). Amplification of the vsr217-malK transcript was then carried out using DreamTaq DNA polymerase (Thermo Fisher Scientific) and primers Vsr217-RT-F and MalK-RT-R (Table 3).
59 and 39 rapid amplification of cDNA ends. Overnight cultures of LGP32 WT and Dvsr217 strains were diluted to an OD 600 of ;0.05 and grown in minimal medium supplemented with 2 g/liter D-maltose. Bacteria were collected by centrifugation when the OD 600 reached ;0.4 and were resuspended in TRI reagent followed by RNA extraction, DNase treatment, and RNA purification. 59 and 39 ends of Vsr217 and malK transcripts were determined using the 59/39 RACE system for rapid amplification of cDNA ends (Invitrogen). RNA primers used are listed in Table 3. In each case, the subsequent PCR products were cloned in the pJET1.2 vector (Thermo Fisher Scientific). Twenty plasmids were prepared from randomly chosen colonies, and inserts were sequenced using oligonucleotides pJET1.2-F/R (Table 3).
RNA stability measurement.
LGP32 WT, Dvsr217, and DpolyT strains were grown overnight in minimal medium supplemented with 2 g/liter D-maltose at 20°C. Bacteria were diluted to an OD 600 of ;0.05 and grown in the same medium until an OD 600 of ;0.4 and rifampicin was added to a final concentration of 500 mg/ml. Bacteria were collected for RNA extraction at 0, 0.5, 1.0, 2.0, and 3.0 min after the addition of rifampicin. Northern blot analysis was performed to determine the amount of Vsr217 and malK transcripts. GFP fluorescence measurement. To monitor green fluorescent protein (GFP) fluorescence, strains to be assayed (derived from E. coli MG1655) were grown overnight in LB supplemented with Cm and Amp. Five microliters of the overnight culture was spotted onto LB agar supplemented with Cm, Amp, and IPTG where indicated, and the plates were incubated at 37°C. Plates were then scanned with an Amersham Typhoon model 5 biomolecular imager (GE Healthcare), using the fluorescence mode (excitation [Exc], 488 nm; emission [Em], 525 nm; filter Cy2 525BP20). Signal intensity of the bacterial spots was determined using ImageJ and the formula (integrated density of the spot 2 [spot area Â mean background signal]), where the background signal was the signal obtained for cells containing the non-GFPexpressing plasmid, to remove the intrinsic fluorescence of bacterial cells. Significance of the relative differences was determined by a Student's unpaired t test, with four biological replicates.
Proteomics. V. tasmaniensis LGP32 and Dvsr217 strains were grown in triplicates in Zobell with glucose or maltose until an OD 600 of ;0.4 (12 samples in total). After centrifugation of 2 ml, cells were washed in Tris/EDTA (pH 7.5) plus 20% sucrose and then resuspended in NuPAGE LDS buffer (Thermo Fisher Scientific) to a final concentration equivalent to an OD 600 of 20. Samples were heated at 95°C for 10 min before loading 10 ml of each sample on a 4% to 12% polyacrylamide gradient gel (Invitrogen). After a 5-min migration to let the samples enter the gels, they were stained in Coomassie blue and then destained.
For tandem mass spectrometry (MS/MS), proteins were subjected to an in-gel enzymatic digestion with trypsin. Nanoscale liquid chromatography-electrospray ionization-tandem mass spectrometry (nLC-ESI-MS-MS; Triple-TOF) was carried out (gradient of 120 min), and the peptides were identified by Mascot and Andromeda and quantified by MaxQuant. Statistical analyses were carried out using Perseus (37). P values were calculated using a Welch's t test.
Bioinformatic predictions. Promoters were predicted using the BPROM software, and Rho-independent terminators were predicted using the FindTerm software (38). Vibrio malK upstream sequences were aligned using Muscle at the EBI (39). Putative Vsr217 targets were identified using CopraRNA, and interaction between Vsr217 and the fbp mRNA upstream region was predicted using IntaRNA (23).

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.