Dear Shu

first and fore mostly see attached document and in particular a .word doc designed by me specifically to train individuals in the 'art' of primer design

Now to address your specific concerns:

There are a number of primer design programs on the internet and yes many give Tm values that can differ by as much as 5C. In essence, this is due to the melting temp prediction algorithm utilised. To elaborate, the melting temp of a primer is not just due to the base composition alone, i.e. sum x(T/A =2C) + y (G/c = 4C). Thus, this old fashioned algorithm is really an historic relic; That is was used by 'veterans' by myself in the early 90s when internet access and open source prediction programs were not available. Thus, IGNORE. Moreover, even if you do calculate putative primer Tm based on base composition alone (as mentioned I would not recommend) the above formula pre supposes average and even base distribution and % of G; C; A and T residues throughout your 21-22mer. For most sequences utilised to make primers this is not true

Most primers have GC rich regions (or clamps) that artificially elevate the Tm relative to the above formula alone. Conversely, AT rich regions will reduce actual Tm relative to the above formula. This phenomenon is known as the 'nearest neighbour' affect'

Crucially, primer Tm is also a function of the local environment - that is PCR buffer and in particular Mg concentration; Sodium concentration; and also dNTP concentration. Generally speaking low ionic environments destabilise any duplex including primer target sequence associations, therefore lowering the Tm where higher strength ionic environments stabilise duplexes and thus raise the Tm.

Certain algorithms will also ask for dNTP concentration, e.g.

https://www.idtdna.com/calc/analyser

That is because higher [dNTP] depletes Magnesium and lower Mg concentration which in turn can alter Tm

Thus, my advice would be to:

1. Use a program such as IDT oligo analyser which takes account of intrinsic base composition plus nearest neighbour affects and also the composition of buffer etc in the local environment, e.g. IDT oligo analyser Oligo 3 and NCBI (which utilises the primer 3 rules)

2. Because other factors like sample purity, secondary structure of the target etc can also impact the actual Tm I tend to take the melting temp predicted by such algorithms and then sit it mid way in a PCR gradient screen of your primers:

That is try PCR with annealing temp T-3C to T + 3C in 1C intervals in an actual PCR screen. One of these temp normally approximating to Tm-2C to the Tm itself will constitute the optimal annealing temp in terms of specificity (single band) and PCR efficiency ( most amount of that band)

 3. In terms of what Tm to design your primers to based on this algorithm try 2C below your desired annealing temp; That is, if you require a primer with an annealing temp of (say) 60C make the predicted Tm 62C based on these software algorithms. However, these are predictions so please verify, preferably by an annealing temp gradient as mentioned 

Dear Shu

first and fore mostly see attached document and in particular a .word doc designed by me specifically to train individuals in the 'art' of primer design

Now to address your specific concerns:

There are a number of primer design programs on the internet and yes many give Tm values that can differ by as much as 5C. In essence, this is due to the melting temp prediction algorithm utilised. To elaborate, the melting temp of a primer is not just due to the base composition alone, i.e. sum x(T/A =2C) + y (G/c = 4C). Thus, this old fashioned algorithm is really an historic relic; That is was used by 'veterans' by myself in the early 90s when internet access and open source prediction programs were not available. Thus, IGNORE. Moreover, even if you do calculate putative primer Tm based on base composition alone (as mentioned I would not recommend) the above formula pre supposes average and even base distribution and % of G; C; A and T residues throughout your 21-22mer. For most sequences utilised to make primers this is not true

Most primers have GC rich regions (or clamps) that artificially elevate the Tm relative to the above formula alone. Conversely, AT rich regions will reduce actual Tm relative to the above formula. This phenomenon is known as the 'nearest neighbour' affect'

Crucially, primer Tm is also a function of the local environment - that is PCR buffer and in particular Mg concentration; Sodium concentration; and also dNTP concentration. Generally speaking low ionic environments destabilise any duplex including primer target sequence associations, therefore lowering the Tm where higher strength ionic environments stabilise duplexes and thus raise the Tm.

Certain algorithms will also ask for dNTP concentration, e.g.

https://www.idtdna.com/calc/analyser

That is because higher [dNTP] depletes Magnesium and lower Mg concentration which in turn can alter Tm

Thus, my advice would be to:

1. Use a program such as IDT oligo analyser which takes account of intrinsic base composition plus nearest neighbour affects and also the composition of buffer etc in the local environment, e.g. IDT oligo analyser Oligo 3 and NCBI (which utilises the primer 3 rules)

2. Because other factors like sample purity, secondary structure of the target etc can also impact the actual Tm I tend to take the melting temp predicted by such algorithms and then sit it mid way in a PCR gradient screen of your primers:

That is try PCR with annealing temp T-3C to T + 3C in 1C intervals in an actual PCR screen. One of these temp normally approximating to Tm-2C to the Tm itself will constitute the optimal annealing temp in terms of specificity (single band) and PCR efficiency ( most amount of that band)

 3. In terms of what Tm to design your primers to based on this algorithm try 2C below your desired annealing temp; That is, if you require a primer with an annealing temp of (say) 60C make the predicted Tm 62C based on these software algorithms. However, these are predictions so please verify, preferably by an annealing temp gradient as mentioned 

Dear Laurence,

Thank you so much for your patient reply! I learnt a lot from your experience.

My forward primer is GCATTTGGTTGTCTTGGTTTGGTATGTC.

My reverse primer is GTTGGGTGTGAGACATAGAGAATAAAAGGG.

I'm using Roche Expend Long Template PCR System to amplify a ~7.7kb fragment from genomic DNA. The Tm for both primers are 63, according to IDT analyzer. But I tried 60, 58 and 55 as annealing temperature and none of them had product! So is there anything special for choosing annealing temperature when PCR from genomic DNA?

Another thing is the Roche Expend Long Template PCR System Kit (the attached file is protocol) doesn't have Na+. So when calculating Tm by IDT analyzer, I just put 0 in that box.

Dear Shu

That temp gradient is fine based on the assumed Tm

Moreover, using the IDT tool mentioned and another tool Oligo calculator:

http://biotools.nubic.northwestern.edu/OligoCalc.html

I cannot find strong evidence of stable hair pin loops and self annealing for the forward and reverse primers respectively (Oligo Calculator tool) and hetero dimers between your forward and reverse primers i.e. primer dimers (IDT tool)

Not absolutely crucial to include sodium in the formula and I have incidentally used Roche expand in the past and using the gradient you describe that should yield bands all else being equal 

In general a lot of polymerase manufacturers provide on line tools for Tm determination and it is often wise to use those optimised for their particular enzymes, if in doubt

The problem may actually be unrelated to primer sequence and one of 2 possibilities spring to mind:

1. The most likely explanation is that even with long range polymerases amplifying anything over 3kb efficiently is inherently difficult. Google factors affecting large amplicon PCR but also see files provided: In essence, secondary structure is inevitable for such large amplicons so factors that reduce such structure, e.g. 5% DMSO in long range PCR reactions or alternatively 0.5_1M betaine (final conc) always help. In addition, extending at 68C instead of 72C can help reduce DNA damage that can limit the efficiency of amplification

2. Do you know anything about the quality of your DNA ? I would check that your DNA prep is capable of amplifying short high copy number fragments like housekeeping genes e.g. GAPDH or B actin; Preferably using primer stocks that have been validated on other DNA preps

If I had to guess however I would say that your problem is likely to be amplification conditions for your large amplicons and/or secondary structure in your target region. I would investigate both

i have little to add to the excellent answers from Laurence but since it is still unclear if primer annealing temperature or other pcr parameters are failing you could design 2 more primers say 500bp in from your end primers and paired with each primer and of higher annealing temperature than your F and R primers. Then gradient pcr will give short amplimers and tell you the highest annealing temperature that your primers will work at then do the long pcr  at the temperature that you know that your primers work at. Just removes one variable

Dear Laurence,

Thanks again for your patient reply.

I will try to use DMSO or betaine since I saw someone else also mention this. The elongation temperature I use is 68C, according to the Roche protocol.

The DNA I used is extracted long time ago and kept in -80C. I don't know the quality of it. But I used the same DNA to PCR a short (300bp) fragment, with a normal PCR polymerase, and it worked. Although there was a slightly smear band in the 300bp region.

I'll give you feedback once I have another try. Thank you!

Dear Paul,

If my forward primer is named F1 and reverse primer is named R1. Do you mean to design another 2 primers, F2 and R2. Both (F1 and R2) and (F2 and R1) will amplify a 500bp fragment? And the annealing temperature for F2 and R2 are  higher than F1 and R1?

Thanks!

hi Shuaichen,

yes that is exactly what I meant  and the highest temperature that you can anneal at and get a good short product is the highest temperature that f1 and r1 can be used at for the long pcr Regards

Paul

It depends on the company you are going to order your primers. For example, if you are going to order from IDT you should use the link below:

https://sg.idtdna.com/calc/analyzer

Generally, you have to look for a Tm calculator tool in the website of your interest company.

Dear Shu

this is my final comment on this subject for now: The fact that a small product will amplify confirms that the problem - as I suspected - is not the quality of your starting DNA

The issue and this is always the issue is amplifying anything with any sort of efficiency of that size; basically anything above 3-5kb

As mentioned secondary structure will be an issue for amplicons of that size; The question then becomes how much secondary structure based on GC content of your target sequence and in particular whether you have long runs of contiguous Gs and Cs which will cause polymerase dissociation and failure to amplify full fragments from large copy numbers of your gDNA

I would therefore make just one or two more suggestions at this juncture:

1. The ability to efficiently replicate large numbers of genomic molecules of that size - the so called processivity of an enzyme - differs greatly between different current generation polymerases; Thus, trying different types of 'long range' polymerases can often help as it addresses one of the fundamental limitations of what you are attempting

2. If you are not already doing so precede your main long range PCR with a single denaturation cycle of 96C for 5 minutes. That will open up as much of your DNA duplexes at the outset as is possible and definitely contribute to PCR efficiency which is imperative to amplicons of this size

3. for a fragment of that size I would advocate 40-50 PCR cycles. Specificity might potentially be an issue (see below) but poisoning of your polymerase and Deamination/ depurination of your DNA DEFINITELY will be. This once again will limit the extent to which you can amplify a fragment of that size in any meaningful amounts. The former is dealt with by extending at 68C (as already discussed) instead of 72C considering the extended extension times you are required to use. The latter is caused by prolonged heating of DNA at 96 and especially 98C for successive cycles. Thus, for denaturing in long range PCR try (say) 94C for 20s ec instead of 96C for 10 sec.

4. By extension the quality of your starting DNA will also impact on PCR efficiency: 2 x column purifications or phenol/trizol chloroform followed by a supplementary column purifications will remove guanidium salts in particular which can and do undermine efficient PCR; a pre requisite for what you are attempting to do

5. Secondary structure is a general issue that is alleviated by use of either 5% DMSO or 0.5M to 1M betaine in your PCR. You MUST do one or other. If by examining the target sequence you find > 60% GC or even below this areas with 5 or more G/C residues constituting 'strong stops' this in itself will NOT be sufficient. In these instances you might want to consider using a specialist GC rich buffer and many companies make them: Sigma, Invitrogen New England Bio labs etc

6. Finally primers and Tm's: I have addressed this issue last as I do not believe it is fundamental to your problem. Nevertheless, I think Pauls suggestion is excellent and in general use of more than 1 primer pair for amplifying a fragment of any size - but especially onerous large fragments - is a useful strategy. use of large primers with Tms of 60-65C and an annealing temp approximating to Tm-2C will also encourage more specific and efficient amplification 

7. If amplification in the context of linear large amplicons fails - in spite of lower denaturation temp (94C); larger cycle numbers (> x40) and modified extension and denaturation temp (68C & 94C respectively) - as a last resort clone fragment in to a plasmid and PCR using M13 forward and reverse primers

Hello, you can use the following online application for the Tm prediction:

http://www.molbiotools.com/primerinspector.php