High Abundance of Transcription Regulators Compacts the Nucleoid in Escherichia coli

ABSTRACT

In enteric bacteria organization of the circular chromosomal DNA into a highly dynamic and toroidal-shaped nucleoid involves various factors, such as DNA supercoiling, nucleoid-associated proteins (NAPs), the structural maintenance of chromatin (SMC) complex, and macrodomain organizing proteins. Here, we show that ectopic expression of transcription regulators at high levels leads to nucleoid compaction. This serendipitous result was obtained by fluorescence microscopy upon ectopic expression of the transcription regulator and phosphodiesterase PdeL of Escherichia coli. Nucleoid compaction by PdeL depends on DNA-binding, but not on its enzymatic phosphodiesterase activity. Nucleoid compaction was also observed upon high-level ectopic expression of the transcription regulators LacI, RutR, RcsB, LeuO, and Cra, which range from single-target gene regulators to global regulators. In the case of LacI, its high-level expression in the presence of the gratuitous inducer IPTG (isopropyl-β-d-thiogalactopyranoside) also led to nucleoid compaction, indicating that compaction is caused by unspecific DNA-binding. In all cases nucleoid compaction correlated with misplacement of the FtsZ ring and loss of MukB foci, a subunit of the SMC complex. Thus, high levels of several transcription regulators cause nucleoid compaction with consequences for replication and cell division.

IMPORTANCE The bacterial nucleoid is a highly organized and dynamic structure for simultaneous transcription, replication, and segregation of the bacterial genome. Compaction of the nucleoid and disturbance of DNA segregation and cell division by artificially high levels of transcription regulators, as described here, reveals that an excess of DNA-binding protein disturbs nucleoid structuring. The results suggest that ectopic expression levels of DNA-binding proteins for genetic studies of their function but also for their purification should be carefully controlled and adjusted.

INTRODUCTION

Regulation of transcription in Escherichia coli involves a repertoire of approximately 300 transcription regulators, more than 90% of which have been functionally validated (1–3). These transcription regulators include single-target regulators such as LacI, solely regulating the lac operon, local regulators with up to 50 target genes, for example, RutR, a pyrimidine utilization repressor (4), and global regulators such as the catabolite activator repressor protein Cra and the pleiotropic regulator LeuO with more than 100 targets (5, 6). Nucleoid-associated proteins (NAPs) constitute another group of DNA-binding proteins; they are abundant and relevant for organization of the genomic DNA as a nucleoid and participate in the regulation of hundreds of targets genes (7).

PdeL, carrying a N-terminal FixJ/NarL/LuxR-type DNA-binding domain and a C-terminal EAL-type c-di-GMP-specific phosphodiesterase domain, is one of the validated transcriptional regulators with a small number of target loci, including the fliFGHIK operon, sslE, rfeE (wecA), and pdeL itself (8, 9). However, under laboratory growth conditions, expression of the pdeL gene, and concomitantly, PdeL protein levels, are low, due to repression of pdeL by the abundant nucleoid-associated and global repressor protein H-NS (9, 10). Genetic analyses, using complementation experiments with moderately elevated PdeL levels or upregulated pdeL mutants, corroborated its function as a transcription regulator and as an active c-di-GMP-specific phosphodiesterase (8, 9, 11). Furthermore, PdeL, as a dual-function protein, may represent a trigger enzyme whose role as a transcription regulator is controlled by c-di-GMP via the phosphodiesterase domain (12).

Considering the dual functions of PdeL, we used a fluorescent protein fusion, PdeL-mVenus, provided by a plasmid to analyze its cellular localization. Serendipitously, we found that ectopic expression of PdeL causes nucleoid compaction even upon weak induction of the P_ara promoter directing expression of pdeL. Weak induction nonetheless led to high levels of PdeL, which is a very stable protein. Further, other transcription regulators (LacI, RutR, RcsB, LeuO, and Cra) cause compaction of the nucleoid as well, when expressed and synthesized at similarly high levels. The data indicate that the mere occupation of the genomic DNA by an abundant DNA-binding protein can have severe effects on nucleoid structuring.

RESULTS

PdeL is nucleoid-associated and causes nucleoid compaction.

Here, we addressed whether the dual function protein PdeL is predominantly nucleoid-associated or localized otherwise. To this end, PdeL-mVenus was provided by a plasmid, since chromosomally encoded PdeL-mVenus is not detectable by fluorescence microscopy due to low expression of the pdeL gene (9). Initially, PdeL-mVenus localization was analyzed using plasmids with P_tac-directed pdeL-mVenus expression in ΔpdeL strain T2057 (Table 1). Fluorescence microscopy revealed a strong effect of PdeL expression on the nucleoid (as visualized by DAPI [4′,6-diamidino-2-phenylindole] staining) and on cell division when grown in LB medium and, to a lesser extent, when grown in tryptone medium (see Fig. S1 in the supplemental material). A DNA-binding mutant of PdeL, PdeL_HTH5M (9), had no such effect (Fig. S1).

TABLE 1.

E. coli K-12 strains

Strain	Genotype	Reference or constructiona
BW30270	MG1655 rph+ dgcJ::S1 flhDC(IS1); motile	37
T1241	BW30270 ilvG+	37
T2057	BW30270 ΔpdeLFRT	9
Donor strains for transduction
T2817	T1241 mukB-3×FLAGkan	T1241/pKD46 × OB68/OB69 (pSUB11)
T2818	T1241 mukB-mNGcm	T1241/pKD46 × OB66/OB67 (pKECY38)
U98	T1241 ΔpdeLcm	9
U65 and derivatives
U65	T1241 Δ(araC-BAD) Δlac(I-ZYA)FRT Pcp8araE ΔaraFGH flhDC+	13
U119	U65 ΔpdeLcm	U65 × T4GT7 (U98)
U121	U65 ΔpdeLFRT	U119 × pCP20
U148	U65 ΔpdeLFRT hupA-mCherrykan	U121/pKD46 × OA484/OA485 (pKECY15)
U159	U65 ΔpdeLFRT hupA-mCherryFRT	U148 × pCP20
U471	U65 ΔpdeLFRT hupA-mCherryFRT mukB-mNGcm	U159 × P1vir(T2818)
U477	U65 ΔpdeLFRT hupA-mCherryFRT mukB-mNGFRT	U471 × pCP20
U466	U65 mukB-3×FLAGkan	U65 × P1vir(T2817)
U473	U65 mukB-3×FLAGFRT	U466 × pCP20
U467	U65 ΔpdeLFRT mukB-3×FLAGkan	U121 × P1vir(T2817)
U474	U65 ΔpdeLFRT mukB-3×FLAGFRT	U467 × pCP20

^aStrains were constructed by transduction, which is stated as “× phage[donor strain]”; λ Red recombineering, stated as “× PCR primer pair (template),” followed by Flp recombinase-catalyzed deletion of the resistance marker, “× pCP20.”

In all further experiments, for being able to adjust pdeL expression levels, we used plasmids carrying pdeL-mVenus (and its variants) under the control of the P_ara promoter in strain U159 (ΔpdeL hupA-mCherry P_cp8araE ΔaraC-BAD ΔaraFGH) (Table 2). Strain U159 permits gradual, nonheterogeneous induction of P_ara by l-arabinose due to mutations of the ara regulon with constitutive expression of araE, encoding an arabinose transporter, and deletion of the arabinose araBAD and araFGH operons, as previously described (13, 14). Further, the HupA fluorescent protein fusions is a well-established marker for nucleoid staining (15, 16).

TABLE 2.

Plasmids

Plasmid	Featuresa	Construction or referenceb
pBAD30	araC Para MCS in p15A-ori; Ampr	38
pCP20	cI857 PR flp in pSC101-repts Ampr	39
pKD3	FRT-cmR-FRT in oriRγ Ampr	33
pKD4	FRT-kanR-FRT in oriRγ Ampr	33
pKD46	araC Para γ-β-exo in pSC101-repts Ampr	33
pSUB11	3xFLAG in pKD4	34
pKESK22	lacIq Ptac MCS in p15A-ori; Kanr	40
pKETS24	lacI PlacUV5 in pSC-ori Cmr	41
pKESL165	lacIq Ptac mVenus in pKESK22	9
pKESL166	lacIq Ptac pdeL-mVenus in pKESK22	9
pKESL209	lacIq Ptac pdeLHTH5M-mVenus in pKESK22	9
pKESL210	lacIq Ptac pdeLEVL-AAA-mVenus in pKESK22	pdeLEVL-AAA (flanking oligonucleotides T925/T972, mutagenesis oligonucleotides OA167/OA168) in pKESL165
pKEHB12	3×FLAG in pBAD30	3×FLAG (annealed oligonucleotidess T687/T906) in pBAD30
pKEHB23	araC Para mVenus in pBAD30	mVenus (pKESL165 EcoRI/XbaI) in pBAD30
pKECY1	araC Para pdeL-mVenus in pBAD30	pdeL (T925/T952) in pKEHB23
pKECY11	araC Para pdeLHTH5M-mVenus in pBAD30	pdeLHTH5M (T925/T952 from pKESL209) in pKEHB23
pKECY15	mCherry in pKD4	mCherry (PCR OA480/OA481) in pKD4
pKECY26	mNeonGreen in pSC-ori Cmr	9
pKECY38	mNeonGreen in pKD3	mNeonGreen (SalI, BamHI from pKECY26)
pKECY43	ftsZ-mNeonGreen in in pSC-ori Cmr	ftsZ (PCR OA807/OA808) in pKECY26 (XbaI, NdeI)
pKECY44	araC Para pdeL in pBAD30	9
pKECY52	araC Para pdeLHTH5M in pBAD30	9
pKECY53	araC Para pdeLEVL-AAA in pBAD30	9
pKECY81	araC Para pdeL-3×FLAG in pBAD30	pdeL (T925/OA116) in pKEHB12
pKECY90	lacI PlacUV5 ftsZ-mNeonGreen in pSC-ori Cmr	ftsZ-mNeonGreen (pKECY43 NcoI, XbaI) in pKETS24
pKECY91	araC Para pdeLHTH5M-3×FLAG in pBAD30	pdeLHTH5M (PCR T925/OA116, pKECY52) in pKEHB12
pKECY92	araC Para pdeLEVL-AAA-3×FLAG in pBAD30	pdeLEVL-AAA (PCR T925/OA116, pKECY53) in pKEHB12
pKECY95	araC Para rcsB-3×FLAG (native-RBS) in pBAD30	rcsB (PCR T358/S866) in pKEHB12
pKECY96	araC Para rcsB-3×FLAG (T7gene10-RBS) in pBAD30	rcsB (PCR OB93/S866) in pKEHB12
pKECY97	araC Para rcsB-3×FLAG (pdeL-RBS) in pBAD30	rcsB (PCR OB94/S866) in pKEHB12
pKECY98	araC Para rcsB (T7gene10-RBS) in pBAD30	rcsB (PCR OB93/T106) in pBAD30
pKECY99	araC Para lacI (T7gene10-RBS) in pBAD30	lacI (PCR OB155/OB156) in pKECY98
pKECY101	araC Para rutR (T7gene10-RBS) in pBAD30	rutR (PCR OB159/OB160) in pKECY98
pKECY102	araC Para leuO (T7gene10-RBS) in pBAD30	leuO (PCR OA005/T558) in pKECY98
pKECY103	araC Para cra (T7gene10-RBS) in pBAD30	cra (PCR OB161/OB162) in pKECY98

^aFeatures include Cm^r (chloramphenicol resistance), Kan^r (kanamycin resistance), MCS (multiple-cloning site), and pSC-rep^ts (temperature-sensitive replication, derivative of pSC101).

^bCloning was verified by sequencing of the cloned PCR fragments.

First, growth of transformants in dependence of the l-arabinose inducer concentration was determined (Fig. S2). The growth analyses revealed that the induction of pdeL expression caused a growth retardation, while induction of expression of the DNA-binding mutant, pdeL_HTH5M, did not impair growth (Fig. S2). Consequently, a concentration of only 2 μM l-arabinose was used for induction of PdeL, which had the least effect on growth (Fig. S2) and still affected nucleoid structuring (see below). In addition, we determined the l-arabinose concentration needed for inducing the synthesis of equal amounts of PdeL and the DNA-binding-defective PdeL_HTH5M by using 3×FLAG alleles carried on the same plasmidic expression system as used for fluorescence microscopy (Fig. S3). PdeL-3×FLAG gene expression was induced with 2 μM l-arabinose, while expression of PdeL_HTH5M-3×FLAG was induced by addition of increasing concentrations of l-arabinose. The amount of PdeL-3×FLAG (induced with 2 μM l-arabinose) and PdeL_HTH5M-3×FLAG (induced with 20 μM l-arabinose) were similar (Fig. S3). The low concentration of l-arabinose that is required for synthesis of significant amounts of PdeL indicates that PdeL synthesis is efficient and that the PdeL protein is stable. Determination of the protein stability of PdeL and PdeL_HTH5M showed that PdeL-3×FLAG was stable for 160 min after inhibition of translation by 100 μg/mL chloramphenicol (Fig. S3). In contrast, the PdeL_HTH5M-3×FLAG steady-state level was lower, and its level decreased approximately 2-fold in the 160 min after inhibition of translation, which suggests that PdeL_HTH5M is less stable than PdeL (Fig. S3). Note that protein stability was determined without induction to keep PdeL-3×FLAG levels sufficiently low for quantification (Fig. S3). The result of the protein stability assay is in accordance with the different concentrations of l-arabinose that are required for similar steady-state protein levels of PdeL-3×FLAG and PdeL_HTH5M-3×FLAG.

For fluorescence microscopy plasmidic P_ara-directed expression of pdeL-mVenus and pdeL_HTH5M-mVenus was induced with l-arabinose after 1 h of growth (t = 1 h), and samples were harvested 1 h and 2 h after induction (Fig. 1). PdeL-mVenus fluorescence was apparent after 1 h of induction and localized to the whole nucleoid (Fig. 1B and Fig. S4). Nucleoid-association of PdeL-mVenus was accompanied by nucleoid compaction and an enlargement of the bacteria. In contrast, PdeL_HTH5M-mVenus was located diffusely in the cell, and after 2 h of induction, aggregates of PdeL_HTH5M-mVenus near the cell pole became apparent. Taken together, the data suggest that the transcription regulator PdeL is nucleoid-associated, and they indicate that DNA-binding by high PdeL levels causes nucleoid compaction.

FIG 1.

PdeL is nucleoid-associated. Shown are representative composite fluorescence microscopy images of transformants of hupA-mCherry ΔpdeL strain U159 with plasmids pBAD30 (control), pKECY1 (PdeL-mV), and pKECY11 (PdeL_HTH5M-mV). Fluorescence of HupA-mCherry is shown in red, and that of PdeL-mVenus or PdeL_HTH5M-mVenus is shown in green. Transformants of hupA-mCherry ΔpdeL strain U159 with plasmids pBAD30 (control), pKECY1 (PdeL-mV), and pKECY11 (PdeL_HTH5M-mV) were inoculated to an OD₆₀₀ of 0.08 and grown in tryptone ampicillin medium at 37°C. After 1 h of growth (t = 1 h), P_ara-directed pdeL expression was induced with l-arabinose (+ara). Similar steady-state protein levels are synthesized upon induction with 2 μM l-arabinose in the case of pdeL-mVenus and 20 μM in the case of pdeL_HTH5M-mVenus, as tested using 3×FLAG variants of PdeL and PdeL_HTH5M (Fig. S3). Samples for microscopy were harvested from uninduced and induced cultures at t = 2 h and t = 3 h. The scale bar corresponds to 2 μm. Full-size images are shown in Fig. S4.

PdeL and RcsB compact the nucleoid and affect localization of MukB-mNeonGreen.

Since the nucleoid structure seems changed upon PdeL-mVenus expression, we tested localization of MukB, a subunit of the structural maintenance of chromosome (SMC) complex and a marker for oriC localization, using chromosomally encoded MukB-mNeonGreen (MukB-mNG) (17–20). In addition, FtsZ-mNG was used as a marker for septum formation (for results, see below). In this experimental approach, transformants expressing nontagged PdeL, the DNA-binding-defective PdeL_HTH5M, and an enzymatically inactive PdeL_EVL-AAA were studied (Fig. 2A). PdeL_EVL-AAA is enzymatically inactive due to mutation of the conserved c-di-GMP-specific EVL motif to three alanine residues (9). PdeL_EVL-AAA impaired growth similarly to the native PdeL protein (Fig. S2). In addition to PdeL, high-level expression of the two-component response regulator RcsB was included, which like PdeL carries a FixJ/NarL/LuxR-type DNA-binding domain, to analyze whether nucleoid compaction by high protein levels is PdeL-specific.

FIG 2.

High levels of PdeL, PdeL_EVL-AAA, and RcsB cause nucleoid compaction and loss of MukB-mNG foci. Transformants of strain U477 (hupA-mCherry mukB-mNG ΔpdeL) with plasmids pBAD30 (control), pKECY44 (PdeL), pKECY52 (PdeL_HTH5M), pKECY53 (PdeL_EVL-AAA), and pKECY98 (RcsB) were grown in tryptone ampicillin medium at 37°C. After 1 h of growth, P_ara-directed expression of pdeL and its mutants as well as of rcsB was induced by addition of l-arabinose, as indicated. Samples were harvested after 2 and 3 h of growth (t = 2h, t = 3h). (A) Protein levels were analyzed by 15% SDS-PAGE and Coomassie staining. Bands corresponding to PdeL proteins and to RcsB are indicated by filled and open triangles, respectively. (B) Representative sectors of microscopy images of transformants expressing no protein (control), PdeL, PdeL_HTH5M, PdeL_EVL-AAA, and RcsB. HupA-mCherry fluorescence is shown in red, and MukB-mNG foci are shown in green. Full-size microscopy images are shown in Fig. S6.

First, synthesis of RcsB levels that are similar to the levels of PdeL were established using plasmids carrying 3×FLAG-tagged rcsB alleles under P_ara control (Fig. S5). In these plasmids the efficiency of translation of rcsB-3×FLAG was varied. RcsB-3xFLAG levels were lowest with its native ribosomal binding site (RBS), higher when rcsB was fused to pdeL’s RBS, and similarly high as PdeL levels when using the phage T7 gene10 RBS and 50 μM l-arabinose for induction (Fig. S5).

Next, fluorescence microscopy was performed using transformants of strain U477 (hupA-mCherry mukB-mNG ΔpdeL) with plasmids coding for PdeL, PdeL_HTH5M, PdeL_EVL-AAA, and RcsB, respectively. These transformants synthesized approximately equal amounts of PdeL, PdeL_HTH5M, PdeL_EVL-AAA, and RcsB upon induction by l-arabinose, as shown by SDS-PAGE (Fig. 2A). Induction of high levels of PdeL, PdeL_EVL-AAA, and RcsB synthesis caused nucleoid compaction, as suggested by the compact HupA-mCherry signal of increased intensity, and a moderate increase of the cell length (Fig. 2B). Furthermore, one or two MukB-mNG foci close to oriC were detectable without induction and in the control (Fig. 2B and Fig. S6), as previously shown for fluorescent protein MukB fusions (18). MukB-mNG foci disappeared upon induction of PdeL, PdeL_EVL-AAA, and RcsB expression but were not changed in case of the DNA-binding-deficient PdeL_HTH5M (Fig. 2B). The disappearance of MukB-mNG foci was not caused by a change in the MukB protein levels, which remained constant (Fig. S7). Taken together, high protein levels of PdeL and RcsB, but not of the DNA-binding-defective PdeL_HTH5M led to nucleoid compaction and loss of MukB-mNG foci.

High levels of PdeL and RcsB proteins lead to misplacement of the FtsZ ring.

As a second marker, we tested whether localization of cell division protein FtsZ is affected by synthesis of high levels of PdeL and RcsB. For visualization of the Z-ring in the ftsZ⁺ background, FtsZ-mNG was provided by a low-copy-number plasmid carrying a P_lacUV5 ftsZ-mNG cassette, as previously described (21). The low activity of the noninduced P_lacUV5 promoter was sufficient for production of detectable amounts of FtsZ-mNG. Fluorescence microscopy was performed with double transformants of the hupA-mCherry ΔpdeL strain U159 with the ftsZ-mNG plasmid and compatible pdeL or rcsB carrying plasmids (Fig. 3). Depending on the progression of the cell cycle, FtsZ-mNG was detectable at the midcell forming the Z-ring in the case of the control and the PdeL_HTH5M DNA-binding mutant (Fig. 3 and Fig. S8). Localization of FtsZ-mNG was different when PdeL and when RcsB were expressed at high levels. Nucleoid compaction led to displacement of FtsZ to positions which are devoid of the nucleoid (Fig. 3 and Fig. S8), in agreement with nucleoid exclusion of FtsZ-ring formation (22). Taken together, PdeL, PdeL_EVL-AAA, and RcsB, but not the DNA-binding-defective PdeL_HTH5M, can lead to FtsZ-mNG displacement (Fig. 3 and Fig. S8).

FIG 3.

PdeL, PdeL_EVL-AAA, and RcsB misplaced from mid-cell FtsZ rings. For visualization of the Z-ring a C-terminally mNeonGreen-tagged FtsZ variant, FtsZ-mNG, was ectopically expressed under the control of P_lacUV5 promoter using low-copy-number plasmid pKECY90. Cotransformants of strain U159 (hupA-mCherry ΔpdeL) with plasmid pKECY90 (P_lacUV5 ftsZ-mNG) and plasmids pBAD30 (control), pKECY44 (PdeL), pKECY52 (PdeL_HTH5M), pKECY53 (PdeL_EVL-AAA), and pKECY98 (RcsB) were grown in tryptone ampicillin chloramphenicol medium at 37°C. After 1 h of growth, P_ara-directed expression was induced by l-arabinose, as indicated, and samples were harvested at t = 2 h and t = 3 h. Shown are representative sectors of microscopy images with HupA-mCherry (red) and FtsZ-mNG (green). Contrast settings were adjusted to 180 to 700 (FtsZ-mNeonGreen) and 200 to 1,500 (HupA-mCherry). Full-size images are shown in Fig. S8.

High levels of the transcription regulators LacI, RutR, LeuO, and Cra also cause nucleoid compaction.

Since high levels of both PdeL and RcsB affect the nucleoid structure and localization of MukB and FtsZ, we tested additional DNA-binding proteins. This included the single-target transcription regulator LacI, the local regulator RutR, and the global regulators LeuO and Cra (6). Plasmidic P_ara-directed expression of these transcription regulators was adjusted to obtain approximately equal amounts of each protein, as validated by SDS-PAGE (Fig. S9). Fluorescence microscopy demonstrated that high levels of all transcription regulators, LacI, RutR, LeuO, and Cra, caused nucleoid compaction as well (Fig. 4 and Fig. S10). In addition, we found that induction of the synthesis of high levels of LacI without and in the presence of IPTG caused the same phenotype (Fig. 5 and S11). Specific DNA-binding of LacI is inhibited by IPTG (23). In all cases the number of MukB-mNG foci was significantly lower (Fig. 4 and 5 and Fig. S10 and S11). Taken together, all tested transcription regulators, LacI, RutR, LeuO, and Cra, caused nucleoid compaction and a decrease of MukB-mNG foci similar to PdeL and RcsB.

FIG 4.

High levels of the transcription regulators LacI, RutR, LeuO, and Cra cause nucleoid compaction and loss of MukB foci. Transformants of strain U477 (hupA-mCherry mukB-mNG ΔpdeL) with plasmids pBAD30 (control), pKECY44 (PdeL), pKECY99 (LacI), pKECY101 (RutR), pKECY102 (LeuO), and pKECY103 (Cra) were grown in tryptone ampicillin medium at 37°C. After 1 h of growth, P_ara-directed expression of transcription regulators was induced with l-arabinose, as indicated. Expression of similar amounts of the transcription regulators was analyzed by SDS-PAGE (Fig. S9). Shown are representative sectors of microscopy images (full-size images are shown in Fig. S10).

FIG 5.

Nonspecific DNA-binding by LacI causes nucleoid compaction. To test whether nonspecific DNA-binding by LacI causes nucleoid compaction, IPTG was added to cultures of transformants of strain U477 (hupA-mCherry mukB-mNG ΔpdeL) with plasmid pKECY99 (LacI). After 1 h of growth in tryptone ampicillin medium at 37°C, lacI expression was induced with 2 μM l-arabinose, and IPTG was supplemented to a final concentration of 1 mM, where indicated. Samples were harvested after 1 h (t = 2 h) and 2 h (t = 3 h) of induction. Shown are representative sections of microcopy images (full-size images are shown in Fig. S11).

For RutR, LeuO, and Cra we also tested FtsZ localization using a low-copy-number plasmid carrying ftsZ-mNG under the control of the LacI-regulated lacUV5 promoter (Fig. 6 and Fig. S12). In these experiments LacI could not be included, since high levels of LacI led to complete inhibition of ftsZ-mNG expression. Fluorescence microscopy of the transformants expressing high levels of RutR, LeuO, and Cra caused misplacement of the FtsZ similarly as PdeL and RcsB (Fig. 6).

FIG 6.

Localization of cell division protein FtsZ upon overexpression of transcription regulators RutR, LeuO, and Cra. Cotransformants of strain U159 (hupA-mCherry ΔpdeL) with plasmids pKECY90 (FtsZ-mNG) and pKECY44 (PdeL), pKECY101 (RutR), pKECY102 (LeuO), and pKECY103 (Cra) were grown in tryptone medium supplemented with chloramphenicol and ampicillin. After 1 h of growth, P_ara-directed expression of transcription regulators genes were induced with l-arabinose, as indicated, and samples were harvested after 2 h (t = 2 h) and 3 h (t = 3 h) of growth. Representative sectors of microscopy images with FtsZ-mNG (in green) and HupA-mCherry (in red). Full-size images are shown in Fig. S12.

Lastly, the amount of PdeL that is synthesized upon induction of P_ara by 2 μM l-arabinose was estimated using purified PdeL-His₆ as a reference (Fig. S13). After 1 h of induction, approximately 40,000 PdeL monomers and after 2 h of induction approximately 150,000 monomers of PdeL are present per cell, which corresponds to 1 PdeL dimer per 250 bp after 1 h and 4 PdeL dimers per 250 bp after 2 h of induction (assuming that a single nonreplicating nucleoid is present).

DISCUSSION

Here, we have shown that high levels of the transcription regulators PdeL, RcsB, LacI, RutR, Cra, and LeuO lead to nucleoid compaction in E. coli. These transcription regulators include single-target and local regulators with only one or few specific DNA-binding sites and global regulators with hundreds of DNA-binding sites in the genome (6). Our data suggest that nucleoid compaction is caused by nonspecific occupancy of the genomic DNA. Comparable observations of nucleoid compaction have been described recently for the phage T4 protein MotB and for the bacterial DNA-binding toxin SymE (24, 25).

The dual-function protein PdeL, a transcription regulator and c-di-GMP-specific phosphodiesterase, is nucleoid-associated. Further, ectopic expression of pdeL directed by weak induction of P_ara resulted in a high cellular protein level, apparently because PdeL is a protein of high stability. The high level of PdeL caused nucleoid compaction. Likewise, the two-component response regulator RcsB, which carries a FixJ/NarL/LuxR-type DNA-binding domain like PdeL, caused nucleoid compaction when expressed at similarly high levels as PdeL. Other transcription regulators led to nucleoid compaction as well, and this is independent of the number of specific DNA-binding sites, as shown with the single-target regulator LacI, the local regulator RutR, and the global regulators Cra and LeuO (5, 6). In the case of LacI, nucleoid compaction is independent of specific DNA-binding as shown using the gratuitous inducer IPTG.

The high level of PdeL with approximately 40,000 molecules per cell 1 h after induction corresponds to 1 dimer per 250 bp, while the approximately 150,000 molecules present after 2 h of ectopic synthesis would theoretically allow complete coverage of the genome. The cellular levels of the other transcription regulators tested in this study are comparable. Occupancy of the whole genome by a DNA-binding protein could hinder DNA replication and/or transcription. In the case of the toxin SymE, it has been shown that toxicity of symE overexpression is presumably based on nucleoid condensation and inhibition of DNA synthesis as well as RNA synthesis (25). Further, these authors have shown that nucleoid compaction caused by synthesis of SymE at high levels is similar to nucleoid compaction caused by overexpression of H-NS (25, 26). Interestingly, in the case of phage T4, a similar mechanism is possibly utilized to reprogram transcription of the whole E. coli genome. Phage T4 encodes the protein MotB, which shortly after infection is synthesized at very high levels corresponding to 40,000 monomers per cell (24). Fluorescence imaging using a MotB-green fluorescent protein (GFP) fusion demonstrated nucleoid compaction (24) similar to the data shown in this work. Interestingly, MotB has dramatic effects on the transcriptome, leading to a relative increase of 1/8 of all transcripts in relation to the whole transcriptome, and of these transcripts ~70% correspond to H-NS-repressed genes (24, 27). Thus, phage T4 apparently employs nucleoid restructuring by the abundance of MotB to reprogram the host transcriptome (24).

Remarkably, nucleoid-associated proteins (NAPs) such as H-NS, HU, IHF, FIS, and others are abundant (28), but at their natural level they do not cause nucleoid compaction. H-NS, present at approximately 20,000 molecules per cell, binds to the minor groove of AT-rich DNA and forms linear and bridged filaments. HU and IHF are abundant DNA-bending proteins that presumably contribute to nucleoid structuring and formation of specific regulatory nucleoprotein complexes. FIS is expressed at very high levels with 50,000 molecules per cell during the early exponential-growth phase only; it is a DNA-bending protein and organizer of plectonemic structures (7, 29). The cellular levels of the abundant NAPs are apparently well adjusted to their function.

Our data suggest that compaction of the nucleoid by high levels of transcription regulators can displace Z-rings from midcell, which is presumably an indirect effect mediated by the nucleoid occlusion system and SlmA protein (22). Similarly, the loss of detectable MukB-mNG foci near the origin of replication can be an indirect consequence of nucleoid compaction, as it is possible that DNA replication is put on hold by excessive occupancy of the genome by high levels of transcription regulators, as shown for SymE (25). In accordance with this, nucleoid compaction is observed first (1 h after induction), followed by the change in positioning of MukB and FtsZ, and subsequent growth retardation.

The serendipitous finding reported here emphasizes the importance of using native protein levels in functional analysis of transcription regulators. Ectopic expression can be required for a functional analysis, as for example, when conditions leading to expression of the gene encoding a transcription regulator and an inducer are unknown. In the case of transcription regulators that are stable, such as PdeL, ectopic expression even for a short time and with a low inducer concentration may yield far too high protein levels. Thus, if an ectopic expression system is used, the rate and duration of synthesis as well as the protein stability and cellular level should be controlled carefully. Another aspect to be considered when using ectopic expression is that the level of a transcription regulator and the number of its specific DNA-binding sites in the genome is balanced (30).

MATERIALS AND METHODS

Bacterial strains, media, and plasmids.

E. coli K-12 strains and their construction are described in Table 1. Strains were constructed by transduction using phage T4GT7 or P1 vir, and by lambda Red-mediated recombineering (31–34). Plasmids are listed in Table 2, and oligonucleotides used for construction of strains and plasmids are listed in Table S1. Bacteria were cultured in LB medium (10 g/L tryptone, 5 g/L yeast extract, and 5 g/L NaCl), tryptone medium (10 g/L tryptone, 5 g/L NaCl), SOB medium (20 g/L tryptone, 5 g/L yeast extract, 0.5 g/L NaCl, 2.5 mM KCl, 10 mM MgCl₂, pH 7.0), or SOC (SOB with 0.4% glucose). For plates, 15 g/L agar was added. Antibiotics were added to a final concentration of 50 μg/mL ampicillin, 15 μg/mL chloramphenicol, and 15 μg/mL kanamycin, as required; IPTG and l-arabinose were added as described in the figures.

Fluorescence microscopy.

For fluorescence microscopy, transformants were inoculated to an optical density at 600 nm (OD₆₀₀) of 0.08 in LB or tryptone medium supplemented with appropriate antibiotics and grown at 37°C with shaking as specified in the figure legends. Plasmid-directed expression of the transcription regulator genes was induced after 1 h of growth by adding IPTG or l-arabinose at the concentration stated in the figure legends. Bacteria were harvested just before induction (t = 1 h) and after induction by pelleting 500 μL of the culture by centrifugation at 5,900 relative centrifugal force (rcf) for 1 min. The bacterial pellets were resuspended in 150 μL of fresh tryptone medium, and 4 μL of these suspensions was spotted onto 1% agarose pads on microscopy slides for microscopy. In the case of DAPI staining, the 4 μL of cells was incubated with 1 μL 10 of mg/mL DAPI for 1 min before spotting on for microscopy. Image acquisition was performed using a Zeiss Axio Imager.M2 microscope with an EC Plan-Neofluar 100×/1.30 oil Ph3 M27 lens objective. Images were captured and processed using ZEN 2012 software (Carl Zeiss Microscopy GmbH, Germany). Excitation times were 500 ms for PdeL-mVenus fusions, MukB-mNG, FtsZ-mNG, and HupA-mCherry. Contrast settings were 1 to –16,384 for phase contrast, 200 to 1,500 for HupA-mCherry, 2,000 to 16,384 for DAPI, 1 to 16,384 for PdeL-mVenus, 180 to 300 for MukB-mNG, and 180 to 1,000 for FtsZ-mNG unless otherwise stated in the figures.

PdeL protein stability analyses.

For determining the protein stability of PdeL and its variants, transformants of E. coli strain U121 (U65 ΔpdeL_FRT) with plasmids pKECY81 (P_ara pdeL-3×FLAG) and pKECY91 (P_ara pdeL_HTH5M-3×FLAG) were inoculated to an OD₆₀₀ of 0.08 in tryptone medium supplemented with ampicillin. The transformants were grown for 2 h at 37°C without induction of P_ara to preserve a low protein level for quantification by Western blotting. Translation was blocked by adding chloramphenicol to a final concentration of 100 μg/mL. Samples of 2-mL volume were taken just prior and at several time points after inhibition of translation. Bacteria were pelleted by centrifugation and resuspended in Laemmli buffer for detection of epitope-tagged protein by Western blotting (35).

Protein detection by Western blotting and Coomassie staining.

SDS-PAGE, Coomassie staining, and Western blotting were performed as described previously (35). Unless otherwise described, bacteria equivalent to an OD₆₀₀ of 0.08 were loaded per lane. Equal loading was validated by 2,2,2-trichloroethane (TCE) staining of total protein (36). Epitope-tagged 3×FLAG proteins were detected using primary antibody anti-FLAG M2 from mouse (diluted 1:4,000, catalogue number F3165; Sigma-Aldrich, Germany) and Alexa Fluor 680 fluorescent dye-labeled secondary anti-mouse antibody from goat (diluted 1:10,000, catalogue number A21057; Thermo Scientific, Germany). Quantification of protein bands was performed with Odyssey V3.0 software for Western blots and with ImageLab (Bio-Rad, Germany) for Coomassie-stained gels.