So much for the mix-up hypothesis. Another possibility is that your prep contains two different plasmids, particularly if you performed the plasmid extraction on a strain that is used for recombinant protein production. It is common for such strains to contain another auxiliary plasmid, e.g. pLysS, or plasmids encoding chaperones, or rare tRNAs etc. One thing you could try would be to transform the current plasmid prep into plasmid-free cells, pick up a few transformants, make minipreps and check by gel electrophoresis and restriction mapping whether some of them carry the plasmid you are interested in.

The middle band that you interpret as the linearized form of your plasmid could also be a concatemer (dimer) of the bottom band, both being covalently closed, circular forms. After cutting with a restriction enzyme for which there is a single site, you get a single band with a size of about 8 kb. This means that the plasmid you are analyzing on the displayed gel and the one you sent for sequencing are not the same (sample mix-up ?). You could further confirm this by performing other digests, including those that should yield more than one fragment, and checking whether the band pattern that you obtain is even partially compatible with the pattern predicted from the sequence (most likely it will not).

Pierre Béguin Thank you for your input but the sample is the same sample I sent to get sequenced. I sent a portion of the sample for sequencing and used the rest to do the above gel electrophoresis after restriction.

Additionally, I asked the sequencing company if they know why this is happening, and they too confirmed that their gel electrophoresis also shows 3 bands with the middle band at around 8k which is not the expected position of the bands from 4k plasmid. They suggested me to do a restriction to linearize the plasmid and run the gel again, but I told them I already did that. They were too unsure why this is happening.

Could the plasmid be forming some kind of weird conformation that impedes its migration down the gel? I'm just imagining things but maybe a portion of the plasmid forms a huge complex hairpin? I don't know... The results don't make sense and it's driving me crazy... I've repeated this three times now..

Anyway thank you very much for your time. Please tell me anything that pops into your mind!

So much for the mix-up hypothesis. Another possibility is that your prep contains two different plasmids, particularly if you performed the plasmid extraction on a strain that is used for recombinant protein production. It is common for such strains to contain another auxiliary plasmid, e.g. pLysS, or plasmids encoding chaperones, or rare tRNAs etc. One thing you could try would be to transform the current plasmid prep into plasmid-free cells, pick up a few transformants, make minipreps and check by gel electrophoresis and restriction mapping whether some of them carry the plasmid you are interested in.

I think That Pierre Béguin is probably correct. If the plasmid has been self ligated then the sequencing primers will anneal in the same sequence ang generally read the same sequence round most of the backbone except where the single cut sites are re ligated. I think that enzyme digesting with REs in the backbone will give expected bands as for 4kb but at least one different size band if the plasmid is self ligated. It would be interesting to see if the sequencing produces only the 4kb sequence or if one of the sequences has additional bases

Thank you Pierre Béguin and Paul Rutland

The strain containing another auxiliary plasmid is something I have not considered. However, the sequencing result shows the plasmid contains the sequence for the protein I need. So the bands must be from the plasmid I am hunting for. Correct me if I'm wrong.

I have some comp cells available so I might as well give what Pierre Béguin suggested a go.

I didn't quite grasp what Paul Rutland said and I have some questions. Are you suggesting that restricted plasmids are self ligating with each other (only 2 at a time)? Even if that were true, I don't understand how the sequence will be different where the single cut sites are re-ligated. After all, it's the same sequence repeated twice right? Also, if this is self ligation problem, I think there should be multiple bands appearing in single cut plasmid gel electrophoresis. For example, single 4kb, dimeric 8kb, trimeric 12kb, and so on. There is no reason for only dimers to appear if the sticky ends are self-ligating.

Hi Keunmin Ken Lee Yes I was suggesting self ligation. In the case of insertion into 2 enzyme cut vector when both ends have re ligated the vector is a closed circle again so there are no 3,4 etc products possible. In the case of a single cut insertion I can only assume that after the first end ligation the close proximity of the free ends of the dimer favours the second ligation of the already joined vectors so 3,4 etc may be possible but are just less likely. I do not think that self ligation is just a duplicate of single vector cutting. Mostly that is true for cut sites in most of the backbone but I think that for a cut site clos to the ligation point points there may be an additional band or size difference from the dimer.

I would double digest the plasmid. Do you know which sites are used for cloning? I would use two reaction sets: In one I would use restriction enzymes that were used for cloning and in other set one enzyme from these two and one enzyme other than the two (I am assuming the sequence is known). This would help you to predict and distinguish between the pattern of bands on the gel and solve the issue

Paul Rutland Again thank you for your explanation. I have solved the problem!

Ameya Dipak Bendre Thank you for your answer. I have ordered different enzymes (blunt end enzyme, one that cuts in the middle of protein codes, etc) and cut the plasmid. However, I found what was wrong elsewhere. I still did the experiment you suggested, and the segments showed bands where expected!

The problem was the Loading Dye. The loading dye I have apparently had staining agents too for UV illumination. I already have RedSafe in the gel for UV visualisation. After I changed the loading dye, the bands started to appear at 4kb mark. Even the previous samples that I had frozen gave the correct bands. My guess is the staining agent in the loading dye was weighing down DNA making it appear more massive than it really is. While it is still puzzling how loading dye with staining agent can make DNA bands appear TWICE the actual size, the bands now appear correctly and I'm glad I got this solved! One thing to note is that uncut plasmid does not change its band pattern. So I guess that means as Pierre Béguin said, the middle band is probably a dimer of the bottom band (all makes sense now!).

Thank you all for your input! Although it is a bit of a disappointing end (my personal mistake..), this has been a good lesson.

Well done Keunmin Ken Lee . I knew that ETBR slowed dna down by about 10% and that sometimes REs if used in excess can stick to dna and make it run slower but have never seen it have such a pronounced effect. What made you think of the loading dye?. What was in the loading dye please and did it have SDS in it?.

This is not a mistake it is a valuable learning situation.

If the uncut plasmid is unchanged then I wonder if the RE digest of the cut samples was not long enough and some enzyme annealed to the dna. Adding SDS would dissociate the enzyme from the dna and sizes could look normal again.

well done and thanks for letting us know the answer

Paul Rutland Sorry for the late reply!

I thought of the loading dye because I was thinking of what could slow DNA down. I kept thinking about where the problem could be. I kept looking at the pictures of my gel then realized sample bands were a different color from the ladder. I thought maybe something in the loading dye could be making things go all funky. The answer was there all along! I just failed to see it fast enough :(.

The product is called EZstain Loading Dye by Enzynomics. It says it is designed to replace EtBr. Doesn't seem to have SDS in it though. I'm calling it a mistake because I should've read the product description before using it. I actually did read it numerous times, but never thought this could be a problem because like you said, the effect is very pronounced. I didn't suspect extra staining would cause such a huge difference. Especially because I thought only so much staining agents (no matter the type) could bind to DNA and the rest would just not bind.

I'm not sure about RE annealing to the DNA. Uncut plasmid now appears the heaviest and REs were never added to these samples. I also inactivate enzymes after each digestion at the required temperature. If this process denatures REs, I think they won't bind to DNA anymore.

Thank you for your help! I wanted to share the conclusion of this mystery because I've seen some posts here that don't have a satisfying closure and it is really annoying!

Keunmin Ken Lee thank you for the answer. It is very useful to learn more about the subject as well as knowing the answer giving good closure as you say. I agree the protein binding is clearly now not the answer. I have looked for information about multiple dye binding because EtBr unwinds the dna and does leave spaces as the strands elongate so possibly redsafe does a similar thing but I have not found anything about other dyes fitting inside the spaces but I am not going to knock a good practical result. Well done and thank you for the update.

PS I certainly would not see this as a mistake but very much a valuable learning experience.

Some enzyme sites have a methylation effect, so it is suggested that the efficiency of enzyme digestion should be considered