A 0000298

8/3/2019 A 0000298

1/7

Bacterial ChromosomeEduard Kellenberger, University of Lausanne, Lausanne, Switzerland

Jing-Yan Li, Kuming Institute of Zoology, Chinese Academy of Sciences, China

Under conditions favouring most active growth, bacteria replicate their DNA and dividewith generation times as short as 2025 minutes; similary rapid rates are practically

unknown in other taxa. In contrast to the typical eukaryotic cell, DNA transcription,

translation and replication occur simultaneously and at nearby locations. The highly

dynamic processes lead to a reticulation of DNA and RNA, permitting optimization of

mRNA interaction (number of contacts) with ribosomes. Increased order in the chromatin

seems to occur only with very much slower overall growth.

Introduction

During the 1950s, a deoxyribonucleic acid (DNA)-specificcytological staining (of the Feulgen type) of chemically-

fixed bacteria revealed in the light microscope, one, two oreven four strongly stained patches per cell. In cells ofhigher organisms, this staining was known to be confinedto the nucleus and defined the chromatin and thechromosomes (Gr. khroma colour). This shared stainingpattern gave rise to the terms bacterial chromosomes,nucleoids or nuclear equivalents, all meaning thesesame, coloured patches. After removing the ribonucleicacid (RNA) by digestion with ribonuclease to leave onlyDNA for staining, these results were further confirmed.More recently, the staining with DNA-specific fluorescentdyes, such as 4,6-diamidino-2-phenylindole (DAPI),substantiates these earlier results and provides an easier

approach. With such dyes, it is now even possible to stainthe DNA of live cells without disturbing their growth(Figure 1). Owing to the limited resolution ($ 0.2 mm), it isimpossible to define exactly the shape of nucleoids in lightmicrographs, particularly because the size and shape ofthese patches is extremely variable and dependent on thepreparation methods and growth media used.

At about the same time, toward the middle of thetwentieth century, bacterial mutants (variants) werediscovered, which breed true, i.e. in which the over-whelming majority of the descendants of an individualmutant cell carry the altered character. In the decadefollowingWatson andCricks double helix proposal (1953)

of a linear genetic code, the widespread acceptance of theDNA molecule as the support for the cells geneticinformation displaced the earlier belief that proteinsfulfilled this role and led to a great increase in researchon DNA, so that knowledge of bacterial geneticsprogressed rapidly.

Microbial taxonomy has been revolutionized throughthe use of new techniques of DNA and RNA sequenceanalysis. By comparing, for example, the detailed basesequence of 16S ribosomal RNAs, Woese proposed that

there were two branches of the former bacteria: thBacteria and the Archaea, of which the latter seem tshare more with Eukarya than the former (as discusse

later in this entry). As very few ultrastructura

Article Contents

Secondary article

. Introduction

. Functions, Size and Nature of the Bacterial

Chromosome and its Packing into the Cell

. New Views of the Bacterial Chromatin UsingCryofixation and Freeze-substitution

. The Bacterial Chromatin is Supercoiled

. Histones, Histone-like Proteins and Nucleosomes

Figure 1 In vivo, fluorescently stained, Escherichia coli. DAPI is added toculture of exponentially growing cells in an amount determinedexperimentally for each strain and medium; being low enough not to

inhibit growth. Viewed under near blue ultraviolet fluorescence and phacontrast. The latter is necessary for visualizing the bacterial body.

ENCYCLOPEDIA OF LIFE SCIENCES / & 2002 Macmillan Publishers Ltd, Nature Publishing Group / www.els.net

8/3/2019 A 0000298

2/7

(i.e. microscopic) studies have yet been made on thenucleoids of archaea, we concentrate mainly on the resultsof the more thoroughly studied bacteria, and Escherichiacoliin particular.

Functions, Size and Nature of theBacterial Chromosome and its Packinginto the Cell

Taking great care to minimize shearing, Cairns (1963)isolated the radioactive isotopically labelled DNA of an E.colicell and spread it onto a photographic plate to recordautoradiographic images. He found large circles with a1300-mm circumference. He also demonstrated the processof replication of such circles, neatly demonstrating that thegenetic apparatus of a bacterium consists of a single linearmolecule of DNA in the form of a (closed) circle. This

conclusion was confirmed independently, by geneticmethods.

Further, a set of essential observations were made byMiller (1973) using electron microscopy: DNA wasisolated from metabolizing cells in such a way that thesynthesized messenger RNA (mRNA) transcript remainedattached to the DNA and formed tree-like patterns,Christmas trees, with an increasing length of the twigstowards one end. When comparing nuclear DNA ofeukaryotes with that isolated from bacteria, Miller foundan essential difference: within the bacteria, the twigs ofmRNA were studded with ribosomes, illustrating previousbiochemical findings that transcription and translation

occur simultaneously in bacteria, on nearby sites. Incontrast, in eukaryotes, the mRNA has to becomedetached from the DNA and to move out of the nucleusinto the cytoplasm before translation can begin.

To accommodate the 1300-mm long chromosomeinto anE. colicell of 23-mm length and some 0.8-mm diameter isoften erroneously considered as a remarkable feat requir-ing a high degree of condensation and compaction. Toquantify these parameters, the packing density was definedby Woldringh and Nanninga as the local concentration ofDNA (mg mL2 1). Estimates of some 20 mg mL2 1 werefound for the DNA of actively growing bacteria(1.5 102 14g mass of DNA of chromosome and

1.4 102 12 mL volume of an E. coli cell with twonucleoids). When similar calculations were made foreukaryotic DNA in the interphase (in which DNA under-goes replication) nucleus of a liver cell, ignoring differen-tially stained eu-and heterochromatin, 2040 mg mL2 1

was also obtained. Metaphase chromosomes, however, aresome ten times more condensed, at 200300 mg mL2 1,and DNA in bacterial viruses can even reach 800 to1000mg mL2 1! These chromosomes are associated withnonreplicating, condensed chromosomes taking up posi-

tions on the equatorial plane of the (three-dimensionaspindle apparatus. We have to conclude that the condensation of the bacterial chromosomes is comparable tthat of the eukaryotic interphase nucleus and thus does norequire any procedure of an exceptional nature.

Is the packing of the chromosome the result of a foldininto a defined three-dimensional structural pattern that i

determined by specific interactions with condensinprinciples, i.e. such as specific, crosslinking proteins? Ois it a more mobile randomly distributed structurereflecting simply those specific metabolic functions activelundergoing transcription-translation as well as thosinvolved in the replication of the chromosome for a givecell environment and growth phase? Highly relevant arrecent physical chemistry experiments involving spontaneous condensation of DNA as a consequence oincreased ion concentrations and of crowding effectThe latter can be simulated in vitro, e.g. by addition opolyethylene glycol (PEG). For the first it is too frequentlforgotten that intracellular concentrations of small mole

cular weight solutes in growing bacteria are generally 2times higher than those of the medium; they are responsibfor the cell turgor pressure. Na1 , generally more abundanextracellularly, is replaced intracellularly by K1 , and Clby negatively charged amino acids (like glutamic acid) another small molecular organic substances, probablreflecting relative levels in the conditions in which bacterievolved. Intracellular, positively charged Mg21 , Ca21

putrescine and spermidine are important as neutralizinpartners of the negatively charged phosphate groups oDNA, which are not occupied by the histone-like, basiproteins. They form associations with lower dissociatioconstants than with the monovalent potassium ions. I

experimental conditions in vitro, resembling intracellulaconditions, the 50-mm long bacteriophage T2 DNAcondenses into individual particles that, with fluorescenstaining, are visible even by light microscopy. Taking intaccount the light microscopes limited resolution, thradius of the particles is estimated to be about 0.3 mm, i.emore than 10 times smaller than the length of thartificially unravelled DNA. With similar experimentbut using electron microscopy, the condensates had a sizcomparable to that of mature phage.

The foregoing findings strongly suggest that achievinthe low packing densities encountered in vivo in growinbacterial cells (20 mg mL2 1), and in the DNA pool o

replicating and transcribing vegetative phage, requires nparticular, highly specific condensing mechanisms. Indeedone purpose of bound proteins might be precisely thprevention of spontaneous aggregation (condensation) oDNA, which, when strong enough, as in eukaryotimetaphase chromosomes, obviously inhibits replicatioand transcription.

The other space-saving organization characteristic oDNA is that of compaction by supercoiling. This discussed below.

Bacterial Chromosome

2 ENCYCLOPEDIA OF LIFE SCIENCES / & 2002 Macmillan Publishers Ltd, Nature Publishing Group / www.els.net

8/3/2019 A 0000298

3/7

New Views of the Bacterial ChromatinUsing Cryofixation and Freeze-substitution

Electron microscopy of thin sections, with a resolutionestimated not to be below some 5 nm, i.e. 40 times that of

the light microscope, should have been able to provide thefine details of the bacterial chromosome that were lacking.In early studies these expectations were only partiallyfulfilled. Thin-section techniques, developed for the studyof eukaryotic cells and tissues, achieved revolutionaryprogress that laid a basis for modern cell biology. The samecytological techniques of fixation and embedding appliedto bacteria yielded disappointing results: the bacterialDNA was in the form of randomly shaped aggregateswithin an otherwise empty zone, at that time named thenuclear vacuole. The absence of a membrane around thisvacuole was later confirmed for bacterial nucleoids inwhatever form they were observed. Improved fixation

prevented the aggregation of the DNA during dehydrationin ethanol so that the chromatin, the DNA-containingplasm, appeared in the form of fine fibrils whose thicknesswas compatible with that expected of DNA. What wastroubling to those involved in such studies was the strongvariation of shape of the nucleoid, which depended on thecytoplasmic fixatives used: with osmium tetroxide fixationit is rather strongly confined; whereas after aldehydefixation it comprises a group of many, much smallervacuoles. A fundamental change came with the introduc-tion of cryotechniques.

Water-containing biological material can be frozen sorapidly that its water is transformed into the vitreous state,

which is amorphous, i.e. noncrystalline. Such frozen andstill hydrated material can be cryosectioned; however, theslices yielded are not as thin as thoseobtained from current

embedding in plastic resins. Techniques for stainincryosections with heavy metals are not yet available animage contrast is therefore too low to reveal fine detaiFortunately, the frozen material at low temperature caalso be freeze-substituted by organic solvents, such aacetone or methanol. At temperatures around2 808Cwhile still solid, the ice is dissolved slowly into the solvent

The ice is thus progressively replaced by the solvent, whiccan in turn be substituted by the liquid resin. When this isubsequently rendered solid, it can be subjected to thislicing or microtomy. This procedure of cryofixation anfreeze-substitution (CFS) is limited to rather thin layers omaterial because the maximum depth of vitreous ice is onlsome 1020-mm thick. The method is thus ideal for severalayers of 0.8-mm bacteria.

Cryofixation is not simply a temperature reversibimmobilization, but has been demonstrated to creatphysical crosslinks, producing a gel that resists subsequenaggregation in organic, water-miscible solvents, even whethey are at room temperature.

Sections ofE. coliand Bacillus subtilis processed by CFreveal additional features: the ribosomes are clearly visibland are not distributed equally throughout. Many ribosome-free spaces of variable size are apparent (Figure2)thacontain fine globular material, which replaces the fibrillaone visible within the nucleoids obtained with thpreviously employed RyterKellenberger (R-K)-technque. The fibrillar material was directly identified, bappearance as containing DNA. For the rather globulacontent of the ribosome-free spaces of CFS-preparebacteria, indirect identification had to be made bimmunocytochemistry. For DNA, a new immunostaiwas discovered which showed convincingly that most,

not all, ribosome-free spaces contained double-strandeDNA (Bohrmann et al., 1991). According to Millerexperiments, referred to previously, the DNA-dependen

Figure2 Serial, longitudinal sections ofEscherichia coli, preparedby cryofixation andfreeze-substitution (CFS)for the electron microscope. Fiveof a seri

of 11 thin sections, taken from the middle part of the cell, are shown. The nonuniform distribution of ribosomes can be distinguished. The bacterialchromatin is in the ribosome-free spaces, as shown by immunostaining (Bohrmann et al. 1991). Bar, 0.5 mm.



8/3/2019 A 0000298

4/7

RNA polymerase must be associated with the metaboli-cally active DNA; this was confirmed by immunolabelling(Du rrenberger et al., 1988). Even by more old-fashionedmethods (osmium tetroxide fixation), the site of RNAsynthesis was found, by autoradiography, to be outside thebulk of the nucleoid (Ryter and Chang, 1975). The veryflexible stem and the twigs of Millers Christmas trees are

distributed all over the cell within the ribosome-free spacesin such a manner that a maximum number of ribosomescan become involved in protein synthesis. According to aproposal of C. Robinow, University of Western Ontario,Canada, this form of the nucleoid, with excrescencesreaching far into the cytoplasm, was described as coral-line.

The problem with these new CSF findings was that, incell sections, the nucleoid no longer appeared to beconfined to a central location, as was the case with osmiumtetroxide fixation, or with the observation of whole cells byelectron microscopy, or, in light microscopy, after stainingor byphase contrast.Knowing that a good thinsection is at

the most 50-nm thick means also that, cut longitudinally, acell ofE. coliis sectioned into about 15 consecutive serialsections. By reconstructing a cell from these serial sections,the problem might be resolved. The ribosome-free spacesof 11 serial sections were transferred to slightly colouredtransparent foil and carefully cut out. The package ofsuperimposed transcripted sections was then photo-graphed by a low-resolution (pinhole) camera and printed(Figure 3). The low-resolution picture of the reconstitutedcell looks exactly like some of the phase contrast lightmicrographs taken from a culture of the same strain. Thisexperiment demonstrates that the ribosome-free spaces arenot entirely randomly distributed within the cell; there is an

increased amount in the central region, corresponding towhat is observed in the light microscope with live, entirecells. Indeed, centrally located, larger ribosome-free spaceshad frequently been observed in near-equatorial sections.These central areas the bulk are considered torepresent those parts of the chromatin, which are not (at agiven moment) involved in gene expression. When theprotein synthesis is inhibited, e.g. by treatment withsublethal doses of chloramphenicol, or by amino acidstarvation of an amino acid-requiring strain, the chroma-tin assembles into a near sphere, frequently showing aribosome-free central core. This typical spherical nucleoidis observed whatever the method of preparation/micro-

scopy technique used. Its physical structure and generationare still not understood.

The Bacterial Chromatin is Supercoiled

Double-stranded DNA, depending upon whether it isunder torsion or not, shows differential binding ofpsoralen. The amount of this substance detected as being

bound is directly related to the degree of supercoilingPettijohn and Pfenniger (1980) have systematically appliethis technique to cells of E. coli. They found negativsupercoiling, i.e. a torsional stress that facilitates openinof the double strands, such as is needed for replication antranscription. We will see further (below) that the DNA othermophilic bacteria is, in contrast, positively supercoiled, so as to inhibit their DNA strands from openingthis would obviously counter the destabilizing effect of thhigh temperatures of their environment.

Left-hand torsion, applied to a thread, will be relaxewhen the thread is put into the form of a left-handesolenoid (Figure 4). This is what we do when we fold ou

garden hose and is what happens when the DNA is wounaround the protein cores of the eukaryotic chromatiforming nucleosomes. In this restrained form, the chromatin is at rest, not subject to twisting forces.

When a circular DNA molecule is supercoiled, then thcircle is reduced (collapsed) into an elongated plectonemisupercoil (Figure 4). While the solenoid allows for substantial shortening (compacting) of a DNA threadthis is possible only to a very limited degree with thplectonemic form. It is still not known in which of the tw

Figure 3 Serial sections, of which five are shown in Figure 2, areschematically redrawnon coloured foilsand, withthe ribosome-free spac

carefully cut out, superimposed to form a package that is a reconstructioof the whole cell. A sharp photographic image of this reconstructed cell

given in (a). The out-of-focus pictures (b1) to (b4) are obtained with apinhole camera; these prints, on hard-grade paper, differ from each otheonly by the exposure time. (c) A light microscope phase contrast

micrograph is shown. According to the concentration of the surroundingrefracting material, phase contrast images vary exactly as do those of (b1to (b4). They demonstrate how the apparent size of the nucleoid is

determined by optical parameters. To a somewhat lesser degree, thisis alstrue for the varying intensities of the fluorescent staining. From Bohrman

et al. (1991) with permission.



8/3/2019 A 0000298

5/7

forms the bacterial chromatin is supercoiled. Both formswould lead to a rather microglobular appearance on highresolution electron micrographs if preserved as such. Afibrous appearance, obtained after chemical fixation, canbe readily accounted for as chemically induced relaxationof the supercoil.

The enzymes responsible for supercoiling are thetopoisomerases, frequently in combination with otherhistone-like proteins. Topoisomerases I and II are ofparticular importance and were discovered first. Topo-isomerase I is able to introduce torsion by opening only onestrand of DNA, whereas topoisomerase II (known asgyrase in bacteria) is able to break both. By immunolabel-

ling, topoisomerase I was localized to the same area as thatof RNA polymerase.

Ruth Kavenoff once succeeded in producing veryelegant electron micrographs of isolated bacterial nu-cleoids (reported in most general reviews on bacterialchromatin), which showed some hundred loops emerginglaterally from a sort of scaffold. Most of these loops wererandomly bent and only few appeared in the form ofplectonemic supercoils. None was solenoidally coiled. Thewhole arrangement showed a surprisingresemblance to the

model of the eukaryotic chromosome proposed bLaemmli: after dissolution of chromatin, the intact scaffolremained behind. Topoisomerase II was demonstrated tbe situated on the scaffold and to be involved in thattachment of the chromatin loops. Considerable effortwere developed to apply Laemmlis biochemical animmunochemical procedures also on the Kavenoff-typof prokaryotic nucleoids. Unfortunately, only two groupachieved something approaching her work in qualityalthough much less convincing, having lost most of thsupercoil of the loops and without succeeding in addinany biochemical and immunological identifications.

Probably stimulated by the pictures of Kavenoff

Pettijohn and coworkers demonstrated the existence oindependent chromosomal segments or domains, possiblin the form of loops. By irradiating the bacterial chromosome in vivo with gamma rays, they demonstrated that thsupercoiling of each individual segment is independentllost through the radiation-induced single-strand nickproduced in each segment (Lyderson and Pettijohn, 1977Speculatively, an attractive model of the bacterial nucleoiis basedon these findings and by analogy with the model oLaemmli: the loops are formed by a crosslinking by gyras

Figure4 Proposedcompactionforms ofDNA.The same lengthof DNAis shownin theformof theloose (a)andcompacted (b)plectonemicsupercoilinIn(c) itis ina solenoidal form.The compaction ratioof thelengthof thestretchedDNAmoleculeto that ofthe supercoiled, compactedformis about9 in(b

and (c), but only 3 in (a). By normalizing the dimensions of the figure such that two windings of (c) correspond to those of a eukaryotic nucleosome, thbending (curvature)is then 0.23 fortheseand 0.25 and0.26 for(b) and(a) respectively. Thelooseplectonemicform andits derivatives have been andst

are extensively studied by electron microscopy and sedimentation rates. For technical reasons, low ionic concentrations are preferred for the electronmicroscopic studies of DNA. Under these conditions it is strongly charged andneighbouring fibresrepulseeachother, so as to produce theloosestructur

In theabsence of a sufficient amountof adequate basic proteins as partners,naked DNAshowsneither thecompacted form (b)nor (c). As yet, solenoidcompaction has been observed only with nucleosomes, where the DNA is wound around a solid core of histones.



8/3/2019 A 0000298

6/7

and other proteins, together forming a scaffold, exactly asdemonstrated for the eukaryotic chromosome. Although anegative result, and thus not published, it is worthmentioning that, by immunocytochemistry, gyrase wasnot found to be enriched in the centre of the globular,chloramphenicol-induced nucleoid (mentioned above),which, in vivo, is the strongest observation in favour of

this putative model. Although the hypothesis of the loopshas not yet been demonstrated, it is interesting to note thatsome repeat sequences in E. coli (BIME-2) have strongcorrelations with the binding and activity of gyrases (Espe liand Boccard, 1997).

The reverse gyrase of thermophilic bacteria generatedinterest in recent years (Bouthier de la Tour etal., 1998), ashave the various helicases.

Histones, Histone-like Proteins andNucleosomes

The name histone-like was chosen because of thechemical properties shared with the eukaryotic histones,mainly their relatively small size, basicity, DNA-bindingproperties and their acid solubility. Implicitly, there is atendency to assume that they have similar functions tothose of the eukaryotic histones, namely to cause thesolenoidal coiling of DNAby forming its centralsolid core.It is well known, however, that the relative amounts ofthese histone-like bacterial proteins are much too low to beable to organize all the chromatin of a cell into nucleo-somes, as a similar structure to that of the eukaryotes, byrestraint of supercoiling to a comparable degree. It is

generally agreed that nucleosomes, if they exist at all inbacteria, must be highly fragile, accounting for theiralternative designation as compactosomes. This fragilityis obviously the reason why they have never yet beenisolated. In the electron microscope, something resemblingthe eukaryotic nucleosomes had been assembled in vitrowith a large excess of the protein HU. When in vivo thehistone-like proteins are bound to DNA, they exertfunctions that are metabolically very active, quite distinctfrom the role of the histones of the eukaryotic nucleosomesin quiescence. In common with the latter, they also modifythe migration rate of small circular DNA during electro-phoresis, a phenomenon which is in agreement with a

beaded structure but there is no proof for it.For the stated reasons, a structural role for the histone-

like proteins could only be inferred indirectly. Byimmunocytochemistry, protein HU was found to belocalized in the area of RNA synthesis, i.e. on the nucleoidprojections and not in the, supposedly inert, bulk(Du rrenberger et al., 1988). By using permeabilized cells,some authors could introduce large amounts of HU intothe cell and found, by fluorescence microscopy, that thebulk of chromatin also contained HU. For us this

observation confirmed that, in the native cell, the bulk othe DNA is not saturated with HU to form nucleosomeswhereas the well-known DNA-binding property of HU iagain validated. The other major histone-like protein ofEcoli, H-NS, was also found mostly on the border betweethe bulk of the nucleoid and the cytoplasm. In a straioverproducing H-NS, it was chiefly present in the bulk o

the nucleoid (for references see Spurio etal., 1992), leadinto the same conclusions as discussed above for HU whepresent in excess. Of interest is the additional observatiothat, in the H-NS excess situation, spherical nucleoidappear. In overproducingconditions the cellslose viabilityand the synthesis of macromolecules, in particular oproteins, is inhibited. The authors interpretation of thesobservations is that H-NS plays a decisive role in thconfinement of the chromosome into the spherical shapeWhy then does direct inhibition of protein synthesis alslead to the same or similar form of nucleoid?

Recent structural studies of eukaryotic histones by Xrays and nuclear magnetic resonance imaging led to th

discovery and definition of a typical substructure, thhistone fold. As far as is known from the still limitenumber of investigations, this fold was not found ihistone-like proteins of bacteria described above, but wapresent, to our great surprise, in DNA-binding proteins omost of the archaea that have so far been carefullinvestigated. The presence of real histones, as defined bthe presence of this typical fold, is correlated with othefeatures that aredistinct from those of thebacteria (Lietal1999): (1) nucleosomes are relatively stable and can bisolated; (2) the chromatin is resistant to aggregatioduring dehydration involved in the preparation for thisections, and therefore must be an HP-chromatin, i.e.

chromatin rich in associated proteins.These distinctive differences between bacteria an

archaea should stimulate the careful investigation oarchaea with modern methods, with a view to obtaininclearer ideas about our unicellular forebears.

References

Bohrmann B, Villiger W, Johansen J and Kellenberger E (199

Coralline shape of the bacterial nucleoid after cryofixation. Journ

of Bacteriology 173: 31493158.

Bouthier de la Tour C, Portemer C, Kaltoum H and Duguet M (1998

Reverse gyrase from the hyperthermophilic bacterium Thermotog

maritima: properties and gene structure. Journal of Bacteriology 18

274281.

Cairns J (1963) The bacterial chromosome and its manner of replicatio

as seen by autoradiography. Journal of Molecular Biology 6: 20821

Du rrenberger M, Bjornsti MA, Uetz T, Hobot JA and Kellenberger

(1988) Intracellular localization of the histone-like protein HU

Escherichia coli. Journal of Bacteriology 170: 47574768.

Espe li O andBoccard F (1997)In vivo cleavage ofEscherichia coliBIME

2 repeats by DNA gyrase: genetic characterization of the target an

identification of the cut site. Molecular Microbiology 26: 767777.



8/3/2019 A 0000298

7/7

Li J-Y, Arnold-Schulz-Gahmen B and Kellenberger E (1999) Histones

and histone-like DNA-binding proteins: correlations between struc-

tural differences, properties and functions. Microbiology 145: 12.

Lyderson K and Pettijohn DE (1977) Interactions stabilising DNA

tertiarystructurein the Escherichia colichromosomeinvestigated with

ionising radiation. Chromosoma 62: 199215.

MillerOL (1973)The visualisation ofgenesin action. ScientificAmerican

228: 2734.

Pettijohn DE and Pfenniger O (1980) Supercoils in prokaryotic DNArestrained in vivo. Proceedings of the National Academy of Sciences of

the USA 77: 13311335.

Ryter A and Chang A (1975) Localization of transcribing genes in the

bacterial cell by means of high resolution autoradiography. Journal of

Molecular Biology 98: 797810.

Spurio R, Du rrenberger M, Falconi A et al. (1992) Lethal over-

production of the Escherichia colinucleoid protein H-NS: ultramicro-

scopic and molecular autopsy. Molecular and General Genetics 231:

201211.

Further Reading

Drlica K and Riley M (1990) The Bacterial Chromosome. Washingto

DC: ASM Press. [Bestcomprehensive book, unfortunately some yea

old.]

Gualerzi CO and Pon CL (1986) Bacterial Chromatin. Berlin: Springe

Nanninga N (1985) Molecular Cytology of Escherichia coli. London

Academic Press.

Nanninga N (1998) Morphogenesis ofEscherichia coli. Microbiology anMolecular Biology Reviews 62: 110129.

Pettijohn DE (1996) The nucleoid. In: Neidhardt FC (ed.) Escherich

coliand Salmonella, vol. 1, pp.158166. Washington, DC:ASMPres

[Comprehensive and short; recent references.]

Robinow CF and Kellenberger E (1994) The bacterialnucleoid revisite

Microbiological Reviews 58: 211232. [Recommended for the mor

technical aspects.]



Documents

A 0000298