Journal of Structural Biology 178 (2012) 260–269

Contents lists available at SciVerse ScienceDirect

Journal of Structural Biology
journal homepage:

Characterization of Escherichia coli nucleoids released by osmotic shock
Anna S. Wegner a,b, Svetlana Alexeeva a,1, Theo Odijk b,c,⇑, Conrad L. Woldringh a,⇑

Swammerdam Institute for Life Sciences, BioCentrum Amsterdam, University of Amsterdam, Kruislaan 316, 1098 SM Amsterdam, The Netherlands Section Theory of Complex Fluids, Kluyver Institute for Biotechnology, Delft University of Technology, Julianalaan 67, 2628 BC Delft, The Netherlands c Lorentz Institute for Theoretical Physics, Niels Bohrweg 2, 2333 CA Leiden, The Netherlands

a r t i c l e

i n f o

a b s t r a c t
Nucleoids were isolated by osmotic shock from Escherichia coli spheroplasts at relatively low salt concentrations and in the absence of detergents. Sucrose-protected cells, made osmotically sensitive by growth in the presence of ampicillin or by digestion with low lysozyme concentrations (50–5 lg/ml), were shocked by 100-fold dilution of the sucrose buffer. Liberated nucleoids stained with 40 ,6-diamidino-2phenylindole dihydrochloride hydrate (DAPI), the dimeric cyanine dye TOTO-1, or fluorescent DNA-binding protein appeared as cloud-like structures, in the absence of phase contrast. Because UV-irradiation disrupted the DAPI-stained nucleoids within 5–10 s, they were imaged by time-lapse microscopy with exposure times less than 2 s. The volume of nucleoids isolated from ampicillin- or low-lysozyme spheroplasts and minimally exposed to UV (<2 s) was on average $42 lm3. Lysozyme at concentrations above 1 lg/ml in the lysate compacted the nucleoids. Treatment with protease E or K (20–200 lg/ml) and sodium dodecyl sulfate (SDS; 0.001–0.01%) caused a twofold volume increase and showed a granular nucleoid at the earliest UV-exposure; the expansion could be reversed with 50 lM ethidium bromide, but not with chloroquine. While DNase (1 lg/ml) caused a rapid disruption of the nucleoids, RNase (0.1–400 lg/ml) had no effect. DAPI-stained nucleoids treated with protease, SDS or DNase consisted of granular substructures at the earliest exposure similar to UV-disrupted nucleoids obtained after prolonged (>4 s) UV irradiation. We interpret the measured volume in terms of a physical model of the nucleoid viewed as a branched DNA supercoil crosslinked by adhering proteins into a homogeneous network. Ó 2012 Elsevier Inc. All rights reserved.

Article history: Received 29 November 2011 Received in revised form 2 March 2012 Accepted 3 March 2012 Available online 6 April 2012 Keywords: Escherichia coli Nucleoid Supercoiling Polymer physics Lysozyme-spheroplast Ampicillin-spheroplast Osmotic shock DAPI/UV-radiation damage Protein cross-links Ethidium bromide

1. Introduction The bacterial nucleoid has been studied extensively both in situ and as an isolated structure. Electron microscope observations of thin sections showed a meshwork of aggregated DNA fibers in well-delineated or dispersed regions in which the DNA must have suffered structural changes due to fixation and dehydration (for reviews see Robinow and Kellenberger, 1994; Woldringh and Nanninga, 1985). Additional information was obtained from isolated nucleoids prepared by the so-called cytochrome-c-spreading technique, showing for the first time supercoiled, branched DNA loops radiating out of a ‘‘spider-like’’ core (Delius and Worcel, 1974; Kavenoff and Bowen, 1976; Meijer et al., 1976). These structures have frequently been interpreted as resulting from protein and RNA cross-links and have led to the so-called ‘‘rosette’’ model
⇑ Corresponding authors. Address: Faculty of Science, Swammerdam Institute for Life Sciences, University of Amsterdam, Science Park 904 (Room C2. 267), 1098 XH Amsterdam, The Netherlands (C.L. Woldringh). E-mail addresses: (T. Odijk), (C.L. Woldringh). 1 Current address: Laboratory of Food Microbiology, Wageningen University, Bomenweg 2, 6703 HD Wageningen, The Netherlands.
1047-8477/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved.

for a folded chromosome (see for review Toro and Shapiro, 2010). The folds or domains are considered to represent independent supercoiled loops, but the nature of possible crosslinks that form the supercoiling barriers and that determine the size of isolated nucleoids has remained obscure. When analyzing the various procedures historically used to release the bacterial nucleoid, we may discern two agents that seem to have been important in determining its size and structure: (i) the presence of detergents and (ii) the concentration of lysozyme in protocols without detergent. Since the introduction of the first protocol by Stonington and Pettijohn (1971), detergents (non-ionic Brij-58 and anionic deoxycholate) at a high salt concentration (1 M NaCl) or spermidine have been used in most studies (Worcel and Burgi, 1972, 1974; Kornberg et al., 1974; Drlica and Worcel, 1975; Meijer et al., 1976; Materman and van Gool, 1978; Murphy and Zimmerman, 2000; Foley et al., 2010). Depending on the lysis temperature, the detergent-salt method produced membrane-attached (at 10 °C) or membrane-free (at 25 °C) ‘‘particles’’ as analyzed by sedimentation through sucrose gradients (Worcel and Burgi, 1974) or by ‘‘visualization’’ with the electron microscope (Delius and Worcel, 1974). The results supported the interpretation that the released nucleoids were intact and supercoiled as

A.S. Wegner et al. / Journal of Structural Biology 178 (2012) 260–269


addition of ethidium bromide caused them to relax at a concentration of 2 lg/ml or re-compact at higher concentrations (>4 lg/ml; Worcel and Burgi, 1972). However, the hydrodynamic properties determined from sedimentation in sucrose gradients showed large variations (Hecht et al., 1977; Odijk, 2002). In part, these could be caused by entanglement with envelope fragments (Meijer et al., 1976; Materman and van Gool, 1978). The theoretical analysis (Odijk, 2002) of the sedimentation data of nucleoids obtained by the detergent method (Hecht et al., 1977) proves that their dimensions are typically of the order of 0.5 lm (see also Hecht et al., 1975). This is much smaller than the nucleoids obtained by the osmotic shock method (Cunha et al., 2001b; this work). The implication is that we have to be careful in interpreting the response to enzymatic treatments in the respective isolation methods. Nucleoids released by the detergent method were found to be sensitive to treatment with RNase (Worcel and Burgi, 1972; Pettijohn and Hecht, 1973) and proteases (Drlica and Worcel, 1975) possibly indicating the existence of crosslinks. Moreover, detergents can be expected to dissociate proteins from the DNA and they could therefore influence subsequent structural transitions upon enzyme treatments. To circumvent such problems, Sloof et al. (1983) introduced an osmotic shock procedure without detergents using protoplasts of Bacillus licheniformis. In a previous study (Cunha et al. (2001a) we compared this method with the lysis with detergents using the same Escherichia coli spheroplast suspension. We observed that both procedures resulted in comparable sizes of the isolated nucleoids visualized by fluorescence microscopy, in spite of the large difference in salt concentrations (1 M NaCl in the detergent method versus 10 mM NaCl in the osmotic shock method). In subsequent studies (Cunha et al., 2001b) we emphasized the osmotic shock method and we determined the compaction of isolated nucleoids by polyethyleneglycol (PEG) and measured the dynamics of fluorescently labeled DNA regions in the isolated structures (Cunha et al., 2005). We found that when released by osmotic shock, the nucleoids expanded like a spring into cloud-like structures with an average volume of about 30 lm3. Romantsov et al. (2007) obtained nucleoids by osmotic shock in the shape of rounded, delineated structures with a volume of about 18 lm3. These nucleoid volumes are well below the theoretical value of un-crosslinked nucleoids expanded by excluded volume effects and estimated to be about ten times larger (Cunha et al., 2001b). What causes this variation in size of isolated nucleoids and what prevents their full expansion? As lysozyme is known to be a basic protein (Kuehner et al., 1999) it could well bind to DNA causing nucleoid compaction. In this work we have therefore refined previous protocols and we have prepared spheroplasts either in the absence of lysozyme, though by growing cells in the presence of ampicillin, or by using much lower concentrations of lysozyme (as low as 0.5–0.05 lg/ ml in the lysate). The goal of nucleoid isolation is to prepare nucleoids that are expanded versions of their counterparts in vivo. Ideally, they should not be perturbed by the compounds used in the isolation protocol. We hope that our new protocols are a further step in realizing this ideal. The present nucleoids obtained by osmotic shock and stained with DAPI, appear to expand faster than previously described (Cunha et al., 2001a,b). We have therefore measured their sizes from images taken by time-lapse microscopy within 2 s of UV-exposure. The cloudlike structures appear to be larger (42 lm3) and to fall apart fast during UV-irradiation as a network of granules with what appears to be Brownian motion superposed. A similar expansion of granular substructures was observed after treatment with protease, sodium dodecyl sulfate (SDS) or DNase. In contrast to studies using the detergent method (Worcel and Burgi, 1972), RNase appeared to have no effect. We show that our current meth-

od of preparing and imaging low-lysozyme or lysozyme-free nucleoids is particularly suitable for establishing their dimensions. These sizes will be discussed in terms of a physical model in which the nucleoid is considered to be a branched DNA supercoil crosslinked by adhering proteins into a homogeneous network (see Fig. 5 and Appendix B for a description of this model). 2. Materials and methods 2.1. Strains and growth conditions Strains E. coli K-12 MC4100 (laboratory strain LMC500) and E. coli 500/pSACT11 were grown at 28 °C in glucose minimal medium containing 6.33 g of K2HPO4.3H2O, 2.95 g of KH2PO4, 1.05 g of (NH4)2SO4, 0.10 g of MgSO4.7H2O, 0.28 mg of FeSO4.7H2O, 7.1 mg of Ca(NO3)2.4H2O, 4 mg of thiamine, 4 g of glucose and 50 mg of lysine, per liter pH 7.0 at 28 °C (doubling time at 28 °C was about 85 min). NaCl was added to adjust the osmolarity of the medium to 300 mosM (Micro-Osmometer, Advanced Instruments). Cell growth was monitored at 450 nm with a spectrophotometer. Exponential growth was maintained by periodical dilutions of the culture. For E. coli LMC500, we estimate the number of chromosome equivalents per nucleoid to be about 1.58 (see e.g. Huls et al., 1999). E. coli 500/pSACT11, grown in the presence of 100 lg/ml ampicillin, was used so as to visualize nucleoids with the fluorescent fusion protein HupA-mRFP (red fluorescent protein) after induction with 0.3 mM IPTG. Strain E. coli K-12 FH2973/pFHC2973 (obtained from F. Hansen, Biocentrum DTU, Denmark. See Nielsen et al., 2006) was grown in the same medium with 0.5% glycerol as carbon source and supplemented with 100 lg/ml ampicillin (doubling time at 28 °C was about 120 min). When 0.58 M sucrose was added to E. coli LMC500 growing in glucose-minimal medium (see below) the doubling time at 28 °C increased to $120 min. When indicated (as TY) the glucose minimal medium was supplemented with 10 g of bactotryptone, 5 g of yeast extract, 5 g of NaCl and 15 mmol NaOH per liter giving a doubling time at 28 °C of about 60 min. 2.2. Plasmid construction A HupA-mRFP protein fusion was constructed using standard laboratory techniques (Miller, 1992). mRFP from pGEM-XbBgmRFPsv (Alexeeva et al., 2010) was used to create a mRFP fusion to the C-terminus of HupA, amplified from chromosomal E. coli MC4100 DNA. See Supplementary data (S2.1) for details. 2.3. Chemicals DAPI was obtained from Sigma and was prepared by dissolving 500 lg/ml in distilled water with the help of sonication. TOTO-1 (dimeric cyanine dye) and FM4-64 were obtained from Molecular Probes (Invitrogen), sodium dodecyl sulfate from Koch-Light Laboratories and glutaraldehyde from BDH (UK). RNase A was purchased from Sigma, bovine serum albumin (BSA) and protease E from Calbiochem, protease K from Invitrogen, pancreatic Dnase I from Roche Diagnostics. 2.4. Preparation of lysozyme-spheroplasts Except for the application of a cold-shock (4 °C) to the cells before lysozyme addition, the protocol was that described essentially by Cunha et al. (2005). In short: 2 ml of E. coli cells grown in glucose minimal medium (Gmin) to an OD450 of 0.2 ($5 Â 107 cells per ml) were centrifuged in an Eppendorf centrifuge for 5 min at 10,000 rpm. The cell pellets were resuspended in 475 ll of


A.S. Wegner et al. / Journal of Structural Biology 178 (2012) 260–269

sucrose-buffer consisting of 0.58 M sucrose, 10 mM Na2HPO4/ NaH2PO4-buffer pH 7.4 (hereafter written as NaPi), 10 mM EDTA and 100 mM NaCl, at a temperature of 4 °C. After this cold-shock, 25 ll of a freshly dissolved lysozyme solution (1 mg/ml, or, as indicated, 0.1 mg/ml) was added giving a final concentration of 50 or 5 lg/ml. Within 5–10 min of incubation at room temperature (RT) a suspension of >95% spheroplasts was obtained. Osmotic shock (see Section 2.6) was applied 10–15 min after lysozyme addition. Additional details of the protocol are given in Supplementary data (S2.2). 2.5. Preparation of ampicillin spheroplasts Ten milliliter of a suspension of E. coli cells growing exponentially in glucose minimal medium (doubling time 90 min at 28 °C) were centrifuged and resuspended in 10 ml of glucose minimal medium containing 0.58 M sucrose. After 4 h of adaptation and growth (doubling time 120 min), ampicillin (pI = 4.95; Hout and Poole, 1969) was added to the culture at a final concentration of 5 mg/ml. Microscopic inspection showed that in 60–90 min most cells either displayed central bulges or had converted to spheroplasts. The suspension (4 ml) was then centrifuged in an Eppendorf centrifuge for 10 min at 2500 rpm and the pellet resuspended at RT in 2 ml of the sucrose-buffer, mentioned above. Faster centrifugation resulted in breakage of spheroplasts and expansion of prematurely liberated nucleoids. The ampicillinspheroplast suspension was kept at RT until osmotic shock. 2.6. Osmotic shock procedure Osmotic shock of either lysozyme- or ampicillin-spheroplasts was performed by adding 10 ll of spheroplast suspension to 990 ll of 10 mM NaPi buffer (pH 7.0) in an Eppendorf tube. This represents a hypotonic shock by a 100-fold dilution of the sucrose buffer giving a final concentration of 1 mM NaCl and of 0.5 or 0.05 lg/ml lysozyme in the lysate. After carefully inverting the tube once, the lysate was kept at RT until staining and microscopic preparation. Nucleoids obtained from either lysozyme- or ampicillin-spheroplasts will be called ‘‘lysozyme-nucleoids’’ or ‘‘ampicillin-nucleoids’’, respectively, throughout. 2.7. Microscopy Lysates containing liberated nucleoids were stained with DAPI at a final concentration of 0.1 lg/ml. For comparison, TOTO-1 was used at a final concentration of 2 lM. The stained suspensions were either observed under a coverslip to observe both ghosts and nucleoids or as a hanging-drop preparation to observe free-floating nucleoids. Application of a cover slip could sometimes cause adherence of nucleoids to the glass surface or nucleoid disruption by liquid motion. For a hanging-drop preparation a coverslip with 20 ll of the lysate was inverted, placed over a hole in a metal object slide and fixed with two stickers. Nucleoid suspensions were photographed with a Coolsnap fx charge-coupled device camera using an Olympus BX60 microscope equipped with a 100Â UplanFL 1.3 oil immersion phase contrast lens. DAPI fluorescence images were taken with an Olympus cube unit (U-MWU) with a single band-pass excitation filter (330–385 nm) and a long-pass barrier filter (>420 nm). (The absorption maximum of DAPI is at 358 nm, close to the near UV-A line at 365 nm of the mercury vapor lamp.) For TOTO-1, HupA-RFP or ethidium bromide (EtBr) fluorescence images, an Olympus U-MNG cube unit was used with an excitation filter of 530–550 nm, a dichroic mirror of 570 nm, and a long-pass emission filter of >590 nm. TOTO-1 was sometimes also excited with an Olympus cube unit U-MNB with a single band excitation filter

(470–490 nm), dichroic mirror of 500 nm, and a long-pass emission filter >515 nm. The size of the isolated nucleoids was estimated as described previously using the Object-Java software package developed in this laboratory and the macro ‘‘Island-method’’ (Cunha et al., 2001b). In short, an ‘‘index volume’’ of the nucleoid was calculated from the area of a circular threshold that contained 50% of the total fluorescence intensity within a sufficiently wide region of interest (radius 100 pixels or 6.7 lm) around the nucleoid (see also http:// To minimize radiation damage of DAPI-stained nucleoids by UV-excitation (Cunha et al., 2001b), the iris diaphragm of the mercury lamp was closed to well within the field of view and the preparation was scanned by hand with the UV-shutter remaining open. As yet unexposed nucleoids coming into the field of view were focused and automatically photographed at a 0.5 s time interval. Sequences of 5–10 images were taken of every individual nucleoid coming into the field of view. The first image that appeared in focus, usually after 1–2 s of exposure, was selected for volume measurement.

3. Results 3.1. Size and stability of liberated nucleoids In protocols for nucleoid isolation using the detergent method, the lysozyme concentration used to obtain spheroplasts was generally 400 lg/ml in the lysate (Worcel and Burgi, 1972). In our previous work applying the osmotic shock method, we used this (Cunha et al., 2001a) or a somewhat lower lysozyme concentration (300 lg/ml, resulting in 3 lg/ml after osmotic shock; Cunha et al., 2001b). That different protocols can result in different properties of spheroplasts and even prevent the liberation of nucleoids upon osmotic shock is described in Supplementary data (Section S2.2; Figs. S1 and S2). It should be noted that different dilutions in the osmotic shock procedure give rise to different final concentrations of NaCl and lysozyme in the lysate. With its isoelectric point (pI) of 11.16 (Kuehner et al., 1999), lysozyme is a strongly charged molecule carrying about nine positive charges under our conditions that could compact the DNA like other multivalent ions. In Fig. 1 we show how the size of liberated nucleoids changes with the final concentration of the lysozyme in the lysate. The NaCl concentration in these lysates varied from 1 to 11 mM. Addition of 200– 1000 mM NaCl only resulted in a slight compaction, but increased the stability of the DAPI-stained nucleoids (in the presence of 1 M NaCl nucleoids were still intact after 20 s UV irradiation; results not shown). High lysozyme concentrations (>1 lg/ml) in the lysate not only determined the size of the DAPI-stained nucleoids but also their stability toward UV irradiation. In order to investigate this influence of lysozyme on nucleoid stability further, we studied lysozyme-free nucleoids obtained from ampicillin-generated spheroplasts (see Section 2; ampicillin pI = 4.95; Hout and Poole, 1969). Snapshot images of such nucleoids usually had much larger sizes (>200 lm3; see Table 1). These dimensions agree quite well with the maximum size of 209 lm3 predicted for the uncrosslinked nucleoid theoretically (see Appendix B). Apparently, these DAPIstained nucleoids were more labile under UV irradiation than lysozyme-nucleoids, and the images were obtained in an incipient stage of disruption by UV radiation (see below). We therefore photographed the free-floating nucleoids using time-lapse imaging in which the first image could be obtained within 1 s of irradiation. In the time sequences of Fig. 2, we discern that lysozyme-nucleoids from 300 lg/ml lysozyme-spheroplasts are stable during at least

A.S. Wegner et al. / Journal of Structural Biology 178 (2012) 260–269


Fig.1. Snapshot images of DAPI-stained nucleoids from E. coli LMC500 cells grown in glucose-minimal medium prepared under a coverslip. The sucrose buffer containing 300 lg/ml lysozyme was diluted 100- to 10-fold giving final lysozyme concentrations in the lysate of 3 (A), 6 (B), 10 (C) and 30 lg/ml (D). Average indexvolumes of nucleoids (see also Table 1) decreased from $30 lm3 (A) to $10 lm3 (D). Magnification bar 2 lm.

from these observations that lysozyme compacts and stabilizes nucleoids upon their release by osmotic shock. In Fig. 2B and C it can be seen that DAPI-stained nucleoids irradiated for 5 s or longer fall apart into granular substructures that apparently show Brownian motion (cf. also Supplementary movie, Fig. S3B). The earliest images of nucleoid sequences (<2 s UV exposure) exhibited a more homogeneous structure, often consisting of large lobular regions as shown in Fig. 3. We do not know whether these lobules can be ascribed to replication and segregation stages of the in situ nucleoid. 3.2. Treatment of nucleoids with protease, SDS, Rnase and DNase Because lysozyme-nucleoids were smaller ($30 lm3) than theoretically predicted (209 lm3; see Appendix B), the possible presence of protein cross-links was investigated by treating lysates with protease E or K and SDS. High-lysozyme-nucleoids (300 lg/ ml; Fig. 2A) were found to be unaffected by incubation with 20 to 200 lg/ml protease E (pI % 4) at 37 °C for 60 min. Addition of the negatively charged BSA (pI = 4.72; Vilker et al., 1981) at the same molarity showed no effect on nucleoid volumes (results not shown). To ascertain the activity of the protease, lysozyme-nucleoids were isolated from E. coli LMC500/pSACT11 cells in which nucleoids are visualized by the DNA-binding protein HupA-RFP. Although the HupA-RFP-fluorescence vanished within several minutes of enzyme incubation after addition of protease E, the nucleoids remained intact for an additional 60 min as became evident from DAPI-stained preparations (results not shown). In contrast, low-lysozyme- (50 lg/ml) and ampicillin-nucleoids did show a doubling in volume after 30 min (Table 2), while the control nucleoids remained stable for at least 24 h (Table 2, row 2) and even 72 h (Table 1, row 7). Protease E-treated nucleoids could still be compacted by addition of 20 lg/ml EtBr (row 12 in Table 2). Nucleoids treated with 20 lg/ml protease K (pI % 4) for 150 min already showed a granular structure within 2 s of UV-exposure, i.e. before disruption by radiation damage (Fig. 4 A, B). Compared to the protease treatment, the effect of addition of SDS was dramatic. Even high-lysozyme-nucleoids (300 lg/ml)

4 s of UV irradiation (Fig. 2A), while ampicillin-nucleoids already start to expand after 2.5 s of illumination (Fig. 2B). The latter also applies to nucleoids obtained from 50 and 5 lg/ml lysozymespheroplasts (0.5–0.05 lg/ml in the lysate; Fig. 2C), indicating that a high lysozyme concentration in the lysate (3 lg/ml) protects nucleoids against radiation damage. It should be noted that TOTO-1-stained nucleoids pre-irradiated with UV for 20 s in the absence of DAPI showed no expansion. This suggests that DAPI renders the DNA much more sensitive to UV, possibly causing direct strand breaks. Such destruction in the presence of DAPI has been described for UV irradiation of isolated chinese hamster cell chromosomes (Buys et al., 1986). Table 1 presents a quantitative overview of the nucleoid volumes obtained in the various protocols. While snapshot images of nucleoids from 300 lg/ml lysozyme-spheroplasts give volumes of 25–27 lm3, ampicillin nucleoids, stained with either DAPI or TOTO-1, were almost 10Â larger. However, volumes of minimally irradiated DAPI-nucleoids from time-lapse sequences of either ampicillin-spheroplasts or low-lysozyme spheroplasts (50 lg/ml) were repeatably in a range of 34–54 lm3 (Supplementary data, Table S1), with an average of 42 lm3. As expected, addition of a final concentration of 3 lg/ml lysozyme to ampicillin-nucleoids did indeed cause a decrease in volume (row 6 in Table 1). We conclude

Table 1 Comparison of index-volumes of nucleoids of E. coli MC4100 from independent growth and isolation experiments. Free-floating lysozyme- or ampicillin-nucleoids in hangingdrop preparations were photographed by snapshot or time-lapse imaging and measured using the ObjectJ macro ‘‘Island method’’. Growth mediuma G-min G-min G-min G-min Gmin + S G-min + S Gmin + S G-min LB + S LB + S G-min + S G-min + S
a b

Sphero- plastb Lys Lys Amp Amp Amp Amp Amp Lys Amp Amp Amp Amp

Lysoz. Conc. c (lg/ml) 300 300 – – – +300h – 50 – – – –

Staining/photographyd DAPI/Snapshot DAPI/Snapshot DAPI/Snapshot TOTO/Snapshot DAPI/T-lapse after 72 h DAPI/T-lapse DAPI/T-lapse DAPI/T-lapse DAPI/T-lapse DAPI/T-lapse DAPI/T-lapse TOTO/T-lapse

Index volume lm3 (CV)e 27 25 (84) 253 (21) 215 (26) 42 (50) 41 (50) 25 (71) 36 (45) 44 (43) 69 (42) 89 (46) 41 (48) 45 (49)

Number nucleoids measuredf – 50 12 26 127 24 10 151 180 28 32 34 16

Compare with Ref. or Fig. (indep. exp’s)g Cunha et al. Fig. 3A – – Fig. 4– (4x) Fig. 3B (6x) – – Fig. 3C –

G-min, glucose-minimal medium; LB, trypton-yeast extract medium;+S, medium supplemented with 0.58 M sucrose. Spheroplast formation: Lys, lysozyme method; Amp, ampicillin method (see text). c Lysozyme concentration in incubation mixture; final concentration in lysate after osmotic shock is x 1/100. d See Section 2 for Snapshot or Time-lapse imaging. e Although volumes are given in lm3 it should be noted that it is an ‘‘index volume’’ (see Section 2) used to compare nucleoid volumes in the different experiments. The radius of the region of interest (ROI) used in the macro ‘‘Island-method’’ was 100 pixels. f The relatively low numbers in individual experiments are due to the fact that from a time-lapse sequence of DAPI-stained, free-floating nucleoids only few nucleoids are measurable (see Section 2). g Indep. exp’s: number of independent growth and lysis experiments that have been averaged. h Lysozyme added before shock; final concentration in lysate is 3 lg/ml.


A.S. Wegner et al. / Journal of Structural Biology 178 (2012) 260–269

Fig.2. Time-lapse sequences of DAPI-stained nucleoids in hanging-drop preparations, expanding during 4–6 s UV-exposure into a network of granules. In each panel the upper number represents the time of irradiation in seconds; the lower number signifies the index volume as determined with the Island-method macro. (A) Nucleoid liberated from a 300 lg/ml lysozyme-spheroplast. (B) Nucleoid from an ampicillin-spheroplast. The first image was captured while the microscope stage was still moving (see Section 2). (C) Nucleoid from a 50 lg/ml lysozyme-spheroplast. Magnification bars 5 lm.

Fig.3. First images (UV-exposure less than 2 s) selected from time-lapse sequences of DAPI-stained nucleoids obtained from ampicillin-spheroplasts of E. coli LMC500, prepared in a hanging drop. At this early stage of exposure the nucleoid seems composed of a few relatively large lobules that rapidly fall apart into a network of granules (cf. Fig. 2). Average volume of nucleoids 42 lm3 (see Table 1). Magnification bar 5 lm.

dispersed within several minutes upon addition of 0.01–0.5% SDS at RT. Table 2 (last rows) shows how nucleoids that are first expanded with 0.01% SDS, can become compacted again by the addition of 20 lg/ml EtBr. It should be noted that DAPI-stained nucleoids treated with SDS showed the same expansion via granular substructures as observed for protease (Fig. 4B). As SDS can be expected to denature and dissociate DNA-binding proteins, we ascribe the limited expansion of nucleoids to the release of protein crosslinks. This is supported by the observation that fixation with 0.25% glutaraldehyde prevented the expansion of DAPI-stained nucleoids by SDS (0.01%; 15 min), but not the subsequent disruption by UV-exposure (results not shown). In studies using the detergent method, RNase was found to unfold the isolated nucleoids (Worcel and Burgi, 1972). Treatment of the present nucleoids isolated by osmotic shock with RNase A concentrations varying from 0.1 to 400 lg/ml only showed a collapse

to spherical structures. Because of the fairly high isoelectric point of RNase A (pI = 9.6; Tanford and Hauenstein, 1956), nucleoids were shocked in 10 mM NaPi at pH 8.5 of the buffer to diminish the positive charge of the RNase. Again, only collapsed nucleoids were observed, which were similar to high-lysozyme nucleoids (cf. Fig. 1D). Addition of the negatively charged BSA at the same molarity showed no compaction (results not shown). On the whole, we know that the nucleoids isolated by the two respective methods are dramatically different in size (Odijk, 2002), so they are not identical entities. Hence the different response to treatment by RNase need not be surprising. Nucleoids subjected to mechanical forces (e.g. by preparation under a coverslip) can become disrupted into thread-like structures. Contrary to expectation, treatment with pancreatic DNase I (pI = 4.7–5.3; Funakoshi et al., 1980) in the presence of 1 mM Mg2+, did not result in such threads, but in nucleoids expanding as a granular network (Fig. 4C), similar to those of protease (Fig. 4B) or SDS treatments. This suggests that the expansion of DAPI-stained nucleoids during prolonged UV irradiation (Fig. 2B, C) likewise results from breaks in the DNA strands.

3.3. Treatment of nucleoids with chloroquine and EtBr Romantsov et al. (2007) report that varying the concentration of chloroquine (10–68 lM) caused changes in volume that reflected the relaxation and positive supercoiling of their nucleoids. In our system, adding chloroquine in a range of concentrations between 12.5 and 500 lM to DAPI-stained nucleoids did not show any significant change in volume (Table 3). Chloroquine has been found to

A.S. Wegner et al. / Journal of Structural Biology 178 (2012) 260–269 Table 2 Index-volumes of nucleoids of E. coli MC4100 treated with 200 lg/ml protease E and SDS (preparations and measurements as in Table 1). Growth medium(a) Gmin Gmin Gmin Gmin Gmin Gmin Gmin Gmin Gmin Gmin Gmin Gmin Gmin Gmin Gmin Spheroplast(b) Lys Lys Lys Lys Amp Amp Amp Lys Lys Lys Lys Lys Lys Lys Lys Lysoz. Conc. (lg/ml) 50 50 50 50 50 50 50 50 50 5 5 5


Staining/ photography(d) DAPI/Tlapse DAPI/Tlapse DAPI/Tlapse DAPI/Tlapse DAPI/Tlapse DAPI/Tlapse DAPI/Tlapse TOTO/Snapshot ‘‘ DAPI/Tlapse DAPI/Tlapse Snapshot DAPI/Tlapse DAPI/Tlapse DAPI/Tlapse

Treatment Control Control after 24 h protE, 30’ protE, 90’ Control protE, 30’ protE, 24 h Control protE, 45’ Control protE, 30’ protE, 30’ + 20 lg/ml EtBr Control +0.01% SDS +0.01% SDS + 20 lg/mlEtBr

Index volume

lm3 (CV)(e)
47 (51) 36 (47) 69 (45) 83 (95) 37 (38) 75 (55) 195 (74) 51 (27) 105 (60) 41 (48) 87 (66) 19 (55) 35 (60) 119 (57) 24 (104)

Number nucleoids measured(e) 26 50 34 15 62 38 7 16 37 40 12 31 56 46 20

See notes of a–e Table 1.

show decreased binding affinity to DNA at higher salt concentrations (Kwakye-Berko and Meshnick, 1989). Because our shockbuffer contained 1 mM NaCl and 10 mM sodium phosphate versus 10 mM NaCl in the procedure of Romantsov et al. (2007), ionic strength cannot explain the discrepancy. Because chloroquine has also been reported to become degraded by UV irradiation (Karim et al., 1994), experiments using DAPI-stained nucleoids were repeated with TOTO-1 staining, but failed to show volume changes, except at the highest concentration (500 lM; Table 3, row 7). In contrast, addition of EtBr did show relaxation (0.5 lg/ml) and compaction (20 lg/ml) of liberated nucleoids (Table 3, row 11 and 12), suggesting that the insensitivity to chloroquine is not the result of the presence of nicks in the DNA. 4. Discussion Bacterial nucleoids released by osmotic shock and stained with DAPI were observed by fluorescence microscopy to be cloud-like structures, sensitive to radiation damage. To characterize the effect of lysozyme, the nucleoids were isolated either in the absence of lysozyme (ampicillin-nucleoids) or at low lysozyme concentrations (0.5–0.05 lg/ml in the lysate). With our standard protocol the majority of the nucleoid mass turns out to be released from the spheroplast, but always remains attached to its ghost. The liberated nucleoids have volumes about 400Â larger than those in vivo and usually show two or more lobules (Fig. 3) that may reflect the state of replication and segregation in the cell. Paradoxically, under conditions that cause the peptidoglycan layer to be well digested, the osmotic shock procedure does not result in nucleoid release, but in so-called balloons (Fig. S1 and S2). These represent expanded but delineated structures that were insensitive to UV exposure. We think that within these structures the swollen nucleoid is still restricted by the outer membrane. Within the context of our standard protocol, however, such structures are never obtained. There are significant differences between the observations of Romantsov et al. (2007) and ours. First, they used a radically different protocol for the osmotic shock of their spheroplasts by including 0.025% formaldehyde in their shock buffer; they also incorporate a much higher concentration of 8 lg/ml lysozyme in their lysate. Their nucleoids are smaller ($18 lm3), appear to be much more delineated than ours, and do not have the cloud-like appearance of the nucleoids here described. Second, they observed an effect of chloroquine on nucleoid volume, whereas we only

Fig.4. Images selected from time-lapse sequences (0.5 s interval) of DAPI-stained nucleoids obtained from 5 lg/ml lysozyme-spheroplasts of E. coli LMC500, prepared in a hanging drop. (A) Time sequence of untreated control nucleoid. Average index volume 57 lm3. Magnification bar 5 lm. (B) Time sequence of nucleoid treated with 20 lg/ml proteinase K for 150 min. Average index volume 103 lm3 (see Table 2). Magnification bar 5 lm. (C) Image selected from a time sequence of nucleoids treated with 1 lg/ml DNase and 1 mM MgCl2 for 3 min, taken after 1.5 s exposure. Note the granular substructure.

found an effect with EtBr. It has been reported that chloroquine binds less efficiently to DNA than quinacrine or EtBr (Jones et al., 1979) and this may bear on the disparity. Further experiments are required to elucidate these differences and to establish whether the ‘‘structural units’’ as characterized by fluorescence correlation spectroscopy (Romantsov et al., 2007) are similar to the granular substructures described here. (These ‘‘units’’ ultimately derive only from the osmotic pressure because a wave vector dependence was absent in their experiments.) Zimmerman (2006a,b) reported on the compaction and degradation of nucleoids prepared by the so-called polylysine-spermidine procedure, a modification of the detergent method. Because this lysis procedure causes associations between nucleoids and cytoplasmic and envelope cell remnants, the final structures


A.S. Wegner et al. / Journal of Structural Biology 178 (2012) 260–269

Table 3 Index volumes of nucleoids treated with chloroquine or EtBr. Growth medium(a) Gmin + S Gmin + S Gmin + S Gmin Gmin Gmin Gmin Gmin Gmin Gmin Gmin Gmin Gmin Gmin Gmin Gmin Gmin Spheroplast(b) Amp Amp Amp Lys Lys Lys Lys Lys Lys Lys Lys Lys Lys Lys Lys Lys Lys Lysoz. Conc.(c) (lg/ml) 5 5 5 5 50 50 50 50 50 50 50 50 50 50 Staining/photo- graphy(d) Dapi/Timelapse Dapi/Timelapse Dapi/Timelapse DAPI/Tlapse TOTO/Snapshot TOTO/Snapshot TOTO/Snapshot DAPI/Tlapse DAPI/Tlapse DAPI/Tlapse EtBr/Snapshot DAPI/Snapshot EtBr/Snapshot DAPI/Tlapse EtBr/Snapshot EtBr/Snapshot EtBr/Snapshot Treatment Control 25 lM Chqu 100 lM Chqu Control 12.5 lM Chqu 50 lM Chqu 500 lM Chqu (258 lg/ml) 250 lM Chqu 500 lM Chqu (258 lg/ml) Control 0.5 lg/mlEtBr 20 lg/mlEtBr + DAPI 100 lg/mlEtBr Control 2 lg/ml EtBr 20 lg/ml EtBr (50 lM) 100 lg/ml EtBr (250 lM) Index volume

lm3 (CV)(e)
58 (42) 52 (54) 54 (59) 35 (60) 52 (41) 42 (59) 39 (57) 32 (72) 16 (72) 46 (27) 82 (48) 7 (68) 1 (85) 41 (48) 46 (68) 18 (45) 1 (98)

Number nucleoids measured(e) 20 28 53 56 44 27 29 34 23 37 45 23 21 40 13 27 20

See notes of a–e Table 1.

resemble lysed cells rather than released nucleoids (see Fig. 11 in Murphy and Zimmerman, 2000). Their conclusions about possible cross-links are difficult to compare with our results on nucleoid stability and expansion. Our results do conform well to the observations of Sloof et al. (1983), whose method was the basis for our procedure in the beginning stages of this work. Although they had a high final concentration of 5 lg/ml lysozyme in the lysate, their results agree with ours in that their nucleoids were sensitive to treatment with proteinase K but insensitive to RNase. They also studied the response of their nucleoids to increasing concentrations of EtBr (elaborating on the work of Worcel and Burgi, 1972). In an analysis by sucrose gradient centrifugation and electron microscopy, Sloof et al., 1983 found a relaxation with 0.5 lg/ml EtBr but a condensation after surface-spreading in the presence of 100 lg/ml EtBr, similar to our fluorescent nucleoids (Table 3). The observation that removal of proteins causes about a doubling in the size of the nucleoids (Table 2; Fig. 4A and B) suggests that protein cross-links play a role in the limited expansion of the nucleoids upon their release by osmotic shock. Such protein-DNA interactions are in accordance with the theoretical inference that a large number of cross-links must be present in isolated nucleoids (Cunha et al., 2001b). The present low-lysozyme and lysozyme-free nucleoids were shown to become compacted by final lysozyme concentrations in the lysate greater than 1 lg/ml (Fig. 2). We think that lysozyme may not only compact but also stabilize a nucleoid both against UV exposure and treatment with protease. The presence of lysozyme in many previous studies using the detergent method (e.g. Drlica and Worcel, 1975) could also explain the varying results concerning nucleoid stability obtained as a function of enzyme, salt and temperature (see for reviews Pettijohn, 1982; Pettijohn and Sinden, 1985). We have argued here that our improved protocol for nucleoid isolation also leads to a more quantitative assessment of nucleoid sizes than in previous experiments. It is then expedient to discuss these dimensions in terms of a physical model of the bacterial nucleoid as presented in earlier work (Odijk, 1998, 2000, 2002; Cunha et al., 2001b). The nucleoid is viewed as a branched plectonemic supercoil with adhering proteins which are assumed to induce crosslinks between two sections of double-stranded DNA (Fig. 5). The supercoil interacts with itself via the excluded-volume effect (and with cytoplasmic proteins, in the case in vivo) and is perturbed by thermal motion. For convenience, we present an outline of the statistical physical model in Appendix B (the equations

referred to below are stated there). The typical index volume Vui of the uncrosslinked nucleoid, i.e. the DNA supercoil if it were free from proteins, is there argued to be about 209 lm3 which is in accord with the maximum sizes achievable in our studies (see Tables 1 and 2), as we have already pointed out above. Both ampicillin and low-lysozyme nucleoids have an average index volume Vni of about 42 lm3 if the irradiation is minimal. What does this small size imply? According to Eq. (4), crosslinking purportedly collapses the supercoil in part; statistical arguments show that the smaller index volume Vni is a measure of the typical number of plectonemic segments N H between two crosslinks. The number of crosslinks N s =N H is predicted to be about 25. This turns out to be of the order of magnitude of the number of ‘‘domains’’ established in strand nicking experiments (Sinden and Pettijohn, 1981; Worcel and Burgi, 1972; Drlica et al., 1978). It would be interesting to monitor Vni as a function of the degree of nicking to see if this correspondence is really causal. In our osmotic compaction experiment a decade ago, we studied the interaction of isolated E. coli nucleoids with polyethylene glycol in order to determine the free energy of a nucleoid as a function of its size (Cunha et al., 2001b). A scaling formalism then also establishes the number of crosslinks which is an order of magnitude greater than that found here on the basis of Eq. (4). At this stage, the discrepency is not so worrisome since numerical coefficients are lacking in both scaling pictures (the one on crosslinks proposed in Cunha et al., 2001b and the other outlined in Appendix B). Another quantity of interest is RH ’ 0:67 lm predicted by Eq. (2). This scale, the distance between crosslinks, is of the order of the wavelength of light. It may be difficult to monitor the correlation structure of the nucleoid by fluorescence microscopy (see Supplementary movie, Fig. S3). As we add protease E to ampicillin nucleoids, the index volume increases from 37 to 75 lm3 after half an hour (see Table 2). Eq (4) then yields N s =N H ’ 4, so the number of crosslinks decreases quite fast. Ultimately, after a day, the typical index volume of the nucleoid has increased to about 195 lm3. Assuming that no nicks have been introduced, this value suggests that all protein appears to have been removed from the supercoil (compare with the theoretical prediction ’209 lm3 for the uncrosslinked superhelix quoted above). Finally, the specific linking difference r is thought to saturate to about 0.15 in absolute magnitude (Pruss, 1985). Accepting this, we compute a supercoil diameter Ds of about 7 nm from the theory of Ubbink and Odijk (1999) at an ionic strength of about 0.005 M (this is an estimate of the ionic strength of our saline phosphate buffer).

A.S. Wegner et al. / Journal of Structural Biology 178 (2012) 260–269


Fig.5. Schematic representation of the E. coli chromosome and of liberated nucleoids. (A) A 5 min-segment of the chromosome (total 100 min) is drawn as a randomly branched chain of 200 Kuhn segments (which is a fractal). Possible crosslinks by DNA-binding proteins (blue circles) are indicated by double arrows. The plectonemic coil has a diameter Ds. (B) Representation of the branched supercoil crosslinked by proteins (blue doublets) with an index volume of 42 lm3. Removal of protein crosslinks causes an expansion (right panel) to a theoretical index volume of 209 lm3 (see text Appendix B). (C) Liberated nucleoids, crosslinked (left panel) and uncrosslinked (right panel), as observed under the microscope; compare with Figs. 3 and 4 C, respectively.

From Eqs (1) and (4), we know that Vn scales as Ds3/5 – the degree of crosslinking is assumed to be independent of the degree of supercoiling – so the index volume Vni is predicted to reach a minimum value of about 13 lm3 (the original diameter is 52 nm, see Appendix B). This estimate agrees with our measurements performed on lysozyme nucleoids treated with ethidium bromide at a concentratioin of 20 lg/ml (see Table 3). At a very much higher concentration of 100 lg/ml, the nucleoid presumably collapses onto itself. The DAPI-stained nucleoids appeared very sensitive to UV irradiation and rapidly fell apart (within 5 s; Figs. 2 and 4) as a network of granular substructures. We could discern temporal fluctuations at short scales superposed on these structures. We believe these could be a signature of Brownian motion, but this would have to be proved within a model. The granules could also be observed after TOTO-1-staining, especially with nucleoids from ampicillin-spheroplasts and when irradiated with the 470–490 nm excitation filter (see also Fig. S3B). Presently, we do not know the significance of the sub-resolution granules that can be observed in all preparations, i.e. also after protease, SDS or DNase treatments.

If the nucleoid does indeed behave like a rubbery gel, the inhomogeneities at the short scale R⁄ would be expected but, again, a detailed theory would need to be correlated with experiment. We hope to investigate this in future work. Although we do not understand the nature of the granules (Figs. 2 and 4 and Fig. S3) appearing in expanding nucleoids under virtually all conditions, our observations suggest that the isolated nucleoid is better described as a homogeneous, core-less DNA network rather than a fanciful ‘‘rosette’’ consisting of individual DNA loops radiating from a central core (Pettijohn, 1982; Toro and Shapiro, 2010). With our improved isolation protocol, nucleoids are now more amenable to quantitative studies such as the influence of specific DNA-binding proteins (like H-NS; ms. in preparation) and the application of microfluidic devices or optical tweezers for further characterization of their physical properties. Acknowledgments The authors thank Karl Drlica for suggestions on the manuscript, Flemming Hansen for giving strain FH2973, Tanneke den Blaauwen and Rogier Stuger for discussions, Norbert Vischer,


A.S. Wegner et al. / Journal of Structural Biology 178 (2012) 260–269

Ronald Breedijk and Jolanda Verheul for technical help, Sonia Cunha for preliminary experiments and Mark Hink for making the movie (Supplementary data, Fig. S3B). This work was supported by the Dutch NWO program ‘‘From Molecule to Cell’’, grant 805-47.032-P. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at Appendix B B.1. Physical model of the nucleoid We here summarize several features of a physical model of the nucleoid as formulated by one of us (Th.O.). The chromosomal DNA helix of E. coli has a contour length of 1.6 mm and is a closed circular twisted curve. Under the usual conditions, the interwound helix is a plectoneme of specific linking difference r = À0.06. However, there are proteins associated with the DNA helix which relieve some of the twisting tension and the effective value of r has been estimated to be reff ’ À0.025 (Bliska and Cozzarelli, 1987). In the physical model of the nucleoid, the latter variable is assumed to be distributed uniformly along the chain. In a concrete model of plectonemic DNA, electrostatic energy arising from the self-energy of the charged DNA is balanced against the bending and twisting energies of the DNA helix together with the configurational entropy (Ubbink and Odijk, 1999). The diameter Ds and effective contour length Ls of the supercoil may be computed as a function of r and the ionic strength I (the diameter is defined in Fig. 5). In polymer theory the notion of statistical segments is often introduced (called Kuhn segments). In the case at hand, the supercoil is viewed as consisting of Ns segments of length As (Cunha et al., 2001b). But the supercoil is branched (see Fig. 5) with a branching density Ks = n3/Ns where n3 is the number of branches presumed to be trifunctional (Vologodskii and Cozzarelli, 1996). The radius of gyration of the uncrosslinked supercoil may be written as (Odijk, 2002; Cunha et al., 2001b)
=2 À1=10 1=5 4=5 Ru ¼ 0:399N 1 Ds As  bN s s Ks 1=2

At large scales, the nucleoid is viewed as a rubbery gel in which the crosslinks are distributed homogeneously. Hence, for the con3 centration of segments we must impose c ’ Nw/ R3 H tbnd N s =Rn where Rn is the radius of gyration of the crosslinked nucleoid. This implies the relations

RÃ ¼

R3 n R2 u  6  2 Rn Vn ¼ Ns Ru Vu


NÃ ¼ Ns


Here, the volume of the nucleoid is Vn = 4pR3 n =3 when it is crosslinked and Vu = 4pR3 u =3 when it is not. We note the strong sixth power law in Eq. (4). In practice, the size of the nucleoid Rni has been monitored by us at 50% threshold intensity. In Cunha et al. (2001b), we have argued that Rn = 1.13 Rni. Moreover, in practice also, the amount of DNA per nucleoid is not that contained in one chromosome but more, namely, 1.58 chromosome equivalents for E. coli LMC500 in the present study (valid for growth in a glucose-minimal medium at 28 °C with Td = 85 min, C = 70 min and D’ = 19 min (Huls et al., 1999)). The ionic strength of the 0.01 M NaPi buffer is taken to be about 0.005 M at pH = 7. Then, the diameter of the plectoneme is about 52 nm at reff = -0.025 as predicted by theory (Ubbink and Odijk, 1999; we derive Ds by interpolation within the tables presented there). The variables Ns, As and Ks are estimated to be 4000, 0.158 lm and 0.71, respectively, (Cunha et al., 2001b). Eq. (1) then yields Ru = 3.31 lm so that the index volume of the uncrosslinked nucleoid is predicted to be Vui = 209 lm3 (i.e. corrected for the number of chromosome equivalents and the 50% threshold factor). References
Alexeeva, S., Gadella Jr., T.W., Verheul, J., Verhoeven, G.S., den Blaauwen, T., 2010. Direct interactions of early and late assembling division proteins in Escherichia coli cells resolved by FRET. Mol. Microbiol. 77, 384–398. Bliska, J.B., Cozzarelli, N.R., 1987. Use of site-specific recombination as a probe of DNA structure and metabolism in vivo. J. Mol. Biol. 194, 205–218. Buys, C.H.C.M., Mesa, J., van der Veen, A.Y., Aten, J.A., 1986. A comparison of the effect of 5-bromodeoxyuridine substitution on 33258 Hoechst- and DAPIfluorescence of isolated chromosomes by bivariate flow karyotyping. Histochemistry 84, 462–470. Cui, S.M., Chen, Z.Y., 1996. Monte Carlo simulations of randomly branched polymers with annealed and quenched branching structures. Phys. Rev. E 53, 6238–6243. Cunha, S., Odijk, T., Sueleymanoglu, E., Woldringh, C.L., 2001a. Isolation of the Escherichia coli nucleoid. Biochimie 83, 149–155. Cunha, S., Woldringh, C.L., Odijk, T., 2001b. Polymer-mediated compaction and internal dynamics of isolated Escherichia coli nucleoids. J. Struct. Biol. 136, 53– 66. Cunha, S., Woldringh, C.L., Odijk, T., 2005. Restricted diffusion of DNA segments within the isolated Escherichia coli nucleoid. J. Struct. Biol. 150, 226–232. Daoud, M., Joanny, J.F., 1981. Conformation of branched polymers. J. Physique 42, 1359–1371. Delius, H., Worcel, A., 1974. Electron microscopic visualization of the folded chromosome of Escherichia coli. J. Mol. Biol. 82, 107–109. Drlica, K., Worcel, A., 1975. Conformational transitions in the Escherichia coli chromosome: analysis by viscometry and sedimentation. J. Mol. Biol. 98, 393– 411. Drlica, K., Burgi, E., Worcel, A., 1978. Association of the folded chromosome with the cell envelope of Escherichia coli: nature of the membrane-associated DNA. J. Bacteriol. 134, 1108–1116. Foley, P.L., Wilson, D.B., Shuler, M.L., 2010. Macromolecular crowding can acoount for RNase-sensitive constraint of bacterial nucleoid structure. Biochem. Biophys. Res. Comm. 395, 42–47. Funakoshi, A., Tsubota, Y., Fujii, K., Ibayashi, H., Takagi, Y., 1980. Simple purification and properties of bovine pancreatic deoxyribonuclease I. J. Biochem. 88, 1113– 1118. Hecht, R.M., Taggart, R.T., Pettijohn, D.E., 1975. Size and DNA content of purified E. coli nucleoids observed by fluorescence microscopy. Nature 253, 60–62. Hecht, R.M., Stimpson, D., Pettijohm, D., 1977. Sedimentation properties of the bacterial chromosome as an isolated nucleoid and as an unfolded DNA fiber. J. Mol. Biol. 111, 257–277. Higgins, N.P., Yang, X., Fu, Q., Roth, J.R., 1996. Surveying a supercoil domain by using the gamma delta resolution system in Salmonella typhimurium. J. Bacteriol. 178, 2825–2835.

ð 1Þ

where we have introduced a length parameter b (note that there is a misprint in Eq. (1) of Odijk (2002)). Eq. (1) represents the largest size the isolated bacterial chromosome could adopt, at least theoretically. Superficially, the half power dependence on Ns in Eq. (1) might lead one to think that the uncrosslinked chromosome is actually a Gaussian random coil but it is not: under ideal condi=4 tions the size would scale as N 1 but the segments are nonovers lapping because they have a nonzero diameter Ds and thus interact via an excluded-volume effect. The exponent ½ of the expanded branched coil stems from the theory of branched polymers (Daoud and Joanny, 1981; Marko and Siggia, 1994; Cui and Chen, 1996). It has often been hypothesized that the bacterial nucleoid could consist of ‘‘domains’’ that may act independently of each other (Worcel and Burgi, 1972; Pettijohn and Hecht, 1973; Sinden and Pettijohn, 1981; Pavitt and Higgins, 1993; Higgins et al., 1996). The supercoil is then supposed to be crosslinked at certain positions which are impediments to supercoil diffusion (see Fig. 5). In the scaling analysis of Odijk (2002), an uncrosslinked blob of radius RH containing N H segments is introduced

RÃ ¼ bN Ã


¼ Ru 

1=2 NÃ Ns

ð 2Þ

A.S. Wegner et al. / Journal of Structural Biology 178 (2012) 260–269 Hout, J.P., Poole, J.W., 1969. The amino acid nature of ampicillin and related penicillins. J. Pharma. Sci. 58, 1510–1515. Huls, P.G., Vischer, N.O.E., Woldringh, C.L., 1999. Delayed nucleoid segregation in Escherichia coli. Mol. Microbiol. 33, 959–970. Jones, R.L., Davidson, M.W., Wilson, W.D., 1979. Comparative viscometric analysis of the interaction of chloroquine and quinacrine with superhelical and sonicated DNA. Biochim. Biophys. Acta 561, 77–84. Karim, E.I., Ibrahim, K.E., Abdelrahman, A.N., Fell, A.F., 1994. Photodegradation studies on chloroquine phosphate by high-performance liquid chromatography. J. Pharm. Biomed. Anal. 12, 667–674. Kavenoff, R., Bowen, B.C., 1976. Electron microscopy of membrane-free folded chromosomes from Escherichia coli. Chromosoma 59, 89–101. Kornberg, T., Lockwood, A., Worcel, A., 1974. Replication of the Escherichia coli chromosome with a soluble enzyme system. Proc. Natl. Acad. Sci. USA 71, 3189– 3193. Kuehner, D.E., Engmann, J., Fergg, F., Wernick, M., Blanch, H.W., et al., 1999. Lysozyme net charge and ion binding in concentrated aqueous electrolyte solutions. J. Phys. Chem. B 103, 1368–1374. Kwakye-Berko, F., Meshnick, S.R., 1989. Binding of chloquine to DNA. Mol. Biochem. Parasitol. 35, 51–56. Marko, J.F., Siggia, E.D., 1994. Fluctuations and supercoiling of DNA. Science 265, 506–508. Materman, E.C., van Gool, A.P., 1978. Nucleoid release from Escherichia coli cells. J. Bacteriol. 133, 878–883. Meijer, M., de Jong, M.A., Woldringh, C.L., Nanninga, N., 1976. Factors affecting the release of folded chromosomes from Escherichia coli. Eur. J. Biochem. 63, 469– 475. Miller, J.H., 1992. A Short Course in Bacterial Genetics: A Laboratory Manual and Handbook for Escherichia coli and Related Bacteria. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Murphy, L.D., Zimmerman, S.B., 2000. Multiple restraints to the unfolding of spermidine nucleoids from Escherichia coli. J. Struct. Biol. 132, 46–62. Nielsen, H.J., Li, Y., Youngren, B., Hansen, F.G., Austin, S.J., 2006. Progressive segregation of the Escherichia coli chromosome. Mol. Microbiol. 61, 383–393. Odijk, T., 1998. Osmotic compaction of supercoiled DNA into a bacterial nucleoid. Biophys. Chem. 73, 23–30. Odijk, T., 2000. Dynamics of the expanding DNA nucleoid released from a bacterial cell. Physica A 277, 62–70. Odijk, T., 2002. Sedimentation of DNA supercoils and bacterial nucleoids. Colloids and Surf. A: Physicochem. Eng. Aspects 210, 191–197. Pavitt, G.D., Higgins, C.F., 1993. Chromosomal domains of supercoiling in Salmonella typhimurium. Mol. Microbiol. 10, 685–696. Pettijohn, D.E., Hecht, R., 1973. RNA molecules bound to the folded bacterial genome stabilize DNA folds and segregate domains of supercoiling. Cold Spring Harbor Symp. Quant. Biol. 38, 31–42.


Pettijohn, D.E., 1982. Structure and properties of the bacterial nucleoid. Cell 30, 667–669. Pettijohn, D.E., Sinden, R.R., 1985. Structure of the isolated nucleoid. In: Nanninga, N (Ed.), Molecular Cytology of Escherichia coli. Academic Press, London, New York, pp. 199–227. Pruss, G.J., 1985. DNA topoisomeras I mutants. Increased heterogeneity in linking number and other replicon-dependent changes in DNA supercoiling. J. Mol. Biol. 185, 51–63. Robinow, C., Kellenberger, E., 1994. The bacterial nucleoid revisited. Microbiol. Rev. 58, 211–232. Romantsov, T., Fishov, I., Krichevsky, O., 2007. Internal structure and dynamics of isolated Escherichia coli nucleoids assessed by fluorescence correlation spectroscopy. Biophys. J. 92, 2875–2884. Sinden, R.R., Pettijohn, D.E., 1981. Chromosomes in living Escherichia coli cells are segregated into domains of supercoiling. Proc. Natl. Acad. Sci. USA 78, 224–228. Sloof, P., Maagdelijn, A., Boswinkel, E., 1983. Folding of prokaryotic DNA: isolation and characterization of nucleoids from Bacillus licheniformis. J. Mol. Biol. 163, 277–297. Stonington, G.O., Pettijohn, D.E., 1971. The folded genome of Escherichia coli isolated in a protein-DNA-RNA complex. Proc. Natl. Acad. Sci. USA 68, 6–9. Tanford, Ch., Hauenstein, J.D., 1956. Hydrogen ion equilibria of ribonuclease. J. Amer. Chem. Soc. 78, 5287–5291. Toro, E., Shapiro, L., 2010. Bacterial chromosome organization and segregation. Cold Spring Harb Perspect Biol. 2, a000349. Ubbink, J., Odijk, Th., 1999. Electrostatic-undulatory theory of plectonemically supercoiled DNA. Biophys. J. 76, 2502–2519. Vilker, V.L., Colton, C.K., Smith, K.A., 1981. The osmotic pressure of concentrated protein solutions: effect of concentration and pH in saline solutions of bovine serum albumin. J. Coll. Int. Sci. 79, 548–566. Vologodskii, A., Cozzarelli, N.R., 1996. Effect of supercoiling on the juxtaposition and relative orientation of DNA sites. Biophys. J. 70, 2548–2556. Woldringh, C.L., Nanninga, N., 1985. Structure of nucleoid and cytoplasm in the intact cell. In: Nanninga, N. (Ed.), ‘‘Molecular cytology of Escherichia coli’’. Academic Press, London, New York, pp. 161–197. Worcel, A., Burgi, E., 1972. On the structure of the folded chromosome of E. coli. J. Mol. Biol. 71, 127–147. Worcel, A., Burgi, E., 1974. Properties of a membrane-attached form of the folded chromosome of Escherichia coli. J. Mol. Biol. 82, 91–105. Zimmerman, S.B., 2006a. Cooperative transitions of isolated Escherichia coli nucleoids: implications for the nucleoid as a cellular phase. J. Struct. Biol. 153, 160–175. Zimmerman, S.B., 2006b. Shape and compaction of Escherichia coli nucleoids. J. Struct. Biol. 156, 255–261.