Abhin is right. The transformation processes in B. subtilis and E. coli are fundamentally different. In B. subtilis, which is naturally competent, linear DNA is actively taken up by the cells. One strand is degraded and the other is recombined into the chromosome. Closed circular DNA is not a substrate for such a process. In E. coli, DNA passes intact through a weakened cell envelope; closed circular plasmid reaches the cytoplasm in a state ready to begin replication, which is why it transforms better. Linearized DNA would need to be religated or recombine in order to establish itself before the intracellular nucleases degrade it.

Bacillus subtilis is a natural competent cell and can retain linear plasmid from lysed cells in natural condition(Natural Competency) but not in case of E.coli, which can be only transformed artificially with the aid of vectors; generally plasmid vectors and plasmid vectors have property of being circular.

In E. coli the circles can be maintained as free-floating Episomes. They get replicated in that form by DNA polymerase that goes around and around... In Bacillus, you want to integrate the plasmid into the genome. You need free ends so that the plasmid can thread itself into the genomic DNA strand as a first step towards recombination.

You can transform E.coli with linear fragments (PCR, ...) but the efficiency is lower.

Transformation of PCR fragments is commonly used in recombineering. Very straightforward.

Generally, wild type E. coli has lots of exonuclease which will degrade linear DNA very quickly so circular plasmids are better for transformation of E coli. However, many B subtilis strains become nature competent at high cell density and they prefer DNA multimers (big fragments, about 8 kb to 12 kb).

Supercoiled DNA transforms E.coli at least a hundred

- fold better than relaxed open-circular DNA. This gives a byas towards uncleaved vector during cloning. This isn't a direct answer to your question but something to bear in mind. Linear DNA can be packed in vitro in cosmids or lambda vectors if the vector carries a lambda cos site. This will circularise in the cell and be propagated there episomally or as phage. The linear DNA has to be ligated at high concentrAtion before packaging.

There should be no difference because of simple reason that before transformation, you will have to make an incision and then do the negative supercoiling.

Abhin is right. The transformation processes in B. subtilis and E. coli are fundamentally different. In B. subtilis, which is naturally competent, linear DNA is actively taken up by the cells. One strand is degraded and the other is recombined into the chromosome. Closed circular DNA is not a substrate for such a process. In E. coli, DNA passes intact through a weakened cell envelope; closed circular plasmid reaches the cytoplasm in a state ready to begin replication, which is why it transforms better. Linearized DNA would need to be religated or recombine in order to establish itself before the intracellular nucleases degrade it.

Both linearized and circular plasmid could be good for transformation but it depend on the situation what you are trans forming because both have advantages and disadvantages. With my experience circular plasmids has more efficiency than the linear one. Working with a plasmid derived from a bacterial source has enough machinery to keep itself replicating and less chance to get into problems with the work seems much simpler.

Linearized plasmid is not much suitable for transformation in E. coli because of exonuclease attacks located in cytoplasm. It is not only hard to transform E. coli with relaxed circular plasmid DNA, especially after recombination-ligation process, but it is also harder to electroporate relaxed circular or newly ligated plasmid DNA into E. coli. If you plan to electroporate your recombinant plasmid, you may subclone or let it replicate in CaCl2 competent E. coli, such as DH5-alpha, DH10 etc. Then isolate supercoiled form of your plasmid and use electroporator for transformation. It would be much better, more recombinant colonies.

As most of them said, Bacillus subtilis can retain linearized DNA because of its natural competence while E. coli a suitable circular DNA. Good luck.

As most of them said

There are some more questions in my mind regarding this. 

1.what about the endonucleases, isn't it cleaves circular DNA also?

2. linear DNA can also be religated by DNA ligase present in the host. Does it occurs?

what exactly is the reason for transformation efficiency being high in circular DNA and low in linear DNA? I am really being in a perplexed state about this topic. Would be really thankful if I could get an answer to this.

 @Partha you asked, "1.what about the endonucleases, isn't it cleaves circular DNA also?”..

E coli and B subtilis inhere endonucleases and exonucleases.

Endonucleases digest both linear and circular DNA.

Exonucleases digest only at DNA ends. Circular DNA has no ends and, therefore, cannot be digested by exonucleases. That is a major reason why circular DNA is digested more slowly than linear DNA within bacterial cells. 

You also asked, “2. linear DNA can also be religated by DNA ligase present in the host. Does it occurs?

Ligase is present within cells at a much lower concentration than exonucleases and endonucleases.  

Digestion by nucleases is ONE reason why circular DNA transforms E coli more efficiently than linear DNA. 

Wikipedia has two entries which describe transformation. They include the following, "The chemical method, however, usually does not work well for linear DNA, such as fragments of chromosomal DNA, probably because the cell's native exonuclease enzymes rapidly degrade linear DNA. In contrast, cells that are naturally competent are usually transformed more efficiently with linear DNA than with plasmid DNA.”

Transformation efficiency:  https://en.wikipedia.org/wiki/Transformation_efficiency

Transformation (genetics):  https://en.wikipedia.org/wiki/Transformation_(genetics)

Each includes links to source material in pier-reviewed journals including data on DNA deletions and rearrangements much greater for linear DNA than for circular DNA when transformation of E coli is attempted.

@Paul Walsh, I have gone through the links that you have provided. Now my confusions are clear.

Thanks for explaining :) 

Regarding transformation of PCR products for subsequent genomic recombination in E. coli, what about recircularizing these PCR products for more efficient transformation and higher recombination efficiency? In my hands the main bottleneck of recombination-based experiments is the poor efficiency of transformation of the exogenous linear donor DNA (PCR products) and I wonder whether prerecircularizing this donor DNA would improve the final outcome. In principle, if recircularized, more donor DNA would be transformed and therefore available for recombination and, as far as I know, the difference between recombination efficiencies of linear vs circular DNA is negligible.

Has anybody ever tried this approach?

That's an interesting idea, Mikel, and I for one would like to know the answer. Might I suggest that you post it as a separate question? 

After having the expression system selected (E Coli, B subtilis, etc) it is important to design the gene that we are going to express according to the following parameters: IAC (codon adaptation index); % GC, codon distribution.This will allow us to evaluate the efficiency of transformation