How Ampure or SPRIselect works

[Genohub]

We, and our NGS service providers are frequently asked about a common clean up / size selection step during library preparation. Specifically, how does PEG precipitate DNA on negatively charged, carboxylated Ampure beads. Thought it would be useful to share our answer here:

http://blog.genohub.com/peg-size-selection-and-precipitation-of-dna-libraries-how-ampure-or-spriselect-works/

- Genohub

[pmiguel]

We, and our NGS service providers are frequently asked about a common clean up / size selection step during library preparation. Specifically, how does PEG precipitate DNA on negatively charged, carboxylated Ampure beads. Thought it would be useful to share our answer here:

http://blog.genohub.com/peg-size-selection-and-precipitation-of-dna-libraries-how-ampure-or-spriselect-works/

- Genohub

Not so much of an answer...

To me the Ampure "mystery" comprises a few components:

(1) What causes the precipitation of DNA (and RNA) in the first place?
(2) What causes the length-specific fractionation of DNA (and RNA)?
(3) Why would DNA and RNA bind to carboxylated beads?

The page you link to covers (1), touches on, but does not explain (3).

I don't think the answer to (3) is known. It was just an empirical observation that proved useful but was never really studied. Why would the negatively charged DNA and RNA molecules bind to negatively charged surface (carboxylated) of the beads? You mention that the precipitated DNA will bind to any available surface. Which, even if true, doesn't explain why it binds particularly well to a carboxylated surface.

I could speculate that there is some sort of non-polar interaction between the carboxylated surface and the DNA (or RNA) that occurs when there is lots of positively charged ions around to block the normally repellent ionic effects. But that is both speculative and provides no actual molecular mechanism for the binding.

Bonus points would accrue for delving into how well Ampure does or does not fractionate protein and other reaction components. Obviously PEG precipitates protein and presumably other macromolecules (eg, polysacharides). Do they bind to the beads?

[Genohub]

The binding of DNA to the carboxylated beads likely has to do with the two factors that cause DNA to aggregate (precipitate) during bead cleanups:

1) neutralizing DNA phosphate charge with a multivalent cation , re-orientating water dipoles around DNA. ~90% of DNA's negative charge must be neutralized for condensation to even happen. I think it's safe to assume that enough Na2+ can be added to neutralize >90%. Once these charges are neutralized, the beads are no longer that repulsive.

2) PEG which excludes water volume to DNA. The effect of excluded volume on DNA is nicely described here: http://shaqfehpc06.stanford.edu/data...DNAjoining.pdf
PEG exclusion of water volume leaves beads as the only thing left to attach, aggregate onto. I view them as a co-precipitant during cleanup.

Good question about Ampure fractionation of proteins. PEG volume exclusion is also a mechanism by which proteins precipitate. And there is some evidence that shows there is a macromolecular size dependence on PEG volume exclusion: http://www.jbc.org/content/256/23/12...ication_detail

Perhaps this relates to length specific fractionation of DNA. I'd like to hear other thoughts.

- Genohub

[pmiguel]

The binding of DNA to the carboxylated beads likely has to do with the two factors that cause DNA to aggregate (precipitate) during bead cleanups:

1) neutralizing DNA phosphate charge with a multivalent cation , re-orientating water dipoles around DNA. ~90% of DNA's negative charge must be neutralized for condensation to even happen. I think it's safe to assume that enough Na2+ can be added to neutralize >90%. Once these charges are neutralized, the beads are no longer that repulsive.

Hmm, I would think there must be some sort of binding attributable to the carboxyl groups themselves. Definitely the implication in the original SPRI paper:
Nucl. Acids Res. (1994) 22 (21): 4543-4544 doi:10.1093/nar/22.21.4543
(PDF is here.) The text includes:

We have noted that under conditions of high polyethylene
glycol(PEG) and salt concentration (10%PEG8000 and 1.2 5
M NaCl final concentrations)(6),DNA would bind to the surface
of carboxyl coated magnetic particles.

However, there is an interesting blog post here on SPRI. Especially some of the comments below the post. There an anonymous commenter claims that the carboxyl groups do not bind the DNA at all, quite the opposite. That they are there only to repel the DNA during elution.

Meh, I don't know if I buy it. I mean, I buy the elution part, but the random glomming onto anything with a high surface area part? Not sure.

Just as a bulwark against bit rot, here is the full comment from the above site:

The comments above assume that the DNA binds "magically" to the beads due to carboxylated coatings. This same myth runs with the use of silica as well. Bottom line is the carboxyls are on the surface to lightly repel the DNA off the surface once the DNA-bead aggregate is place in an aqueous solution that allows the DNA to fully hydrate back into solution phase.

The other reason for carboxyl groups is shield the DNA from coming in contact with sticky plastic surface (or silica)- i.e. use of nylon, for Southern blots or polystyrene for ELISA.

The process of DNA capture on a bead surface is the same as process of precipitating DNA for centrifugal collection. The long complex polymer of DNA remains in solution phase due to the structured water layer or cloud that forms around the DNA polymer primarily due to hydrophilic/ionic interaction of (deoxy)ribosugar-phosphate backbone of the nucleic acid. When "water hungry" alcohols or polyalcohols (PEG) are added to very high concentrations, the water cloud is disrupted, becomes significantly thinner, which results in the DNA transitioning from solution phase to a flocculated or precipitated phase. DNA transitions from a long polymer water cloud, coated polymer to water-stripped condensed "sticky ball". These DNA "balls" will stick to or aggregate with each other or to any surface that the ball collides into, surface such as that of a bead, fiber, or membrane. This flocculation process can be induced with adding salts to high concentrations, i.e. sodium perchlorate, ammonium acetate, lithium chloride, and with proteins the use of alcohols or high concentration of sulfate salts.

In reference to carboxyl and bead surface; if the surface was coated with amines, the DNA (via phosphate backbone) would adsorb to the surface via ionic bonds and not elute from bead surface. This ionic adsorption would occur with metal oxide surfaces including semi metal oxides such as silica if Lewis acid sites are exposed - which is common with most materials.

Nucleic acids have been isolated using plastic, silane, silica, ferrite, or even clay beads/particles. For all cases if the surface is pretreated with a starch, proteins or some other slightly negative charged coating, which shields the surface from DNA, the DNA easily elutes from the surface. Thus by coating the silica or plastic surface with carboxyl groups, the DNA when it transitions to fully hydrated form during elution phase will be ionically repelled by carboxyl coated surface.

Side note, smaller nucleic acid polymers are more soluble in these alcohol solutions due to their smaller size - form smaller condensed balls. Cold or time will allow the smaller balls to aggregate form larger balls and increase surface area of the ball to stick to a solid support or be spun out by centrifugation.

So I hope this helps in dismissing the myth that DNA has special ionic attraction to carboxyl or silicon oxide surface and that the process is due to physical chemistry of flocculation / precipitation process.

2) PEG which excludes water volume to DNA. The effect of excluded volume on DNA is nicely described here: http://shaqfehpc06.stanford.edu/data...DNAjoining.pdf
PEG exclusion of water volume leaves beads as the only thing left to attach, aggregate onto. I view them as a co-precipitant during cleanup.

Good question about Ampure fractionation of proteins. PEG volume exclusion is also a mechanism by which proteins precipitate. And there is some evidence that shows there is a macromolecular size dependence on PEG volume exclusion: http://www.jbc.org/content/256/23/12...ication_detail

Perhaps this relates to length specific fractionation of DNA. I'd like to hear other thoughts.

- Genohub

--
Phillip

[austinso]

Hmmm...

PEG brings things out of solution but it is the salt-bridging of the phosphate backbone with Si-OH or carboxylate groups by Na+ that is required for purification.

Non-specific adsorption also occurs (mediated through bases) to any "rough" surface, but that is counter-productive to recoveries, which is why a lot of effort is spent trying to make surfaces less sticky.

And yes, the size fractionation windows (which are very non-linear) are dependent on the size of PEG used and applies generally to polymers. So you will get different cutoffs depending on whether one uses PEG-4000 or PEG-8000, or PEG-20,000. Rule of thumb is that 30 kDa is 50 bp dsDNA, which is the bottom end of PEG precipitation.

PEG has some really interesting properties for sure...

[Simone78]

that´s a lot of useful information!
we are actually making our own Ampure beads for single-cell libraries using PEG 8K, following what described by Rohland & Reich in their Genome Res. paper from 2012 (by the way, they work beautifully and they cost 1/10 compared to the commercial beads).
To get different cutoffs we just vary the amount of PEG, but changing the length of the polymer is an interesting alternative!