Internal transcription termination widely regulates differential expression of operon-organized genes including ribosomal protein and RNA polymerase genes in an archaeon

Abstract Genes organized within operons in prokaryotes benefit from coordinated expression. However, within many operons, genes are expressed at different levels, and the mechanisms for this remain obscure. By integrating PacBio-seq, dRNA-seq, Term-seq and Illumina-seq data of a representative archaeon Methanococcus maripaludis, internal transcription termination sites (ioTTSs) were identified within 38% of operons. Higher transcript and protein abundances were found for genes upstream than downstream of ioTTSs. For representative operons, these differences were confirmed by northern blotting, qRT-PCR and western blotting, demonstrating that these ioTTS terminations were functional. Of special interest, mutation of ioTTSs in ribosomal protein (RP)-RNA polymerase (RNAP) operons not only elevated expression of the downstream RNAP genes but also decreased production of the assembled RNAP complex, slowed whole cell transcription and translation, and inhibited growth. Overexpression of the RNAP subunits with a shuttle vector generated the similar physiological effects. Therefore, ioTTS termination is a general and physiologically significant regulatory mechanism of the operon gene expression. Because the RP-RNAP operons are found to be widely distributed in archaeal species, this regulatory mechanism could be commonly employed in archaea.


INTRODUCTION
The pioneering study of Jacob and Monod in 1961 generated the operon hypothesis (1)(2)(3), which states that genes involved in related functions are generally organized within operons and co-transcribed, providing an efficient means to synchronously coregulate transcription of multiple genes and produce equal amounts of subunits for assembling complex es ( 2 , 4-7 ). The pr esence of operons also distinguishes the genomic organization of pr okaryotes fr om that of eukaryotes. While differential pr otein pr oduction is commonly found for genes within operons encoding multisubunit complexes, such as the ribosome, secretion machinery, ATP synthase, and antiviral defense systems (8)(9)(10)(11), the mechanisms involved are not well documented.
Selecti v e mRNA processing was found to regulate the differential production of the cellulase subunits in the Clostridium cellulolyticum operon ( 12 ) and ribosome subunits in an archaeal operon ( 13 ). Based on informatic analyses of bacterial genomes, transcriptomes and ribosomal profiling da ta, dif ferential transla tion ef ficacy depending mainl y on mRN A secondary structure and codon usage was presumed to be a key determinant in differential protein pr oduction fr om genes within bacterial oper ons to maintain appropriate stoichiometries of proteins r equir ed for formation of complexes (14)(15)(16)(17)(18).
Using SMRT-Cappable-seq, the full-length transcriptomic map of Escherichia coli was recently obtained. It not onl y accuratel y defined genome-wide transcription units (TUs) and operons, but also re v ealed e xtensi v e transcription read-through at over 40% of the transcription termination sites (TTSs), which especially complicated the operon transcription profiles ( 19 ). Extensi v e transcription read-through was also more recently reported in se v eral other bacteria ( 16 , 20 ), indicating the complexity of the transcriptional landscape in bacteria. In previous study, using Term-seq to sequence the genome-wide TTSs, we found that the majority of operons in the methanogenic archaeon Methanococous maripaludis , one r epr esentati v e of the thir d life form Archaea, exhibited incomplete termination at the TTSs. Depending on the terminator strength, only 30-80% terminations were observed for the majority of TTSs. Most terminations depended on the terminator strength ( 21 ) and also a ne wly discov ered transcription termination factor aCPSF1. As this protein was essential, depletion of the cellular le v els of aCPSF1 caused a genome-wide reduction in transcription termination, se v ere growth inhibition, and generation of a chaotic transcriptome ( 21 , 22 ). TTSs were also observed in the intergenic regions (IGRs) within some putati v e operons in the Term-seq map, and Illuminaseq data detected different transcription le v els of the genes within operons. This suggested that transcription termination might occur within many operons and could exert a regulatory role in fine-tuning expressions of the operon genes.
This study extends the above observations. The combination of PacBio-seq, dRNA-seq, Term-seq and Illumina-seq of the M. maripaludis transcriptome thoroughly deciphered the operon organization and surprisingly found that the internal operon intergenic terminators, here named as ioTTS terminators, are widely distributed. The ioTTS terminations were found to be functional, causing reduced transcription of down-stream genes and reduced production of the encoded proteins. Mutation of r epr esentati v e ioTTS terminators increased le v els of the transcripts and encoded proteins of the downstream genes. Furthermore, these mutations often se v erely inhibited growth, demonstrating the importance of ioTTS terminations in overall cellular fitness. Of special interest, mutation of the ioTTS terminators in ribosomal protein-RN A pol ymerase (RP-RN AP) operons also decreased the cellular RNAP complex abundances and the transcription and translation velocities, demonstrating a key role in controlling production of the RNAP complex.
Gi v en that similar RP-RNAP operons are ubiquitous in archaea, this work suggests tha t ioTTS termina tion could be a novel and potentially ancient mechanism for tuning differ ential expr ession of the operon genes in ar chaea.

Strains, plasmids and culture conditions
Strains and plasmids used in this study are listed in Supplementary Table S1. M. maripaludis and its derivatives were routinely cultured in anaerobic McF medium under a gas phase of N 2 / CO 2 (80:20) as previously described ( 23 ). As aCPSF1 was depleted to a low abundance (20% compared to the wild type) at 22 • C and much lower than the abundance at 37 • C ( ∼50%), the transcriptomic analysis and the experiments with the aCPSF1 depletion strain were all performed at 22 • C unless specially indicated otherwise. These were the same conditions used for the previously collected Term-seq, differential RNA-seq (dRNAseq) and Illumina-seq data ( 21 , 22 ). For solid medium, 1.5% agar was added. For Fe-limited growth, the concentrations of Fe (II) were adjusted by varying the le v el of ammonium iron(II)sulfate in McF medium. Growth was monitored by determining the optical density at 600 nm (OD 600 ). Unless indicated otherwise, 2.5 g ml −1 puromycin was added for genetic selections and to maintain recombinant plasmids in M. maripaludis . E. coli strains were routinely grown in LB Miller broth at 37 • C with shaking at 200 rpm. When needed, final concentrations of antibiotics and supplements used were ampicillin at 100 g ml −1 , kanamycin at 50 g ml −1 , isopropylthiogalactoside (IPTG) at 0.1 mM.

PacBio sequencing of the transcriptome
Long reads of the re v erse transcribed M . maripaludis mR-NAs were sequenced on the PacBio platform as described previousl y ( 19 ). Briefly, total RN A was extracted from mid-exponential M. maripaludis cultures using TRIzol TM reagent (Invitrogen) as described previously ( 13 , 22 ). The extracted RNA had an RNA integrity number (RIN) determined by Bioanalyzer (Agilent) ≥9.0. Then 5 g of the extracted RNA was capped through a capping reaction, and then a poly-A tail was added using E. coli P oly(A) P olymerase (New England Biolabs). rRNA was depleted using Ribo-Zero ™ rRNA Removal Kit (Epicentre). The capped and tailed RNA was purified using AM-Pure beads and eluted in TE buffer, and the full-length primary transcripts were further enriched using streptavidin magnetic beads (New England Biolabs). Polyadenylation (A-tailing) ensured priming by the poly dT primer at the 3 end of transcript and re v erse transcription of the first cDN A strand. Pol yG was then added to the cDNA 3 end for second-strand synthesis using Terminal Tr ansfer ase (TdT, New England Biolabs). Unenriched fulllength transcripts pr epar ed without the capping r eaction and streptavidin enrichment were used to construct a control library for verifying the positions and abundances of transcription start sites (TSSs) and transcription termination sites (TTSs) of operons. Library construction and P acBio-sequencing wer e performed at Frasergen Bioinformatics Technology Co., Ltd (Wuhan, China). As described previously ( 19 ), the PacBio-sequencing reads were filter ed to r emove chimeras and trimmed for 3 end polyA and 5 end polyC prior to mapping on the M. maripaludis genome.

Operon identification by bioinformatic analysis of multiple transcriptomic data
Operons were identified based on the mapping of the P acBio r eads in combina tion with da ta of the TSSs and TTSs that were previous identified at single-base resolution via dRNA-seq and Term-seq ( 22 ), respecti v ely. Transcription units (TU) containing at least a TSS at the 5 end, a TTS at the 3 end, and a continuous coverage by a P acBio r ead for the full length of the RNA was defined as an operon. In addition, a downstream gene encoded on the same strand must possess its own TSS. The transcript boundaries and abundances were then verified by data from the previous Illumina-sequenced transcriptome ( 22 ). TSSs and TTSs identified in the intergenic regions within an operon were named the internal operon TSSs (ioTSSs) and internal operon TTSs (ioTTSs), respecti v ely. Using WebLogo v2.8.2, the terminator motifs preceding the ioTTSs were analyzed from 36 nt upstream to 1 nt downstream ( 24 ) (Dataset S1). The ioTTS-based differ ential expr ession ratio (TDER) of operon genes was calculated from the ratio of the transcript abundances of the ioTTS-upstream genes divided by the abundances of the downstream genes. TDERs of > 2, < 2 & > 1.5 and < 1.5 were defined as having high-, mid-, and low significance, respecti v ely. Transcription termina tion ef ficacy (TTE) of an ioTTS terminator was defined as 1-TDER −1 in the M. maripaludis wild-type transcriptome. aCPSF1 dependency of TDER is (TDER of the aCPSF1 depletion strain aCPSF1 ) / (TDER of the wildtype strain) (Dataset S2).

Northern blot assay
Northern blot assays were performed as previously described ( 22 ). Briefly, total RNA was extracted from the midexponential cells using TRIzol ™ reagent (Invitrogen). After quantification using a Nano photometer spectrophotometer (Implen), RNA was dena tured a t 65 • C for 10 min in loading buffer containing 95% (v / v) formamide, and 5-10 g was loaded in each lane of a 6% polyacrylamide gel with 7.6 M ur ea. Electrophor esis was performed in 1 × TBE buffer. A single-stranded RNA (ssRNA) ladder (New England Biolabs) served as a size marker. After separation, RNAs wer e transferr ed onto Hybond-N + membranes (GE Healthcare) by electroblotting and crosslinked to the membrane using UV. Next, membranes wer e pr ehybridized at 42 • C in prehybridized buffer (5 × SSC, 5 × Denhardt s, 50% v / v formamide deionized, 0.5% w / v SDS, 200 g ml −1 pr e-denatur ed Salmon sperm DNA)for 4 h, followed by hybridization for 12 h with 2-10 pmol of biotin-labeled DNA probes listed in Supplementary Table S4. After three rounds of washing for 10 min in 1 ×, 0.2 × and 0.1 × SSC-0.1% SDS solutions, band signals were visualized using a chemiluminescent nucleic acid detection module (Thermo Scientific) according to the manufacturer's protocol.

Quantification of transcript abundances by quantitative RT-PCR (qRT-PCR)
Total RNA was extracted from the mid-exponential cells as described above, and 500 ng of RNA was used to generate complementary DNAs using Re v erTra Ace ® qPCR RT Master Mix with gDNA Remover (Toyobo). Quantitati v e PCR amplifications were performed with Mastercycler eprealplex2 (Eppendorf AG, Hamburg, Germany) and a parallel amplification without RT reaction was used as a control to assess the absence of DNA in the RNA samples for each gene (Supplementary Figure S1). The standar d curv e f or each gene was generated using 10-f old serially diluted PCR products as template. The primers used are listed in Supplementary Table S3. Transcript abundances were then normalized by comparison to the abundance of the 16S rRNA. All assays were performed on triplicate samples and repeated at least three times.

Recombinant protein purification
To obtain the recombinant M. maripaludis proteins in E. coli , the respecti v e open reading frames (ORFs) for the Histagged RpoP, Rpl37Ae, Rpl21e and AfuA and GST-tagged RpoF and AfuC proteins were cloned into the expression vector pET28a or pGEX 4T-1 via stepwise Gibson assembly using ClonExpress MultiS One Step Cloning Kit (Vazyme). The expression plasmids were then transformed into E. coli BL21(DE3)pLysS, which was cultured at 37 • C in LB broth containing 50 g ml −1 kanamycin or 100 g ml −1 ampicilin. When the OD 600 reached 0.6-0.8, 0.1 mM isopropylb-D -thiogalactoside (IPTG) was added. After 16 h at 22 • C, cells were harvested by centrifugation and stored at -80 • C. The harvested cells wer e r esuspended in binding buffer (for his-tagged proteins: 0.5 M NaCl, 20 mM imidazole, 5% v / v glycerol, 20 mM HEPES, pH 7.5; for GST-tagged proteins: 0.5 M NaCl, 26 mM Tris-base, 0.97 mM DTT, 5% v / v glycerol, pH 7.5) and lysed by sonication, and the supernatant was collected by centrifugation. The his-tagged or GST-tagged recombinant proteins were purified from the supernatant by passage through a His-Trap HP column or a GSTrap HP column (GE Healthcare) according to the manufacturer's protocol ( 25 ). Purified proteins were analyzed by SDS-PAGE, and the protein concentration was determined using a BCA protein assay kit (Thermo Scientific).

Western blot assays
Western blots were performed to determine the cellular protein abundances in the wild-type and genetically modified strains of M. maripaludis . The polyclonal r abbit antiser a against the purified proteins were generated by MBL Interna tional Corpora tion. Mid-exponential cells of M. maripaludis were harvested and resuspended in Lysis Buffer [50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 10 (w / v) glycerol, 0.05% (v / v) NP-40 detergent] and lysed by sonication. The cell lysate was centrifuged at 14,000 g for 15 min a t 4 • C , and proteins in the supernatant were separated on 15% SDS-PAGE and transferred to a nitrocellulose membrane. The antisera to the proteins were diluted 1:5000. A horseradish peroxidase (HRP)-linked secondary conjuga te a t a 1:5000 dilution was used for the imm unoreaction. Imm une-acti v e bands were visualized by an Amersham ECL Prime western blot detection reagent (GE Healthcare). For quantifying the cellular content of a protein, cell-free extract of M. maripaludis a t indica ted amounts was electrophoresed on SDS-PAGE synchronously with loading the corresponding purified recombinant proteins from E. coli as r efer ences. Density of each protein band was quantified from photo gra phs using ImageJ software, and the cellular abundance of a protein was calculated by comparison to the respecti v e recombinant protein.

Construction of strains with mutations in the ioTTS terminators
The M. maripaludis strains with ioTTS terminator mutations were constructed using the recently de v eloped CRISPR-Cas9 genome-editing toolkit ( 26 ). Based on pMEV4-Cas9 that carries the Streptococcus pyogenes Cas9 and a small guide RN A (sgRN A) sequence under control of constituti v e M. maripaludis promoters, the plasmids for ioTTS terminator mutations (Supplementary Table S1) were constructed in three steps. First, a sgRNA sequence of 20-bp targeting the terminator region was designed and inserted into the sgRNA expression region of pMEV4-Cas9. Second, a donor DNA carrying the upstream and downstream homologous arms flanking the sgRNA targeting region was inserted into the donor region of pMEV4-Cas9. Third, the sequence for the mutated ioTTS terminator was inserted into the donor sequence by stepwise Gibson assembly using the primers listed in Supplementary Table S2. PCR amplification was performed using the high-fidelity KOD-plus DN A pol ymerase (TOYOBO, Ja pan). DN A fragments and the amplified plasmid backbone were ligated by stepwise Gibson assembly using Clon-Express MultiS One Step Cloning Kit (Vazyme). The sequences of the constructed plasmids were confirmed by DNA sequencing prior to PEG-media ted transforma tion into M. maripaludis . Positi v e mutants were selected first by puromycin resistance to obtain the transformants and then 8-azahypoxanthine resistance to remove pMEV4-Cas9 plasmid as described previously ( 27 ).

Construction of RpoP and RpoF o ver expr ession str ains
The RpoP and RpoF ov ere xpression strains in M. maripaludis were constructed using the expression plasmid pMEV4 that carries the puromycin resistance gene pac and the terminator T MMP1559 . DNA fragments containing the ORF regions of M. maripaludis rpoP or rpoF were PCR amplified using primers listed in Supplementary Table S2 and then inserted into pMEV4 under control of the constituti v e promoter of MMP0386 through Gibson assembly to obtain plasmids P MMP0386 -rpoPpac -T MMP1559 and P MMP0386 -rpoFpac -T MMP1559 (Supplementary Table S2). Similarly, a pMEV4 plasmid carrying the mCherry gene, P MMP0386 -Mcherrypac -T MMP1559, was constructed as a control. The plasmids were each transformed into M. maripaludis via the PEG-mediated method to produce the ov ere xpression strains pMEV4-P 0386 -rpoP , pMEV4-P 0386 -rpoF and pMEV4-P 0386mc herr y (Supplementary Table S1) through puromycin selection.

Over all tr anscription r ate estimation by 3 H-uridine labeling
The relati v e ov er all tr anscription r ate was compared by determining the rate of 3 H-uridine incorporation into the exponential cultures of M. maripaludis grown a t 22 • C . Portions of exponentially growing cultures at OD 600 of ∼0.45, 600 l, were collected at 0, 0.2-, 0.5-, 1-, 1.5-and 2-min postaddition of [5,6-3 H]-Uridine (PerkinElmer) at a final concentration of 105 Ci / ml. Total RNA was extracted using TRIzol TM reagent (Invitrogen) and dissolved in 700 l Ecoscint A (National Diagnostics). 3 H isotope incorporation was determined by liquid scintillation counting (Perkin Elmer). Relati v e ov er all tr anscription r a te was calcula ted from the linear increase in cpm incorporation with time.

Nascent protein synthesis determined by puromycin incorporation
Puromycin incorporation was employed to determine the abundance of nascent pr oteins ( 28 ). Pur omycin (catalog no. P8833; Sigma) was added to a final concentration of 1% (w / v) to exponentially growing cultures at OD 600 = 0.5. After 10 min a t 37 • C , cells were harvested, resuspended in Lysis Buffer, and lysed by sonication. Total protein in cell-free extracts (CFE) was quantified using a BCA protein assay kit (Thermo Scientific). Fi v e microgram of CFE protein was western-blotted to quantify the puromycin incorporated using the monoclonal anti-puromycin antibody at a 1:10 000 dilution (Millipore Company, Darmstadt, Germany) and HRP-conjugated, anti-mice secondary antibody (Abmart) at a 1:2000 dilution.

Size e x clusion chromatography
Cells of M. maripaludis in the exponential growth phase at OD 600 of ∼0.45 were harvested and resuspended in Lysis Buffer, lysed by sonication, and centrifuged at 12,000 g for 30 min at 4 • C. The supernatant was incubated at 37 • C for 30 min with and without treatment of DNase I (100 U) and RNase A (100 U). A total of 500 g of the pretrea ted superna tant was fractiona ted through the size exclusion chromato gra phy on a Super de x 200 10 / 300 GL column (GE Healthcare), and the Mr of each fraction was estimated by gel filtration molecular weight markers (Kit No.: MWGF1000 of Sigma-Aldrich). Finally, each fraction was collected for protein identification by western blotting.

Internal-operon termination sites (ioTTSs) and internaloperon start sites (ioTSSs) are common in M. maripaludis
PacBio-seq enabled mapping the full length of transcripts in M. maripaludis . W hen integra ted with pre viously pub lished dRNA-seq data of transcription start sites (TSSs) ( 29 ), Term-seq data of transcription termination sites (TTSs) ( 21 , 22 ), and Illumina-seq data of transcript abundances ( 22 ), the operon map of M. maripaludis was characterized at single-base r esolution (Figur e 1 A). In total, 884 transcriptional units (TUs) were identified among the total of 1722 coding genes, in which half of the TUs (410 / 884) were operons, ie. comprising more than one gene (Figure 1 B and C). The operon length was positi v ely correlated with the number of genes, and the median length was 1794 nt or close to the median length of operons composed of two genes (Supplementary Figure S2A). Remar kab ly, internal operon TSSs (ioTSSs) and TTSs (ioTTSs) were found in the intergenic regions (IGRs) within 53% and 38% operons, respecti v ely. Four operon types were identified (Figure 1 A and C). Type I did not contain either ioTSSs or ioTTSs. Type II contained at least one ioTTS. Type III contained at least one ioTSS. Type IV contained both ioTTS and ioTSS (Dataset S1). The median lengths of type II and III operons were also longer than that of type I operons ( Supplementary Figure S2B), and the longer operons possessed more ioTTSs or ioTSSs (Supplementary Figure S2C and D). ioTTSs correlated with differential expression of upstream and downstream genes ioTTSs were found in 38% (157 / 410) of operons. The ioTTSs were characterized by their transcription differential expression ratios (TDERs) of the flanking genes or the transcript abundance of the upstream gene divided by that of the downstream gene. TDERs > 1.5 were considered biolo gicall y significant, and TDERs > 2 were considered as highl y significant biolo gicall y (see below). For the operons containing ioTTSs, 72% (150 / 210) of the TDERs possessed mid-or high-significance (Figure 1 D). Thus, most of the ioTTSs were functional, suggesting that they are the cause for the une v en transcription of the flanking genes within the operons.
In M. maripaludis , transcription termination depends upon the general transcription termination factor aCPSF1 and occurs at a cis-element with a polyuridine (polyU)motif ( 21 , 22 ). To determine if the mechanism of ioTTS termination resembles to that at the end of TUs, the role of aCPSF1 and terminator motif were examined. The TDERs for many of the ioTTSs were much lower in the mutant aCPSF1, which possessed only low le v els of aCPSF1 (20% of wild type) by growing at 22 • C but not 37 • C (Supplementary Figure S3A). Similarly, terminators with a higher termination efficiency had a higher dependency on aCPSF1 than the weak terminators (Supplementary Figure S3C). These observations were consistent with an important role for aCPSF1 at ioTTSs. In addition, the motifs preceding the ioTTSs with biolo gicall y significant TDERs featured pronounced polyU-tracts at the termination sites (Supplementary Figure S3B). In summary, termination at ioTTSs also depended on the presence of both polyU tracts and the transcription termination factor aCPSF1, similar to the r equir ement for a trans-acting factor and cis-element for archaeal transcription termination at the end of TUs ( 21 ).
To confirm these general properties of ioTTSs, three operons each consisting of ≥ 2 genes and one ioTTS wer e selected. P acBio-seq identified one ioTTS in the trigenic operon MMP1635-1633 that encodes a thior edoxin / glutar edoxin homolog, DsrE-like protein, and a conserved hypothetical protein. A sharp decrease in transcription of MMP1634-1633 immediately downstream of the ioTTS was observed, and this decrease was greatly alleviated in the aCPSF1 strain (Figure 2 A). Similar r esults wer e found by Illumina-seq. MMP1634 had a 2.7-fold and 1.8-fold lower transcription abundance than MMP1635 in the wildtype and aCPSF1 strains , respecti v ely (Figure 2 B). Northern blotting confirmed these results. A probe targeting MMP1635 detected both the expected three-cistron transcript > 1000 nt in length and the monocistronic transcript of MMP1635 in the wildtype str ain. In contr ast, depletion of aCPSF1 resulted in an increase of the three-cistron transcript and diminution of the monocistronic transcript (Figure 2 C). Similarly, the MMP1634 probe only detected the three-cistron transcript, whose abundance increased follo wing gro wth of the mutant a t 22 • C , a condition where the le v els of aCPSF1 wer e r educed to 20% of the wildtype ( 22 ). qRT-PCR also detected a 3-fold and 1.6-fold lower abundance of the transcripts downstream of the ioTTS in the wildtype and aCPSF1 strains, respecti v ely (Figure 2 D). Differential transcription of the ioTTS flanking genes was also displayed for other operons (Supplementary Figure S4) and similarly verified in another two operons, MMP0290-0291 (Supplementary Figure S5) and MMP1190-1189 (Supplementary Figure  S6). Collecti v ely, these e xperiments v erified the partial termina tion a t ioTTSs by multiple experimental methods at multiple r epr esentati v e operons.

Role of ioTTS termination in differential expression of genes within operons encoding protein comple x es
In M. maripaludis , a number of operons encode multisubunit enzymes and other protein complexes. In some cases, the stoichiometries of the subunits differ, and ioTTS termination could play a role in the differential expression of their genes. To explore the significance of ioTTS termination in expression of operons encoding protein complexes, the operon afuABC encoding the Afu ABC transporter for Fe(II) uptake was selected. An ioTTS, assigned as ioTTS afuAB , was found in the IGR between afuA and afuB . While, no evidence for ioTSS or antisense RNA was detected in this operon. Illumina-seq and PacBio-seq detected a higher abundance of transcripts for afuA than afuBC (Figure 3 A), and qRT-PCR assayed > 9-fold higher le v els of afuA transcripts than those of afuB in the wild type (Figure  3 B). Quantitati v e western-b lotting also f ound 8-f old higher le v els of the proteins AfuA than AfuC (0.0354 pmol / g vs. 0.0044 pmol / g total cell proteins) (Figure 3 C and D). In contrast, m utants w here the ioTTS afuAB terminator pol yUtract was disrupted using the CRISPR-Cas9 toolkit de v eloped for M. maripaludis ( 26 ) possessed significantly elevated le v els of afuB transcripts (Figure 3 B) and a 3 −5-fold increase in the AfuC but not AfuA protein content (Figure 3 C). In addition, the mutant grew more poorly than the wild type, especially at low concentrations of Fe (II), suggesting that the iron transporter in the mutant was functionally defecti v e (Figure 3 E). Thus, the ioTTS was necessary for the corr ect expr ession of the genes in the afuABC operon, the r esulting differ ential ab undances of the encoded sub units, and presumably the correct or optimal assembly of the Afu ABC transporter, as reflected in the poor growth of the mutant.
To further evaluate the physiological role of ioTTS terminators, ioTTSs were inserted into an operon that did not have one. The 10-gene operon encoding the ATP synthase complex was identified by both Illumina and PacBio transcriptome sequencing as lacking an ioTTS (Supplementary Figure S7A). Furthermore, transcript abundance was the same in the wild type and aCPSF1 mutant, confirming that no internal transcription termination occurred within the operon. Insertion of a strong terminator from MMP0204 (T1) or a medium-strength terminator from MMP0229 (T2) into the IGRs of atpK -atpE and atpF -atpA reduced transcription of the ioTTS downstream genes and also disrupted the coordinated transcription of the ten genes in the operon (Supplementary Figure S7). Growth of the M. maripaludis mutants was also r educed, compar ed from 0.15 h −1 of the wild type to 0.07 h −1 and 0.08 h −1 of the T1 and T2 terminator mutants, respecti v ely (Supplementary Figure S7D). In conclusion, ioTTSs are not only necessary for the proper expression of the genes in some operons, but also they are detrimental when positioned incorrectly in the intergenic regions of other operons.

ioTTS termination controls expression of the ioTTS downstream RNAP subunits in RP-RNAP operons
M. maripaludis possesses an RN A pol ymerase (RN AP) homologous to the eukaryotic RNAP II and composed of 12 subunits and encoded in fiv e operons (Supplementary Figure S8). One operon, rpoHB2B1A1A2, encodes the four catalytic subunits and an auxiliary subunit. The genes rpoDLNP, encoding four assembly subunits, and rpoEFK, encoding three auxiliary subunits are located in operons mostly comprising ribosomal protein (RP) genes, forming RP-RNAP operons. Because the le v els of the RPs and RNAP proteins ar e differ ent, differ ential transcription of the genes might be necessary for correct protein production. Interestingly, in these operons the RNAP genes are mostly located downstream of a RP gene, and an ioTTS separates them (Supplementary Figure S8). Accordingly, higher transcript abundances were found for the upstream RP than the downstream RNAP genes (Figure 4 A, Supplementary Figures S9A and D, and S10A, and Dataset S1). Northern blotting and qRT-PCR verified the differential transcriptions of the downstream and upstream genes, and this difference was reduced in the aCPSF1 mutant ( Figure 4 and Supplementary Figure S9). Thus, termination at the ioTTS appeared to be at least partly responsible for the correct expression of the genes in the RP-RNAP operons.
To examine ioTTS termination further, a mutation was generated in the ioTTS rpl37 / rpoP of the operon containing rpl37Ae and rpoP (Figure 4 B). The ioTTS rpl37 / rpoP terminator polyU-tract was replaced with non-U sequence to generate the mutant Mu-T p (Figure 4 B). Northern blotting with four probes targeting the genes downstream of the two ioTSSs and flanking ioTTS rpl37 / rpoP detected fiv e TUs of different lengths in wild type and the Mu-T p mutant, with a two-cistron transcript TU1 including rpl37Ae and a sixcistron transcript TU2 including rpl37 and rpoP being most abundant. In the Mu-T p mutant, the abundance of TU1 was markedly reduced while that of TU2 was elevated. In addition, the 7-cistron TU4, which comprises all the genes in this operon, was also detected in the mutant (Figure 4 B). Similarly, qRT-PCR detected a 4.5-fold higher le v el of transcripts of rpl37Ae than rpoP in the wild type, but only 2f old and 2.7-f old higher in the Mu-T p and aCPSF1 mutants, respecti v ely (Figure 4 C). Likewise, western blotting   3-fold higher le v els of Rpl37Ae (0.86 pmol / g) than RpoP (0.06 pmol / g) in the wild-type strain. Howe v er, 1.6-fold higher le v els of RpoP were detected in the Mu-T p mutant than in wild type, which consequently reduced the dif ferential ra tio of Rpl37Ae versus RpoP from 14.3fold to 8.9-fold (0.88 pmol / g vs. 0.097 pmol / g). Therefor e, these r esults confirmed the role of the ioTTS in r educing both transcription and protein production of downstream genes within the RP-RNAP operon (Figure 4 D).
Involvement of ioTTS termination in expression of RP and RNAP genes was also investigated in a tricistr on oper on containing rpl21e-rpoF . This operon contains ioTTS rpl21 / rpoF in the rpl21e-rpoF IGR and two ioTSSs upstream of rpl21e (Supplementary Figure S10A). Using probes targeting each gene, northern blotting detected fiv e TUs, with the 3-cistronic TU1, TU4 containing rpl21e-rpoF, and TU5 containing only rpl21e being abundant in wild type (Supplementary Figure S10B). Following mutation of the ioTTS rpl21 / rpoF , abundances of the TU2 and TU4 containing rpl21e-rpoF were elevated while the abundances of TU5 and TU3, which contained only rpl21e , were reduced. Thus, ther e appear ed to mor e r ead-through of the ioTTS in the m utant. Similarl y, qRT-PCR f ound 2.5-f old more transcripts of rpl21e than rpoF in the wild type, but only 1.4-and 1.5-fold more in the ioTTS rpl21 / rpoF terminator and aCPSF1 mutants following a ∼2-fold increase in the rpoF transcripts in both mutants ( Supplementary Figure S10C). By western blotting, Rpl21e (0.67 pmol / g) was detected to be 4.3-fold more abundant than RpoF (0.157 pmol / g) in wild type, but the ratio was reduced to 3-fold in theTTS rpl21 / rpoF mutant due to a 30% increase of RpoF (Supplementary Figure S10D). Thus, ioTTS termination played a role in controlling protein production as well as transcript abundance of the rpl21e-rpoF operon.
Collecti v ely, these e xperiments demonstrated that ioTTSs play important roles in the expression of RP-RN AP operons, especiall y in controlling the differential transcription and protein production of the two types of genes encoding two essential cellular macromolecular genetic machineries for translation and transcription.

Functional significance of ioTTS termination in expression of RNAP subunits
Following the demonstration of the role of ioTTS termination in RNAP gene transcription, the physiological significance of ioTTS termination on RNAP subunit production was investigated. Compared to the growth rate of the wild type of 0.14 h −1 , the ioTTS rpl37 / rpoP mutant Mu-T p grew with a half of the rate or 0.07 h −1 at 37 • C ( Figure  5 A). While the ioTTS rpl21e / rpoF terminator mutant Mu-T f exhibited the same growth rate as the wild type at 37 • C, its growth rate was reduced 5-fold to 0.01 h −1 at 22 • C (Figure 5 B). Similarly, rpoN and rpoK were downstream of the ioTTS rps9 / rpoN and transcribed less than the upstream RP genes (Supplementary Figure S11A and B). A mutation of the ioTTS rps9 / rpoN resulted in a reduced growth at 37 • C (Supplementary Figure S11C) as well. To verify that the elevated production of RNAP subunits was responsible for the growth retardation, rpoP and rpoF were expressed on shuttle v ectors. Ov ere xpression was verified by west-ern blotting, and RpoP and RpoF protein productions increased 1.5-fold and 2.1-fold in the strains pMEV4( rpoP ) and pMEV4( rpoF ), respecti v ely. Growth of the ov ere xpression strains was also retarded and similar to those of the ioTTS terminator mutants ( Figure 5 ). Thus, an excess of RN AP subunits a ppeared to cause the growth defects observed in the ioTTS mutants.
RNAP plays a critical role in transcription, and the overall transcription rate of the Mu-T p was compared with that of the wild type. Based on the 3 H-uridine incorporation r ate, the over all tr anscription r ate of the Mu-T p mutant was determined to be about half that of the wild type at 22 • C (Figure 5 C). Because of its rapid growth, it was impractical to measure the rate of 3 H-uridine incorporation a t 37 • C . The ra te of transla tion was also impacted in the Mu-T p mutant. Methanococci are sensiti v e to puromycin, an aminonucleoside antibiotic and tRN A analo gue that covalently binds to the nascent polypeptide chain during acti v e translation. Thus, it can be used to indicate newly synthesized protein ( 30 ). Using the anti-puromycin antibody, western blotting determined significantly decreased puromycin-integration in the Mu-T p mutant compared with that of the wild type ( Figure 5 D).
In conclusion, ioTTS termination in RP-RNAP operons plays important roles in the normal rates of RNAP production with profound consequences on growth and rates of transcription and translation.

Importance of intracellular levels of RNAP subunits in RNAP assembly
To examine why ioTTS termina tion-modula ted RNAP protein expression impacted rates of transcription, the assembl y of RN AP was survey ed in the two str ains with elevated le v els of RpoP, the ioTTS rpl37 / rpoP mutant Mu-T p and the ov ere xpression strain pMEV4( rpoP ). RN AP assembl y in cell extracts was tested by size exclusion chromatography coupled to western blotting using antibodies against RpoA", RpoP and RpoD. Elution of the fully assembled RNAP with a predicted molecular weight (Mr) of ∼350 kDa was detected in fractions ∼9-10 ml from all tested strains ( Figure 6 ). Howe v er, compared with wild type, lower amounts of the fully assembled RNAP complex were detected in both Mu-T p and pMEV4( rpoP ) strains. In contrast, both ov ere xpression strains contained ∼2-fold higher le v els of free RpoP (eluting at 18-19 ml or ∼6 K d ), lower le v els of free RpoD (eluting at 14.5-15 ml or ∼20 K d ) and potential RN AP assembl y intermediates, presumabl y including RpoNPDL (eluting at 13.5-14.5 ml or ∼34 K d ) ( Figure 6 A and B).
In conclusion, elevation of the levels of a RNAP subunit impaired the assembly of the RNAP complex, and termina tion a t the ioTTS within the RP-RNAP operon was one mechanism to control the production of the appropriate levels of RNAP subunits.

Wide distribution of the RP-RNAP operons and ioTTS termination among archaea
Distribution of operons with the rpl37Ae -rpoP and rpl21e -rpoF gene organization among di v erse archaea was investigated. Using the M. maripaludis rpoP as a probe, operons bearing rpl37Ae -rpoP were found in the Euryarchaeota and superphylum DPANN (Figure 7 A) and were often associated with the conserved genes psmA, IMP4 , pcc1 and pfdB , which encode proteasome alpha subunit, rRNA ma tura tion protein IMP4, KEOPS subunit Pcc1, and prefoldin beta subunit, respecti v ely. Genes encoding exosome subunits Rrp4, Rrp42 and Rrp41 were also found upstream of rpl37Ae in many Euryarchaeota and DPANN genomes. Genes encoding EF-1a and Rps10 replaced rpl37Ae upstream of rpoP in Thaumarchaeaota. Interestingly, stand-alone rpoP was found in all Crenarchaeota species, and probably in Ca. Prometheoarchaeum syntrophicum of the Asgard group (Figure 7 A). Therefore, the gene organization of rpl37Ae -rpoP was widely but not uni v ersally conserv ed among the archaea.
The operon organization of the RP-RNAP operons across archaeal lineages suggests that an evolutionary transition in genomic arrangements may have occurred. Large operons (up to 10 genes) are found in Euryarchaeota and DPANN species, while shorter operons consisting of 3-4 genes occur in Haloarchaea and Nanoarchaeota. Interestingly, in TACK species, the archaeal clade belie v ed as the closest relati v es of eukaryotes ( 31 ), the rpoP gene is present as stand-alone. This evolutionary trend may reflect a potential link between prokaryotic gene cluster organization and the transition to the single-cistron arrangement seen f or eukary otic genes. Similarly, the gene organization rpl21e -rpoF was widely distributed but not uni v ersal. It was found in the majority of groups that contained rpoF homologs, including the Euryarchaeota, the majority of Crenar chaeota and DPANN ar chaea, and Ca. P. syntrophicum (Figure 7 B).
Reexamination of the available transcriptomic data of archaea, higher le v els of transcription of rpl37Ae than that of rpoP was found in Thermococcus k odakar ensis , T. onnurineus , Methanosarcina acetivorans , and Methanococcoides burtonii (Dataset S3). Thus, ioTTS termination may be common within this operon in the Euryarcheaota and awaits further examination in other archaea.

DISCUSSION
Transcription termination plays critical regulatory roles in gene expression, such as in regulating sRNA function (32)(33)(34)(35), riboswitch action (36)(37)(38), r epr essing x eno genic DN A and pervasi v e transcriptions (39)(40)(41)(42)(43) and maintaining chromosome integrity and transcription and translation coupling ( 38 , 44 , 45 ). Howe v er, the process of transcription ter-mination and its significance in regulating physiology remain largely unclear in the third form of life, ar chaea. Pr eviously, we identified the genome-wide transcription termination sites and demonstrated that transcription termination is r equir ed for programming an ordered transcriptome of M. maripaludis ( 22 ). This study reported that the internal operon (ioTTS) termination tunes differential transcriptions of operon-organized genes. Based on the multi-transcriptomic da ta, ioTTS termina tion was found to be common in M. maripaludis and have been experimentally verified to be physiological significant. Either destroying the intrinsic ioTTS terminators in selected operons, such as Afu ABC transporter and RP-RNAP operons, or artificially inserting an ioTTS into the operon encoding ATPase complex not only disrupted the inherent stoichiometric expression of the operon genes, but also impaired the physiological functions of the complex (Figures 2 -6 and Figs. S7, S9, S10, S11). Ther efor e, the differ ential transcription of the operon-organized genes modulated by ioTTS termination is a physiological demand.
Clustering of functionally related genes in operons allows for coregulation and especially e v en gene e xpression in prokaryotes. Howe v er, achie ving precise, and sometimes une v en, stoichiometries in protein complexes requires finetuning mechanisms. Transla tion ef ficiency, influenced by mRNA secondary structure and codon usage ( 15 , 17 , 18 ), as well as posttranslational mechanisms including differential protein degradation ( 46 , 47 ), play roles in maintaining protein stoichiometries. Regulatory elements, present in or targeting mRNA 5 or 3 untranslated regions (5 UTR or 3 UTR) or regions internal to ORFs, and RN A pol ymerase pausing also contribute to fine-tune gene expression ( 16 , 20 , 48-50 ). In this study, we experimentally demonstrate the physiological significance of ioTTS termination as an additional effecti v e regulatory mode that fine-tunes the stoichiometric expression of operon genes. Unlike the transcription polarity effect observed in bacteria, which arises from the dissociation of transcription and translation when translation is halted, usually at nonsense mutations or ribosome pausing, and resulting in premature transcription termination so affecting the expression of downstream genes (51)(52)(53)(54)(55), ioTTS termination occurs under normal physiological conditions. Through terminating partial transcriptions from the operon promoter and allowing part of transcription readthrough, ioTTS termination controls transcription of the genes downstream of ioTTSs to be lower than that of the upstream ones and the subsequent protein pr oduction, assembly of pr otein complexes, and gr owth. ioTTS termination can not only explain 30-80% transcription termina tion ef ficiency (TTE) determined in many of TTSs in M. maripaludis ( 21 , 22 ) but also can echo the finding that e xtensi v e transcription read-through was found at over 40% of TTSs in E. coli and se v eral other bacteria ( 16 , 19 , 20 ), in which similar transcription differences occurred at genes flanking the ioTTS within operons. Importantly, ioTTS termination was found to be a novel regulatory mechanism in controlling the pr ecise expr ession of the RNAP subunits in the RP-RNAP operons. Mutation of the ioTSS terminators in the RP-RNAP operons, rpl37Ae-rpoP, rpl21e-rpoF, and rps9-rpoNK , elevated transcription of the ioTSS downstream rpo genes, but instead reduced the cellular RNAP complex contents, and transcription and translation velocities (Figures 4 , 5 , ( 6 , S9, S10, and S11), and the consequent growth of M. maripaludis . Proper cellular contents of RNAP subunits appear to be r equir ed in assembling the complete RNAP complex, and termina tion a t ioTTSs appear to be one of the control strategies. The in vivo assembly process of archaeal RNAP complex and the regulation of its expression and assembly all remain unknown. The archaeal RN AP assembl y process was predicted based on in vitro reconstitution experiments ( 56 ). In which, the subunits RpoD, RpoL, RpoN, and RpoP are first fit together as an assembling platform ( 57 , 58 ), and on it the catalytic subunits RpoB , RpoB , RpoA and RpoA followed by RpoH and RpoK are assembled. Finally, the accessary subunits RpoE and RpoF are recruited ( 58 , 59 ). In the current study, through antibodies targeting the catalytic and assembly platform subunits, RpoA and RpoD / RpoP subunits respecti v ely, we found that elevation of the RNAP subunit RpoP in both the ioTTS rpl37Ae-rpoP terminator mutation (Mu-T p ) and the pMEV4-rpoP ov ere xpression strain, instead decreased the cellular content of the complete RNAP complex and a potential assembly intermediate (Figure 7 ). This indica tes tha t appropria te contents of one or more RNAP subunits or certain cellular ratio among the subunits is r equir ed for the effecti v e assemb l y of the archaeal RN AP comple x. Supporti v ely, a precise content of Rpb10, a eukaryotic RNAP subunit, is r equir ed in eukaryotic RNAP I and RNAP III assembly as indicated by genetic studies (60)(61)(62). Rpb10 is the homolog of archaeal RpoN, and functions as the RN AP assembl y platform by serving as a structural adaptor to fit the ␤and ␣-like catalytic Rpo2 (RpoB) and Rpo3 (RpoL) subunits ( 63 , 64 ). Recent studies indica ted tha t the cellular concentra tions of Rpb10 are precisely controlled through a reciprocal regulation of an RNA binding protein Rbs1 and a Upf1 helicase (60)(61)(62). Analo gousl y, the precise contents of RpoP and RpoN, the platform subunits, could also be key to the archaeal RNAP assembly, and ioTTS termination can be one of the control strategies.
ioTTS termina tion in coordina ting expression of the archaeal and bacterial RP-RNAP operons can exert an important role in physiology, r epr esenting a novel regulatory mode that not only ensures the proper le v els of the suite of subunits that assemble into the complete RNAP complex but also in coordinating the production of the systems to couple translation and transcription. Coupled transcription and translation in prokaryotes is usually achie v ed by ribosomes loaded on the transcripts being transcribed, and the leading ribosome e v en binding to the processi v e RNAP ( 65 ). In acti v e E. coli , ribosomal proteins account for up to 40% of the total cellular proteins, and multiple ribosomes can synchronically translate one transcript ( 66 , 67 ). Thus, it reasonably predicts that higher abundances of ribosomal than RNAP proteins are in cells, and this was confirmed in M. maripaludis where the ratios of Rpl37Ae to RpoP and Rpl21e to RpoF are determined to be 14.3:1 and 4.3:1, respecti v ely. While, changing the ratios by mutation of ioTTS terminators reduced transcription and translation velocities and also growth of M. maripaludis (Figures 4 , 5 , S10 and S11). This suggests that the appr opriate pr oportion between the subunits of the translational and transcriptional macromolecules should be the physiological demands and are controlled at least in part by termination at ioTTSs.
Gi v en the wide distribution of the RP-RNAP operons in ar chaea and pr esumably differ ential transcription of RNAP and RP subunit genes in the operons of some archaeal species (68)(69)(70)(71) and dataset S3, the ioTTS termination Nucleic Acids Research, 2023, Vol. 51, No. 15 7865 regulation might be a common strategy employed by archaea. Ther efor e, the operon organization of RP and RNAP subunits, the two macromolecular machineries that are involved in translation and transcription, could be evolutionaril y ad vantageous to ensure the optimal coupling of the two processes, and ioTTS termination could be a widely employed regulatory mode for controlling the r equir ed levels of them in cells.
In conclusion, this work reports tha t ioTTS termina tion functions as a new regulatory mode in coordinating differ ential expr ession of genes in ar chaeal operons that encode protein complexes consisting of une v en stoichiometry of subunits or involving in different biological processes. The biological significance of this type of coordina ted dif fer ential expr ession is particular ly illustr ated by the RP-RN AP operons, w here disruption of this coordination reduced RNAP complex content and as a consequence tr anscription, tr anslation, and growth. The ioTTS termination not only demonstrates the complexity of prokaryotic transcription regulation mechanisms but illustrates a novel mode of controlling the precise RNAP subunit content. Thus, ioTTS termination-controlled RP-RNAP operon expression is likely widely employed in prokaryotes.