The subcellular architecture of the xyl gene expression flow of the TOL catabolic plasmid of Pseudomonas putida mt-2

Despite intensive research on the biochemical and regulatory features of the archetypal catabolic TOL system borne by pWW0 of Pseudomonas putida mt-2, the physical arrangement and tridimensional logic of the xyl gene expression flow remains unknown. In this work, the spatial distribution of specific xyl mRNAs with respect to the host nucleoid, the TOL plasmid and the ribosomal pool has been investigated. In situ hybridization of target transcripts with fluorescent oligonucleotide probes revealed that xyl mRNAs cluster in discrete foci, adjacent but clearly separated from the TOL plasmid and the cell nucleoid. Also, they co-localize with ribosome-rich domains of the intracellular milieu. This arrangement was kept even when the xyl genes were artificially relocated at different chromosomal locations. The same happened when genes were expressed through a heterologous T7 polymerase-based system, which originated mRNA foci outside the DNA. In contrast, rifampicin treatment, known to ease crowding, blurred the confinement of xyl transcripts. This suggested that xyl mRNAs intrinsically run away from their initiation sites to ribosome-rich points for translation—rather than being translated coupled to transcription. Moreover, the results suggest that the distinct subcellular motion of xyl mRNAs results both from innate properties of the sequence at stake and the physical forces that keep the ribosomal pool away from the nucleoid in P. putida. This scenario is discussed on the background of current knowledge on the 3D organization of the gene expression flow in other bacteria and the environmental lifestyle of this soil microorganism. IMPORTANCE The transfer of information between DNA, RNA and proteins in a bacterium is often compared to the decoding of a piece of software in a computer. However, the tridimensional layout and the relational logic of the cognate biological hardware i.e. the nucleoid, the RNA polymerase and the ribosomes, are habitually taken for granted. In this work we inspected the localization and fate of the transcripts that stem from the archetypal biodegradative plasmid pWW0 of soil bacterium Pseudomonas putida KT2440 through the non-homogenous milieu of the bacterial cytoplasm. The results expose that— similarly to computers also—the material components that enable the expression flow are well separated physically and they decipher the sequences through a distinct tridimensional arrangement with no indication of transcription/translation coupling. We argue that the resulting subcellular architecture enters an extra regulatory layer that obeys a species-specific positional code that accompanies the environmental lifestyle of this bacterium.

ABSTRACT. Despite intensive research on the biochemical and regulatory features of the archetypal 1 catabolic TOL system borne by pWW0 of Pseudomonas putida mt-2, the physical arrangement and 2 tridimensional logic of the xyl gene expression flow remains unknown. In this work, the spatial distribution 3 of specific xyl mRNAs with respect to the host nucleoid, the TOL plasmid and the ribosomal pool has 4 been investigated. In situ hybridization of target transcripts with fluorescent oligonucleotide probes 5 revealed that xyl mRNAs cluster in discrete foci, adjacent but clearly separated from the TOL plasmid 6 and the cell nucleoid. Also, they co-localize with ribosome-rich domains of the intracellular milieu. This 7 arrangement was kept even when the xyl genes were artificially relocated at different chromosomal 8 locations. The same happened when genes were expressed through a heterologous T7 polymerase-9 based system, which originated mRNA foci outside the DNA. In contrast, rifampicin treatment, known to 10 ease crowding, blurred the confinement of xyl transcripts. This suggested that xyl mRNAs intrinsically run 11 away from their initiation sites to ribosome-rich points for translation-rather than being translated 12 coupled to transcription. Moreover, the results suggest that the distinct subcellular motion of xyl mRNAs 13 results both from innate properties of the sequence at stake and the physical forces that keep the 14 ribosomal pool away from the nucleoid in P. putida. This scenario is discussed on the background of IMPORTANCE. The transfer of information between DNA, RNA and proteins in a bacterium is often 20 compared to the decoding of a piece of software in a computer. However, the tridimensional layout and 21 the relational logic of the cognate biological hardware i.e. the nucleoid, the RNA polymerase and the 22 ribosomes, are habitually taken for granted. In this work we inspected the localization and fate of the 23 transcripts that stem from the archetypal biodegradative plasmid pWW0 of soil bacterium Pseudomonas 24 putida KT2440 through the non-homogenous milieu of the bacterial cytoplasm. The results expose that-25 similarly to computers also-the material components that enable the expression flow are well separated 26 physically and they decipher the sequences through a distinct tridimensional arrangement with no 27 indication of transcription/translation coupling. We argue that the resulting subcellular architecture enters 28 an extra regulatory layer that obeys a species-specific positional code that accompanies the 29 environmental lifestyle of this bacterium. 30 The TOL system encoded by plasmid pWW0 of Pseudomonas putida mt-2 is to this date the most 3 thoroughly studied example of biodegradative system in soil microorganisms. The primary function of this 4 catabolic device is enabling carrier bacteria to grow on toluene, m-xylene, p-xylene and other related 5 aromatics through a set of enzymes encoded by upper and lower plasmid-borne operons ( Fig. 1A; 1-3). 6 While catabolic traits of this sort are not uncommon in many other environmental isolates, what makes 7 the TOL system special is the extraordinary regulatory intricacy that controls expression of the xyl genes 8 and their high-level interplay with the host's physiological regulons. The many mechanisms unveiled over 9 the years in this respect seem to include nearly every device known in the prokaryotic word for ruling the 10 gene expression flow (4-6). This state of affairs has made the TOL plasmid and its bearer (P. putida 11 KT2440) a beneficiary of the suite of conceptual and material tools of contemporary Systems and 12 Synthetic Biology (7, 8). In particular, the wealth of experimental data on expression of the xyl genes has 13 enabled the understanding of the cognate regulatory network as a complex device that processes inputs 14 into outputs following a layer of logic gates implemented with promoters, transcriptional factors and 15 sRNAs (9, 10). It could thus be argued that we know at this point much about the genetically-encoded 16 software of the system and the relational logic that rules its performance. In contrast, we know virtually 17 nothing of the physical arrangement of the hardware that sustains the same process. In particular, the 18 gross spatial disposition of the molecular actors that execute the transfer of information from the xyl genes 19 to production of catabolic enzymes is unknown. 20

21
The same sort of questions has been tackled in other prokaryotic systems. There seem to be at least two 22 different patterns of transcript localization. In one scenario, the mRNA remains close to the site of 23 transcription. This is the case of the mRNAs of groESL and creS of Caulobacter crescentus and lacZ of 24 E. coli, which appear to remain in the vicinity of the corresponding genomic DNA loci where they are 25 transcribed (11). This suggests that for mRNAs to be translated they need to be associated to the 26 ribosomes while being produced (as the prevailing view of generalized transcription-translation coupling 27 would imply) or short after their creation by RNAP (11,12). An alternative scenario involves migration of 28 mRNAs to specific spots of the cytoplasm for translation close to the site(s) where the gene products are 29 needed. While this setting argues against generalized transcription/coupling translation, there is solid 30 evidence of its occurrence e.g. the bglGFB operon, whose products were located in sub-cellular localizations where cognate proteins were expected to work (13). This example may not be anecdotal as 1 transcriptome-wide scale studies of mRNA localization in E. coli revealed that a large share of transcripts 2 were found in cell domains (e.g. membranes, cytoplasm, poles) that coincided with the function of the 3 cognate proteome (14, 15). Furthermore, the mRNA of archetypal membrane proteins LacY and TetA 4 are distinctively close to the cell envelope (16,17). In other instances, a signal-recognition particle (SRP) 5 is involved in the movement of mRNAs to the membrane, as translation of some mRNAs encoding inner-6 membrane proteins produces a signal peptide that recruits such SRP and the complex leads the transcript 7 to its intracellular address (15, 16,18). Finally, Rho-dependent transcription termination in Bacillus subtilis 8 is somewhat weak, and runaway, untranslated mRNAs are abundant (19). In sum, the fate of each 9 transcript seems to be both gene (i.e. sequence)-dependent and species-dependent. 10

11
In this work we have inspected the localization of mRNAs initiated in the catabolic promoters of the TOL 12 plasmid pWW0 of P. putida mt-2. The starting point for tackling the issue is the earlier observation that 13 RNA polymerase (RNAP) of P. putida fully co-localizes with the chromosomal DNA of the nucleoid while 14 being entirely apart from the bulk of the ribosomal pool (20). This observation suggested that rather than 15 being coupled to translation ab initio, many (if not most) of the mRNAs initiated at chromosomal promoters 16 need to move to ribosome-rich, nucleoid-free domains of the intracellular space for translation. Note that 17 in the case of the xyl promoters, transcription initiates in an extrachromosomal element, and therefore 18 the 3D itinerary of the corresponding mRNAs could be different. As shown below, by merging genetic 19 analyses with in situ RNA-FISH and DNA-FISH technology we could faithfully locate the relative positions 20 of the nucleoid DNA, the pWW0 plasmid, the xyl transcripts and the ribosomes and predict the motion of 21 cognate mRNAs through the cell interior. The results exposed an unexpected degree of physical partition 22 among the material actuators of the gene expression flow that may help P. putida to deal with its typical 23 environmental settings. specific mRNAs of P. putida mt-2 stemming from the TOL operons of plasmid pWW0 we adopted an 29 RNA FISH approach (Fig. 1B). To this end, the strain was cultured in M9 minimal medium with succinate 30 as the sole carbon source until the cells reached exponential phase. At this point the cultures were exposed (or not) to saturating vapors of m-xylene to induce transcription of the catabolic xyl genes. After 1 2 h, samples were collected and fixed with formaldehyde for hybridization with specific probes as 2 described in the Materials and Methods section. For this, two sets of 48 CAL Fluor Red 610-tagged 3 fluorescent oligonucleotides (20 nt long; Supplementary Table S2) were synthesized that covered, 4 respectively, the leading 1418 bps of the upper TOL operon transcript spanning the whole of xylU and 5 part of xylW (encoding benzyl alcohol dehydrogenase) and the front segment of the TOL lower operon, 6 encompassing 1203 pb of the xylX gene (alpha subunit of toluate 1,2-dioxygenase). After hybridization 7 with these oligo sets, the samples were washed and red signals inspected with fluorescence microscopy. 8 9 As shown in Fig. 1C, distinct, discrete fluorescent foci were clearly noticed under the microscope (1-2 per 10 cell) from the cultures subject to m-xylene exposure upon in situ hybridization with upper-pathway or 11 lower-pathway specific oligonucleotides. In contrast, no signals were detected in bacteria grown in M9-12 succinate without aromatic effector ( Supplementary Fig. S1A). These observations accredited that the 13 designed probe sets and the methodology proper were working as expected, as there were no signals 14 that could be attributed to hybridization with DNA. Therefore, the foci appeared to represent upper or 15 lower xyl transcripts. To further benchmark the experimental approach, cells were treated with either that rules the interplay between the substrates, regulators and catabolic operons of the TOL system ( Fig.  22 1A). We could thus safely consider that the foci inside cells shown in Fig. 1C were the result of the bona 23 fide hybridization of the fluorescent probes to specific mRNA of xylUW or xylX. Closer inspection of the 24 images revealed that no red output ever appeared dispersed throughout the cell, but always as discrete 25 foci (Fig. 1B). Yet, their position in respect to the cell shape varied among cells and signals were located 26 near the center, the poles, the contour or the septum of the cells (Fig. 1C). We thus set out to characterize 27 this asymmetrical localization of the xyl transcripts in respect to the nucleoid occupation and to the TOL 28 plasmid, as explained below. 29 30 xyl mRNAs occupy subcellular nucleoid-free regions. In order to identify the relative localization of 1 the xyl transcripts in respect to the nucleoid, following hybridization with the fluorescent probes described 2 above the bacterial genome was stained with 4', 6-diamidino-2-phenylindole (DAPI) in cells exposed to 3 m-xylene. The results of this procedure with the upper and lower pathway TOL probes (xylUW and xylX) 4 respectively are shown in Fig. 2. Images were processed and signals analyzed with the CellShape 5 software. For interpreting these results, note that [i] previous work showed that the DNA of the nucleoid 6 overlaps spatially with P. putida's RNAP (20) and [ii] DAPI binds both chromosomal and plasmid DNA. 7 Inspection of the cells under the microscope ( Fig. 2A) revealed that signals stemming from xylUW or xylX 8 transcripts (red foci) were located in sites inside cells with a low DAPI signal, i.e. in the nucleoid-free 9 regions. In order to quantity the phenomenon, >100 pictures of individual DAPI-stained and fluorescent 10 oligo-hybridized cells were separately recorded with the blue and red channels of the fluorescence 11 microscope and they were automatically inspected with the CellShape image analysis tool (23; Fig. 2B). 12 The outcome of this analysis is shown in Fig. 2C. Very few red foci (< 1%) overlapped with the densest 13 DAPI signals and as little as 15% of fluorescent spots from either xylUW or xylX RNA were detected at 14 the peripheral regions of genomic DNA. Instead, the vast majority of the remaining foci were placed away 15 from the blue signal (Fig. 2C). It thus looked like the bulk of TOL transcripts were located within subcellular 16 regions with no or little overlap with the DNA signal and therefore virtually devoid of RNAP (20). This 17 suggests that once formed, xyl mRNAs could migrate for translation to a site different from the place 18 where transcription is initiated. Note that the pathway substrate (m-xylene) dissolves in the cell membrane 19 (24). Moreover, the xylM product (hydroxylase component of the leading pathway enzyme xylene 20 monooxygenase) is located in the membrane (25). It is not uncommon that genomic sites encoding 21 envelope-associated proteins are transcribed coupled and co-translationally inserted to the membrane 22 through a so-called transertion mechanism (16). In the case of the TOL transcripts it looked instead that 23 movement away from the DNA begins after transcription has terminated and the mRNA has disengaged 24 from the nucleoid. To clarify this, we examined the relative position of the xyl mRNA in respect to their 25 actual origin inside the cells (i.e. the TOL plasmid) as explained below. corresponding sequences could then be exposed through FISH with a special type of 6-2 Carboxyfluorescein (6-FAM)-labeled 18 nt oligonucleotides. These encoded the tetO operator and bore 3 a distinctive chemical configuration (so called locked nucleic acid LNA structure) that increases specificity 4 for target DNA (11; Supplementary Fig. 2A). Samples were thus first hybridized with xyl-specific (red 5 fluorescence) and pWW0-specific (green fluorescence) probes and then stained with DAPI. As a result 6 of this procedure, both red, green and blue fluorescent signals could be mapped in cells exposed to m-7 xylene, corresponding to xylUW or xylX mRNA, plasmid, and nucleoid, respectively ( Inspection of colored spots merged within the contour of single cells exposed two predominant spatial 4 arrangements of the different signals. In one case, red and green spots virtually co-localized (e.g. Fig.  5 3B), while in most others, the two signals were close to each other but clearly separated ( length mRNAs could be found well away from the plasmid while still tethered to the template. However, 12 it is known that the mRNA from the meta pathway is quite short lived and starts being degraded before it 13 is fully synthesized (32). Note also that mRNA tends to form secondary structures that shorten the 14 distances between the 5′ and 3′ ends (33). On this basis, we argue that separation of red and green spots 15 in the images shown in Fig. 3, Fig. 4 and Supplementary Fig. S4 reflect an authentic migration of the TOL 16 transcripts away from their DNA template. The next obvious question was whether this was result of TOL 17 genes being encoded in a type of plasmid that largely stays in the periphery of the nucleoid (see above). 18 Alternatively, the unexpected localization of xyl mRNAs could stem from intrinsic properties of the 19 transcripts themselves. In order to sample these possibilities we entered various types of perturbations 20 in the system as described below. in P. putida Paw140 as compared to the lower numbers of the pWW0-bearing P. putida mt-2 strain (Fig.  2 2C), surely due to a higher transcriptional activity. This is not surprising as the pedigree of strain P. putida 3 Paw140 involved random chromosomal insertion of TOL genes and selection for best growers on m-4 xylene, what plausibly favored implantation in regions of high transcriptional activity (37-39). Yet, when 5 RNA-red signals were compared to those of the DAPI-stained nucleoid, most of them were observed 6 away of the denser DNA regions. Specifically, < 3% of the red spots of either xyl mRNAs overlapped the 7 more compacted chromosome, while 74% of xylUW and 87% of xylX signals were enriched in the 8 peripheral space of the cell (Fig. 5B and 5C). Preferential localization of xyl mRNAs was thus maintained 9 regardless of the replicon (plasmid of chromosome) that bears the cognate genes. Moreover, the higher 10 incidence of TOL mRNAs in the nucleoid-free subcellular regions corroborates that xyl transcripts move 11 away from their transcription site.  Note also that there were RNA dots stemming from the lower pathway in the wild-type P. putida mt-2 host 24 bearing the modified plasmid pTOL-PuxT7 ( Supplementary Fig. S5B). This is because m-xylene-25 activated XylR caused overexpression of XylS, which in turn suffices to activate the Pm promoter even 26 in the absence of any effector (5, 43, 44). As a consequence, when pTOL-PuxT7 was placed in P. putida 27  (Fig. 1A). But in contrast to the effect of the native 1 effectors in the naturally-occurring system, the surrogate control of the upper route by T7 polymerase 2 originated a higher number of distinct foci per cell. While this reflected the strength of the PT7 promoter 3 as compared to Pu (which propagated into a more potent activity of the Pm promoter as well), it is 4 noteworthy that the red signals always appeared focused (i.e. constrained within a subcellular domain) 5 rather that diffused through the cytoplasm. Furthermore, they again materialized mostly in the nucleoid 6 free regions. Automated quantification of signals detected with the red and blue channels in individual 7 cells with pixel precision revealed that 91% of xylUW and 98% of the xylX were excluded from the DAPI-8 stained field (Fig. 6C). Given that such transcripts originate in single promoters per bacterium and that

Disruption of intracellular crowding enables xyl transcripts to diffuse throughout the cytoplasm 23
The data above accredit that xyl mRNAs leave the proximity of the nucleoid towards the ribosome-24 enriched subcellular domains. Although a role for specific RNA binding proteins cannot be ruled out as 25 drivers of the process, a simpler explanation is that the corresponding sequences endow the xyl 26 transcripts with physical properties that make them to be quickly discharged from the vicinity of the 27 nucleoid, especially if the transcribed sequences are large. In fact, it seems that smaller RNAs use to 28 appear uniformly distributed throughout the bacterial cell, while longer molecules typically display more 29 limited dispersion (47). Large mRNAs can hardly co-localize with densely packed DNA regions due to 30 straight physical forces e.g. excluded volume effects (48, 49). As a consequence, the more crowded the bacterial cytoplasm is the less diffusible cellular components and molecules are, logically, in a size-1 dependent fashion (50, 51). 2

3
In order to test whether such a mutual exclusion between different intracellular domains accounted for 4 the unusual behavior of the xyl mRNAs we treated P. putida KT2440•T7 (pTOL-PuxT7) cells with 5 rifampicin. In one hand, since T7 RNAP is not sensitive to this antibiotic, the upper TOL transcript (xylUW) 6 can still be produced (the xylX gene of the lower pathway predictably cannot, though). On the other hand, 7 it is known that diffusion rate of ribosomal proteins is faster (11,12) and the cDNA of the nucleoid expands 8 after treatment with the drug (52, 53). Finally, the dearth of ribosomes available for interacting with RNAs 9 due to inhibition of 16s rRNA synthesis with (54) may also ease diffusion of otherwise tethered transcripts. the mRNAs were visualized with FISH as before. Expectedly, we could hardly detect any xylX mRNA 15 signals ( Supplementary Fig. S6). In contrast, when cells were hybridized with the xylUW probe under the 16 same conditions, red signals did appear ( Fig. 7 and Supplementary Fig. S5). Yet, inspection of individual 17 bacteria revealed a considerable variability in the intensity of the signals and their intracellular distribution. 18 Some cells had their whole contour non-uniformly filled with red color, while other showed a number of 19 discrete red signals with lower intensity (Fig. 7 and Supplementary Fig. S7). This outcome is not 20 altogether unexpected, as the strong T7 promoter in the plasmid (55) drains cellular resources (56, 57) 21 for the sake of transcription of the upper pathway, thereby originating noise. This effect can be 22 exacerbated in a plasmid, as the loss of gyrase production with the drug and the ensuing accumulation 23 of local supercoiling can lead to stochastic transcriptional bursts (58-60). In any case, we argue that the 24 lack of distinct localization patterns in rifampicin-treated cells reflects the diffusion of the TOL transcripts 25 under the conditions. Moreover, some images showed enrichment of xylUW mRNAs in the cells' internal 26 periphery in bacteria where the chromosome was otherwise highly compacted ( Supplementary Fig. S7, 27 marked). In these cases, it looked like mRNA was unable penetrate the denser DNA regions but could 28 freely diffuse in the rest of the cytoplasmic space. Taken together, analysis of images shown in Fig. 7  29 and Supplementary Fig. S7 indicated that the xylUW transcript of rifampicin-treated cells lost its restraint 30 in discrete foci and could then freely circulate as the cellular crowding decreases upon antibiotic addition.
In sum, the data suggests that under native conditions xyl mRNAs become localized away from the 1 nucleoid because of their entrapment with the translational machinery and the physical forces that 2 determine phase separation between the different components of the gene expression flow. As shown 3 above, if such a separation is perturbed upon rifampicin addition, xyl transcripts can then circulate through 4 the whole cytoplasmic interior. 5 6 Conclusion. This study shows that xyl mRNAs are constrained within subcellular regions with limited 7 mobility rather than being freely diffusible inside P. putida mt-2 cells. Most xyl transcripts were detected 8 in the space peripheral to the nucleoid, i.e. in a subcellular domain with virtually no DNA but enriched in 9 ribosomes (20). That TOL mRNAs were detected away from the genetic loci where they originated 10 suggests that they can migrate to translational sites from the 3D spot of the cell where they were 11 synthesized. Our data show also that physical forces and translation is crucial for bringing about this 12 scenario, which disappeared when in-house transcription was halted to reduce cellular crowding and 13 ribosome engagement. This picture departs from the standard view of transcription-translation coupling, 14 where RNAP and ribosomes get together in the so-called expressome complex (45, 46) which, obviously, 15 must operate in close proximity to the nucleoid's DNA encoding the gene(s) at stake. crescentus forces ribosomes to mRNAs and their translating ribosomes to remain close to their gene loci 28 in the chromosomal DNA (11,65,66). In contrast the low NC ratio of E. coli allows segregation of 29 ribosomes and mRNA away from their transcriptional sites. Pseudomonas sp. has also a low NC ratio 30 (65) and the ribosomes are spatially segregated from the nucleoid (20). It is thus conceivable that TOL transcripts are not coupled to translation as they are produced but they instead move after complete 1 transcription towards ribosome-rich cytoplasmic spots. 2 3 What could be the advantage of this setting of the xyl gene expression flow in P. putida as compared to 4 an alternative subcellular architecture? This species has bona fide nusA and nusG homologues encoded 5 in its genome. This indicates that transcription/translation coupling can indeed occur (45, 46), although 6 its global incidence as compared to E. coli or Bacillus (19) is unknown. It may well happen that having 7 movable xyl mRNAs eases assembly of the biodegradative complex encoded by the TOL operons for 8 degrading aromatic compounds. This is because the cognate catabolic enzymes, in particular those 9 which are membrane-bound (see above) could be synthesized at an optimal location site close to the site 10 of action. Moreover, thereby synthesized enzymes could be produced close to each other and enable 11 metabolic channeling (67). Finally, while Rho and transcription/translation coupling helps avoiding RNAP Culture conditions Unless otherwise indicated, E. coli and P. putida were routinely grown at 37°C and 21 30°C, respectively, in Luria-Bertani (LB) or M9 minimal medium (6 g l -1 Na2HPO4, 3 g l -1 KH2PO4, 1.4 g 22 l -1 (NH4)2SO4, 0.5 g l -1 NaCl, 0.2 g l -1 MgSO4·7H2O) with 10 mM succinate. Bacteria were cultured in 100-23 ml Erlenmeyer flasks with shaking at 170 rpm with 20 ml of the medium specified for the corresponding 24 experiment. Whenever necessary kanamycin (Km, 50 µg ml -1 ), ampicillin (Ap, 150 µg ml -1 ), gentamycin 25 (Gm, 10 µg ml -1 ) or chloramphenicol (Cm, 30 µg ml -1 ) was added to cultures of bacterial cells for ensuring 26 plasmid retention and maintenance of manipulated genotype. For the induction of TOL catabolic genes 27 during RNA-FISH experiment, P. putida strains, carrying the xyl genes either in the pWW0 plasmid or in 28 the chromosome, were overnight cultured in succinate amended M9 medium and then the bacterial 29 culture were 100-fold diluted in the same medium and grown until exponential phase (OD600 = 0.3-0.5). 30 (1/2 dilution in dibutylphthalate, which is a non-effector for TOL genes) in a flask for 2h. To validate 1 whether the output of the FISH approach matches the known TOL catabolic system, we cultured the mt-2 2 strain with other vaporous effectors such as toluene, o-xylene or soluble substrates such as benzoate 3 (5 mM) and 3MBz (5 mM). The transcription of the xyl genes by bacterial RNA polymerase was halted 4 by adding 200 µg ml -1 rifampicin (Rif) to the grown medium of the cognate sensitive strains.  Table S2). The resulting PCR 17 product was cloned into the pEMG by using restriction enzymes such as EcoRI and BamHI and ligation 18 within the same site of the vector yielded the plasmid pEMG-PuxT7, which was kept into the E. coli DH5a 19 lpir strain. This plasmid was transferred to P. putida mt-2 by triparental mating using the E. coli HB101 20 (pRK600) as helper strain (70, 71). Next, the pSW plasmid that expresses I-SceI endonuclease under 21 the Pm promoter (72) was introduced by electroporation into the pEMG-PuxT7 cointegrated strain and 22 thus had resistance for both Km and Ap. The clones were grown in LB medium with Ap (500 µg ml -1 ) and 23 3MBz (15 mM) to activate the Pm promoter, allowing I-SceI expression. The cells were plated on LB agar 24 and we confirmed the promoter replacement in the TOL plasmid, by testing the loss of the pEMG-PuxT7, 25 encoded Km resistance gene. Km-sensitive clones were selected and PCR further confirmed the 26 emergent colonies with T7F/ PuXT7-TS2R primer pairs (Supplementary Table S2).Then, the manipulated 27 plasmid pTOL-PuxT7 was either maintained in the mt-2 strain or transferred into the KT2440•T7 strain 28 by conjugation. 29 To label the pWW0 plasmid with the tandemly repeated tet operators, we inserted the arrays into the 1 orf105 locus of the plasmid. To this end, the 570-bp of PCR product was obtained using primer pairs 2 105F/ 105R (Supplementary Table S2) corresponding to the 5' portion of the orf105 gene of the pWW0 3 plasmid. Then, the amplified fragment was cloned into the HindIII and NotI cloning site of the pP30D-4 FRT-tetO vector (26), generating pP30D-FRT-tetO-orf105. The construct was either kept into the CC118 5 strain and subsequently transferred into P. putida mt-2 strain, in turn the backbone of the vector, including 6 the tetO arrays, was integrated into the pWW0 plasmid, generating the strain mt-2 (pTOL-tetO). Next, 3 µl of the mixed solution was applied onto poly-L-lysine coated cover slip, and stored at room 18 temperature to be dried on the cover slip. After putting the cover slip into a rack, this assembly was 19 immersed sequentially in methanol for 10 min and acetone for 30s at -20 °C. Once the coverslip was dry, 20 it was kept at 37 °C in 10% formamide solution (10% formamide, 2X saline-sodium citrate buffer (SSC) 21 in DEPC treated water, 2 mM VRC). After 60 min, the solution was removed and 50 µl of the hybridization 22 solution (10 % formamide, DEPC treated 2X SSC, 10 % dextransulphate, 2 mM VRC, 40 U RNase 23 inhibitor, and 250 nM CAL Fluor Red 610 labeled FISH probes; Stellaris TM , Biosearch Technologies, 24 Supplementary Tables S3) was spotted onto the cover slip and the sample was incubated in a dark and 25 humid chamber at 42 °C for the hybridization process. After overnight, the sample was then washed 26 twice with the solution (10% formamide and DEPC treated 2X SSC) for 15 min at 37 °C and DAPI (2.5 27 ug ml -1 ) staining was performed in second washing step. After brief rinse with PBS, the coverslip was 28 assembled with slide glass including antifade reagent Prolong (Invitrogen) and sealed by clear nail polish. 29 The specimen was visualized by fluorescent microscope. To identify cellular position of the TOL plasmid 30 carrying the tandem copies of the tet operators, DNA-FISH was exploited with the 6-carboxyfluorescein (FAM) labeled locked nucleic acid (LNA) tetO probe (5' 6-FAM-CTCTATCACTGATAGGGA; Bionova). It 1 is the same strategy as RNA-FISH, but the hybridization was carried out at two different temperatures: 2 at 95 °C to denature DNA for 2 min and followed by at 42 °C overnight in a dark and humid chamber. 3 The fluorescent signals representing xyl mRNAs and the tetO-tagged DNA were achieved by employing 4 sequentially combined RNA/ DNA-FISH. We followed the same procedure the RNA-FISH with xyl probe 5 sets prior to performing the DNA-FISH with the tetO probe. 6 7 Microscopy and image analysis. Microscopy was performed using an Olympus BX61 apparatus 8 equipped with ´100 phase contrast objective and a DP70 camera of the same brand. Signals for red-9 RNA, DAPI, and green-DNA were obtained using wide field excitation with following filters; MWIY2, U-  Quantification of mRNA-red foci with respect to the nucleoid occupation. Note that the number of red stop 12 per cell increased by ~ 30% in the Paw140 strain compared to that in the mt-2 strain (Fig. 2C). Scale bar, 2.5 µm. (C) Quantification of overlapping regions between mRNA-red signals and DAPI-blue 12 signals with sub-pixel precision. Over 100 exemplary cells were considered for the colocalization 13 analysis. 14 15 16