Central Role of Sibling Small RNAs NgncR_162 and NgncR_163 in Main Metabolic Pathways of Neisseria gonorrhoeae

ABSTRACT Small bacterial regulatory RNAs (sRNAs) have been implicated in the regulation of numerous metabolic pathways. In most of these studies, sRNA-dependent regulation of mRNAs or proteins of enzymes in metabolic pathways has been predicted to affect the metabolism of these bacteria. However, only in a very few cases has the role in metabolism been demonstrated. Here, we performed a combined transcriptome and metabolome analysis to define the regulon of the sibling sRNAs NgncR_162 and NgncR_163 (NgncR_162/163) and their impact on the metabolism of Neisseria gonorrhoeae. These sRNAs have been reported to control genes of the citric acid and methylcitric acid cycles by posttranscriptional negative regulation. By transcriptome analysis, we now expand the NgncR_162/163 regulon by several new members and provide evidence that the sibling sRNAs act as both negative and positive regulators of target gene expression. Newly identified NgncR_162/163 targets are mostly involved in transport processes, especially in the uptake of glycine, phenylalanine, and branched-chain amino acids. NgncR_162/163 also play key roles in the control of serine-glycine metabolism and, hence, probably affect biosyntheses of nucleotides, vitamins, and other amino acids via the supply of one-carbon (C1) units. Indeed, these roles were confirmed by metabolomics and metabolic flux analysis, which revealed a bipartite metabolic network with glucose degradation for the supply of anabolic pathways and the usage of amino acids via the citric acid cycle for energy metabolism. Thus, by combined deep RNA sequencing (RNA-seq) and metabolomics, we significantly extended the regulon of NgncR_162/163 and demonstrated the role of NgncR_162/163 in the regulation of central metabolic pathways of the gonococcus.

IMPORTANCE Neisseria gonorrhoeae is a major human pathogen which infects more than 100 million people every year. An alarming development is the emergence of gonococcal strains that are resistant against virtually all antibiotics used for their treatment. Despite the medical importance and the vanishing treatment options of gonococcal infections, the bacterial metabolism and its regulation have been only weakly defined until today. Using RNA-seq, metabolomics, and 13 C-guided metabolic flux analysis, we here investigated the gonococcal metabolism and its regulation by the previously studied sibling sRNAs NgncR_162/163. The results demonstrate the regulation of transport processes and metabolic pathways involved in the biosynthesis of nucleotides, vitamins, and amino acids by NgncR_162/163. In particular, the combination of transcriptome and metabolic flux analyses provides a heretofore unreached depth of understanding the core metabolic pathways and their regulation by the neisserial sibling sRNAs. This integrative approach may therefore also be suitable for the functional analysis of a growing number of other bacterial metabolic sRNA regulators. of the sibling sRNAs has been demonstrated experimentally (19)(20)(21) for a subset of targets.
To analyze the role of the sibling sRNAs in the neisserial metabolism in further detail, we combined transcriptome analysis and the investigation of carbon fluxes based on metabolomics and stable isotope incorporation experiments. By this integrative approach, several new target genes, including positively regulated targets, were identified and the sibling sRNAs were shown to interfere with sugar catabolism, the TCA cycle, serine-glycine metabolism, and the transport of glycine, phenylalanine, and branched-chain amino acids (BCAA).

RESULTS
Global transcriptome analysis reveals new members of the NgncR_162/163 regulon. We and others previously identified target genes of the sibling sRNAs NgncR_162/163 and their homologues in N. meningitidis by in silico analysis of the sRNA-mRNA interaction (19,20) and comparative mass spectrometric analysis of cell lysates from wild-type bacteria and an sRNA double deletion mutant (21). To obtain more complete insight into the NgncR_162/163 regulon, we now compared the gene expression profiles of N. gonorrhoeae wild-type MS11 and the sRNA DD162/163 double deletion mutant (19) grown in rich proteose peptone medium (PPM) by deep RNA sequencing (RNA-seq). Genes differentially expressed in the two strains were assessed using DESeq2 (22). A number of the 94 protein coding genes (54 upregulated, 40 downregulated) exhibited expression ratios below 0.75 and above 1.5 (adjusted P value [q] of ,0.05), which was considered differentially expressed (Table 1). Differentially expressed genes mostly belonged to the following functional categories (COG): amino acid transport and metabolism (n = 10), energy production and conversion (n = 7), transcription (n = 6), and unknown function (n = 36) (Fig. 1a). Putatively regulated open reading frames (ORFs) encoding hypothetical proteins of unknown function tend to be very short and are most frequently located in the N. gonorrhoeae MS11 homologues of the maf genomic islands of N. meningitidis encoding secreted polymorphic toxins (mafB), their specific immunity proteins (mafI), and alternative MafB C-terminal domains (MafB-CT) or MafI modules (23). However, functional enrichment analysis using a hypergeometric test did not indicate significant overrepresentation of these COG classes. From the NgncR_162/163 target genes validated previously by real-time quantitative PCR (qRT-PCR) (19), the NGFG_01721 and NGFG_02049 genes, prpC, ack, and gltA were detected by RNA-seq (Fig. 1b), whereas prpB, gdhR, fumC, and sucC did not meet the applied cutoff, reflecting differences in the sensitivity of the two RNA quantification techniques. Furthermore, seven putative N. gonorrhoeae noncoding RNAs (24) were found to be differentially transcribed in the DD162/163 mutant (Table 1).
Fifteen protein coding putative NgncR_162/163 target genes with predicted metabolic functions were selected for target validation via qRT-PCR analysis (Fig. 2a) in the DD162/163 mutant and the complemented mutant carrying both sRNA genes integrated into the iga-trpB locus (DDc162/163) (19). Differential expression in the DD162/ 163 mutant could be confirmed, except for that of the sugar transporter NGFG_02263. Gene ilvB (NGFG_02042) from the BCAA biosynthetic pathway was significantly downregulated in the DD162/163 mutant, but its expression did not revert in the DDc162/163 complemented mutant. Validated target genes encode proteins putatively involved in the transport of amino acids or peptides (NGFG_00045 and NGFG_01564 [members of the neurotransmitter sodium symporter {NSS} family], NGFG_00093, and NGFG_01937), citrate (NGFG_00249), and inorganic phosphate (NGFG_01353), as well as NGFG_01722, whose gene is cotranscribed with the putative alanine symporter NGFG_01721 gene and encodes D-alanine dehydrogenase (dadA). Furthermore, differential expression of lactate permease (NGFG_01471) showing a ratio slightly above the applied cutoff in RNA-seq (0.765779; q = 0.0191) could be validated by qRT-PCR (Fig. 2a). Another validated target, aconitate hydratase (acn), belongs to the methylcitric acid cycle, which was previously shown to be controlled by the sibling sRNAs (19,20). Interestingly, Gene gcvH (NGFG_01514) was found to be upregulated in the absence of the sibling sRNAs. GcvH is a component of the glycine cleavage system that converts glycine to CO 2 , NH 3 , and 5,10-methylene-tetrahydrofolate (5,10-MTHF), suggesting a role for the sibling sRNAs in purine, histidine, thymine, panthothenate, and methionine synthesis via control of the supply of C1 units. Furthermore, RNA-seq suggested a contribution of the sibling sRNAs in the control of iron- sulfur cluster synthesis and homeostasis (iscR, dnrN) and anaerobic respiration (aniA, norB, NGFG_01516). Notably, 16 of the putative targets of the sibling sRNAs listed in Table 1 as well as lactate permease were also differentially expressed in an hfq deletion mutant of N. meningitidis (25). Since the sibling sRNAs were shown to bind Hfq in both N. meningitidis and N. gonorrhoeae and their stability was largely diminished in its absence (20; E. Heinrichs and T. Rudel, unpublished), deregulation of transcript levels in the hfq mutant is likely to be the consequence of posttranscriptional regulation mediated by the sibling sRNAs. In the case of NGFG_01471, differential expression in the DD162/163 mutant can be explained by an indirect effect, because lactate permease was shown to be regulated by GdhR (26), which itself is a target of the sibling sRNAs (19).
Transcription of three genes encoding putative noncoding RNAs (NgncR_199, NgncR_201, and NgncR_210) was also analyzed by qRT-PCR, but differential expression could not be validated using this method (data not shown).  Table 1. (a) Pie chart depicting the biological role of putative protein-encoding sibling sRNA target genes identified by RNA-seq. Numbers within or next to the segments of the graph represent the percentages of putative sibling sRNA targets (n = 94) belonging to the respective functional categories listed on the right. Numbers given in parentheses in the pie chart legend represent the numbers of putative targets from the respective categories which are repressed or activated by the sibling sRNAs (repressed, n = 54/activated, n = 40). (b) Venn diagram representing intersecting sets of Neisseria sibling sRNA target genes identified in this study (yellow) and in the works of Pannekoek et al. (21) (blue) and Bauer et al. (19) (red). Pannekoek et al. investigated the sibling sRNA regulon in N. meningitidis; for a better comparison, N. gonorrhoeae orthologs of the N. meningitidis targets are listed in the figure. Numbers in the diagram represent the number of targets identified in the respective study, as well as the number of common targets which are also listed next to overlapping segments. Common targets of the three studies are highlighted in bold.

Metabolic Control by Neisseria Sibling sRNAs mBio
To test the functional redundancy of NgncR_162 and NgncR_163 in the regulation of newly identified targets, we investigated as an example the impact of the individual sibling sRNAs as posttranscriptional regulators of gcvH, NGFG_01937, NGFG_00249, NGFG_00093, acn, dadA, and NGFG_00045. Complementation of the DD162/163 mutant with NgncR_163 restored mRNA abundance to wild-type levels, whereas complementation by NgncR_162 was less efficient for some of the investigated targets. Functional redundancy was observed only in the case of NGFG_00093, acn, and dadA target genes (Fig. 2b). These data suggest that the individual siblings act in a hierarchical manner on certain targets while exhibiting complete functional redundancy on others.
sRNA-target interaction is predicted to occur within the 59 UTR and coding region of the regulated mRNAs. All targets identified previously (19)(20)(21) are negatively regulated by the sibling sRNAs via mRNA-sRNA duplex formation, resulting in obstruction of the RBS. The sRNA-mRNA interaction involved the single-stranded region of stem-loop 2 (SL2) and in some cases the single-stranded region connecting SL1 and SL2 (SSR1) (see Fig. S1a in the supplemental material). Complementarity between the validated new target mRNAs and the sibling sRNAs was analyzed using IntaRNA (27) (Fig. S1 and S2). In genes which were found differentially expressed in the absence of the sibling sRNAs by RNA-seq were analyzed by qRT-PCR in strain MS11, the DD162/163 sRNA double deletion mutant, and the DDc162/ 163 complemented strain cultivated in PPM. The ratios (fold change) of the transcript amount relative to that of the wild-type MS11 (mutant versus wildtype; wild-type normalized to 1) are depicted. (b) In the case of gcvH, NGFG_01937, NGFG_00093, NGFG_00249, dadA, acn, and NGFG_00045, transcript levels were also assessed in derivatives of MS11 DD162/163 complemented with either NgncR_162 (DDc162) or NgncR_163 (DDc163). ack was used as a control, since the sibling sRNAs were previously shown to display functional redundancy on this target (19). The indicated ratios represent the mean of the results of qRT-PCR experiments performed in triplicate using cDNAs obtained from at least three independent RNA preparations. Error bars indicate standard deviations. Statistical significance was determined using Student's t test analysis (*, P , 0.05; **, P , 0.01; ***, P , 0.001).
accordance with negative regulation of dadA, NGFG_00249, NGFG_01937, NGFG_00093, and glyA (see below), obstruction of the RBS by the SL2 loop or SSR1 sequence was predicted for both sRNAs (Fig. S1b). Surprisingly, also in the case of the positively regulated target norB, a region of complementarity to SL2 overlapping the RBS was detected. Furthermore, the 59 UTR of dnrA shows complementarity to the SL2 sequence of both sRNAs, but in the case of NgncR_163, a different interaction site within the coding sequence (CDS) which engages the SL1 loop sequence was predicted. In the case of the other members of the NgncR_162/163 regulon, base-pairing interactions are likely to occur within the CDS (Fig. S2). IntaRNA analysis detected similar regions of complementarity between NGFG_00045, iscR, and gloA mRNAs and the SSR1 and SL2 sequences of both sRNAs. NgncR_162 exhibited extended complementarity with gloA mRNA, which also engages the single-stranded region connecting SL2 and SL3 (SSR2) and part of the SL3 sequence. In the case of gcvH, NGFG_01564, NGFG_01353, and acn, IntaRNA predictions differed for the two sRNA molecules. NGFG_01564 and gcvH mRNAs exhibited segments of 8 to 13 nucleotides (nt) with base-pairing capability to the SL2 sequence of both sRNAs or the SSR1 sequence of NgncR_162. The SL1 and SSR1 sequence of NgncR_162 showed partial complementarity to a 28-nt region within the CDS of NGFG_01353, while NgncR_163 was predicted to interact with a different region of NGFG_01353 via the SL2 and SSR2 sequence (which, however, is conserved in both sRNAs [ Fig. S1a]). Within the acn CDS (1,920 bp), several regions of complementarity with the sibling sRNAs were detected (IntaRNA energy scores, 24.46 to 210.35 kcal/mol), one of which is shown as an example in Fig. S2. Thus, putative hybridization regions can be predicted for most of the identified target mRNAs. Furthermore, IntaRNA analysis with the N. meningitidis sibling sRNAs NmsR A (NgncR_162) and NmsR B (NgncR_163) revealed similar predictions for the orthologs of the glyA (NMB1055), dadA (NMB0176), NGFG_00249 (NMB1794), NGFG_00093 (NMB1964), NGFG_01937 (NMB2136), norB (NMB1622), iscR (NMB1378), gloA (NMB0340/NmsR B ), and gcvH (NMB0575) putative targets.
Validation of sRNA-target interactions on the protein level. Validation of sRNAtarget interactions within the 59 UTR was performed in Escherichia coli by using a twoplasmid gfp reporter system (28). The region covering the 59 UTR and the first 8 to 32 codons of the NGFG_01937, NGFG_00249, and NGFG_00093, and dnrN genes was fused to gfp-SF encoding a superfolder variant of green fluorescent protein (GFP) in vector plasmid pXG10-SF (28). Similarly, the intergenic region between the NGFG_01721 and dadA genes as well as the first 40 codons of dadA was cloned into the intercistronic gfpfusion vector pXG30-SF (28). E. coli Top10 cells were cotransformed with the target-gfp fusion plasmids, and plasmids expressing NgncR_162, a derivative of NgncR_162 with a mutated SL2 loop sequence (19), or a nonsense sRNA (29) and reporter gene expression was monitored by Western blotting. It should be noted that E. coli expressing NgncR_162 exhibited moderate growth retardation, as was also observed previously (19). While expression of NGFG_00093-gfp remained unaffected in E. coli (data not shown), downregulation of gfp expression was observed in the presence of NgncR_162 in the case of NGFG_01937 and NGFG_00249 and was abolished by mutagenesis of the SL2 loop sequence (Fig. 3a). GFP levels decreased to about 70% and 60%, respectively, of that observed in the negative control. Expression of dadA-gfp was diminished by 60% in the presence of NgncR_162, but in this case, the mutated SL2 loop sequence in sRNA NgncR_162m1 did not restore dadA-gfp expression to the level observed in the negative control (Fig. 3a). This observation is in accordance with the predicted sRNA-target interaction which, in the case of dadA, engages the SSR1 sequence (Fig. S1b). Interestingly, in accordance with transcriptome analysis, but despite the fact that sRNA binding is predicted to affect RBS accessibility (Fig. S1b), expression of dnrN-gfp was upregulated 2fold in the presence of NgncR_162 (Fig. 3a). The mechanism of this unexpected regulation remains to be investigated further. Immunoblot results obtained from E. coli lysates could also be confirmed by direct fluorescence measurement (data not shown).
To investigate sRNA-mediated regulation of the protein level also in N. gonorrhoeae, NGFG_00093 and NGFG_00249 targets were chosen. Translational target-gfp fusions under the control of the respective target promoter were integrated into the intergenic region between the iga and trpB genes in wild-type MS11 and the DD162/163 mutant, and gfp expression in the mutants was monitored by Western blotting. Deep sequencing of the N. gonorrhoeae transcriptome had revealed only weak transcription of the NGFG_00249 gene (24), and consistently, NGFG_00249-GFP could not be detected (data not shown). Although NgncR_162 had no impact on the expression of NGFG_00093-gfp in E. coli, increased amounts of the fusion protein were detected in the absence of the sibling sRNAs in N. gonorrhoeae (Fig. 3b), confirming negative regulation via obstruction of the RBS. The observation that sRNA-mediated regulation failed in E. coli but was detectable in N. gonorrhoeae had previously also been made in the case of the sibling sRNA target gdhR (19). To validate predicted sRNA binding within the coding region, gcvH was selected (Fig. S2); this is the last gene in a tricistronic transcript composed of gcvT (NGFG_01512), the NGFG_01513 gene (encoding a hypothetical protein), and gcvH (24). Since transcript levels of gcvT and the NGFG_01513 gene remained unchanged in . For each target-gfp fusion, GFP levels (fold change) relative to E. coli cells harboring the negative-control plasmid pJV300 are indicated. (b) GFP expression was monitored by Western blotting in N. gonorrhoeae strains MS11 93gfp (lane 1) and MS11 DD93gfp (lane 2) harboring a translational NGFG_00093-gfp fusion inserted between the iga and trpB genes. The result from a representative experiment is shown (n = 2), and the relative change in GFP protein levels in the absence of the sibling sRNAs is indicated. (c) Expression of gcvH carrying a C-terminal FLAG tag was analyzed in the presence and absence of the sibling sRNAs in strains MS11 gcvH-F (lane 1) and MS11 DDgcvH-F (lane 2). The result from a representative experiment is shown (n = 2), and the relative change in GFP expression is indicated. For Western blot analysis, equal amounts of protein were separated on 12% polyacrylamide gels and hybridization of membranes was performed with monoclonal antibodies directed against GFP (a and b) or FLAG (c). Hybridization with a monoclonal antibody directed against HSP60 was performed as a loading control. the sRNA DD162/163 double deletion mutant, we reasoned that sRNA-mediated indirect effects on transcription initiation are unlikely to account for the differential expression of gcvH in the absence of the sibling sRNAs. Therefore, derivatives of wild-type MS11 and the DD162/163 mutant expressing a C-terminally FLAG-tagged GcvH protein were constructed. In accordance with our transcriptome data (Table 1), GcvH was found to be upregulated 1.5-fold in the DDgcvH-F mutant (Fig. 3c).
The NGFG_00045 mRNA level is directly affected by the sibling sRNAs. RNAs downregulated in the absence of NgncR_162/163 identified in this study represent a new class of targets. To demonstrate the validity of this class of targets, NSS family transporter NGFG_00045 was chosen to investigate this positive regulation by the sibling sRNAs in more detail. Northern blot analysis was performed on RNA extracted from MS11, the DD162/163 mutant, and the DDc162/163 complemented mutant. RNAseq had revealed that the NGFG_00045 gene is cotranscribed with a gene encoding a hypothetical peptide of 30 amino acids (24), and a transcript of the expected size was observed in wild-type MS11 and the complemented mutant but was barely detectable in the DD162/163 mutant (Fig. 4a). The hybridization pattern of RNA from the DD162/ 163 mutant was unaltered, indicating that the sibling sRNAs are not involved in prominent processing events (Fig. 4a). Next, we replaced the NGFG_00045 promoter region (24) with the promoter of a gonococcal opa gene in MS11 and the DD162/163 and DDc162/163 mutants. qRT-PCR and Northern blot analysis showed that transcript amounts were still affected by the absence of the sibling sRNAs ( Fig. 4b and c). However, when the NGFG_00045 upstream region including promoter and 59 UTR was fused to gfp, almost equal amounts of mRNA were detected in the presence and absence of the sibling sRNAs (Fig. 4d). Since expression of GFP could not be detected in these mutants, probably due to a weak RBS in the NGFG_00045 59 UTR, we introduced the consensus E. coli Shine-Dalgarno sequence via site-directed mutagenesis. As observed for gfp mRNA, protein levels were not affected in the absence of the sibling sRNAs (Fig. 4e). These data demonstrated that the positive regulation of NGFG_00045 is not indirectly mediated via a transcriptional regulator that is the target of the sRNAs and that the NGFG_00045 59 UTR is not involved in posttranscriptional regulation. Rifampin assays performed on cultures of wild-type MS11, the DD162/163 mutant, and the DDc162/163 complemented strain suggested a 30% reduction in NGFG_00045 transcript stability in the absence of the sibling sRNAs (Fig. S3). The mechanism of NGFG_00045 target activation by the sibling sRNAs remains to be investigated in more detail.
Amino acid uptake is altered in the DD162/163 mutant and an NGFG_00045 deletion mutant. Since several targets of the sibling sRNAs were predicted to be involved in transport processes, amino acid uptake was analyzed in the wild-type, the DD162/163 double deletion mutant, and the complemented strain to identify effects associated with sibling sRNA expression. First, the growth behavior of the respective N. gonorrhoeae strains was monitored in chemically defined CDM10 medium, demonstrating slightly reduced growth of the sRNA DD162/163 double deletion mutant (Fig. S4a). Furthermore, Northern blot analysis revealed abundant sRNA expression in gonococci grown in CDM10, with only a minor reduction compared to that of bacteria from PPM culture (Fig. S5). For the analysis of spent culture supernatants, gonococci were grown to logarithmic phase (optical density at 550 nm [OD 550 ], 0.6) and amino acid levels were determined by ultraperformance liquid chromatography-mass spectrometry (UPLC/MS). Growth of wild-type MS11 resulted in 90% depletion of glutamine and glutamate and about 50% and 40% reduction of proline and asparagine concentrations, respectively, in the spent culture medium (Fig. S6). Consumption was marginal in the case of glycine, alanine, arginine, tyrosine, tryptophan, valine, and histidine (less than 5% of the supplied amounts) and moderate (ranging from 5% to 30%) for other amino acids. Surprisingly, the concentration of aspartate seemed to be even higher in the culture supernatant (Fig. S6), suggesting a lack of aspartate transport, which might be compensated by efficient biosynthesis of oxaloacetate from phosphoenolpyruvate (PEP) via PEP carboxylase (30); (see below) or deamination of asparagine. Compared to the wild-type, the sRNA DD162/163 double mutant showed significant differences in the consumption of proline, glycine, alanine, serine, and threonine. More specifically, consumption of glycine, alanine, and proline was increased 9-, 4-, and 1.2-fold, respectively, while uptake of serine and threonine was diminished about 1.5-fold. As expected, the amino acid profile of the culture supernatant of the DDc162/163 complemented strain resembled that of the wild-type (Fig. 5).
To investigate the contribution of sibling sRNA targets to amino acid transport, knockout mutants for NGFG_01721, which belongs to the alanine or glycine:cation symporter (AGCS) family, and for the NSS family transporters NGFG_00045 and NGFG_01564 were created. Growth of the mutants was indistinguishable from that of the wild-type in rich medium (data not shown), but both the D45 and D1564 mutants exhibited a clear growth defect in CDM10 medium (Fig. S4b). The D1721 knockout mutant showed an intermediate phenotype resembling the growth of the sRNA double grown to logarithmic phase were analyzed by Northern blotting using a NGFG_00045-specific radiolabeled probe, which was generated by PCR using primer pair qRT45-1/qRT45-2 (see Table S2 in the supplemental material). Probing for 5S rRNA was used as a loading control. The positions of radiolabeled marker RNAs of 1,500 and 100 nt are indicated on the left side of the panels. (c) Transcription of NGFG_00045 under the control of the P opa promoter in MS11 P opa 45, MS11 DDP opa 45, and MS11 DDP opa 45c was analyzed by qRT-PCR. The ratios of the transcript amount relative to that of P opa 45 (normalized to 1) are depicted. The indicated ratios represent the mean of results of qRT-PCR experiments performed in triplicate on cDNAs obtained from three independent RNA preparations. Error bars indicate standard deviations. Statistical significance was determined using Student's t test analysis (***, P , 0.001). Metabolic Control by Neisseria Sibling sRNAs mBio deletion mutant in CDM10 (Fig. S4b). In the NGFG_00045 D45 knockout mutant, proline uptake was massively increased, while the uptake of leucine, isoleucine, and valine was abolished. Again, the observed effects were reversed in the respective Dc45 complementation mutant (Fig. 5). Analysis of the amino acid compositions of culture supernatants of D1721 and D1564 mutants did not reveal statistically significant differences from that of wild-type MS11 (data not shown). In the D45 mutant, increased consumption of proline, which can be efficiently converted to glutamate via the bifunctional proline dehydrogenase/pyrroline-5-carboxylate dehydrogenase NGFG_01376, apparently compensates for the defect in isoleucine uptake caused by the inactivation of NGFG_00045, which is likely to be a BCAA transporter (for more details, see below). Elevated proline uptake in the DD162/163 mutant is consistent with downregulation of NGFG_00045 in this mutant. Furthermore, increased glycine uptake in the DD162/ 163 mutant strongly argues in favor of NGFG_01721 being predominantly a glycine transporter, since this protein is massively upregulated in the absence of the sibling sRNAs (Table 1) (19). Glycine cleavage, which is affected by the sibling sRNAs via target gene gcvH, yields 5,10-MTHF, which together with another molecule of glycine can then be converted to serine by the serine hydroxymethyltransferase GlyA (reviewed in reference 31) (see Fig. 10). In N. meningitidis, glyA expression was reported to be derepressed in the absence of hfq and the sibling sRNAs, respectively (21,25). Since serine uptake was diminished in the DD162/163 mutant and in our RNA-seq analysis glyA was weakly upregulated with a q value close to significance (fold change, 1.29; q = 0.059), we validated glyA transcript levels by qRT-PCR. In fact, upregulation of glyA was observed in the sRNA double deletion mutant, while the wild-type transcript level was restored in the complemented strain (Fig. 6a). Furthermore, a translational glyA-gfp fusion was downregulated in the presence of NgncR_162 in E. coli (Fig. 6b), confirming negative regulation of glyA by the sibling sRNAs. This is in accordance with predicted binding of the sibling sRNAs to the RBS of the glyA mRNA (Fig. S1b). Based on these multiple evidences, we conclude that derepression of NGFG_01721, gcvH, and glyA in the sRNA double mutant enhances serine biosynthesis, resulting in a decreased demand for serine uptake. Isotopologue profiling reveals a bipartite metabolism in N. gonorrhoeae. From all these findings, it became apparent that the sibling sRNAs in N. gonorrhoeae are involved in the regulation of the central carbon metabolism in response to available nutrients in the environment. The central carbon metabolism in N. gonorrhoeae has so far only been investigated using enzyme assays with cell extracts or genome annotations (32,33). To generally define the central carbon metabolism of N. gonorrhoeae during growth and to further substantiate the effects of the sibling sRNAs, we here employed stable isotope incorporation experiments with subsequent isotopologue profiling.
To this aim, the bacteria were grown for 4 h in duplicate to logarithmic phase (OD 550 = 0.6) in the chemically defined medium CDM10 containing fully 13 C-labeled glucose or proline ([U-13 C 6 ]glucose or [U-13 C 5 ]proline). After harvest, the cells were mechanically disrupted (fraction 1) or hydrolyzed under acidic conditions (fraction 2). Protein-derived amino acids (in fraction 2) and fatty acids (in fraction 1) were silylated and then applied to gas chromatography (GC)-MS analysis (three technical replicates). From the relative masses detected for the specific fragments in the MS spectra, the 13 C-excess values of amino acids and fatty acids were calculated using the software package Isotopo (34). Here, the overall 13 C-excess values show 13 C contents beyond the natural 13 C abundances in the respective molecules. Moreover, using the same software, the isotopologue compositions were determined displaying the relative fractions (%) of isotopologues (M 1 1, M 1 2, M 1 3, . . ., M 1 n) for each molecule under study. Herein, M denotes the molecular mass with only 12 C in the carbon backbone of the molecule and n specifies the number of 13 C-atoms.
Using 13.9 mM [U-13 C 6 ]glucose as a supplement to the medium, high 13 C-excess values were detected in alanine (41.3% 13 C excess) with about 90% M 1 3 (i.e., displaying the fraction of the U-13 C 3 -labeled species) and valine (44.7% 13 C excess) with about 70% M 1 5 ( Fig. 7a and b). This confirmed the efficient uptake and utilization of [U-13 C 6 ]glucose to afford fully labeled pyruvate, which was then converted into the detected [U-13 C 3 ]alanine and [U-13 C 5 ]valine specimens, respectively (see also Fig. 7c). Although the labeling patterns from [U-13 C 6 ]glucose do not allow a distinction between glycolysis and the Enter-Doudoroff (ED) pathway for glucose degradation, the ED pathway is the predominant pathway for pyruvate formation from glucose in N. gonorrhoeae, based on enzyme assays (32). In addition, glycolysis seems to be nonfunctional due to the absence of the gene for phosphofructokinase in the genome of N. gonorrhoeae (33). Tyrosine ( 13 C excess of 48.6%) and phenylalanine ( 13 C excess of 19.5%) also reflected the efficient incorporation of exogenous [ 13 C]glucose via the ED pathway and the pentose phosphate pathway (PPP) producing M 1 4 erythrose-4-phosphate (E-4-P) and M 1 3 phosphoenolpyruvate (PEP) as precursors. More specifically, the formation of the detected M 1 9 isotopologues in tyrosine and phenylalanine can be easily explained by the assembly of the aromatic amino acids from [U-13 C 4 ]E-4-P and two molecules of [U-13 C 3 ] PEP via the shikimate pathway (Fig. 7c). These observations are in full accordance with the reported genome sequence (33) and enzyme assays of sugar-degrading pathways (32).
Although serine is typically derived from 3-phosphoglycerate via 3-hydroxypyruvate, the genome of N. gonorrhoeae does not encode an enzyme for the 3-phosphoglycerate reduction into 3-hydroxypyruvate. Indeed, only a minor 13 C excess (1.4%) was observed for serine from the protein hydrolysate in the [U-13 C 6 ]glucose experiment (Fig. 7a). In conclusion, serine was derived mainly from unlabeled serine and other unlabeled components in the medium. This was also in accordance with high serine uptake observed through culture supernatant analysis (Fig. S6). Not surprisingly then, almost no 13 C excess was detected in glycine (0.7%), since glycine was derived from serine or was also directly taken up from the medium in an unlabeled form. Similarly, due to the absence of the histidinol phosphatase gene in the genome sequence, no 13 C excess was detected in histidine.
Since only low 13 C enrichments were detected in amino acids using intermediates of the TCA cycle in these experiments, we assumed that the TCA in N. gonorrhoeae is mainly fueled by unlabeled compounds from the medium (i.e., not from [U-13 C 6 ]glucose). The DD162/163 mutant had revealed an increased uptake of proline (Fig. 5), which indicated that this amino acid could serve as a precursor for glutamate and TCA cycle intermediates via 2-oxoglutarate. Therefore, a labeling experiment starting with 0.4 mM [U-13 C 5 ]proline in the medium was performed. Indeed, [U-13 C 5 ]proline added to the culture medium was efficiently taken up and utilized, leading to 80% M 1 5 in the isotopologue profile of glutamate (17.7% 13 C excess) and about 80% M 1 4 in the isotopologue profiles of aspartate (20% 13 C excess) and threonine (5.7% 13 C excess) (Fig. 8). In contrast, only a low 13 C excess was detected in alanine, glycine, valine, and isoleucine (below 5%), whereas serine did not contain 13 C beyond the natural abundance content. Based on this observation, labeled glycine could hardly be formed from the apparently unlabeled serine. Rather, glycine was potentially synthesized from (labeled) threonine via the threonine utilization (Tut) pathway (reviewed in reference 31), as the genome of N. gonorrhoeae contains distant homologues of threonine dehydrogenase and 2-amino-3-ketobutyrate CoA ligase from E. coli.
The different 13 C enrichments from the labeling experiments with [ 13 C]glucose and [ 13 C]proline indicated that N. gonorrhoeae used different substrates simultaneously for its growth, suggesting a bipartite metabolic network, as shown in Fig. 9. In this model, glucose is used as a substrate providing precursors for anabolic purposes, i.e., for sugar components and some aromatic amino acids derived from the PPP, amino acids derived from pyruvate, and fatty acids derived from acetyl-CoA. In contrast, glucose does not serve as a major substrate to feed the TCA cycle. Rather, proline or related compounds drive the TCA cycle and contribute mainly to energy metabolism.
Based on this metabolic model, we then characterized the impact of the sibling sRNAs and targeted transport proteins on the core metabolism of N. gonorrhoeae. This was done by supplementing the respective mutant strains with [U-13 C 6 ]glucose or [U-13 C 5 ]proline during growth, as explained above for the parent strain.
NGFG_00045 and NGFG_01564 transport BCAA and phenylalanine, respectively. The sibling sRNA NGFG_00045 and NGFG_01564 target genes presumably encode amino acid transporters (Table 1; Fig. 5). Indeed, the 13 C excesses of alanine, isoleucine, leucine, and valine differed significantly between the N. gonorrhoeae D45 mutant and the wild-type during growth with [U-13 C 6 ]glucose (Fig. 10a). While the 13 C excess increased massively in the BCAA isoleucine, leucine, and valine (116% to 135%), it decreased in alanine (29%). The same effect, albeit to a much lower extent, was observed during growth of N. gonorrhoeae D45 with [U-13 C 5 ]proline (Fig. 10b). Here, other amino acids also showed minor differences, but the most significant effect was observed in glutamate and aspartate, with increased 13 C excess values of 8 to 9% in the mutant. These complementary effects observed with different labeled substrates again corroborated the validity of the bipartite metabolic model introduced above (Fig. 9). Based on these observations, the NGFG_00045 gene encodes a BCAA transporter, as de novo synthesis of this class of amino acids, especially from [U-13 C 6 ]glucose, was highly increased in the D45 mutant strain. BCAAs use pyruvate as a building block (cf. Fig. 9); their increased biosynthesis depleted the pool of pyruvate available for alanine synthesis and thereby reduced the 13 C excess in this amino acid. On the other hand, the synthesis of aspartate and glutamate was potentially upregulated to meet the increasing demand for aspartate as a precursor in isoleucine biosynthesis (cf. Fig. 9) and for nitrogen donors in the transamination step of BCAA biosynthesis (Fig. 10c).
In the case of the NGFG_01564 mutant (D1564), the labeling experiment with [U-13 C 6 ] glucose produced a huge increase in the 13 C excess of phenylalanine (140%) compared to that of the wild-type (Fig. 10a). Other significant differences (>1%) included a decrease in the 13 C excess of alanine, valine, aspartate, and threonine (26.2 to 21.6%). These effects could again be explained by a decreased pool of labeled PEP and pyruvate, as [ 13 C]PEP was consumed for de novo phenylalanine biosynthesis in the D1564 mutant  Metabolic Control by Neisseria Sibling sRNAs mBio strain. As explained earlier, synthesis of aspartate from glucose heavily depends on the PEP carboxylase reaction in N. gonorrhoeae and aspartate is a precursor for threonine biosynthesis (cf. Fig. 9). Surprisingly, the 13 C content of tyrosine was not significantly altered in the D1564 mutant strain (Fig. 10a), despite its origination from the same metabolic precursors as phenylalanine. Based on these observations, the NGFG_01564 gene encodes a phenylalanine transporter (Fig. 10d). The identification of a previously noncharacterized BCAA and phenylalanine transporter as targets of the sibling sRNAs demonstrates the straightforward approach of our strategy to include 13 C-based metabolomics as a tool to functionally analyze bacterial regulatory RNAs.
Sibling sRNAs interfere with the TCA cycle, amino acid transport, and metabolism. Looking at the sibling sRNA double mutant strain N. gonorrhoeae DD162/163, a variety of effects were observed in the labeling experiments. When supplementation was with [U-13 C 6 ]glucose (Fig. 10a), the 13 C excess in alanine, tyrosine, and valine decreased (25 to 220%), while the 13 C excess in isoleucine, lysine, and threonine increased (11.7 to 14%). In the experiment with [U-13 C 5 ]proline, the double mutant strain showed an increased 13 C excess in aspartate, glutamate, and again in threonine and isoleucine (15 to 17.6%). Alanine and glycine showed decreases in 13 C excess of 1.1% and 0.7%, respectively (Fig. 10b).
As suggested earlier, the sibling sRNAs downregulate the TCA cycle activity via negative control of various enzymes in the cycle (19,21). Consequently, in the double mutant strain, the TCA cycle was more active and utilized the supplied [ 13 C]proline tracer at higher rates, thereby producing higher 13 C excess values in aspartate and glutamate, which are derived from oxaloacetate and proline, respectively (cf. Fig. 9). By closer examination of the isotopologue profiles, this effect can be distinguished from the concomitant increased proline metabolization as a consequence of decreased BCAA uptake due to downregulation of NGFG_00045 (Fig. S7).
In the experiment with [ 13 C]glucose, utilization of this substrate via the ED pathway as well as the PPP seems to be decreased, leading to lower 13 C excess values in tyrosine and alanine. However, there was no significant effect observed in phenylalanine. Since the sibling sRNAs upregulate the phenylalanine transporter NGFG_01564, there was lower import of this amino acid in the double mutant strain. Therefore, the 13 C excess was much less diluted by the import of exogenous phenylalanine than with the wildtype. The same case can be made for the BCAAs valine, leucine, and isoleucine, which were much less affected than alanine in the double mutant strain, although they are also fully or partially derived from pyruvate. Herein, the sibling sRNAs upregulated the BCAA transporter NGFG_00045, and consequently, reduced import was observed in the double mutant strain. However, genes in BCAA biosynthesis were also downregulated in the DD162/163 mutant (Table 1), which could explain the lower 13 C excess in valine from [ 13 C]glucose. Reduced glucose usage could be due to reduced glucose uptake, since mRNA levels of the putative glucose transporter NGFG_02263 were reduced in the sibling sRNA mutant (Table 1). Furthermore, increased uptake of unlabeled alanine by the sRNA double mutant due to the upregulation of NGFG_01721 (Fig. 5) is likely to contribute to reduced 13 C accumulation in alanine.
Threonine presents an interesting case; when [ 13 C]proline was supplemented, the 13 C excess of threonine increased but to a smaller extent than that of its metabolic precursor aspartate. However, when [ 13 C]glucose was used as a substrate, threonine showed a greater increase in 13 C excess than aspartate. Increased 13 C excess might be related to threonine acting as a precursor for glycine, which can be converted to serine by GlyA (reviewed in reference 31). Increased uptake of unlabeled glycine in the DD162/163 mutant (due to the upregulation of NGFG_01721) and elevated glyA expression resulted in an increased rate of serine production. These effects led to decreased 13 C excess in glycine and also mitigated the need for threonine as a precursor for glycine/serine, thereby leading to the accumulation of labeled threonine in the double mutant strain. This is in accordance with the labeling experiments, since in the double mutant strain, the 13 C excess in threonine was significantly increased while the 13 C excess in glycine was significantly decreased with supplementation of [U-13 C 5 ]proline (Fig. 10b). The metabolic effects of sibling sRNA deletion are summarized in Fig. 10e.

DISCUSSION
The functional analysis of regulatory RNAs is hampered by the fact that large regulons are often governed by them. Target identification by in silico predictions, quantification of mRNA or protein abundance in the presence or absence of the sRNA, or RNA-seq-based approaches exploiting direct sRNA-target mRNA interactions frequently yield numerous candidates, which are then used to predict the function of the sRNA. Since sRNA-mediated modifications of mRNA or protein abundances are often moderate and metabolic pathways are frequently controlled by posttranslational modifications rather than protein abundance, very little is known about the true regulatory impact of the rapidly growing number of sRNAs with predicted metabolic targets (36)(37)(38)(39)(40)(41). We here performed a combined RNA-seq and metabolomics approach and thereby significantly extended the regulon controlled by the N. gonorrhoeae sibling sRNAs NgncR_162/163.
Our metabolomics analyses focused on differentially transported metabolites (amino acids), leading to identification of functions of heretofore uncharacterized NgncR_162/163 targets like the potential BCAA and phenylalanine transporters (NGFG_00045 and NGFG_01564, respectively). Metabolic flux analysis using isotopologue profiling confirmed the role of identified targets in metabolic pathways like the TCA and unveiled their role in orchestrating the substrates used in a newly identified bipartite metabolic network. The combination of transcriptome and carbon flux analyses used here thus provides an excellent approach toward a real understanding of the outcome of riboregulation targeting metabolic processes in bacteria.
New members of the NgncR_162/163 regulon predominantly encode proteins involved in nutrient uptake and serine/glycine metabolism. This is reminiscent of E. coli/Salmonella enterica serovar Typhimurium GcvB, which is conserved in members of the Gammaproteobacteria (36) and directly controls more than 50 mRNA targets comprising mostly periplasmic amino acid binding ABC transporter components, amino acid permeases (among others, the glycine permease encoded by cycA), and enzymes involved in amino acid metabolism (42)(43)(44). GcvB is abundant during exponential growth in nutrient-rich medium but hardly detectable in bacteria from stationary phase or upon culture in minimal medium (36,45,46). The growth phase-dependent expression pattern of GcvB results from the accumulation of the GcvB sponge SroC in stationary phase, which triggers the RNase E-dependent degradation of GcvB via a base-pairing interaction (43,47). In addition, expression of GcvB is under the control of GcvA, the transcriptional regulator of the glycine cleavage operon gcvTHP (48), and is induced when glycine is available. As with GcvB, the Neisseria sibling sRNAs are most abundant during exponential growth under nutrient-rich conditions, while sRNA levels decline in stationary phase (see Fig. S8 in the supplemental material). Interestingly, in Rhizobiales, a plethora of ABC transport systems is regulated by sibling sRNAs (AbcR1 and AbcR2) (reviewed in reference 37).
Newly identified targets are under both negative and positive control, indicating that the sibling sRNAs not only act by obstructing the RBS as demonstrated previously (19)(20)(21) but also employ other means of target regulation triggered by binding of the mRNA within the CDS (Fig. 4; Fig. S2). In silico-predicted regions of complementarity between the newly identified target mRNAs and the N. gonorrhoeae sibling sRNAs engage sequence motifs which are shared by both sRNA molecules (Fig. S1 and S2). However, in contrast to the well-established prpB, prpC, ack, and NGFG_01721 targets, full posttranscriptional regulation of which requires only one sibling (19; J. Helmreich and D. Beier, unpublished), complete functional redundancy is not observed for new members of the NgncR_162/163 regulon. Single complementation of MS11 DD162/ 163 under steady-state conditions revealed a higher impact of NgncR_163 on target regulation (Fig. 2b), which might be explained by the higher abundance of NgncR_163 (19). Such a diverse regulatory impact of siblings that differ in their abundances but exhibit the same base-pairing capability for the target has also been observed in the case of posttranscriptional regulation of the salmochelin siderophore receptor IroN by Salmonella Typhimurium RhyB1 and RhyB2 (49). Our findings therefore suggest hierarchical target control based on the abundance of the Neisseria sibling sRNAs. However, besides growth phase-dependent differences in sRNA abundancy, which similarly apply to both siblings (Fig. S8), environmental conditions affecting expression of NgncR_162/163 could not yet be identified.
From isotopologue profiling experiments, we deduced a bipartite metabolism in N. gonorrhoeae, where energy metabolism is driven mainly via amino acids like glutamate and proline, which feed the TCA cycle, while glucose degradation via the ED and PPP pathway mostly provides intermediates for anabolic pathways (Fig. 9). This model is well in line with the observation that glutamate, glutamine, and proline (and also asparagine) are consumed to the highest extent when gonococci are grown in chemically defined media (Fig. S6). Similar models for a bipartite metabolic network have already been presented for other pathogenic bacteria such as Listeria monocytogenes, Legionella pneumophila, Coxiella burnetii, Chlamydia trachomatis, and Helicobacter pylori (50)(51)(52)(53)(54). Downregulation of the citrate synthase GltA (and the citrate transporter NGFG_00249) by the sibling sRNAs (19,21) might promote the channeling of glutamate-derived 2-oxoglutarate into the TCA cycle. It is interesting to note that gltA is also a target of Pasteurella multocida GcvB (38). In the sRNA double mutant, activity of the ED and PPP pathways was seemingly dampened, while the TCA cycle was more active, which is in accordance with the fact that several TCA cycle genes are under negative control of the sibling sRNAs (Table 1) (19,21). Interestingly, expression of the ED pathway enzymes glucose 6-phosphate 1-dehydrogenase (zwf) and 6-phosphogluconolactonase (pgl) in meningococci was reported to be negatively regulated by the transcription factor GdhR (55), which itself is under negative control of the sibling sRNAs in N. gonorrhoeae (19). Furthermore, in meningococci, it has been shown that ED pathway genes are upregulated in the presence of glucose, while TCA cycle genes and gdhR are downregulated (56). However, zwf and pgl were not differentially expressed in a gdhR mutant of N. gonorrhoeae (26), and consistently, we did not detect changes in zwf mRNA amounts in the DD162/163 mutant by qRT-PCR (data not shown). Nevertheless, the labeling experiments clearly support the notion that the two putative modules of the bipartite metabolic network in N. gonorrhoeae are under opposite regulation by the sRNAs (Fig. 10e). This suggests a role for the sRNAs in optimizing growth of the pathogen while adapting to different environments during infection.
In this study, we demonstrate that the sibling sRNAs modulate BCAA (NGFG_00045)/ phenylalanine (NGFG_01564) and glycine/alanine (NGFG_01721) import in a reciprocal manner (Fig. 5 and 10). Derepression of glycine import in the absence of NgncR_162/163 is concomitant with upregulation of gcvH and glyA (Fig. 2a, 6, and 10e), suggesting an important role for the sibling sRNAs in the regulation of serine-glycine metabolism, which in turn impacts the biosynthesis of nucleotides, vitamins, and other amino acids via the supply of C1 units. In fact, based on isotopologue profiling data, we propose that the threonine utilization (Tut) cycle (31,57) is active in Neisseria to enable serine biosynthesis from threonine in the absence of 3-phosphoglycerate dehydrogenase, which is not encoded in the genome of the pathogenic Neisseria species. Both serine and threonine were efficiently taken up by wild-type gonococci during growth in chemically defined medium, while the glycine concentration remained almost unchanged (Fig. S6). In contrast, in the sRNA double mutant, the need for serine uptake was compensated to a certain extent by increased glycine uptake and cleavage via the glycine cleavage system, followed by a rise in serine biosynthesis from glycine and 5,10-MTHF due to upregulation of GlyA. This was accompanied by increased 13 C enrichment in threonine, which was converted to glycine to a lesser extent (Fig. 10). The fact that threonine is a precursor to serine might even provide a hint for the biological relevance of positive regulation of the BCAA transporter NGFG_00045 by the sibling sRNAs: since threonine is also a precursor in the biosynthesis of isoleucine, increasing the uptake of isoleucine might spare threonine for use in serine synthesis.
Expression of lactoylglutathione lyase, also named glyoxalase I (GloA), is positively regulated by the sibling sRNAs (Table 1; Fig. 2a). The glyoxalase system mediates the detoxification of the highly reactive compound methylglyoxal, which is a by-product of glycolysis and gluconeogenesis during the conversion of triose phosphate isomers. First, methylglyoxal and glutathione are converted to S-lactoylglutathione by glyoxalase I, which is then cleaved by glyoxalase II, yielding D-lactate and glutathione (reviewed in reference 58). Interestingly, methylglyoxal is also formed during the catabolism of threonine via threonine dehydrogenase in the Tut cycle, due to the decarboxylation of a-amino-b-ketobutyrate and the subsequent oxidation of the intermediate aminoacetone (59). Upregulation of glyoxalase I is concomitant with downregulation of glycine uptake by the sibling sRNAs, which we propose to result in increased Tut cycle activity.
Other new members of the NgncR_162/163 regulon are related to the synthesis and maintenance of iron sulfur (Fe-S) centers (Table 1; Fig. 2a). DnrN is supposed to be involved in the repair of Fe-S centers damaged by oxidative or nitrosative stress (60). Interestingly, dnrN and two other positively regulated NgncR_162/163 targets, norB and aniA, were shown to be under the control of the same transcription factor, NsrR (61). IscR is the transcriptional regulator of the iscRSUA operon encoding enzymes for Fe-S cluster biosynthesis. IscR itself contains an [Fe2-S2] cluster, and holo-IscR was shown to directly repress expression of the iscRSUA operon in E. coli and pathogenic bacteria in order to maintain proper Fe-S cluster homeostasis (reviewed in reference 62). Both dnrN and iscR were deregulated in a Dhfq mutant of N. meningitidis (25), indicating posttranscriptional regulation. Since dnrN and iscR are inversely regulated by the sibling sRNAs, with dnrN being activated and iscR being repressed, metabolic enzymes containing Fe-S clusters might be of particular importance under conditions in which the sibling sRNAs are abundant. Besides glucose and pyruvate, gonococci can use lactate as a carbon and energy source, since electrons from the oxidation of both L-and D-lactate feed directly into the respiratory chain (63). Interestingly, one of the two gonococcal L-lactate dehydrogenases (LutACB) contains Fe-S clusters (64,65) and lactate uptake is indirectly controlled by the sibling sRNAs via GdhR, which is a repressor of the lctP gene encoding lactate permease (26). Lactate permease is considered a virulence factor of N. gonorrhoeae, since lctP-deficient mutants are attenuated in a murine model of lower genital tract infection (66). It should be noted that RNA-seq also suggested the positive regulation of the NGFG_02263 gene, encoding the ortholog of the meningococcal sole glucose transporter (67), by the sibling sRNAs. In N. meningitidis, deletion of hfq resulted in downregulation of the glucose transporter transcript, arguing in favor of direct or indirect sRNA-mediated expression control (25). However, the NGFG_02263 putative target could not be validated by qRT-PCR.
In conclusion, the work presented here expands our knowledge about the mechanisms of action of the sibling sRNAs NgncR_162/163 and their regulon. The data demonstrate that the siblings do not exhibit a complete functional redundancy and that they can act both as negative and positive regulators, thus applying different mechanisms of action, which, however, need to be characterized in more detail in the future. Moreover, the combined results of RNA-seq analysis and isotopologue profiling point to the operation of a bipartite central carbon metabolism in N. gonorrhoeae and a role of the sibling sRNAs in the regulatory networks which govern these central metabolic pathways and link them to the requirement of nutrient uptake and, in particular, amino acid uptake. Thus, the sibling sRNAs appear to play a superior role in the regulatory hierarchy of central metabolic pathways of the gonococcus.

MATERIALS AND METHODS
Bacterial strains and growth conditions. The N. gonorrhoeae mutants used in this study were derived from wild-type strain MS11 (GenBank accession number NC_022240.1) and are listed in Table S1 in the supplemental material. N. gonorrhoeae was grown on GC agar (Oxoid) plates with 1% vitamin mix (19) [1 g], NaCl [5 g]) containing 1% vitamin mix and 0.04% (wt/vol) NaHCO 3 . Growth in chemically defined medium was conducted in CDM10 (68) with slight modifications (L-glutamate, 0.0445 g/L; L-aspartate, 0.02 g/L). For metabolic labeling experiments L-proline and glucose were replaced by [U-13 C 5 ]L-proline (0.05 g/L; 0.4 mM) and [U-13 C 6 ]D-glucose (2.5 g/L; 13.9 mM), respectively. Bacteria were grown to an OD 550 of 0.6 for 4 h in order to achieve steady-state conditions in protein-derived amino acids and fatty acids, which were subjected to isotopologue profiling. Media were supplemented with kanamycin, erythromycin, or spectinomycin at final concentrations of 40 mg/mL, 7 mg/mL, or 50 mg/mL, respectively, when required. Escherichia coli TOP10 (Thermo Fisher Scientific) and E. coli DH5a (69) were cultured in lysogeny broth (LB). When required, antibiotics were added to the following final concentrations: ampicillin, 100 mg/mL; kanamycin, 30 mg/mL; chloramphenicol, 30 mg/mL.
Construction of N. gonorrhoeae mutants. PCR primers for the amplification of DNA fragments used for mutant construction are listed in Table S2. For the synthesis of N. gonorrhoeae-specific fragments, chromosomal DNA of strain MS11 was used as the template. Clonings were performed in E. coli DH5a.
(i) MS11 93gfp, MS11 DD93gfp, MS11 249gfp, and MS11 DD249gfp. N. gonorrhoeae mutants carrying translational target-gfp fusions integrated into the intergenic region between the iga and trpB genes were obtained by transformation of strain MS11 or the DD162/163 mutant with DNA fragments composed of part of the trpB gene, the target-gfp fusion under the control of the promoter of the respective target gene, an erythromycin resistance cassette, and part of the iga gene. To generate appropriate DNA fragments, two consecutive steps of overlap extension PCR were performed. First, a 590-bp fragment from the 39 end of the trpB gene (amplified with primer pairs trbB5/trpBF93 and trpB5/trpBF249) was combined with DNA fragments comprising the upstream region and 23 or 8 NGFG_00093 or NGFG_00249 codons, respectively, which were amplified using primer pairs 93up-5/93up-3 and 249up-5 and 249up-3. Subsequently, the resulting fragments were combined with a segment comprising gfp, ermC, and iga DNA, which was amplified from chromosomal DNA of mutant MS11 1721-gfp (19) with primer pair iga5/Lgfp5.
(ii) MS11 gcvH-F and MS11 DDgcvH-F. In mutants MS11 gcvH-F and MS11 DDgcvH-F, a C-terminal FLAG tag was added to gcvH via allelic-exchange mutagenesis. An appropriate DNA fragment for the transformation of N. gonorrhoeae MS11 and MS11 DD162/163 was obtained via overlap extension PCR by a combination of DNA segments comprising gcvH (amplified with primer pair gcvH-F1/gcvH-F2), the sequence encoding 3ÂFLAG followed by ermC (amplified with primer pair gcvH-3/gcvH-4), and the intergenic region between gcvH and the NGFG_01515 gene as well as part of the NGFG_01515 gene (amplified with primer pair gcvH-5/gcvH-6).
(iv) MS11 P 45 gfp and MS11 DDP 45 gfp. The fusion of the NGFG_00045 upstream region, including the 59 UTR, to the gfp-mut2 gene was constructed by combining DNA fragments amplified with primer pairs 45gfp-1/45gfp-3 and 45gfp-2/45gfp-7 via overlap extension PCR. Plasmid pKEN (72) was used as the template for the amplification of gfp-mut2. The resulting DNA fragment was combined with a DNA fragment comprising ermC and 500 bp from the NGFG_00045 downstream region, which was obtained by overlap extension PCR using DNA fragments amplified with primer pairs 45-5ermC-13/45-3ermC-23 and 45gfp-6/45mut-5, respectively. Transformation of the resulting DNA fragment into N. gonorrhoeae MS11 and MS11 DD162/163 yielded strains MS11 P 45 gfp and MS11 DDP 45 gfp, in which the NGFG_00045 gene is replaced by gfp. A consensus Shine-Dalgarno sequence was introduced in the NGFG_00045 59 UTR by performing overlap extension PCR using DNA fragments amplified with primer pairs 45gfp-1/ 45gfp-8 and 45gfp-9/45mut-5 from chromosomal DNA of strain MS11 P 45 gfp. The combined DNA fragment was transformed into N. gonorrhoeae MS11 and MS11 DD162/163 to yield strains MS11 P 45 gfpSD and MS11 DDP 45 gfpSD.
(v) MS11 D45 and MS11 Dc45. To create the D45 mutant, a DNA segment covering 289 bp from the upstream region and the sequence encoding the first 80 amino acids of NGFG_00045 was replaced by an erythromycin resistance cassette via allelic-exchange mutagenesis. In the DNA fragment used for transformation of MS11, the ermC gene (amplified with primer pair 45-5ermC-13/45-3ermC-3) is flanked by sequences comprising the NGFG_00044 39 end and part of the intergenic region between the NGFG_00044 and NGFG_00045 genes (amplified with primer pair 45-5UTR-1/45-5UTR-23) and encoding amino acids 81 to 239 of NGFG_00045 (amplified with primer pair D451/D45-2). For complementation, overlap extension PCR was applied to insert a kanamycin resistance cassette (amplified with primer pair D45-4/D45-5) between the upstream fragment used for construction of the D45 mutant (amplified with primer pair 45-5UTR-1/D45-3) and a DNA segment covering the promoter region, the 59 UTR, and 785 bp from the 59 end of the NGFG_00045 gene (amplified with primer pair D45-5/D45-6). The resulting DNA fragment was transformed into MS11 D45 to yield MS11 Dc45.
(vii) MS11 D1721. In the D1721 mutant, the NGFG_01721 ORF as well as 248 bp from its upstream region was replaced by a spectinomycin resistance cassette. To construct this mutant, a SacI/PstI fragment comprising part of the upstream NGFG_01720 ORF and the intergenic region was amplified with primer pair 1721up1/1721up2 and was cloned together with a spectinomycin resistance cassette into pSL1180. The spectinomycin cassette expressing aadA1 under the control of the Neisseria P opa promoter was amplified from MS11 DtfpR2 (73) using primer pair spec2(PstI)/Popa5(KpnI). The assembled fragments were subsequently amplified with primer pair 1721up1/D1721-4 and were combined via overlap extension PCR with a DNA fragment derived from the downstream region of NGFG_01721 (amplified with primer pair D1721-3/D1721-2). Transformation of MS11 with the full-length DNA fragment yielded strain MS11 D1721.
Construction of plasmids for sRNA target validation in E. coli. Validation of sRNA-target interactions was performed in E. coli using a GFP-SF-based reporter system (28). DNA fragments comprising the 59 UTR and the first 8 to 32 codons of NGFG_00249, NGFG_00093, NGFG_01937, and dnrN targets were amplified with appropriate primer pairs (Table S1) and were cloned into the BfrBI-and NheI-digested plasmid pXG10-SF (28) to yield plasmids pXG-249, pXG-93, pXG-1937, and pXG-1146, respectively. Forward primers correspond to the annotated transcriptional start site in the case of NGFG_00093 and NGFG_01937 (24) or were chosen arbitrarily when the 59 end of the transcript had not been mapped. Plasmid pXG-863 carrying a glyA-gfp fusion was obtained by cloning a DNA fragment comprising the intergenic region between NGFG_00864 and glyA and encoding the last 8 amino acids of NGFG_00864 as well as the first 40 codons of glyA into the intercistronic fusion vector pXG30-SF (28). Plasmid pXG-1722 carrying a dadA-gfp fusion was constructed accordingly via ligation of a DNA fragment comprising the intergenic region between NGFG_01721 and dadA and encoding the last 24 amino acids of NGFG_01721 as well as the first 40 codons of dadA. Plasmid pJV-162 expressing NgncR_162 and plasmid pJV-162m1 expressing the sRNA with a mutated SL2 sequence have been described previously (19).
RNA preparation, RNA-seq, and analysis of RNA-seq data. N. gonorrhoeae MS11 and MS11 DD162/ 163 were grown to an OD 550 of 0.5 in PPM. RNA was prepared using the miRNeasy micro kit (Qiagen) according to the manufacturer's instructions, followed by DNase I treatment. RNA integrity was checked using a Bioanalyzer (Agilent). After enrichment of mRNA using the Universal RiboDepletion kit (siTOOLs Biotech), cDNA preparation was performed with the NEBNext Ultra directional library preparation kit for Illumina (NEB). The cDNA was sequenced on a HiSeq 3000 (Illumina), yielding 100-bp paired-end reads. Reads with a minimum length of 15 bp after removal of low-quality ends and adapters using Cutadapt (74) were mapped to the N. gonorrhoeae MS11 genome (genome assembly ASM15685v2) (75). Read mapping was conducted using Bowtie2 (76). Genes were quantified using featureCounts (77). DESeq2 (22) was used to identify differentially regulated transcripts.
Northern Blot analysis and real-time quantitative PCR. Northern blot analysis was performed as described previously (19). Quantification of signal intensities was performed using ImageJ (78). For realtime quantitative PCR (qRT-PCR) experiments, 1 mg of RNase-free DNase-treated RNA was reverse transcribed with random hexamer primers using a RevertAid first-strand cDNA synthesis kit (Thermo Scientific). All qRT-PCRs were performed in triplicate in a 20-mL mixture containing cDNA (5 mL of 1:20 dilution), PerfeCTa SYBR green FastMix containing ROX reference dye (Quanta Biosciences), and 18 pmol of primer (Table S2). Amplification and detection of PCR products were performed with a StepOne Plus qRT-PCR system (Applied Biosystems) using the following procedure: 95°C for 10 min and then 40 cycles of 95°C for 15 s and 60°C for 60 s, followed by dissociation curve analysis. The relative expression levels of the genes studied were normalized to the 5S rRNA gene. Data were analyzed using the DDC T method (79). If not stated otherwise, at least three qRT-PCR experiments were performed in triplicate with cDNA that was reverse transcribed from independent RNA preparations.
Immunoblot analysis. For the analysis of GFP expression in E. coli, bacteria were grown to an OD 600 of 1.0 in LB. Bacteria from a culture volume of 2 mL were pelleted and resuspended in 200 mL of Laemmli buffer. N. gonorrhoeae was grown to an OD 550 of 0.5 in PPM. Cells from 1 mL of culture were harvested by centrifugation and resuspended in 50 mL of Laemmli buffer. Samples were incubated for 5 min at 95°C. Western blot analysis of the samples was performed as described previously (19). Quantification of signal intensities was performed using ImageJ (78).
Analysis of amino acid composition of (spent) culture medium. Amino acids in bacterial culture medium were measured by UPLC-electrospray ionization-tandem mass spectometry (UPLC-ESI-MS/MS) using an Acquity UPLC combined with a Quattro Premier triple-quadrupole mass spectrometer (Waters, Milford, MA, USA). Derivatization, chromatographic separation, and successive detection of amino acids were carried out as described by Salazar et al. (80), with modifications.
After sterile filtration, 100 mL of medium was diluted with 100 mL of methanol containing norvaline as an internal standard, with a final concentration of 1 mM. Twenty microliters of this mixture was used for subsequent amino acid derivatization using the AccQÁTag Ultra derivatization kit (Waters, Milford, MA, USA) in accordance with the manufacturer's instructions.
Chromatographic separation was carried out using an ethylene-bridged hybrid (BEH) C 18 column (2.1 by 100 mm, 1.7-mm particle size; Waters) equipped with a VanGuard precolumn and an in-line particle filter. Elution was performed with 100% eluent A for 1 min, followed by a binary solvent gradient to 30% of eluent B within 12 min at a flow rate of 0.4 mL/min. Eluent A consisted of 0.1% formic acid in water, and eluent B consisted of 0.1% formic acid in acetonitrile.
The ESI source was operated in the positive mode at a source temperature of 120°C with the capillary voltage set to 3 kV, the cone voltage at 30 V, and the desolvation gas at 850 L/h at 400°C. Compounds were detected by multiple reaction monitoring (MRM; [M1H] 1 ! m/z 171) with a dwell time of 25 ms and a collision energy of 18 V and using argon as the collision gas at a flow rate of 0.3 mL/min. Data acquisition and processing were carried out using MassLynx and QuanLynx (Waters, Milford, MA, USA; version 4.1).
Sample preparation for 13 C analysis of protein-bound amino acids. The analysis of protein-bound amino acids was done as previously described (81). About 1 mg of lyophilized bacterial cell pellet was hydrolyzed overnight at 105°C after the addition of 500 mL HCl (6 M). The hydrolysate was dried under a gentle stream of nitrogen at 70°C and dissolved in 200 mL of acetic acid (50%). For the isolation of protein-bound amino acids, a cation exchange column of Dowex 50WX8 (H 1 form; 7 by 10 mm; 200 to 400 mesh, 34 to 74 mm) was washed with 1,000 mL of methanol (70%) and 1,000 mL of H 2 O (double distilled). After application of the sample, which was dissolved in acetic acid, to the column, the column was first evolved with 1,600 mL of H 2 O (bidest.). Subsequently, the amino acids were eluted with 1,000 mL of aqueous ammonia solution (4 M). After the ammonia eluate was dried under a gentle stream of nitrogen at 70°C, the isolated amino acids were incubated with 50 mL of N-methyl-N-tert-butyldimethylsilyltrifluoroacetamide (MTBSTFA) containing 1% tert-butyldimethylchlorsilane and 50 mL acetonitrile (anhydrous) for 30 min at 70°C. The N-tert-butyldimethylsilyl (TBDMS) derivates of the amino acids were analyzed by GC-MS.
Sample preparation for 13 C analysis of fatty acids. The analysis of fatty acids was done as previously described (82). In brief, about 5 mg of the lyophilized bacterial cell pellet was dissolved in 1 mL of cold methanol and 800 mg of glass beads was added. Cells were mechanically disrupted and lysed using a Ribolyser system (Hybaid) with three cycles of 20 s at 6.5 m s 21 . Then, the samples were centrifuged for 10 min at 7,000 rpm, and the supernatant was subsequently dried under a gentle stream of nitrogen at room temperature. For derivatization, the dry residue was incubated with 50 mL of MTBSTFA containing 1% tert-butyldimethylchlorsilane and 50 mL acetonitrile (anhydrous) for 1 h at 70°C. The TBDMS metabolites were analyzed by GC-MS.
GC-MS measurement parameters. For the analysis of TBDMS-amino acids, a QP2010 Plus gas chromatograph-mass spectrometer was used as previously described (80). The column was heated to 150°C, kept at 150°C for 3 min, heated to 280°C with a temperature gradient of 7°C min 21 , and kept at 280°for 3 min. For analysis of TBDMS-fatty acids (52), the column was kept at 100°C for 2 min and subsequently heated to 234°C (3°C min 21 ). The column was then heated at 1°C min 21 to 237°C. Finally, the column was heated to 260°C (3°C min 21 ).
Each sample was measured in triplicate in order to account for technical errors. GC-MS data were processed with Shimadzu LabSolution software v4.20. For the calculation of 13 C excess values and isotopologue profiles, Isotopo software was used (34).
Data availability. RNA-seq data obtained in this study have been deposited in the GEO database under accession number GSE177032.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.

ACKNOWLEDGMENTS
Roy Gross is acknowledged for critical reading of the manuscript. We thank Lars Barquist for functional enrichment analysis of RNA-seq data. We declare that there are no conflicts of interest.