E.coli strains are usually leaky, you can fix that by using the BL21 (or others) carrying the pLys S or E plasmids.

For the low levels, you might have to figure out if your protein is unexpressed or expressed as insoluble fraction and if the induction of the expression is toxic for the cell. In this case, I can suggest you to try the C41 or C43 strains as they have been selected to be able to express toxic proteins (one subunit of the ATPase). One could also try arabinose induction which is more smooth.

For protein design changes, we would need to know more about the expression (is it a membrane, secreted, soluble protein?, are there some signal sequences at the N terminal, etc....)

Cheers and good luck

E.coli strains are usually leaky, you can fix that by using the BL21 (or others) carrying the pLys S or E plasmids.

For the low levels, you might have to figure out if your protein is unexpressed or expressed as insoluble fraction and if the induction of the expression is toxic for the cell. In this case, I can suggest you to try the C41 or C43 strains as they have been selected to be able to express toxic proteins (one subunit of the ATPase). One could also try arabinose induction which is more smooth.

For protein design changes, we would need to know more about the expression (is it a membrane, secreted, soluble protein?, are there some signal sequences at the N terminal, etc....)

Cheers and good luck

How do you know that the band you see in both uninduced and induced samples is the protein you are trying to express?

I agree with Alejandro. Can you be sure that the protein you are seeing before induction is your Leishmania protein? Have you tried growing your BL21 cells without the expression plasmid to compare? Can you verify that you are getting leaky expression by running a western plot for your protein on un-induced cells? I think it is more likely that you are not expressing your protein at all, but the above tests will establish where you are expressing any protein before induction.

If you find you are having leaky expression this should still not prevent you from getting good induced expression. This is why I think you probably aren't getting expression at all - well its my guess. But maybe its true your protein just wont express. You can try adding certain tags which can help expression. GST fusions often work, and MBP fusions can help as well. I have had success fusing difficult proteins to GB1.

I recall I have used expression plasmids called pBAD in the past. They are very tight (not leaky) but I don't remember anything else about them right now.

Good luck!

Prob a leaky promoter because you are using IPTG. Change to arabinose for tigher control. Also, The 'ATG' means little here, you may want to look into your codon usage and IF you find something out of the ordinary for E.coli use pRARE or pRARE2 as a helper plasmid (they both encode rare t-RNAs). Also, as eluded to above, is it a membrane protein? Because these can be a problem with expression levels...other than that, does it stop growing? because your protein load could be effectively a poison to your host (or the proteins function...). FYI dropping expression to 30 degrees or lowering IPTG concentration helps for expression some times as well...lots to do for you, many controls.

I assume you have done a western right? making sure your band IS your protein?

Agree with Nicolas Bocquet, some of the systems are leaky so try with pLysS or E plasmid. Codon usage may be a problem for your low expression levels or toxicity of the protein but easy to check codon usage. Try expressing at a lower temperature for a longer time, this may increase your yield.

It could be that the promoter is a little bit leaky, it happens quite often. Another answer is that the band that you see is not actually your protein but something aspecific or maybe that there is something aspecific of the same MW in your uninduced sample. A good negative control would be to try to induce your E.coli strain without any plasmid in. If the band disappear it could likely be a leaky promoter, if it's not the case you could have a problem of expression and maybe try to check as well your IB.

Did you add a tag into your protein?. You can identify your expressed protein by identifying the tag attached to your recombinant protein (His-tag/Anti-His Antibodies). Specific antibodies against your tag will help you a lot to solve the question. By the way, a protein band in the same position than your expressed protein does not necessary means that is the same protein!.

I agree with pretty much everyone so to sum up, these are the things I would test first before jumping into changing the system dramatically (and considering that you have done nothing of the above):

- Do a Western Blot (not a Coomasie, your protein should at least be his-tagged) of both soluble and insoluble fractions. BL21 are awesome bacteria to express recombinant proteins because they are deficient in the main cell proteases. However, that also means that the chances of protein aggregation (especially with a eukaryotic protein) are high from high to very high. In reality I have never been able to express a non-bacterial protein in BL21 without some degree of protein aggregation.

- Decrease the concentration of IPTG; while 1mM is used widely, sometimes the expression of the recombinant protein adds an extra - extra - burden to the cell and it does not induce well so what you may have right now is a combination of very low levels in the induced culture and leakage in the uninduced. Try 0.1mM or 0.5mM, and see how it works.

- Try to induce with less IPTG at lower temperatures but longer times: try 0.1mM at 28C for 5-6 hours instead of 1mM at 37C for 2-3. And here I am assuming you are inducing at 37C. If what you are actually having is a problem of protein aggregation and/or IPTG-induced toxicity, this should be a very good way to try to solve it.

Check your induction times, I have experienced better expression of some difficult to express proteins at higher ODs. say like OD600 of 0.8-1.0.

Also take samples at different time points, your protein might be expressed but degraded/exported after sometime. So check after few hours of induction and compare it after 12 h or 24h etc. So you might need to harvest the culture very early or very late.

Reducing the growth temperature after induction also helps. (as mentioned in above post) e.g 30C, 24C or 17C. I normally do some pilot expressions (5-10 ml) for optimization.

Some proteins can be expressed better with the help of Chaperons, They help in proper folding of the proteins. I didn't get success in my case but others have reported success with co-expressign chaperons. There are kits of different plasmids (from TAKARA) bearing chaperon genes and they can be co-expressed.

I would suggest BL21 are your best bet. However you have not mentioned what was the detection method, was it western blot or just coomassie staining? If your conclusion is based on commassie I strongly suggest running a longer gel to have a better resolution in that region. Probably your protein of interest is masked by over expression other E.coli proteins. Expressing at low temp for longer time has helped for one my constructs using 1.0 mM IPTG (especially membrane associated protein, check nature of your protein)... also follow Nuria's suggestion of titrating IPTG. good luck!!

I agree with all the excellent suggestions given above. In one of the lectures I recently attended at Univ. of Toronto's Molecular Techniques course, a Professor recounted his experience of finding same protein band in both induced and uninduced lanes. It turned out after doing a carefully designed western blot with a lot of good positive and negative control lanes and a reliable antibody that his induced protein had been modified in some way in E.coli to a HIGHER molecular weight .

Just thought you may want to consider that... look closely at protein bands HIGHER up from 35kDa and see if there is a thick band higher up than 35kDa in the INDUCED lane and if this band is ABSENT from the uninduced lane.

Maybe you will end up shouting "Aha!! Eureka !! "

Some members in my group also had similar problem, and they solved it like this: first, make sure it is your protein by western blot or better MS; then try to use glucose as an inhibitor to inhibit the leaking; and you can also try pBAD/TOP10 expression system, pBAD is a tightly controlled vector, so the leaking problem should be solved in this way.

Good luck!

Leaky expression is common during bacterial protein purification.

My suggestion

1. Check your plasmid construct first if everything is alright

2. Just express it in bacteria (both induced and non induced with IPTG), test you protein expression in western blot (Careful some time it will give huge signal you cannot control in WB use very little amount of protein), As a control your can simply use untransformed bateria or transformed with vector only.

3. Avoid using too old IPTG

4. If everything above is alright then you need to troubleshoot with the low expression.

Low expression

1. First check you protein is in soluble or insoluble fraction (Lyse bacteria spin, use both supernatant and pellet seperately run on gel to test)

2. If it is in insoluble fraction try various condition to make it expression in soluble fraction for this you have to try various concentration of IPTG, Changing the incubation time (For example high IPTG for short time or low IPTG longer time), Culturing temperature (37'C, 30'C..)

3. If it is in insoluble aggregate you can try denaturing method (probably using urea in your lysis buffer) for you purification, and you need to renatured again at the end.

3. If non of the above is not working you need to codon optimize for E.Coli expression or you need to use some helper plasmids which provide rare tRNAs

Goodluck!!

leaky expression. I had this situation before. You won't get too much protein from leaky expression, IPTG inductions is still required.

Most likely you have a leaky promoter. Use IPTG induction with tightly controlled T7 promoter system.

As our colleagues have suggested you have a leaky promoter and a basal level of constitutive expression, that is assuming the band you are seeing on gels is the recombinant protein of interest ! Assuming this is the case,then iIf your vector has a lac-based promoter you should check the following. Does the multi-copy vector also carry the lacIq mutant repressor to facilitate repression...? If it is on the chromosome the lacIq can quickly get titrated out. Also, many people are not aware of this, but you can make use of Catabolite Repression by adding glucose to the growth media which will repress the promoter further and prevent/ reduce promoter leakiness.

To my knowledge this promoter is leaky, and in the cases like yours when promoter is leaky and expression is low (is codon usage optimal for E. coli?) you wouldn't be able to see the difference between induced and non-induced band (which you have to confirm with Western: I think this vector is used for His-tagging) Also, you have to do expression trial every time you work with new protein: to induce small amount of bacteria in different way (37 degrees few hours: at this temperature you should not go over 6h; or at 18 or 20 C degrees 16 hours) and with different concentrations of IPTG: usually from 0.1M in few steps to 1M. You can then load same amount of in water boiled bacteria on SDS-PAGE (I also used 8M urea) from each expression conditions and compare them (even better do the Western). This should give you the best conditions for expression of your protein. I even know the case where expression 2 days with 2 inductions with 0.1 M IPTG at 18 C Degrees helped (because protein was going mostly to inclusion bodies: this way, small amount of IPTG helped because it gave time to the protein to be properly folded and increased the amount of soluble protein in total). If this doesn't work, you can change the construct; put your tag to different end of protein, change the tag (use GST, SUMO etc) Good luck!

I agree with all of above suggestions! Here is my 2 cents. You might consider trying different fusion tags, such as sumo/His double tag (Enzymax) and it will add about 20kda to your protein. So, your fusion protein will be at ~55kDa position and it is easy and cost efficient to remove the double tags after purification. You can use His tag for purification and SUMO tag for fusion tag cleavage by Ulp1 protease. Here is the why:

1. Enhance protein expression and solubility because of the SUMO sequence.

2. More efficient and cost effective protease comparing to Thrombin, TEV, and EK, and only need 1ul (12U) to remove fusion tags from 1mg purified protein with >95% cleavage at 4*C for 16 hours.

Ulp1 is a highly active SUMO protease that specifically recognizes the tertiary structure of ubiquitin-like protein, SUMO, rather than a short sequence as other commonly used proteases such as Thrombin, TEV, and EK. After digestion, the His-tagged Ulp1 can be removed by His tag affinity chromatography.

For more details:

http://www.enzymax.net/sumo%20protease.htm

Good luck:-)

pQE41 vector contains T5 promoter. This means that this promoter is recognized by the E.coli RNA polymerase, not T7. In the pQE41 T5 promoter has only one site for lacI operator and no additional gene for lacI repressor. This is a very leaky system that can be readily induced by lactose contamination in the growth media (especially if casaminoacids is among the compounds). This system is also can give very little yield, because it utilizes host RNA polymerase, and T5 promoter should compete for the enzyme with thousands of the host genes.

As many others here suggested, use T7 promoter containing plasmid with additional lacI copy and use BL21 strain or its derivatives with the gene for T7 polymerase integrated into its genome. After you obtain sufficient quantity of protein then verify its solubility and other properties needed for your downstream applications.

You might also want to try to express your target protein using 'Auto Induction' method. This method can be developed for any expression system in which the expression of target proteins are induced by a shift in a metabolic state as a result of growth of a culture. using this method, you take advantage the high culture densities that produce more of your target protein per volume of culture than common IPTG induction. This article will answer all your questions about protein expression by auto-induction in E.coli.

(Protein Expr Purif. 2005 May;41(1):207-34.) by F.William Studier

Hope thats helps,

Good luck,

I agree with most of the suggestions you have received already... IPTG-inducible promoter systems have never worked right for us either... I agree with Duncan that switching to a pBAD / arap system is much better... we have gotten our best results with these systems with the caveat that we are expressing Salmonella proteins in Salmonella or E. coli...

My two cents: The transcriptional control of your construct is not tight, which may be the result of low levels of lacI repressor (lacIq has ten-fold higher expression, try a vector which has it), poor binding to the operator sequence (which may be enhanced by tandem operator sites), or a really high background transcription (try using vectors with heterologous promoters like T7, which may be used in BL21 strains).

We encountered the same problem and found different media can also cause this problem. We resolved this problem by different commercial medium.

Leaky promoter??

Do you have a negative control that contains the same plasmid, but without this protein? The band you see could be non-specific. It can be important to have a negative control carrying a null (no insert) vector; for example, from many plasmids selected for chloramphenicol resistance, chloramphenicol acetyltransferase can make up a large fraction of total protein and could be confused for your protein of interest.

It's difficult to give more advice without more information on your expression system or controls. If this plasmid contains the T5 promoter tested here - http://parts.igem.org/Part:BBa_K592008 - then your results might be typical; very change in induced conditions.

How do you test for the expression of your 35kDa protein? western blot or just comparing it with empty vector tranformed E.coli? If you are sure that your protein is expressed by not induced by IPTG then the issue is different.

As others advised, if it it is a membrane protein, sometimes expression is difficult. Try to clone it under some other tight promoter to avoid leaky expression. Also, as per my experience play around with low IPTG concentration for induction and at low temperature or duration of induction make it short.

Hi

In addition to what other people have suggested, like changing to arabinose promoter. I would suggest that you include 0,5 % Glucose in your culture medium. As you cells are growing before induction, they will use Glucose as the a carbon source (though not necessarily primarily) As they use glucose this will result in the silencing/suppression of the operon. Once you add IPTG, your operon will be opened up. I have successfully used this started to express many of my recombinant protein, even if I am using an Arabinose operon

All the best

Please paste your sequence (from the start codon till the stop codon) in the below link (Genescript - rare codon analysis tool) and select the host organism that you want your protein to be expressed in (in your case E.coli) and you will identify the rare codons that could hamper the expression efficiency of your protein.

http://www.genscript.com/cgi-bin/tools/rare_codon_analysis

In addition you can also use Graphical codon usage analyzer to determine the quality of both non-optimized and optimized sequence for E.coli.

http://gcua.schoedl.de/

This will take only few minutes to check your non-optimized sequence.

To optimize sequence, use Gene optimizer (from GeneArt, Life Technologies) to optimize the sequence and then paste your sequence again in the above 2 links to find out whether it is optimized. This tool works best compared to other sequence optimizer tools available online. I checked the quality of optimization of other tools and this was my experience.

You can also add desired restriction sites on both sides to assist with cloning in your protein expression vector and check again for optimization and then order for gene synthesis. Make sure your protein sequence does not get altered while doing this. Add additional nucleotides if necessary.

I followed the above approach for my protein and worked out well. If it does not work, which I think is highly unlikely, and in addition you can also adopt the above suggestions of reducing induction temperature to 18 or reducing IPTG to 0.5mM or including 0.5% glucose and also determine the right time point after protein expression to lyse your bacteria.

Leaky expression by your promotor

As others have stated, the T7 lac promoter is leaky and will produce some expression without IPTG. If you want to minimize this leaky expression, you should use a bacterial strain that also has pLysS, a plasmid encoding T7 lysozyme. Not only does lysozyme help to lyse your cells in downstream purification steps, but it also binds to the T7 lac promoter and keeps it more tightly off before induction with IPTG.

I have in the past compared +/- pLysS expression, and it did decrease background expression for me.

Here is an example of a strain:

http://www.promega.com/products/cloning-and-dna-markers/cloning-tools-and-competent-cells/bacterial-strains-and-competent-cells/bl21_de3_plyss-competent-cells/

Oh - and if the protein you're expressing is toxic to cells, having tighter control with pLysS can also help keep your cells happy before you induce which might give you better expression overall.

You either have a leaky system or an E.coli protein of the same MW. As someone else mentioned do you know by western that the uninduced protein is truly your protein? Empty vector is a good controI. If you know it is your protein why is that a problem, as long as you get lots more when induced. Sounds like there isn't a large increase in protein of that size with IPTG induction. instead of inducing with IPTG,use a Magic media recipe. This medium keeps the T7 lac promoter turned off until all carbon sources have been utilized.You should get much higher induction and let it run longer. It adds lactose and glycerol as energy source. Lactose can substitute for IPTG by de-repressing the same lac operator, but is usually prevented from doing so by other media compounds. These are depleted during growth at which time induction will begin. The target protein is induced automatically, without the need to monitor growth and add inducer at the proper time. Studier called this process auto-induction.

F. William Studier, Protein Production by Auto-Induction in High-Density Shaking Cultures .Protein Expression and Purification 41: 207-234 (2005) and http://www.bnl.gov/biology/people/studier.asp

Joan Hare is absolutely correct. It is either 'leakiness' of the promotor (which is actually quite hard to circumvent). Amanda Solem's suggestion of using one of the pLysS or pLysE strains (where the T7 Lysozyme is expressed, and binds to any T7 produced because of 'leaky' expression, and inhibits it.

BUT - Joan's comment about you possibly being mislead by an E. coli protein of the same size is absolutely something you need to check. This is the major worry here.

As she states - if it really is leaky expression of your protein, which increases when you induce, why does it matter?

If the band intensity does not increase when you induce, sounds more like an equal sized E. coli protein!

About 35kDa - is there a heat shock protein or chaperonin that size? It could be produced in increased amounts due to the protein you are expressing being misfolded. Auto-induction as suggested is what I would try, and also expressing the protein slowly in case some unknown required cellular process is a limiting factor to getting properly folded protein.

I find this thread very interesting. We  have a protein which is giving us a possibly related problem.

We have a small (~50 residues) soluble protein which we have cloned into a pET-19b plasmid (this has ampicillin resistance, lacI, and an origin of replication). The protein has a HIS tag.  Anyway, some 33 mutants of this protein expressed with no problem at all.

Two of the mutant plasmids had problems though. The two transformed perfectly well into our methylation strain, DH5α. We plated (into TB plates with ampicillin), restreaked, and then innoculated into liquid media (also TB with ampicillin). Then we miniprepped to extract the methylated plasmids, and transformed into BL21 for expression. 

However one of the mutants failed to make any BL21 colonies at all on the plates, even after two tries. The other mutant failed on two attempts, and formed ONE and only one colony (again, I am talking about BL21 colonies) on one of the attempts.  This colony was expanded in liquid media and appears to have expressed its protein just fine, meaning that OD increased in the normal mount of time. (though we have not yet measured its stability or other characteristics).

I wonder, is it possible that BL21 has leakier expression than DH5α? And that these two proteins are somehow toxic to the cell? Also, in the one case in which a single colony formed and we were able to grow it in liquid media, is it possible that this problem affects growth on agar but not in liquid media? Maybe it is only toxic in the very early growth stage?

Sam

Sam,

Verify that the minipreps did indeed yield plasmid DNA. If they did and everything was OK in the transformation (checked using a control plasmid), then the simplest explanation is that the mutants are somehow toxic to the host.

The simplest solution is to decrease basal expression as much as possible. Try adding glucose to the plates/cultures (anywhere in the 0.2 to 2% range) to see if you get transformants. Glucose inhibits basal T7 RNA pol expression by inducer exclusion (see e.g. PMID 9133663). You can also try the Walker strains (their names are C41 and C43, I think) in combination w/ glucose. They have tighter control of basal T7 RNA pol expression, due to mutations in the lacUV5 promoter.

DH5 alfa will, of course, be less leaky than BL21(DE3). It does not have the gene for T7 RNA polymerase, hence no expression from the T7 promoter in your plasmids.

Hope it helps!

Hi Alejandro,

Thank you for pointing out the inducer exclusion effect, which this mere structural bioinformaticist did not know about. The absence of T7 RNA pol could explain the thriving DH5α's.

My colleague Nissbeck tells me that the presence of a single colony rather than many tiny colonies argues for low methylated plasmid yield (your first suggestion) rather than protein toxicity. What is odd is that we repeated the miniprep for one of the two, and this did not fix the problem. It seems unlikely that we would do this wrong twice for the same mutant even when aware of a problem, but I guess it could happen. Measuring plasmid concentration and adding glucose are both easy experiments so I guess I could try both.

Did you mean glycerol rather than glucose?

Thanks again for your educational response.

Sam

No, I meant glucose. Glycerol should not work, at least not in theory (no inducer exclusion, and no catabolite repression, either).

We measured the plasmid concentration by nanodrop for the three plasmid solutions and got around 50 ng/uL.

We decided to compare against a couple of mutants that had no problem forming colonies. They had higher concentrations. It also appeared that the concentration was related monotonicaly to the number of colonies. so we ended up making 11 nanodrop measurements to try to get a trend. It appears that nanodrop concentration is correlated with number of colonies, and concentrations around 50 are in the danger zone — this seems to be more or less the minimum concentration needed to get any colonies at all.

it seems to us that if the concentration is around 50, then we should transform with 2 uL of methylated plasmid, rather than 1 uL. Actually it seems safer to just use 2 uL always, so we don’t have to make a nanodrop measurement. What do you think of this solution?

Me acabo de dar cuenta que estas en la Habana. Es el centro que descubrio la terapia para pie diabetico, no? como les va por alla?

Hola Sam

¡Tu español es excelente! Sí, es el mismo centro en que estás pensando :-)

If the transformation frequency of the cells is low (as is often the case with BL21(DE3))  then yes, 50 ng might not do the trick. At that DNA conc. range, it is safe to transform with volumes up to 10 uL if you are using around 100 uL of cells/transf.

Hey Sam!

I hope your experiments are working!  Here are some little ideas about your plasmid concentration.

 in the nanodrop,be wary of the accuracy of low concentration readings.  (For me 50ng/ul is getting close to the edge, but probably means you've got something.). To be sure, you should look to see if you have a salt peak (a big peak around 200-230nm). A big salt peak can trail into your DNA absorbance - in some cases it might make your DNA concentration appear slightly higher than it really is.  In other cases your 'DNA' absorbance can be almost entirely salt!

Also, you want to make sure it is plamid DNA, not degraded DNA.  You can check some on a gel to make sure it is intact (and semiquantitative ladders also may let you guesstimate low concentration samples that confuse the nanodrop).  You might also try a fluorescent method that can more accurately detect ONLY double stranded DNA like a qubit or similar product if there is one nearby.

Best of luck!

Hi Alejandro,

The transformation efficiency of BL21 is on the order of 10^7 colony forming units per ug of plasmid.I should get half a million cfu's from 50 ng.  This doesn't make sense to me. Most petri dishes done under such conditions would have on the order of 50 colonies. Am I missing a factor of 10000?

My colleague decided our transformation technique was just too sloppy, he got the colonies to grow by keeping to the protocol within +/- 10 seconds. I guess the thaw, heat shock, etc, is more sensitive to time than I had appreciated.

Saludos desde la tierra invernal. Espero tengan hoy bonito dia para ir a la playa!

Hi Amanda,

Good to hear from you again! You are right, the nanodrop has lots of information that we missed by just writing down the concentration.

Hi Sam - I'm so glad you solved your problem!  Hooray for Nissbeck and Alejandro!  Yes, transformation protocols are precise in timing - some steps are more important than others to time precisely (as you probably know already!).  The most important step to do precisely is the heat shock - here seconds matter.  In fact, so can the type of tube you use!  When you thaw on ice with plasmid before heat shock, you can sometimes shorten the time by a few minutes (occasionally a company will show a graph on transformation efficiency vs that preincubation time) with only a small decrease in efficiency.  So that could be something your colleague found to be important when you needed max efficiency.  The recovery phase before plating is tricky, because the requirement for the full hour depends on the antibiotic you're using and how it works.  For instance, with amp you can cheat and plate a little early, with kan you can't.

EnjoySweden!

Thanks guys!  Never realized social media could be so useful. To me it was all about pictures of kittens.

Hi everyone, I've come across this thread as it reflects the problems that I have been having. I have cloned my gene of interest (amplified from gDNA) into the pET-21b vector in frame with the 6xHis tag. I am expressing the protein (about 30 kDa) in CodoPlus RIPL cells using 0.1mM IPTG. After running a SDS and a WB I can see my protein present, however with several concerns. I should mention that I am 99.9% sure this is my protein of interest as I have also analysed it by MS. So the the gel and blots show the following:

1) significant expression of protein prior to induction

2) after lysing induced cells, protein appears to only be present in the pellet rather than the supernatant suggesting insolubility and/or aggregation....?

Looking at the comments on this post, I can see that perhaps I should try to either use glucose to suppress basal expression prior to adding IPTG or try autoinduction? Does anyone have any other reflections? I can also mention that I have tried cloning into a different pET vector and expressing with Rosetta cells, but with the same results!

Any help will be much appreciated...

Thanks

Julia