NGS library generation: two different adapters

[MarijkeB]

Hi all!
I am new to NGS technology and I am trying to understand all the steps involved in the protocol going from seq library preparation to the actual NGS run on Hiseq instrument. During library prep, two different adapters are ligated at the DNA fragments. As far as I understand, this is needed for the bridge amplification later on in the flow cell, since this surface is covered with the complements of the two adapters. What I do not understand is why two different adapters are needed and not only one is sufficient? Maybe this is a very naive question and I am overlooking something very basic... but I don't get it.
I made a simplified drawing to illustrate my question (see attach).

Thanks!

[SNPsaurus]

One problem with your drawing is that the fragment would not be:

AGGT--TGGA

The adapters are double stranded, so would be

AGGT
TCCA

And when ligated would be:

AGGT----ACCT
TCCA----TGGA

and the strand binding to the flow cell would be:
AGGT---ACCT

But the bigger issue with single adapters is that you don't want to sequence both sides of the fragment within a cluster as you would end up with a jumble of signals.

If you have 5'Ax---T
It sticks to an adapter and copies
A---xT

then you sequence off of A... half the fragments will sequence the x side and the other half the other side.

[MarijkeB]

Hi SNPsaurus, thanks for the information about the fact that I have to consider double-stranded adapters. However, I don't understand what you mean with the fact that the fragment will be sequenced from opposite sides using the sequencing primer...? I again tried to visualize this in the attached file: in the first round of bridge formation, there will be no bridge formation since the end of the fragment is not complementary to the adapter bound on the flow cell. After amplifying the tethered fragment though, bridge formation becomes possible. If then the sequencing primer is added, the primer can bind to all the fragments and will start the sequencing in the same direction for all bound fragments. Please, tell me where I am reasoning wrong... thanks!

[SNPsaurus]

On the Illumina flow cell, there are two kinds of oligos that can bind both adapters. I was thinking about that case of two flow cell oligos with the single adapter. But you are asking why not just use a single adapter and single oligo on the flow cell, right? I think I need some coffee first! But let's see...

In your diagram, you have fragment ACGT----ACGT binding to the flow cell, and the flow cell oligo copying that. But, ACGT----ACGT is 5'ACGT----ACGT3', so would not act as a template.

Fragment 3'TGCA----TGCA5' needs to bind to a 5'ACGT oligo, which extends to make 5'ACGT---ACGT3'. But other than your 5'-3' error, the rest works out as you drew. I suspect then that the problem is that the fragment is likely to form a hairpin with itself:

5'ACGT----\
3'TGCA----/

rather than making the bridge, which would make cluster generation difficult. Illumina does lots of molecular gymnastics on the flow cell, cleaving the P5, etc, so either they missed something straightforward or it is a problem!

[MarijkeB]

The coffee worked :-)
I can agree with the hairpin formation when using a single oligo on the flow cell. And now I also understand your explanation about getting sequence reads starting from both fragment sides when using two oligos bound to the flow cell :-)
Thanks a lot for your help!