NAD+ capping of sibD transcripts in E. coli is mediated by its minimal promoter and enhanced by ppGpp
Wuzhen Liu, Kefan Hu, Shiqi Nie, Zhang Feng, Qiongfang Li, Hailei Zhang, Shumin Liang, Jialin Peng, Shoudong Zhang, Zongwei Cai, Yiji Xia
TL;DR
This study shows that NAD+ capping of RNA in E. coli is controlled by a specific promoter and enhanced by a bacterial alarmone called ppGpp.
Contribution
The study identifies the minimal promoter of sibD as sufficient for NAD capping and reveals that ppGpp and DksA enhance this process.
Findings
The 35-bp minimal promoter of sibD is sufficient for NAD capping of transcripts.
ppGpp and DksA synergistically enhance transcription of NAD-capped and uncapped RNAs from sibD and related genes.
A ppGpp⁰ mutant shows reduced NAD-RNA production from sibD and homologous genes.
Abstract
Recently, nicotinamide adenine dinucleotide (NAD+) and other nucleotide analogs have been identified as non-canonical RNA caps in both prokaryotes and eukaryotes. In Escherichia coli, NAD capping has been shown to be influenced by environmental conditions in a gene-specific manner, yet its regulatory mechanisms remain poorly understood. We previously reported that most transcripts produced by sibD are NAD-capped during the stationary phase. In this study, we found that the 35-bp minimal promoter of sibD is sufficient for its NAD capping. When this minimal promoter was applied to express genes not typically producing NAD-RNAs, their transcripts could also be NAD-capped. These findings strongly support that NAD capping in E. coli occurs during transcription initiation mediated by the promoter and RNA polymerase. Additionally, the bacterial alarmone ppGpp and the small protein DksA, both…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9
Figure 10- —Research Grants Council of Hong Kong
- —GRF10.13039/100001641
- —CRS10.13039/100009326
- —CRF10.13039/100002002
- —AoE
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsRNA modifications and cancer · RNA and protein synthesis mechanisms · Bacterial Genetics and Biotechnology
Introduction
In eukaryotes, mature messenger RNA (mRNA) generally features a methylated guanosine moiety at its 5′ end, known as the m^7^G cap. This cap structure plays crucial roles in maintaining RNA stability, mRNA processing, transporting RNA from the nucleus to the cytosol, and initiating translation [1–3]. Unlike eukaryotic RNA, bacterial RNA lacks such a cap structure. However, in 2009, certain RNAs in Escherichia coli were found to carry the NAD (nicotinamide adenine dinucleotide) moiety at their 5′ ends [4]. This cap structure was subsequently identified in archaea [5], yeast [6, 7], plants [8–11], and mammalian cells [12]. Additionally, several other noncanonical cap structures, including NADH, dpCoA, FAD, NpnN, and cell wall precursors (UDP-glc and UDP-glcNAc), were identified in both prokaryotes and eukaryotes [13–20].
Among these non-canonical caps, NAD-capped RNAs (NAD-RNAs) have been relatively well-studied. However, the biological functions of NAD-RNAs remain elusive. In E. coli, NAD-capped RNAs have been found to be more stable than the uncapped triphosphate RNAs [21]. In contrast, NAD-capped mRNA in human cells was found to be more susceptible to degradation compared to their m^7^G-capped counterparts [12]. In bacteria, it appears that the NAD cap did not affect the translation process [22], whereas NAD-capped luciferase mRNAs were found not to be translated in mammalian cells [12]. Conversely, other studies suggest that NAD-capped mRNAs are enriched in the polysome fraction under active translation in plant cells [8]. Recently, NAD-RNAs from E. coli were found to serve as natural substrates for an (ADP)-ribosyltransferase (ART) of T4 phage, linking the RNA to ribosomal protein S1 during phage infection [23].
Various RNA polymerases, such as T7 RNA polymerase [24] and those from E. coli and mitochondria of yeast and humans, are capable of producing NAD-RNAs through in vitro transcription (IVT) in the presence of NAD [14, 25]. This supports the hypothesis that the NAD cap is incorporated into RNAs as the first nucleotide in place of adenylate during transcription initiation. Structural analysis of the NAD-RNA/bacterial RNA polymerase (RNAP) complex further confirms this hypothesis [22, 26]. Moreover, altering the transcription start site (TSS) from the original +1A to +1G completely inhibits the synthesis of cellular NAD-RNAs [14]. The −1 site adjacent to the TSS is also crucial for the NAD incorporation process [14, 27]. Additionally, a consensus sequence H_-3_R_-2_R_-1_A_+1_S_+2_W_+3_W_+4_ (where the +1A at TSS underlined, H represents A, T, or C; R stands for G or A; S indicates G or C; and W consists of A or T) has been identified to favor NAD capping [28]. This consensus motif aligns well with genes encoding NAD-RNAs [21, 27–29]. On the other hand, the report on the identification of several small nucleolar RNAs (snoRNAs), which undergo 5′ trimming during maturation, as NAD-RNAs in human cells, raises a possibility that certain NAD-RNAs might be produced via post-transcriptional enzymatic reactions [12], although the direct evidence of post-transcriptional NAD capping is still lacking.
Using the NAD-tagSeq II method for NAD-RNA profiling, we have previously found that three sib family genes, sibC, sibD, and sibE, produced a majority of their transcripts as NAD-RNAs during the stationary phase, whereas their NAD-RNA proportion was much lower at the exponential phase [29]. The sib genes encode small antitoxin RNAs that function as antisense RNAs for suppressing the expression of ibs toxins in E. coli [30–32]. The finding indicates that NAD capping is influenced by environmental factors. However, *cis-*elements and trans-acting factors that mediate NAD capping remain to be defined.
In this study, we primarily used sibD as a model to investigate the cis elements that might control its NAD capping. We discovered that the 35-bp minimal promoter of sibD is sufficient to drive NAD capping of its own transcripts as well as other RNAs. Furthermore, the alarmone ppGpp and the small protein DksA were found to increase the production of both NAD-capped and uncapped SibD RNAs in vitro, as well as to enhance NAD capping of SibD in vivo. ppGpp and dksA are known to bind to E. coli RNAP to mediate transcriptional initiation [33–41]. This study advances our understanding of the molecular mechanisms that regulate NAD capping.
Methods
Bacteria growth and media
All E. coli strains used in this study were derived in the K12 MG1655 strain background, and their detailed genotypes are listed in Supplementary Table S1. Escherichia coli cells were grown at 37°C in LB medium (1% w/v peptone, 0.5% w/v yeast extract, 1% w/v NaCl).
Plasmids construction and PCR mediated site-directed mutagenesis
All plasmids and oligonucleotides used in this study were listed in Supplementary Tables S2 and S3, respectively. The sibD gene (a DNA fragment from −189 to +345 relative to its TSS) and the sibE gene (from −190 to +340) were synthesized by BGI, Shenzhen, and verified by Sanger sequencing. Polymerase chain reaction (PCR) amplification of the sibD and sibE genes was conducted from these synthesized genes, while amplifications of other genes were performed from the genomic DNA of E. coli K12 MG1655 using gene-specific primers. The PCR products were cloned into the pMD19-T vector (Takara) and validated by Sanger sequencing.
A one-step overlap extension PCR (OE-PCR) was employed to generate the plasmid pWZ19 using primers pWZ19_overlap_F and pWZ19_overlap_R, as described in cited references [42–46] (Supplementary Fig. S2A). Positive single colonies containing the pWZ19 plasmid were verified through plasmid isolation and sequencing. The sibD gene fragment (from −189 to +165), digested with BglII and SalI, was inserted into the pWZ19 vector to produce the plasmid pWZ19-sibD. To express RNAs from the pET28 vector backbone, BglII and SalI were used to remove the T7 RNAP expression elements, and the trimmed pET28 vector backbone will be subjected to ligation with sibD gene fragment (from −189 to +165). To generate pBBR-sibD construct, the same sibD gene fragment was cloned into a low-copy plasmid pBBR1MCS-2 vector [47].
Generation of ∆sibD/ibsD, ∆dksA, ppGpp0, and other mutants with genetic modifications
A CRISPR-Cas9 gene editing system was employed to generate the ∆sibD/ibsD, ∆dksA, ∆relA, ppGpp^0^, sibD P-35-trpT, sibD P-35-ryjA, sibD P-35-sroC, and sibD P-35-symR mutants in the E. coli genome, following established protocol [48]. In this section, the construction of the ∆sibD/ibsD knockout mutant is described as an example. The plasmid pACRISPR-sibD Oligo*-∆sibD/ibsD* was constructed to produce guide RNA targeting the sibD gene and to provide a ∆sibD/ibsD template for DNA repairing. This plasmid, along with pCasPA, was transformed into wild-type E. coli cells. 0.3 mM Isopropyl β-D-thiogalactoside (IPTG) and 5 mM arabinose were added to induce the expression of guide RNAs and the Cas9 protein for in vivo gene editing. Colonies contain the ∆sibD/ibsD mutation were screened by PCR and validated by Sanger sequencing.
To generate the ppGpp^0^ mutant, the spoT gene was deleted in a ∆relA genetic background using the CRISPR-Cas9 system as described earlier, resulting in the ∆relA∆spoT double mutant (hereafter referred to ppGpp^0^), which lacks the ability to synthesize ppGpp. To construct the pACRISPR template for the sibD P-35-trpT mutant, we inserted a DNA sequence containing the 266-bp rrnB terminator (including both T1 and T2 terminators) to mediate termination of the upstream rrnC operon, followed by sibD P-35 sequence positioned immediately upstream of the trpT gene body. A similar strategy was applied to generate the sibD P-35-sroC mutant strain. For the sibD P-35-ryjA and sibD P-35-symR strains, the 35-bp sibD minimal promoter was placed immediately upstream of the TSS of each gene.
Expression and purification of His-tagged DksA proteins
Rosetta (DE3) cells with pET28a-DksA were cultured to an OD_600_ of ∼0.4, then induced with 0.3 mM IPTG at 28°C for 3 h. The cells were then collected and lysed by sonication in a lysis buffer (10 mM Tris–HCl, pH 7.4; 150 mM NaCl; 0.5% Triton X-100; 10 mM imidazole) and centrifuged. The supernatant was incubated with HisPur™ Ni-NTA Resin (Thermo Fisher) for His-tagged protein binding. The beads were subsequently washed with buffer (10 mM Tris–HCl, pH 7.4; 150 mM NaCl; 0.5% Triton X-100) in a series containing 30, 40, and 60 mM imidazole. His-tagged proteins were then eluted with the elution buffer (10 mM Tris–HCl, pH 7.4; 150 mM NaCl; 300 mM imidazole). Imidazole contaminants in the eluted proteins were then removed using Amicon® Ultra Centrifugal Filters (10 kDa MWCO). The purified proteins were dissolved in storage buffer (10 mM Tris–HCl, pH 7.4; 150 mM NaCl; 50% glycerol) to 1 μg/μl and stored in a –80°C freezer. The expression and purification of rpoD (σ^70^) and other five sigma factors were also conducted using similar procedures.
RNA isolation and northern blotting
For RNA isolation, bacterial cells in mid-log phase (OD_600_ ∼0.6) or stationary phase (OD_600_ ∼2.3) were immediately killed with an equal volume of RNA stabilization buffer (5% phenol, 95% ethanol) to preserve RNA integrity. The cells were then centrifuged, and RNA isolation was performed using TRIzol reagent (CWBIO) following the manufacturer’s instructions. The RNA pellets were then dissolved in DEPC-treated water.
Northern blotting analysis was carried out as described in cited references with minor modifications [49, 50]. Briefly, 10 μg of denatured RNAs were loaded onto a 6% polyacrylamide gel electrophoresis (PAGE) gel containing 7 M urea for separation. After electrophoresis, the RNAs were transferred to a nylon membrane and crosslinked using UV light. The membrane was hybridized with digoxin-labelled specific DNA probes in a hybridization buffer (Thermo Fisher) at 42°C overnight. Post-hybridization, the membrane was blocked and washed with the DIG Wash and Block Buffer Set (Roche). Band signals were detected using Anti-Digoxigenin-AP Fab fragments (Roche) following the manufacturer’s instructions.
NAD cap detection by mass spectrometry from both purified SibD RNAs and E. coli total RNAs
To isolate and purify sibD transcripts from wild-type E. coli cells, a biotin-labelled sibD-specific DNA probe was used to hybridize SibD RNAs. The hybridized RNAs were enriched and purified with streptavidin beads (NEB), followed by nuclease P1 digestion and liquid chromatography–mass spectrometry (LC–MS) analysis to detect released NAD caps.
To compare NAD cap content among different E. coli strains, stationary-phase cells of the indicated strains were subjected to total RNA isolation. Each RNA pellet was washed with 75% ethanol at least four times and then filtered using an ultra-centrifugal column (Amicon, 10 kDa MWCO) at least five times to remove free NAD contamination, yielding ultra-pure RNA samples. Nuclease P1 was then used to digest these ultra-pure RNA samples before conducting LC–MS analysis to detect NAD content.
NAD-RNA biotinylation with ADPRC-SPAAC reaction
The ADPRC-SPAAC reactions assay involves two steps, as described previously [29]. First, 3-azido-propanol (Sigma–Aldrich) was used to label the NAD cap via the ADPRC enzyme (Sigma–Aldrich) at room temperature for an hour. The azido-modified NAD cap is then tagged with a biotin group through a SPAAC-type click-chemistry reaction at 37°C for an hour. Finally, the biotinylated RNAs are dissolved in DEPC-treated water.
NAD cap detection with NADbio-northern blotting, NADbio-streptavidin blotting, and APB gel blotting analysis
To detect NAD caps on SibD or other NAD-RNAs with NADbio-northern blotting, total RNAs from E. coli cells or synthetic RNAs from IVT products were subjected to the ADPRC-SPAAC reaction assay to label NAD caps with biotin groups. The biotinylated NAD-RNAs are then enriched using streptavidin beads and eluted with TRIzol reagent. Individual NAD-RNAs from the final eluate are detected by northern blotting with a digoxin-modified specific DNA probe.
To detect NAD caps on sibD or other NAD-RNAs with NADbio-streptavidin blotting, total RNAs from plasmid-containing E. coli cells or synthetic RNAs from IVT products were subjected to ADPRC-SPAAC reaction for biotinylation of NAD-RNAs. RNAs were then separated on a 7 M urea gel and transferred to a nylon membrane. HRP-streptavidin (Thermo Fisher) was then applied to detected biotin groups on biotinylated RNAs from the membrane via blotting.
To detect and compare NAD-capped version RNAs with total RNAs from isolated RNAs. A TAE-PAGE gel with 0.5% APB (3-(Acrylamido) phenylboronic acid) (Aladdin) was prepared as published [51]. 10 μg of total RNAs were then separated on the APB-containing gel. After electrophoresis, the RNA bands were transferred to a nylon membrane. A digoxin‐labelled specific probe was used to detect NAD-RNA via northern blotting. The NAD capping ratio for each NAD-RNA in each sample was calculated with the following formula:
Capping ratio (%) = (Band intensity of capped band) × 100 / (Total band intensities of (capped + uncapped)).
In vitro transcription with T7 RNA polymerase
The T7 ϕ2.5 promoter (TAATACGACTCACTATTA) was applied to synthesize SibD RNAs and other model RNAs with T7 RNA polymerase. To prepare the synthetic 112 nt model RNA with 5′-NAD or 5′-ppp linkages, a DNA fragment with the sequence ATTCAGTAATACGACTCACTATTA CTTGTTTTGGTGTCTGCGCTCCTCCTTGCCTGTTTCCTCGGTTCTTTGTGTTGGTTGCTCTGTGTTCCTTCGTTTTTCCGCCCTGCTTGGCGGTTTTTTCGTTTTCTGTGC (with the T7 ϕ2.5 promoter sequence bolded) was used as template for IVT. Similarly, PCR products containing the sibD and sfGFP genes were used as DNA templates for synthesizing ppp-SibD, NAD-SibD, ppp-sfGFP mRNA, and NAD-sfGFP mRNAs. To synthesize the 38 nt Model RNA, both the T7_38nt_F oligo with the sequence TCCAGTAATACGACTCACTATTA GGGTGGTGTGTGTTGTTCTTTGGTGTCTTTGCTTGG (with the T7 ϕ2.5 promoter sequence bolded) and the T7_38nt_R oligo with the complementary sequence were annealed into double-strand DNAs in a PCR machine, yielding the DNA templates for IVT use. DNA templates for transcription from other genes were prepared by PCR using primer pairs listed in Supplementary Table S3.
IVT assays were then performed to produce NAD-capped SibD RNAs and other NAD-RNAs following methods from our previous publication [29]. Transcription was conducted in a 400 μl mix: 40 mM Tris–HCl (pH 8.0), 1 mM spermidine, 22 mM MgCl_2_, and 0.01% Triton X-100, 1 mM rNTPs, 5 mM NAD (Sigma–Aldrich) and supplied with 1 μg DNA templates, 250 U T7 RNAP (NEB), and murine RNase inhibitor (NEB). The transcription reaction was conducted for 4 h, and DNA templates in the transcription products were removed by treatment with DNase I (NEB). The resulting DNA-free RNA products were precipitated with isopropanol and dissolved in DEPC-treated water.
In vitro transcription with E. coli RNA polymerase in the presence or absence of cellular factors
To generate the sibD template for IVT with E. coli RNAP, the primers sibD P-35_F and SibD145nt_R were used to generate DNA templates for generating SibD RNAs with E. coli RNA polymerase (NEB).
To determine whether ppGpp (Jena Bioscience), pppGpp (Jena Bioscience), or protein DksA impact the NAD capping of SibD RNA, a 200 μl transcription mix was prepared with the following recipe: 40 mM Tris–HCl (pH 8.0), 1 mM spermidine, 22 mM MgCl_2_, and 0.01% Triton X-100, 1 mM rNTPs, and 5 mM NAD, along with 1 μg DNA templates, 10 U E. coli RNAP holoenzyme, murine RNase inhibitor, and specified types and amounts of cell factors. For synthesizing the ppp-SibD using the E. coli RNA polymerase, the same transcription mix was prepared without the adding of NAD. The reactions were conducted for 2 h, followed by chloroform purification and isopropanol precipitation. DNase I was then used to remove DNA template contaminants, and the DNA-free RNAs were further analysed using northern blotting and APB gel blotting.
To perform IVT assays with various sigma factors, a similar assay was conducted, except the E. coli RNAP holoenzyme was replaced by the same amount of E. coli RNAP core enzyme (NEB) and in the absence of ppGpp or DksA. The 10 U of E. coli RNAP core enzyme was saturated with 10 μg of purified sigma factor protein before being added to the transcription mix. SibD RNAs in the final products were detected with northern blotting and APB gel blotting. Three independent biological replicates were performed for each IVT condition. The capping ratio of SibD RNA was calculated for each replicate, and the mean values with standard deviations were plotted as line charts. Statistical significance was assessed using an unpaired two-tailed Student’s t-test.
RNA sequence and transcription start site identification
To identify the RNA sequence of sibD and RNA I transcripts expressed from plasmids, 1 mg of total RNAs from ∆sibD/ibsD cells containing either pWZ19 or pWZ19-sibD were subjected to ADPRC-SPAAC reactions to label NAD-RNAs with biotin groups. The biotinylated NAD-RNAs were then enriched and purified with streptavidin beads (NEB), followed by RNA elution with TRIzol reagent (CWBIO). The eluted RNAs were then separated on a PAGE gel containing 7 M urea, and RNA patterns were visualized with Redsafe (Geneon) gel staining. RNA bands corresponding to SibD and RNA I were excised from the gel and ground into small pieces, then soaked in RNA storage buffer [0.3 M NaOAc, pH 5.3, 0.5% sodium dodecyl sulphate (SDS)] with gentle shaking. After overnight incubation, the RNAs were recovered by chloroform extraction and isopropanol precipitation.
The recovered RNAs were then subjected to a decapping assay by incubating 20 ng of RNA with 0.5 μl NudC (M0607S, NEB) and 0.5 μl RppH (M0356S, NEB). For SibD RNAs, the decapped SibD RNAs were circularized with 30 U T4 RNA ligase (Thermo Fisher) and converted into complementary DNA (cDNA) through reverse transcription with SuperScript™ II Reverse Transcriptase (Invitrogen). The circularized DNA fragment containing the junction formed by the 5′ and 3′ ends of sibD was amplified using sibD-specific primers and cloned into the pMD19-T vector (Takara). The NAD capping site of sibD and its sequence were identified by Sanger sequencing. For RNA I, the decapped RNA products were ligated to an RNA adapter (TCTACACTCTTTCCCTACACGACGCTCTTCCGATCUUU), followed by reverse transcription using an RNA I-specific primer to yield cDNA of RNA I fused with the adapter. The cDNA was then amplified by PCR using adapter-specific and RNA I-specific primers, and the resulting fragments were cloned into the pMD19-T vector (Takara) for further sequence identification.
To define the TSS of SibD purified from IVT products. An RNA adapter (TCTACACTCTTTCCCTACACGACGCTCTTCCGATCUUU) was ligated to the NudC/RppH-decapped RNAs. Reverse transcription was performed with the sibD specific primer, followed by PCR amplification with the adapter-specific primer and sibD specific RT primer. The RNA sequence and TSS on sibD gene were further identified by Sanger sequencing.
Transient expression and detection of FLAG-tagged RelA 455aa and DksA proteins in E. coli
For transient expression of FLAG-tagged 455-aa N-terminus of RelA and DksA proteins, overnight cultures of wild-type E. coli cells harbouring either pBAD33-RelA 455aa-FLAG or pBAD33-DksA-FLAG plasmids were diluted into 5 ml of fresh LB medium to an initial OD_600_ of 0.1. Cultures were incubated at 37°C with shaking at 200 rpm for 2 h, until the OD_600_ reached ~0.6. Protein expression was induced by adding arabinose to a final concentration of 10 mM. After 1 h of induction, 300 μl of cells were collected and pelleted by centrifugation. As a control, cells treated with an equal volume of water instead of arabinose were collected. Cell pellets were resuspended in 100 μl of 1 × SDS loading buffer (63 mM Tris–HCl, pH 6.8, 10% glycerol, 1% SDS, 0.005% bromophenol blue, 5 mM DTT), and samples were boiled at 95°C for at least 10 min before loading onto the SDS–PAGE gel. Proteins were separated by SDS–PAGE and transferred to a nitrocellulose membrane (Bio-Rad). FLAG-tagged RelA 455aa and DksA proteins were detected using monoclonal ANTI-FLAG M2-peroxidase (HRP) antibody (Sigma–Aldrich).
Phenotypic analysis of ∆dksA and ppGpp0 mutants
Overnight cultures of E. coli strains were washed three times with sterilized water, then resuspended and adjusted to a final OD_600_ of 0.05. Serial tenfold dilutions were performed to obtain suspensions at OD_600_ values of 0.05, 0.005, 0.0005, and 0.00005. For spot assays, 5 μl of each dilution was plated onto LB rich medium and M9 minimal medium plates. Each condition was replicated three times. The growth status of each strain was monitored, and photographs were taken at defined time intervals.
sfGFP quantification in wild-type and ppGpp0 backgrounds
To compare growth and sfGFP intensity for each soluble RNA (sRNA)-encoding gene promoter-driven sfGFP reporter between wild-type and ppGpp^0^ backgrounds, serial dilution spot assays were performed on LB plates as described above. sfGFP fluorescence of each colony spot was imaged using a ChemiDoc MP imaging system (Bio-Rad), and the fluorescence intensity was quantified using ImageJ software. Data are presented as means ± standard deviation. Statistical significance was tested using Duncan’s multiple range test, and statistically distinct groups are indicated by different letters for each sample.
Results
Detection of specific NAD-RNA with NADbio-northern blotting in E. coli
The sibD gene of E. coli encodes small RNA that functions as an antisense transcript to inhibit expression of toxin IbsD, which is encoded by the opposite DNA strand [30, 31]. Using NAD tagSeq II, a method for genome-wide profiling of NAD-capped RNAs (NAD-RNAs), we previously identified sibD as one of several genes in the E. coli K-12 strain whose transcripts are predominantly NAD-capped during the stationary phase [29]. To validate this sequencing result, we purified sibD transcripts from total E. coli RNA through hybridization using biotin-conjugated sibD-specific probes. The purified RNAs were then digested with nuclease P1 to release cap structures and individual nucleotides. As shown in Fig. 1A, LC–MS analysis confirmed the presence of NAD caps in SibD RNAs.
Development and validation of NADbio-northern blotting analysis. (A) sibD gene produces NAD-RNAs. The NAD cap content in various RNA samples was determined with LC–MS. The samples analyzed included purified sibD transcripts from the E. coli K12 strain, along with 10 nM NAD+ standard. One hundred nanograms of ppp-SibD RNA and NAD-SibD RNA from IVT products were used as negative and positive controls, respectively. (B) A schematic illustration of the workflow for the NADbio-northern blotting analysis. NAD and biotin are highlighted in red and blue, respectively. ‘DIG’ highlighted in yellow indicates the digoxin tag in the DNA probes. (C) Detection of NAD-RNAs by NADbio-northern blotting with synthetic 5′-ppp-SibD and NAD-SibD. HRP-streptavidin blotting was performed to monitor biotinylation of RNAs mediated by the ADPRC-SPAAC reaction. In addition, a digoxin-tagged sibD DNA probe was used to detect sibD transcripts. ‘ADPRC+’ indicates the biotinylation of NAD-RNAs via the ADPRC-SPAAC reaction with ADPRC, while ‘ADPRC-’ denotes the ADPRC-SPAAC reaction without ADPRC. (D) Detection of several NAD-capped cellular RNAs with NADbio-northern blotting analysis. Total RNAs isolated from the wild-type strain were subjected to NADbio-northern blotting analysis. Individual RNAs in the eluate were detected with digoxin-conjugated gene-specific probes. The bands corresponding to individual NAD-RNAs in ADPRC+ lanes were marked with red arrows.
To detect NAD-RNA from individual genes in a more cost-effective way, we developed the NADbio-northern blotting (NAD-RNA biotinylation coupled with northern blotting analysis). In this method, the NAD cap in RNA samples is biotin-labelled via an ADPRC-catalysed reaction followed by SPAAC click chemistry (ADPRC-SPAAC reaction) [52]. Biotinylated NAD-RNAs are then enriched and purified with streptavidin beads. Finally, a DNA probe conjugated with digoxin is used in northern blotting to detect specific RNAs in the eluate (Fig. 1B). As NAD-RNAs are typically low abundant [21, 29], biotin tagging of NAD-RNAs and enrichment through streptavidin allow detection of the enriched NAD-RNAs through northern blotting analysis.
We first used synthetic RNAs to test the NADbio-northern blotting method. NAD-capped and uncapped (5′ ppp-) SibD RNA were synthesized through in vitro transcription. After ADPRC-SPAAC reactions, the biotinylation of RNAs was detected using the HRP-streptavidin blotting. Only NAD-SibD, but not ppp-SibD was detected (Fig. 1C), indicating that ADPRC-SPAAC-mediated biotinylation is specific to NAD-capped RNAs. In addition to SibD, two other in vitro synthesized RNAs, including a 112-nt NAD-RNA previously used as a model RNA molecule [26] and NAD-sfGFP mRNA, could also be specifically detected via NADbio-northern blotting analysis (Supplementary Fig. S1A).
Next, to determine whether NADbio-northern blotting can detect endogenous NAD-RNAs, we applied the method to detect known NAD-RNAs in total RNAs extracted from wild-type E. coli. Several NAD-RNAs previously identified by NAD tagSeq II [29], including SibC, SibD, SibE, and ChiX, were successfully detected by this method. In contrast, no signal was detected for SibA and SibB RNAs (Fig. 1D), which were previously reported to carry no detectable level of NAD caps [29]. These results indicate that NADbio-northern blotting is a reliable method for determining whether specific RNA species are NAD capped in vivo.
Detection of plasmid-derived NAD-RNAs with NADbio-streptavidin blotting analysis
Although NADbio-northern blotting can be used to determine whether a specific RNA is NAD-capped, we found that its sensitivity is still too low to detect NAD-RNAs from most loci in E. coli. This could be explained by the fact that transcripts with an NAD cap typically account for <1% of the total transcripts for most genes in E. coli [29], making it difficult to enrich sufficient biotinylated NAD-RNAs for northern blot detection. We hypothesized that genes carried on plasmids could produce higher levels of NAD-RNAs, facilitating easier detection and serving as an effective alternative genetic system to elucidate genetic and physiological factors influencing NAD capping in E. coli.
To test this hypothesis, we cloned the sibD gene (spanning −189 to +165 relative to its TSS) into the high-copy pWZ19 plasmid to create the pWZ19-sibD construct and transformed it into the ∆sibD/ibsD deletion strain. We then developed a Western blotting-like method, NADbio-streptavidin blotting (NAD-RNA biotinylation coupled with HRP-streptavidin blotting), for detecting and quantifying NAD-RNAs from the sibD gene in the plasmid (Fig. 2A). In this method, NAD-RNAs in total RNA samples are biotin-labelled with ADPRC-SPAAC reactions, similar to NADbio-northern blotting. After biotinylation, the RNA samples are separated by gel electrophoresis, transferred to a membrane, and probed with HRP-conjugated streptavidin. Compared to NADbio-northern blotting, this method eliminates the RNA enrichment with streptavidin beads and DNA-probe hybridization steps, making it less time-consuming and more cost-effective for detecting high abundant NAD-RNAs produced from a plasmid clone.
NAD-SibD RNA could be produced at high levels from plasmid clones and detected with NADbio-streptavidin blotting analysis. (A) Schematic workflow for detecting NAD-SibD with NADbio-streptavidin blotting, with the biotin group of biotin-NAD-RNAs highlighted in red. Three main steps are illustrated in this method, including RNA isolation from plasmid-containing cells, NAD cap biotinylation with ADPRC-SPAAC reaction, and NAD-RNA detection with HRP-streptavidin blotting. (B) Detection of the NAD-SibD RNAs with NADbio-streptavidin blotting. ppp-SibD indicates synthetic SibD RNA with 5′ triphosphate linkage, NAD-SibD refers to synthetic SibD RNA with NAD caps. ‘ADPRC+’ indicates the biotinylation of NAD-RNAs via the ADPRC-SPAAC reaction with ADPRC, while ‘ADPRC-’ denotes the ADPRC-SPAAC reaction without ADPRC. (C) Detection of NAD-SibD from the sibD gene in three different plasmids with NADbio-streptavidin blotting. Ten micrograms of total RNAs were used and detected with NADbio-streptavidin blotting for each lane. Bands of biotinylated NAD-SibD (labelled as bio-SibD) and biotinylated NAD-RNA I (labelled as bio-RNA I) were marked with arrows. EV refers to the empty vector. Two exposition times, namely 30 and 300 s, were applied to detect abundance of SibD RNAs. 5S rRNAs were detected as loading controls.
This method was able to detect as little as 0.5 ng of synthetic NAD-SibD RNAs and NAD-capped SibD RNAs from 2 µg of total RNAs extracted from the pWZ19-sibD strain (Fig. 2B). In the RNA samples from the pWZ19-sibD strain, two prominent biotinylated bands were detected (Fig. 2B and C). We collected these two bands and used rapid amplification of cDNA ends (RACE) followed by sequencing (Supplementary Figs S3 and S4), a method modified from the CapZyme-seq, to identify them [28, 53]. The upper band represents the full-length sibD transcript (145 nt), while the lower band comprises two different RNA species: a truncated SibD variant (lacking the 3′ sequence, ∼10% of the total transcripts) and RNA I (Supplementary Fig. S3). RNA I is transcribed from the sequence of the ColE-like plasmid replication origin, and NAD-capped RNA I has previously been identified as the most abundant NAD-RNA in strains carrying these plasmids [21, 22, 26]. This RNA I band was also detected in the strain carrying the empty pWZ19 and empty pET28 plasmids (Fig. 2C).
Although cellular NAD-RNAs produced from chromosomal genes can also be biotinylated during ADPRC-SPAAC reaction, their levels were apparently too low to be detected by this method (Fig. 2C). ImageJ quantification revealed that biotinylated full-length NAD-SibD accounted for around 70% of the combined bands’ intensity (Fig. 2C). Furthermore, LC–MS analysis showed a five-fold increase in total NAD cap content in the pWZ19-sibD strain compared to the strain carrying the empty pWZ19 plasmid and a 23-fold increase over the wild-type strain without the plasmid (Supplementary Fig. S1B). Additionally, the RNA band corresponding to the full-length sibD transcript could be directly visualized when SibD RNA was overexpressed using the pWZ19 vector and separated on a PAGE gel, followed by nucleotide dye staining (Supplementary Fig. S5). These results indicate that the plasmid-carried gene can produce a very high level of NAD-capped transcripts.
To further validate the broader applicability of the NADbio-streptavidin blotting method, we detected NAD-RNAs from sibD gene carried in the plasmids with varying copy numbers: high-copy pWZ19 (∼200 copies/cell), moderate-copy pET (20–30 copies/cell), and low-copy pBBR (7–8 copies/cell) (Supplementary Fig. S2B). NAD-SibD was detectable in all three cases with NADbio-streptavidin blotting (Fig. 2C).
NAD capping of Sib RNAs is dependent on the adenosine at the transcription start site
It is generally believed that the NAD cap is formed when RNA polymerase utilizes NAD instead of ATP during the transcription initiation process [26], as both molecules contain an adenosine monophosphate (AMP) group that pairs with thymine (T) to initiate transcription (Fig. 3A). If this is the case, the TSS should normally be the base A to produce NAD-capped transcripts, which is indeed the case for the sibD gene, as it has a +1A at the TSS. To test this, we converted the TSS of sibD from +1A to +1G in the pET-sibD and pWZ19-sibD plasmids (Fig. 3B). The NADbio-streptavidin blotting results showed that the bands representing biotinylated NAD-SibD RNAs were absent in the sibD +1G mutation strains, while the levels of total transcripts were not significantly affected by the +1G mutation (Fig. 3B), indicating that +1A is essential for NAD capping of SibD RNA.
The +1A TSS is essential for NAD capping of all Sib RNAs. (A) Schematic illustration of the molecular structures of ATP and NAD, with AMP groups highlighted in purple. (B) The NAD-capped SibD transcript was not detected when its TSS was mutated from +1A to +1G. All sibD gene isoforms were expressed from the pET or pWZ19 plasmids in a ∆sibD/ibsD genetic background. NAD caps on sibD and the sibD +1G variant were initially detected with NADbio-streptavidin blotting (top panel). APB gel electrophoresis (second panel) and standard gel electrophoresis were then performed for northern blotting (labelled as ‘normal gel’, third panel). ‘ADPRC+’ indicates the biotinylation of NAD-RNAs via the ADPRC-SPAAC reaction, while ‘ADPRC−’ denotes the ADPRC-SPAAC reaction without ADPRC. The positions of biotinylated NAD-SibD (labelled as bio-SibD) and biotinylated NAD-RNA I (labelled as bio-RNA I) RNAs were marked with arrows. (C) Detection of NAD caps for all five Sib RNAs and their variants in pWZ19 plasmids with NADbio-streptavidin blotting. The positions of biotinylated NAD-Sib RNAs (labelled as bio-Sib RNA) were marked with red arrows. EV refers to the pWZ19 empty vector. 5S rRNA was detected as a loading control.
Additionally, 3-acrylamidophenylboronic acid (APB) gel electrophoresis coupled with northern blotting was performed to detect and compare NAD-capped and uncapped SibD RNA. The boronic acid in the APB gel reversibly interacts with the cis‐diol group of NAD, causing NAD-capped RNA to migrate more slowly than uncapped transcripts, resulting in two distinct bands [27, 51, 54]. This analysis revealed that approximately half of SibD transcripts were NAD capped, whereas the NAD cap was absent in the sibD + 1G strains, although the total transcript level was not significantly affected by the mutation (Fig. 3B).
Two other sib genes, namely sibC and sibE, were also previously found to produce a high proportion of their transcripts as NAD-capped RNAs [29]. As expected, both sib genes have their annotated TSS as adenosine (+1A). On the other hand, another two sib genes, namely sibA and sibB, have a TSS as guanosine (+1G) and were not found to produce any NAD-capped RNAs in our NAD tagseq II results [29]. To determine whether replacing the +1A with +1G in sibC and sibE would also block the NAD capping process, and whether changing the +1G to +1A in sibA and sibB would lead to the production of NAD-capped RNAs, we cloned all sib genes into the pWZ19 vector and introduced the point mutations at their TSSs, generating versions of all sib genes with either +1A or +1G at their TSSs. As shown in Fig. 3C, we did not detect the NAD cap in the + 1G versions of any Sib RNAs. In contrast, NAD-capped transcripts were detected in all + 1A versions with NADbio-streptavidin blotting. These results further confirm that + 1A is essential for NAD capping of all Sib RNAs in E. coli.
The minimal promoter of sibD is sufficient for NAD capping of SibD and other RNAs
The verification of the +1A position is crucial for NAD capping of all Sib RNAs in vivo (Fig. 3), indicating a strong coupling of NAD capping with transcription initiation. Therefore, cis elements in the promoter may play a significant role in modulating the NAD capping process. To identify these cis elements, pWZ19-sibD plasmid and its variants with a series of truncations in the sibD promoter regions were generated and transformed into the ΔsibD/ibsD strain (Fig. 4A). As shown in Fig. 4B, the sibD variant with the shortest promoter, sibD P-35, which contains the 35-bp minimal promoter, produced a similar amount of NAD-SibD RNAs as the longest full-length promoter (sibD Pfl) tested, suggesting the sufficiency of the minimal promoter for NAD capping.
The 35-bp sibD minimal promoter is sufficient to produce NAD-RNA from sibD variant genes. (A) Schematic of pWZ19 plasmid constructs for sibD, sibD +1G and multiple promoter truncation mutations. TSS indicates the transcription start site of sibD. (B) Detection of NAD caps in SibD RNA with NADbio-streptavidin blotting and APB gel coupled with northern blotting. Bands of biotinylated NAD-SibD and its variants (labelled as bio-SibD) and biotinylated NAD-RNA I (labelled as bio-RNA I) were marked with arrows in the HRP-streptavidin blotting panel. The band indicating NAD-SibD in the APB gel was marked as ‘capped’, whereas the band referring to uncapped SibD RNA was labelled as ‘uncapped’. Capping ratios were calculated based on the intensities of capped versus uncapped bands. (C) Schematic of pWZ19 plasmid constructs for several deletion designs within the sibD gene body, with the trpA terminator (highlighted in yellow) replacing the original sibD terminator in the deletion 4 construct. The positions of probe 1 and probe 2 for sibD are indicated. (D) Detection of NAD caps in strains harboring sibD gene body region deletions with NADbio-streptavidin blotting (top panel). Bands corresponding to biotinylated NAD-SibD and NAD capped chimeric SibD variants in the HRP-streptavidin blotting panel were indicated with red arrows. The biotinylated RNA I band, which is 108 nucleotides in size and labelled as ‘bio-RNA I’, serves as a size marker. Additionally, northern blotting was performed to detect the expression of sibD variants with two separate sibD probes (second and third panels).
To assess whether any cis elements exist within the sibD gene body that contribute to NAD capping, we created deletions in the gene body on the pWZ19-sibD plasmid. One of these deletions, termed deletion 4, involved replacing the sibD terminator with the trpA terminator to mediate transcription termination of the chimeric RNA [55–58] (Fig. 4C). None of these deletions significantly affected NAD capping, as confirmed by NADbio-streptavidin blotting, indicating that the sequence in the sibD gene body does not play a crucial role in modulating the NAD capping process (Fig. 4D). These results prompted us to test whether the sibD minimal promoter could drive NAD-RNA production from genes that don’t produce detectable levels of NAD-capped transcripts. We selected tRNA trpT and small RNA ryjA, both starting with A, but not identified to be NAD capped by NAD tagSeq II [29]. The sibD minimal promoter was used to drive expressions of trpT and ryjA gene bodies and cloned them into the pWZ19. The chimeric constructs both produced high levels of NAD-capped transcripts, revealed by both NADbio-streptavidin blotting and APB gel blotting (Fig. 5).
The 35-bp sibD minimal promoter can drive NAD capping of TrpT and RyjA RNAs. (A) Detection of NAD-RNAs from trpT and ryjA driven by its full-length and 35-bp minimal promoters of sibD. NAD caps in SibD, TrpT, and RyjA RNAs were detected with NADbio-streptavidin blotting analysis. The overall transcript levels of these three RNAs were examined with gene-specific probes by northern blotting. Pfl represents the full-length sibD promoter (189 bp), while P-35 denotes the 35-bp minimal promoter of sibD. EV refers to the ∆sibD/ibsD strain carrying an empty pWZ19 vector. gtrpT and gryjA indicate constructs that contain genomic DNA fragments of the trpT and ryjA genes contained their own promoters in the pWZ19 vector. (B) Detection of NAD caps on TrpT and RyjA RNAs with APB gel electrophoresis coupled with northern blotting. The capped and uncapped bands were marked with short lines. 5S rRNA was detected as loading control.
We also tested additional genes that normally do not produce detectable levels of NAD-capped transcript by placing them under the control of the sibD minimal promoter. These genes include those encoding 5S ribosomal RNA, small RNA SroC, and sfGFP mRNA. It was also found that the sibD minimal promoter drove strong NAD capping of these transcripts (Supplementary Fig. S6). Intriguingly, full-length sfGFP mRNAs were barely detected, with a high presence of smeared bands indicative of shorter NAD-capped RNA fragments, as shown by the NADbio-streptavidin blotting result (Supplementary Fig. S6A). Compared with the other genes, sfGFP is much longer. This raises a possibility that NAD-capped mRNAs could be pre-terminated during transcription, or they are processed into truncated fragments post-transcriptionally.
Several small RNA-encoding genes in the E. coli chromosome produces NAD-capped transcripts when driven by the sibD minimal promoter
To investigate whether the sibD minimal promoter can drive NAD capping of RNAs from genes in the chromosome that typically do not produce NAD-capped transcripts, CRISPR-Cas9 gene editing was used to place the sibD P-35 immediate upstream of each of four small RNA-encoding genes in the genome (Fig. 6A). For the ryjA and symR genes, the 35-bp DNA region immediately upstream of each gene’s TSS was replaced with the sibD P-35 promoter at their respective genomic loci (Fig. 6A). TrpT and SroC are normally transcribed as part of a long operon transcript, and mature RNAs are generated post-transcriptionally through RNA cleavage and trimming [59–62]. For these two genes, we also inserted the rrnB terminator immediate upstream of the sibD P-35 promoter (Fig. 6A) to ensure the gene is specifically transcribed from sibD P-35, instead of its own operon promoter.
Several RNAs transcribed by the sibD minimal promoter from E. coli chromosomal DNA could be NAD capped. (A) Schematic illustration of gene editing designs in the E. coli genome for four small RNAs expression driven by the sibD minimal promoter. The gene body of trpT is highlighted in purple, sroC in green, and ryjA and symR in blue. The sibD minimal promoter (sibD P-35) is labelled as a short red line, and the rrnB terminator is highlighted in yellow. (B) Detection of NAD caps in SibD, TrpT, RyjA, SroC, and SymR RNAs with NADbio-northern blotting in the wild-type and indicated mutant strains. ‘ADPRC+’ indicates the biotinylation of NAD-RNAs via the ADPRC-SPAAC reaction with sufficient ADPRC, while ‘ADPRC−’ denotes the ADPRC-SPAAC reaction without ADPRC. 5S RNAs were detected as loading controls. (C) Detection and quantification of NAD-RNAs from SibD, TrpT, RyjA, SroC, and SymR using APB gel blotting. Capping ratios were calculated based on the band intensity of the capped transcripts relative to the total transcripts (both capped and uncapped transcripts) in the APB gel.
NADbio-northern blotting was then first used to detect NAD caps in SibD, TrpT, SroC, RyjA, and SymR RNAs from both wild-type and mutant strains. As shown in Fig. 6B, NAD-capped SibD bands were present in all ADPRC+ lanes across all tested strains, confirming the NAD caping of SibD in these samples. In addition, NAD capping of all four other sRNAs were also observed in the ADPRC + lanes of the sibD P-35-driven mutants (Fig. 6B), verifying that the sibD minimal promoter can induce production of NAD-capped transcripts from these genes. By contrast, we did not detect NAD-capped transcripts from these four genes by NADbio-northern blotting in the WT strain (Fig. 6B).
Furthermore, APB gel blotting was performed to detect and quantify NAD capping of the four sRNAs driven by the sibD minimal promoter. As shown in Fig. 6C, ~20% of ryjA and sroC transcripts were NAD capped, while SymR RNA exhibited the highest NAD capping level among all four tested sRNAs, with 26% of transcripts carrying the NAD cap. The trpT gene, when expressed from the pWZ19 plasmid, was previously verified to generate a significant proportion of NAD-capped transcripts under the control of the sibD minimal promoter (Fig. 5). However, when trpT was expressed from its native genomic locus with the same promoter, the abundance of NAD-capped transcripts was too low to be detected by APB gel blotting (Fig. 6C), even though NADbio-northern blotting confirmed the presence of NAD-capped TrpT in this sibD minimal promoter-driven strain (Fig. 6B). These findings suggest that certain genomic sequences surrounding the trpT gene may influence its NAD capping efficiency.
The alarmone ppGpp and small protein DksA synergistically enhance levels of both NAD capped and uncapped Sib RNAs in vitro
Minimal promoters in E. coli are typically recognized by the sigma factor (σ)/RNA polymerase (RNAP) complex to initiate transcription. Certain sigma factors play crucial roles in regulating transcription to adapt to varying stresses, such as nutritional starvation during the stationary phase [63–66]. To determine which sigma factor might mediate transcription of sibD, we expressed and purified six sigma factors from E. coli and used them to initiate transcription of SibD RNA in vitro. We discovered that only the housekeeping sigma factor RpoD (σ^70^) could drive the production of both NAD capped and uncapped (5′ppp-) SibD RNAs (Supplementary Fig. S7), indicating that RpoD is the primary sigma factor for initiating transcription of sibD, likely also for NAD capping of SibD RNA.
Subsequently, to determine whether specific nucleotides in the 35-bp sibD minimal promoter critically contribute to its high NAD capping levels, we generated a series of mutations within this promoter region (Supplementary Fig. S8A). While most mutations showed no significant effect on either overall transcription or NAD capping efficiency, two particular mutations, namely mP12 (G_−2_A_−1_ → C_−2_T_−1_) and mP13 (C_+2_A_+3_ → G_+2_T_+3_), significantly reduced NAD capping of SibD (Supplementary Fig. S8B). This finding is consistent with previous reports that sequences flanking the TSS influence NAD incorporation efficiency in NAD-RNA [27, 28].
The discovery that the sibD minimal promoter influences NAD capping implies that certain factors binding to the σ/RNAP complex may mediate NAD capping by altering the complex’s conformation, potentially affecting the selection of NAD as the initiation nucleotide. We hypothesized that certain factors that bind to the σ/RNAP complex during transcription initiation could play a role in mediating NAD capping. To test this hypothesis, we focus on the alarmone ppGpp and the small protein DksA. Both molecules are known to bind to RNAP complexes to affect transcription in response to stress conditions [37, 67–73].
We first used an IVT system to test whether the presence of either factor could alter NAD capping of SibD RNA. When ppGpp was added to the transcription mix at varying concentrations in the presence of NAD, it increased the levels of both NAD-capped and uncapped (5′ppp-) SibD RNAs (Fig. 7A and C; Supplementary Fig. S9). Similarly, lower concentrations of DksA (50–200 nM) significantly enhanced the production of both RNA forms (Fig. 7A and C). Although a slight reduction in SibD RNA levels was observed at the highest DksA concentration (500 nM), the production of NAD-capped SibD RNA still increased compared to samples without DksA (Fig. 7A and C; Supplementary Fig. S9).
The transcriptional production of NAD-SibD was synergistically enhanced by ppGpp and DksA in vitro. (A) Detection of NAD capping in SibD RNA from IVT products with ppGpp or DksA at concentrations ranging from 0 to 2000 μM (for ppGpp) and from 0 to 500 nM (for DksA). ‘w/o NAD’ indicates transcription mix without NAD, ‘+NAD’ refers to transcription mix containing NAD. Synthetic NAD-SibD and ppp-SibD RNA produced by T7 RNA polymerase were included as controls. (B) NAD capping of SibD was synergistically enhanced by ppGpp and DksA in vitro when both were present. DksA was added to the transcription mix to 50 nM, followed by ppGpp at varying concentrations (0–2000 μM). Alternatively, 200 μM ppGpp was added first in the transcription mix, followed by DksA at varying concentrations (0–200 nM). ‘w/o NAD’ indicates transcription mix without NAD, ‘+NAD’ refers to transcription mix containing NAD. Synthetic NAD-SibD and ppp-SibD RNA produced by T7 RNA polymerase were included as controls. (C, D) Relative transcription yields of NAD-SibD and ppp-SibD RNAs are plotted based on the band intensities in the ‘+NAD’ samples from the APB gel in panels (A) and (B), respectively. The bands were analyzed with ImageJ software and were normalized against the intensity of the ppp-SibD band in the sample lacking both ppGpp and DksA.
Given that both DksA and ppGpp are known to bind to the different sites of E. coli RNAP [39, 74, 75], we investigated their combined effect on the NAD capping of SibD RNA with IVT assay. As shown in Fig. 7B and D; Supplementary Fig. S10, in the presence of 50 nM DksA, the levels of both NAD-capped and uncapped (5′ppp-) SibD RNAs increased significantly upon the addition of 200 μM ppGpp. Subsequently, the levels of both RNA forms continued to rise at a relatively slower rate, ultimately reaching approximately eightfold higher levels compared to the sample lacking both ppGpp and DksA. A similar trend was observed when ppGpp was held at 200 μM and the concentration of DksA increased (Fig. 7B and D; Supplementary Fig. S10). Furthermore, the co-presence of ppGpp and DksA was also observed to significantly increase the production of NAD-capped SibC and SibE RNAs (Supplementary Fig. S11). These results indicate that ppGpp and DksA synergistically enhance the transcriptional production of NAD-capped Sib RNAs in vitro.
Although production of NAD-capped SibD RNA was significantly increased in the presence of ppGpp alone or together with DksA protein, the NAD capping ratio of SibD RNA did not significantly change with increasing ppGpp, as the levels of uncapped SibD RNA also increased to a similar extent. Notably, a significant increase in the NAD capping ratio was observed only when 500 nM DksA was added to the transcription mix, compared to samples not treated with ppGpp or DksA (Fig. 7A and Supplementary Fig. S12).
(p)ppGpp (ppGpp and pppGpp) is primarily produced by the enzyme RelA, and pppGpp shares most of the same targets and acts through mechanisms similar to ppGpp in E. coli [74, 76–79]. To determine whether pppGpp has the same effect on NAD capping of SibD as ppGpp, we performed IVT assays. As shown in Figs. 8A and B, pppGpp significantly enhanced the transcription levels of both 5′-ppp-linked SibD and NAD-capped RNAs to the extent with ppGpp in vitro. When DksA was added together in the transcription mix, a greater increase in SibD RNA production was observed with ppGpp compared to pppGpp (Fig. 8B and D). In addition, DksA and (p)ppGpp were previously discovered to play important roles in regulating RNA polymerase promoter selection and impacting the transcription initiation process [35, 38, 80]. To examine whether both factors would alter the TSS selection during transcription initiation, we identified the TSS of SibD RNAs from transcription products generated in the presence of ppGpp, pppGpp, DksA, and co-presence of ppGpp and DksA. We found that the TSS of SibD did not change upon incubation with these components compared to SibD RNAs from untreated samples (Supplementary Fig. S13).
pppGpp enhances the transcriptional production of both NAD-capped and uncapped RNA from the sibD P-35 templates, both alone and in combination with DksA. (A) IVT of 5′ ppp-SibD RNA was enhanced by (p)ppGpp and DksA alone and together with DksA. The sibD P-35-sibD DNA fragment was used as a template for transcription assay, with ppGpp or pppGpp added at varying concentrations (0–2000 μM). Alternatively, DksA was added to the transcription mix at 50 nM, followed by ppGpp or pppGpp at the same range of concentrations. SibD RNA from IVT was detected by northern blotting. Synthetic NAD-SibD and ppp-SibD RNAs produced by T7 RNA polymerase were included as controls. (B) Production of both NAD-capped and uncapped SibD in vitro was enhanced by (p)ppGpp and DksA alone and synergistically. The sibD P-35-sibD DNA fragment was used as a template to produce NAD-capped SibD RNA in vitro, and NAD-capped SibD was detected by APB gel blotting. (C) IVT from the sibD P-35-symR DNA template was enhanced by ppGpp and DksA alone and synergistically. NAD-capped SymR was detected by APB gel blotting. ‘+NAD’ indicates that NAD was added to the IVT reaction mixture, while ‘w/o NAD’ denotes that NAD was omitted. (D) Relative transcriptional production levels of NAD-SibD and ppp-SibD were plotted based on the band intensities in the normal gel shown in panel (B). Band intensities were analyzed with ImageJ software and normalized against the intensity of the SibD band in the sample lacking both ppGpp and DksA. (E) Relative transcriptional production levels of NAD-SymR and ppp-SymR RNAs are plotted based on the band intensities in the ‘+NAD’ samples from the APB gel shown in panel (C). The bands were analyzed using ImageJ software and normalized against the intensity of the ppp-SymR band in the sample lacking both ppGpp and DksA.
Furthermore, to determine whether the enhancement in NAD capping levels by ppGpp, pppGpp, and DksA in vitro depends on the sibD minimal promoter, we introduced the sibD P-35-symR chimeric gene as a template in the IVT assays. The SymR RNA transcribed from the sibD P-35-symR locus in the E. coli genome was previously verified to carry a high proportion of NAD caps (Fig. 6). As shown in Fig. 8C and E, both 5′-ppp-SymR and 5′-NAD-SymR transcript levels significantly increased in the presence of either ppGpp, pppGpp, or DksA, with an even more extensive enhancement observed when both ppGpp and DksA were present in the transcription mix. These observations align well with findings from the sibD P-35-sibD template, verifying that (p)ppGpp and DksA enhance production of SibD and SymR RNAs in a manner dependent on the sibD minimal promoter.
Absence of the alarmone (p)ppGpp reduces NAD capping of the sib family genes in E. coli
To determine whether (p)ppGpp has any effect on NAD capping of the NAD-RNA-encoding genes sibC, sibD, and sibE in vivo, we first transiently overexpressed a truncated active form of the ppGpp synthetase protein RelA with the pBAD expression system. This truncated form comprises the N-terminal 455 amino acid residues of RelA (referred to as RelA 455aa) and was found to increase (p)ppGpp levels when transiently induced in wild-type E. coli strains [81, 82]. A similar approach was used to overexpress the DksA protein in vivo. Western blotting results showed that both proteins were successfully induced to moderate levels within 1 h of induction by arabinose (Supplementary Fig. S14A). APB gel blotting was then employed to detect changes in the NAD capping ratio of SibD RNA before and after arabinose induction. As shown in Fig. 9A and Supplementary Fig. S14B, overexpression of RelA 455aa significantly elevated the total transcription level of SibD, resulting in an ~2.5-fold increase compared to the uninduced sample. The level of NAD-capped transcripts increased more as the NAD capping ratio increased from 9% to 27%. Overexpression of DksA did not significantly affect the total abundance of SibD RNA; however, the NAD capping ratio of SibD increased from 7% to 20%. Collectively, these results indicate that (p)ppGpp and DksA can enhance the NAD capping of SibD RNA in vivo.
Effects of (p)ppGpp and DksA on transcription and NAD capping of certain small RNAs in E. coli cells. (A) The NAD capping level of SibD increased upon transient induction of RelA 455aa and DksA. Both RelA 455aa and DksA were expressed from plasmids under the control of the pBAD promoter. NAD capping of SibD was assessed by APB gel blotting. The total level of SibD RNA in each lane was quantified from the normal gel using ImageJ software and normalized to the intensity in the first EV lane. The NAD capping ratio was calculated as the percentage of the intensity of the NAD-capped band relative to the sum of the intensities of both the capped and uncapped bands in the APB gel. ‘Arabinose−’ indicates RNA samples without arabinose induction, while ‘Arabinose+’ signifies that arabinose was added to induce the expression of RelA 455aa and DksA. ‘EV’ indicates strain carrying the empty pBAD33.1 vector. The tmRNA was used as a loading control and each blotting has three independent replicates. (B) Detection of NAD-capped transcripts of five sRNAs with NADbio-northern blotting analysis, including four known NAD-RNAs: SibC, SibD, SibE, and GcvB. The tmRNA was used as a loading control. (C–G) Detection of total transcripts and NAD-capped transcripts of five sRNAs, namely SibA, SibC, SibD, SibE, and GcvB, respectively. The total abundance of individual RNA was determined by electrophoresis on a standard PAGE gel followed by northern blotting (labelled as normal gel), while the NAD-capped transcripts were identified with APB gel blotting (labelled as APB gel). The non-NAD-RNA SibA was included as a negative control. The NAD capping ratio was calculated based on the band intensity of the NAD-capped version relative to the total transcription levels (NAD-capped version plus uncapped version). Two types of synthetic RNAs for each sRNA, namely with 5′-ppp- and 5′-NAD modifications, were used as controls.
To further determine the roles of (p)ppGpp and DksA in NAD capping, we generated a ppGpp^0^ mutant by deleting the (p)ppGpp synthetic genes relA and spoT, thereby blocking (p)ppGpp biosynthesis, and also a ∆dksA deletion mutant. Both ppGpp and DksA are known to be essential for E. coli growth in minimal medium; however, certain RNAP mutations were reported to bypass this phenotype by mimicking the effect of ppGpp binding [68, 83]. Therefore, we assessed the growth phenotypes of these mutants on M9 minimal medium plates. As expected, both the ppGpp⁰ and ∆dksA mutants displayed growth comparable to the wild-type strain on LB rich medium but were unable to form colonies on M9 plates. In contrast, the ∆relA mutant exhibited a growth rate similar to the wild-type strain on both medium plates (Supplementary Fig. S15). These results confirmed that the ppGpp^0^ and ∆dksA mutants generated in our study contained the expected deletions and did not acquire unforeseen compensatory mutations in the genome.
Subsequently, we employed NADbio-northern blotting to examine and compare the NAD capping of these sRNAs in the ppGpp^0^ mutant, the ∆dksA mutant, and wild-type strains. As shown in Fig. 9B, NAD-capped SibC and SibE RNAs were clearly detected in the wild-type strain but were barely detectable in the ppGpp^0^ mutant. Moreover, a significant reduction in NAD-SibD abundance was also observed in the ppGpp^0^ strain (Fig. 9B). Additionally, APB gel blotting analysis revealed that the NAD capping ratios of SibC, SibD, and SibE RNAs were also significantly reduced in the ppGpp^0^ mutant compared to the wild-type (Fig. 9D–F and Supplementary Fig. S16). For all tested sib genes, including sibA, which does not encode NAD-capped transcripts, expression levels were generally decreased in the absence of intracellular (p)ppGpp (Fig. 9C–F and Supplementary Fig. S16), suggesting that (p)ppGpp may act as a positive regulator of their transcription in vivo. In contrast, both the total transcript levels and NAD-capped transcripts of gcvB, another known NAD-RNA-producing gene in E. coli, were significantly increased in both the ppGpp^0^ and ∆dksA mutants compared to the wild-type strain, and its NAD capping ratio was also elevated in the ppGpp^0^ mutant (Fig. 9G). These results suggest that the effects of (p)ppGpp and dksA on NAD capping vary among different genes.
We then used an sfGFP reporter to examine the transcription strengths of these sRNA promoters. As shown in Fig. 10, sfGFP signal intensities significantly decreased in the ppGpp^0^ strain for the sibC and sibD promoters, suggesting that (p)ppGpp is essential for maintaining relatively high transcription levels of both RNAs in vivo. In contrast, a significant increase in sfGFP intensity was observed for the gcvB promoter when cells lacked (p)ppGpp, indicating that (p)ppGpp suppresses gcvB expression in vivo (Fig. 10B and C). These results verify that promoters of these sRNAs are direct targets of (p)ppGpp in vivo and suggest that (p)ppGpp modulates the cellular levels of Sib RNAs and GcvB by altering the transcription strengths of their promoters, rather than through post-transcriptional regulation.
Transcription strengths of several NAD-RNA-derived promoters were regulated by (p)ppGpp in vivo. (A) The schematic illustration of sfGFP reporter system for comparing transcription strengths of certain sRNAs between the ppGpp0 mutant and the wild-type strain. (B) Comparison of the transcription strengths of the promoters of sibC, sibD, and gcvB between the ppGpp0 mutant and the wild-type strain on LB plates. Each sfGFP construct was transformed into the wild type and ppGpp0 mutant, respectively. Overnight cultures were then initially diluted to an OD600 of 0.05, followed by further serial dilutions. All dilutions were dotted onto LB plates. Photographs of both bright-field and GFP fluorescence were taken at regular intervals. (C) sfGFP signal intensities from panel (B) were quantified and presented as a bar chart. Statistical analysis of the three replicates of sfGFP signal intensities was performed using Duncan’s multiple range test, and letters indicating significant differences were labelled above each bar.
Discussion
Recently, several nucleotide analogues, including NAD, NADH, dpCoA, and FAD, which function as cofactors in enzymatic reactions, have been discovered to modify RNAs as noncanonical cap structures at their 5′ ends across prokaryotic and eukaryotic species [4, 13, 18, 84]. Although NAD-RNA was discovered over a decade ago, their biological functions and the mechanism that control NAD capping remain largely unknown.
In this study, we developed various in vivo and in vitro methods to analyze the NAD capping process in E. coli. Using sibD as a model, we have dissected the cis elements that might contribute to NAD capping and found that the 35-bp minimal promoter of sibD is sufficient in mediating NAD capping of SibD RNA. Additionally, when the sibD minimal promoter was used to drive expression of several genes that typically do not produce NAD-capped transcripts, these genes were able to produce NAD-RNAs. These results highlight the importance of the promoter elements in NAD capping and strongly support the notion that the NAD-capped RNAs are formed during the transcription initiation.
It is known that the NAD cap can be formed when RNA polymerase uses NAD, in place of ATP, to initiate transcription. As expected, when the TSS was changed from +1A to +1G, NAD capping of SibD and other RNAs was abolished. On the other hand, replacing +1G with +1A in sibA and sibB genes introduced NAD capping to their RNAs. These results further confirm that NAD capping occurs during transcription initiation. However, one might argue that NAD cap may be formed post-transcriptionally by certain enzyme when the first transcribed nucleotide is A, but not G. This hypothesis may be supported by the findings that 3′-dephospho-CoA (dpCoA) capping can occur through post-transcriptional modification by an enzyme involved in CoA biosynthesis [85]. However, no enzyme in the NAD biosynthesis pathway appears capable of mediating NAD capping in E. coli [86]. Furthermore, RNAs like TrpT and RyjA, which typically are not NAD-capped despite starting with A, become highly NAD capped when driven by the sibD minimal promoter. This result argues against post-transcriptional formation of the NAD cap, at least for these transcripts, and further indicates that the NAD cap is added during transcriptional initiation in E. coli.
We have previously found that growth conditions can have profound influence on the ratio of NAD-RNA to total transcripts of sibD and many other genes [29], indicating that in addition to *cis-*elements, trans-acting factors play key roles in mediating NAD capping. In this study, we found that the small molecules ppGpp, pppGpp, and protein DksA, each known to bind to the bacterial RNAP complex, enhanced the production of NAD-capped Sib RNAs during IVT. However, the total transcript levels from these genes were also similarly increased in the presence of these factors, making it difficult to determine whether (p)ppGpp or DksA specifically enhances NAD capping, or whether the increased abundance of NAD-capped transcripts was simply a consequence of overall transcriptional upregulation. To further investigate their potential roles in NAD capping in vivo, we transiently expressed the (p)ppGpp synthetase and DksA, and observed an increase in the NAD capping ratio of SibD. Moreover, we generated the ppGpp^0^ mutant strain, which lacks (p)ppGpp synthesis. In this mutant, the ratio of NAD-capped version transcripts versus overall transcript levels from sibC, sibD, and sibE were more significantly reduced, indicating that ppGpp^0^ plays a positive role in mediating NAD capping of these genes. On the other hand, deletion of dksA had no significant effect on NAD capping of these genes. While our in vitro data indicate that (p)ppGpp primarily increases total SibD RNA production without affecting the NAD capping ratio, multiple lines of in vivo evidence demonstrate that both (p)ppGpp and DksA enhance NAD capping. This apparent discrepancy suggests that an additional, yet unidentified factor may collaborate with (p)ppGpp and DksA to promote NAD capping in vivo. As a result, in vitro, where this factor is absent, only transcriptional yield is increased without a corresponding enhancement in NAD capping efficiency.
The alarmone (p)ppGpp drastically accumulates under nutrient deficiency, stationary phase, and other stresses, mediating the ‘stringent response’ that orchestrates transcriptional reprogramming as well as other metabolic processes [67, 73, 87, 88]. (p)ppGpp regulates transcription initiation by binding to RNAP with the assistance of DksA, thereby enhancing the expression of stress-responsive genes and down-regulating many housekeeping genes [39, 68, 70, 71, 89]. The reduced production of NAD-capped transcripts of these sib genes in the E. coli (p)ppGpp^0^ mutant is consistent with our previous finding that E. coli cells produced significantly more NAD-capped transcripts and higher NAD capping ratios from these genes during the stationary phase than in the exponential phase [29]. In contrast, (p)ppGpp and DksA appear to be negative regulators for NAD capping of gcvB, as we found that NAD capping of the gcvB gene was enhanced in the (p)ppGpp^0^ mutant and ∆dksA mutant compared to the wild-type strain, suggesting that the effects of ppGpp and dksA on NAD capping vary among different genes.
In summary, our study identified the *cis-*elements and trans-acting factors that mediate NAD capping of the sib family genes. These findings provide insights into how growth conditions and other environmental factors may influence NAD capping, likely by affecting the RNA polymerase complex and modulating transcriptional initiation. A similar mechanism may also mediate NAD capping of other genes, although the specific *cis-*elements and trans-acting factors involved may differ. The biological consequences of NAD capping regulation on specific RNAs remain to be elucidated in future studies.
Supplementary Material
gkag102_Supplemental_File
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Meyer KD, Jaffrey SR. The dynamic epitranscriptome: n 6-methyladenosine and gene expression control. Nat Rev Mol Cell Biol. 2014;15:313–26. 10.1038/nrm 3785.24713629 PMC 4393108 · doi ↗ · pubmed ↗
- 2Kapp LD, Lorsch JR. The molecular mechanics of eukaryotic translation. Annu Rev Biochem. 2004;73:657–704. 10.1146/annurev.biochem.73.030403.080419.15189156 · doi ↗ · pubmed ↗
- 3Ramanathan A, Robb GB, Chan S-H. m RNA capping: biological functions and applications. Nucleic Acids Res. 2016;44:7511–26. 10.1093/nar/gkw 551.27317694 PMC 5027499 · doi ↗ · pubmed ↗
- 4Chen YG, Kowtoniuk WE, Agarwal I et al. LC/MS analysis of cellular RNA reveals NAD-linked RNA. Nat Chem Biol. 2009;5:879–81. 10.1038/nchembio.235.19820715 PMC 2842606 · doi ↗ · pubmed ↗
- 5Gomes-Filho JV, Breuer R, Morales-Filloy HG et al. Identification of NAD-RNA species and ADPR-RNA decapping in Archaea. Nat Commun. 2023;14:7597. 10.1038/s 41467-023-43377-x.37989750 PMC 10663502 · doi ↗ · pubmed ↗
- 6Walters RW, Matheny T, Mizoue LS et al. Identification of NAD+ capped m RN As in Saccharomyces cerevisiae. Proc Natl Acad Sci USA. 2017;114:480–5.28031484 10.1073/pnas.1619369114 PMC 5255579 · doi ↗ · pubmed ↗
- 7Zhang Y, Kuster D, Schmidt T et al. Extensive 5′-surveillance guards against non-canonical NAD-caps of nuclear m RN As in yeast. Nat Commun. 2020;11:5508.33139726 10.1038/s 41467-020-19326-3PMC 7606564 · doi ↗ · pubmed ↗
- 8Wang Y, Li S, Zhao Y et al. NAD+-capped RN As are widespread in the Arabidopsis transcriptome and can probably be translated. Proc Natl Acad Sci USA. 2019;116:12094–102.31142655 10.1073/pnas.1903682116 PMC 6575598 · doi ↗ · pubmed ↗