How Nextera transposons attach different oligos to different terminus of DNA?

[Nikvailo]

As seen in image, while one transposon (e.g. TR1) adds Read1_Seq_primer (Blue), other one (e.g. TR2) adds Read2_Seq_primer (Green). Souldn't it be random? In this scenerio, same primer (e.g Read1_Seq_primer) can be added in both sides.

How one transposon (TR1) know that other terminus is processed with other transposon (TR2)? Not the same transposon (TR1)?

Is there only one type of transposon? Or is it a mixture of two different transposon?

[Simone78]

After the synthesis the Tn5 transposase is mixed with 2 different indexed oligos (N5xx and N7xx if you use the Illumina oligos). On average, 50% of the molecules (monomers) will carry the N7xx and 50% the N5xx index. But the transposase works as a dimer and the dimers are formed in solution in the presence of a buffer containing MgCl2. When monomers carrying the 2 different indices are mixed all the following combinations are possible:
- 25% of the dimers will carry 2 N7xx oligos
- 25% of the dimers will carry 2 N5xx oligos
- 50% of the dimers will carry N7xx/N5xx oligos.
Obviously, when you are doing dual-indexed sequencing you are interested in this last combination only. It is therefore clear that, out of all the dimers, only 50% will have the "right" combination of adaptors. And there is no way how to get 100% of the dimers with the N7xx/N5xx combination without the other "byproducts" (as far as I know).

[Nikvailo]

@Simone78

Thank you so much. This is exactly what I asked.

[wraithnot]

After the synthesis the Tn5 transposase is mixed with 2 different indexed oligos (N5xx and N7xx if you use the Illumina oligos). On average, 50% of the molecules (monomers) will carry the N7xx and 50% the N5xx index. But the transposase works as a dimer and the dimers are formed in solution in the presence of a buffer containing MgCl2. When monomers carrying the 2 different indices are mixed all the following combinations are possible:
- 25% of the dimers will carry 2 N7xx oligos
- 25% of the dimers will carry 2 N5xx oligos
- 50% of the dimers will carry N7xx/N5xx oligos.
Obviously, when you are doing dual-indexed sequencing you are interested in this last combination only. It is therefore clear that, out of all the dimers, only 50% will have the "right" combination of adaptors. And there is no way how to get 100% of the dimers with the N7xx/N5xx combination without the other "byproducts" (as far as I know).

They can favor the amplification of the heterodimers carrying N7xx/N5xx oligos with suppression PCR: N7xx/N5xx http://www.nature.com/app_notes/nmet...eth.f.272.html

The single-stranded products that result from tagging with two N7xx or two N5xx can form hairpins and the amplification of the heterodimers with N7xx/N5xx can be favored under the right conditions.

[Simone78]

They can favor the amplification of the heterodimers carrying N7xx/N5xx oligos with suppression PCR: N7xx/N5xx http://www.nature.com/app_notes/nmet...eth.f.272.html

Exactly, I forgot to mention that, thanks!

[nucacidhunter]

When monomers carrying the 2 different indices are mixed all the following combinations are possible:
- 25% of the dimers will carry 2 N7xx oligos
- 25% of the dimers will carry 2 N5xx oligos
- 50% of the dimers will carry N7xx/N5xx oligos.
Obviously, when you are doing dual-indexed sequencing you are interested in this last combination only. It is therefore clear that, out of all the dimers, only 50% will have the "right" combination of adaptors. And there is no way how to get 100% of the dimers with the N7xx/N5xx combination without the other "byproducts" (as far as I know).

I agree with dimer ratios, but I do not think that only tagmentation by heterodimers results in fragments flanked by P5 and P7 adapter overhangs which are amplified and indexed during PCR and can form cluster for sequencing. Each fragment is result of tagmentation by two independent transposon dimers. A fragment tagged in 5’ end with P5 overhang (in this thread context it could be by N5XX dimer or a N5XX/N7XX heterodimer) can be tagged in 3’ end with P7 overhang either by a N7XX dimer or a heterodimer. This is demonstrated in post #1 above illustrating homodimers only.

[ps79]

On the subject of suppression PCR in Nextera library preps, does anyone know if there is anything special in the Illumina primer cocktails or in the PCR mix that encourages this? I’m seeing evidence of AA (2 N5xx ends) and BB (2 N7xx ends) products amplifying in an experiment I’m trying where I use transposomes from a Nextera XT kit, but then I use custom N5 and N7 oligos and Phusion polymerase for the 12 cycles of PCR to append on the P5 and P7 sequences and amplify the library.

My understanding is that ssDNA with AA or BB ends should form stem-loop structures, which under the right conditions should prevent primer binding and thus inhibit amplification. If I’m imagining this correctly, the intramolecular binding should be much stronger than the binding of the N5xx or N7xx primers - 33 nt of complementarity for A-A ends (19bp ME and the A-specific 14nt), and 34 nt of complementarity for B-B ends (19bp ME and the B-specific 15 nt) vs just 14 or 15 nt of complementarity for the N5xx and N7xx primers on the ends of the fragments. This will prevent the primers from binding, and therefore prevent amplification of the undesired products.

I wonder if this is why the recommended annealing temperature for the PCR reaction is low (by high fidelity enzyme standards) at 55°C to bind up these undesired products into hairpins that are refractory to amplification?

Does anyone have any suggestions why I am getting these AA and BB products and how to reduce their abundance?

[nucacidhunter]

On the subject of suppression PCR in Nextera library preps, does anyone know if there is anything special in the Illumina primer cocktails or in the PCR mix that encourages this? I’m seeing evidence of AA (2 N5xx ends) and BB (2 N7xx ends) products amplifying in an experiment I’m trying where I use transposomes from a Nextera XT kit, but then I use custom N5 and N7 oligos and Phusion polymerase for the 12 cycles of PCR to append on the P5 and P7 sequences and amplify the library

Would you elaborate on your evidence.

[ps79]

Sure. As I am using custom primers, I wanted to make sure I wasn't getting undesired products, so I set up control PCR reactions with only the N5xx oligo or only the N7xx oligo on some of my tagmentation reaction. I am getting a smear at about the expected size of the library - approximately 1/10 as much as when I use both the N5xx and N7xx primers together in the PCR, but definitely present.

[nucacidhunter]

I have tried Nextera XT prepared library PCR with single and two primers and there was not any amplification in single primer reactions indicating effective suppression effect with Illumina reagents and PCR condition. However, a similar library prepared with Illumina index primers but a third party polymerase had 1/4 amplification of two primer reactions with single primers. This indicates that PCR condition with third party polymerase need to be significantly optimised for optimum suppression effect. Fragments with the same adapter sequences in termini will not cluster. But, they cause quantification error and clustering have to be optimised empirically.

[ps79]

Thanks for the replies, nucacidhunter. I am aware that these products won't cluster, and there are indeed ways to work around the unreliable quantification with respect to optimizing clustering of desired products, but I'd prefer a more elegant solution! I've since done the same experiment you describe and seen the same thing. I guess I'm just hoping someone has some thoughts on the actual mechanistic reasons why Illumina kit polymerase mix behaves differently than third party polymerases. Is it just a matter of buffer compositions - i.e. can I achieve the same thing just playing with salt and annealing, or is there something proprietary in the mix (additional oligos for example) that enhances the suppression effect? Wild guesses welcome!

[lpcastro]

On the subject of suppression PCR in Nextera library preps, does anyone know if there is anything special in the Illumina primer cocktails or in the PCR mix that encourages this? I’m seeing evidence of AA (2 N5xx ends) and BB (2 N7xx ends) products amplifying in an experiment I’m trying where I use transposomes from a Nextera XT kit, but then I use custom N5 and N7 oligos and Phusion polymerase for the 12 cycles of PCR to append on the P5 and P7 sequences and amplify the library.

My understanding is that ssDNA with AA or BB ends should form stem-loop structures, which under the right conditions should prevent primer binding and thus inhibit amplification. If I’m imagining this correctly, the intramolecular binding should be much stronger than the binding of the N5xx or N7xx primers - 33 nt of complementarity for A-A ends (19bp ME and the A-specific 14nt), and 34 nt of complementarity for B-B ends (19bp ME and the B-specific 15 nt) vs just 14 or 15 nt of complementarity for the N5xx and N7xx primers on the ends of the fragments. This will prevent the primers from binding, and therefore prevent amplification of the undesired products.

I wonder if this is why the recommended annealing temperature for the PCR reaction is low (by high fidelity enzyme standards) at 55°C to bind up these undesired products into hairpins that are refractory to amplification?

Does anyone have any suggestions why I am getting these AA and BB products and how to reduce their abundance?

Hi ps79,

I am about to start trying tagmentation without any kit. As I could see you had a lot of experience on this protocol. Could you give me some advices?

Ligia

[finswimmer]

Hello,

in nearly every picture I see to tagmentation, there is a gap between one adapter and the DNA. The other adapter seems to be directly added to the 3' end. How is this gap filled? It must be before the PCR as with denaturation this piece of sequence would flow away.

fin swimmer

[Simone78]

Hello,

in nearly every picture I see to tagmentation, there is a gap between one adapter and the DNA. The other adapter seems to be directly added to the 3' end. How is this gap filled? It must be before the PCR as with denaturation this piece of sequence would flow away.

fin swimmer

I think other threads have probably additional information but I can summarize some things here.
You are correct, the gap needs to be filled and it can be done in 2 ways:
1- by a Polymerase with strand-displacement activity, like KAPA HiFi. The gap is filled, the small piece is displaced and then the Polymerase starts with the regular PCR reaction. This is how Illumina does it but they probably use Phusion Pol instead.
2- by a Klenow enzyme w/o strand displacement activity. In this case the Klenow is just filling the gap and the PCR should be done with a separate enzyme. This is useful when you don´t want to remove the small piece added by the Tn5 because, for example, it´s methylated and by replacing it you would lose the information about the methylation of the C nucleotides in that stretch. I refer you to Adey & Shendure, Genome Res 2012 for a detailed explanation. For a detailed protocol see also Wang, Gu, Adey,....Nature Protocols 2013.

Best,
Simone

[finswimmer]

1- by a Polymerase with strand-displacement activity, like KAPA HiFi. The gap is filled, the small piece is displaced and then the Polymerase starts with the regular PCR reaction. This is how Illumina does it but they probably use Phusion Pol instead.

Ah, great! That's probably the reason why there is a 72°C step for three minutes right before the normal pcr cycle in the TruSight/NextTera Kits from Illumina.

Thanks a lot!

fin swimmer