Have you been able to transform the NEBuilder® Positive Control with success? Have you tried other e. coli like NEB Stable? 

No colonies does not mean your assembly failed.  It means your transformation process isn't efficient enough to tell.  Even in a completely failed assembly you should have scores of colonies from vector carryover and false assemblies.  What do you mean by you "controlled" for it?  Do these cells give you 100s of colonies for other ligation reactions? What is your efficiency with supercoiled fresh miniprep?  

@Nicholas Tremblay, yes, transforming the E. coli with the Gibson assembly positive control was a success. No, I haven't tried other E. coli. I'll see if I can.

@John Schloendron, I see. Well having controlled for them means I have assembled the Gibson positive control with success and transformed the E. coli successfully with other plasmids (pQE80L for example). Actually I haven't paid close attention to the cell's competency. I'll do that.

In the mean time, can you advise me on the overlap region? What's the range for length, GC content and Tm? Other than that and secondary structures, do I need to worry about self-dimerization of the overlap region?

I have some suggestions for you.

1. Technically, 80bp overlap should be no problem for GA. The problem is in transformation (assembly) as John Schloendorn mentioned above. 

For Gibson assembly, you must consider the concentration of each DNA fragments according to the gibson assembly manufacturer recommendation, rather than just consider the quantity like 2 or 5ul. If you are using NEB, then for 1.5 -1.9bp DNA  you will need 0.04-0.05 pmols for each fragments. In case you don't know how to calculate it, here is the formula: pmols = (weight in ng) x 1,000 / (base pairs x 650 daltons). And according to NEB, it is recommended for using no more than 1 pmols in TOTAL per reaction for 4 fragments assembly.

2. I wonder if one of your fragment has origin for replication in E.coli? Just wonder.

3. If you consider to do PCR to amplify each fragments with 30bp overlap and worry about nucleotide substitution, you will better use high-fidelity polymerase like gFlex or KoD. And make sure you follow the manufacturer recommendation as I mention in point 1 (especially for asembly).

Regards,

Wahyu

@Yudhistira, yes I have considered the number of molecule for each fragment. I used 0.25 pmol of each fragment totalling in 1 pmol. Yes, one of the fragments has E. coli origin of replication, the pUC type. And yes, we're considering high fidelity, although I'm still worried about substitution (or shouldn't I be?)

Hi,

you shouldn´t worry about amplification errors as long as you´re using a proof-reading polymerase. I am doing 6-7 kb amplifications regularly and as long as the GC-content is in a normal range, there never occured a problem from PCR. What i had problems with were wrong primers or in vivo cloning strategies (TAR), where the host organism introduced mutations. PCR-amplify your fragments, do a DpnI-digestion and check on a gel for integrity and if there´s really only one single band, then do a PCR cleanup and measure DNA concentration. I also did Gibson assembly with 4 fragments during my master´s thesis a lot, also with sub-optimal concentrations, and it worked 90% of the time. Design a colony PCR strategy, which gives you only bands when at least 2 fragments covalently joined, otherwise it is sometimes possible to get false-positive results from re-amplifying DNA that is on the outside of the cells from the transformation.

Good luck!

Then I don't understand what do you mean by "tried first with 2 microliters, and then next with 5". I did Gibson assembly with 75-80bp overlapping fragments and follow the NEB recommendations thoroughly, it worked fine for me.

1. If you are that worry about mutation, I wonder how did you generate your current 4 fragments and how can you sure those fragments do not contain wrong nucleotide sequence? are they digested from individual plasmid or they are PCR products? if from plasmids, have you read/confirm their nucleotide sequence before use for GA? if those fragments were from PCR, did you purified them (by gel/another kit) before GA?(unpurified PCR products will reduce GA efficiency) and how can u sure no nt substitution on those PCR products?

2. If u believe to do PCR amplifying fragments with 30bp-80bp excess will solve this, and worrying about nt substitution despite using HF polymerase, I recommend these options:

Option 1. Perform PCR to amplify those 4 fragments and subcloning them individually. Then read the sequence of each cloned PCR product to confirm there are no nt substitution. After confirmed, do digestions and gibson assembly. This method can be done ONLY if it's possible to find/attach unique restriction sites in each fragment. If no,

Option 2. (standard method). Perform PCR, purify PCR products and do Gibson assembly. Purify 4-8 out of all resulting colonies and read the sequence of each plasmid. In my experience performing this method, I found no mutation occurred and the plasmid worked well for expression. I used Tks Gflex from takara.

I would strongly recommend subcloning first. Get the NEB PCR cloning kit and use Q5 HiFi polymerase. If you do it right (calculate conc., 3:1 insert:vector, NEB cells from the kit), you should get 95% clones right. From there, re-amplify your fragment by PCR (Q5) and then try the assembly.

Looking at the number of fragments, you might be out of luck. Reduce it by subcloning pairs (the same DNA for GA with 2 fragments) and then try GA. For long constructs, 40-80 bp is sometimes required. Try using PAGE-purified primers for single primers or split them 50/40 (10-15 bp overlap) using HPLC-purified primers.

NEB admitted that assembling such long fragments using their kit was beyond their capacity. The largest they got was a single 7.3 kb fragment into a vector to get 12-15 kb total.

If I were you, I would balance rewards vs time and money spent. Sometimes, it is not worth it and you should change strategy. For starters, use Gene Art Seamless PLUS kit with their Higher Order assembly option.

The other option is doing it stepwise by using PCR overlap extension. You have to subclone fragments using NEB PCR kit (highly recommended) and then try assembling parts or play with them.

In-fusion system will work by recombination but it has a 10 kb size limit for insert +vector. I know some people who ditched GA and went to the In-fusion cloning because of the same problems I see in forums. NEB has many outstanding products but some require a touch-up to make them commercial.