![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| GUEST COMMENTARY |
Department of Biochemistry, Molecular Biology and Biophysics and Biotechnology Institute, University of Minnesota, St. Paul, Minnesota 55108
Geometrical and, to some extent, physical and informational properties of covalently closed DNA molecules (or molecules that can be ideally viewed as closed) are determined by their connectivity, i.e., the number of times one strand of DNA is topologically linked to another. Any link that cannot be removed by merely sliding one strand over another but can be removed only by breaking a strand or two is a topological link. All living forms evolved to maintain the numbers of links, per unit of DNA length, most suited for their respective environments and growth conditions. The study by Champion and Higgins (4) offers a tantalizing insight into how two very closely related bacterial species control the topological states of their DNA.
Over the years, the properties of closed circular DNA have been systematically investigated and successfully modeled (5), and some relevant features of such molecules are descriptively presented below. Incision of covalently closed double-stranded DNA will produce a relaxed molecule in which, under standard conditions, one strand will wind around another with a frequency of about one turn per 10.5 base pairs. The length of the molecule (in base pairs) divided by the number of base pairs per turn yields the expected linking number of topologically relaxed DNA. Resealing such incised DNA results in the formation of a discrete distribution of molecules with different linking numbers, centered around the relaxed form (Fig. 1). The molecules with linking numbers greater than that of the relaxed form are known as overwound, and those with linking numbers smaller than that of the relaxed form are known as underwound. In such a distribution, which can be approximated by a bell curve, the probability of occurrence of molecules with increasing degrees of under- and overwinding drops exponentially. The shapes of both types of molecules can be described in three dimensions only by a characteristic coiling of the axis of the double helix (Fig. 1). Since the axis of the double helix of a relaxed, covalently closed double-stranded DNA molecule has the shape of a coil, sufficiently represented in two dimensions, further coiling of that coil is referred to as supercoiling. The greater the difference between the linking number of a given topological species of DNA molecules and that of its respective relaxed form, the more supercoiled the species is. Interestingly, covalently closed circular double-stranded DNA molecules isolated as plasmid or viral DNAs from almost all studied organisms are supercoiled more than their respective relaxed and resealed forms (see, for example, reference 1). In other words, the distributions of linking numbers in populations of plasmids isolated from living cells are different from the corresponding distributions for relaxed DNA molecules. Given this observation, supercoiling is considered to be a ubiquitous property of covalently closed double-stranded DNA molecules in vivo.
|
It has also been established that in the absence of the topA gene, the distribution of linking numbers in a population of some plasmid DNA molecules becomes heavily skewed toward extremely underwound, hyper-negatively supercoiled species (e.g., reference 17). Thus, the connection between the observable steady-state supercoiling and the opposing effects of TopA and gyrase has been established. The intolerance to the topA loss in E. coli is generally attributed to excessive underwinding of the DNA template attained in the presence of fully active gyrase, whose activity is unbalanced in topA-deficient mutants. Since in the mutants of both species the topA gene is missing, Champion and Higgins reasoned that the difference must be in the gyrase side of the equation. (Of course, the tolerance is an organismal-scale phenomenon and may be explained by some yet poorly understood differences in how two species deal with the consequences of excessive DNA underwinding.) In the study, the authors focused on the GyrB subunit of gyrase as a follow-up to an earlier work from the same group. Salmonella GyrB is 96.6% identical to its E. coli homolog; however, their results demonstrated that Salmonella GyrB, as well as a catalytically defective single-residue mutant which can be reasonably well tolerated by Salmonella, compromised the ability of E. coli to grow. The most straightforward explanation of this observation is that E. coli is sensitive to the levels of supercoiling activity of gyrase and that when the activity drops below a certain threshold, efficient growth cannot be sustained. This is presumably because DNA becomes insufficiently negatively supercoiled. But what is insufficient for E. coli may be well within the acceptable range for Salmonella; after all, the evolutionary paths of the bugs diverged many millions of years ago. Indeed, plasmids isolated from Salmonella are less underwound than those isolated from E. coli (4, 20), suggesting that the balance between relaxation and supercoiling in Salmonella is shifted, relative to that in E. coli, to a somewhat more relaxed state.
Qualitatively, the existence of the species-specific balance states can be rationalized as resulting from E. coli having either lower relaxation or higher supercoiling activity than Salmonella. Although the nonviability of E. coli carrying a supercoiling-deficient mutant allele of gyrB argues in favor of E. coli sensitivity to the supercoiling activity per se, the issue of balancing can be addressed only by a reciprocal swapping of orthologs between the species.
Steady-state supercoiling of plasmid DNA, which is commonly analyzed by the band-counting method with a gel following electrophoresis in the presence of a DNA intercalator in one (10) or two (12) dimensions, is observed in vitro after removal of all proteins. In vivo, however, the supercoils of underwound DNA can be bound by proteins which wrap DNA around themselves, effectively constraining the loops of supercoiled DNA just like histones in eukaryotes, or supercoils can be unconstrained by proteins, freely participating in various DNA transactions and stress-induced reactions. One such reaction is a transition from the right-handed B form of double helix to the left-handed Z form of DNA in sequences containing GC tracks. The transition occurs at some threshold level of torsional stress in underwound DNA with a certain linking number deficit; all DNA molecules with equal or greater degrees of underwinding will extrude the track into the Z form (15). Such extrusion, in turn, will consume some free supercoils, effectively shifting the distribution of linking numbers toward less underwound DNA species. When gyrase acts upon this "relaxed" substrate, it will introduce additional negative supercoils whose presence can be visualized on a two-dimensional gel along with the break in distribution associated with the B-to-Z transition. This is in comparison to the distribution of linking numbers in the population of the same plasmid without the GC segment. Champion and Higgins found that the surplus of negative supercoils, pumped into the GC-containing DNA by gyrase, was significantly less in Salmonella than in E. coli, so much that the Salmonella distribution did not seem to have DNA species that were sufficiently stressed to undergo the conformational transition. Thus, not only is the level of steady-state supercoiling in Salmonella less than that in E. coli, but the extent of unconstrained, free supercoiling is also less: free supercoiling seems to account for more than half of the apparent supercoiling in E. coli and for less than half of that in Salmonella. It follows that the majority of negative supercoils in Salmonella, unlike that in E. coli, are constrained by proteins, and it might be expected that dispensing with the proteins which do that will have a more adverse effect on Salmonella than on E. coli. The results of the study and other evidence are consistent with this hypothesis. Salmonella was much more sensitive than E. coli to the presence of MukB and H-NS, which are known to constrain underwound DNA in vivo and in vitro (16, 22). At the same time, the lack of SeqA, which can sequester also positive supercoils (11), among other things, had no discernible effect on Salmonella growth.
Of course, neither histones nor any other proteins capable of constraining negative supercoils can change the topological characteristics of DNA on their own, as measured by shifts in distributions of linking numbers. These proteins can wrap DNA only in such a way that generates positive torsional stress in a protein-free portion of covalently closed DNA, equivalent to overwinding. When the stress is removed by the topoisomerases, which can react on positively supercoiled DNA, DNA becomes relaxed while bound by the proteins, and when the proteins are removed, DNA assumes its underwound conformation. Since the potentials to remove positive supercoils are indistinguishable between E. coli and Salmonella (17), it is quite plausible that the variations in the sequence of the Salmonella gyrase affect only its supercoiling activity, not its ability to relax overwound, positively supercoiled DNA.
It is also plausible, albeit not as easily falsifiable, that Salmonella reliance on constrained supercoiling represents an evolutionary choice consistent with the Salmonella environment. The principal ecological difference between the intracellular pathogen Salmonella and the commensal bacterium E. coli is that Salmonella has to survive in macrophage phagolysosomes (3), where it is subject to extensive oxidative stress (7). It has been demonstrated that wrapping DNA in nucleosomes protects DNA from oxidative damage (13). It remains to be seen whether the method(s) of supercoil constraint "practiced" by Salmonella has a similarly protective effect.
 |
FOOTNOTES |
|---|
Published ahead of print on 15 June 2007.
The views expressed in this Commentary do not necessarily reflect the views of the journal or of ASM.
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Appl. Environ. Microbiol. | Infect. Immun. | Eukaryot. Cell |
|---|---|---|
| Mol. Cell. Biol. | J. Virol. | Microbiol. Mol. Biol. Rev. |
| ALL ASM JOURNALS |
