Bead cleanups

Bead cleanups are a nifty way of selectively binding and washing certain DNA fragments (or other molecules) and then releasing them into the proper buffer.  You might use them for DNA extractions, cycle sequencing cleanup, NGS preps to remove small fragments, specialty applications when binding biotin-labels, or many other applications that might not involve nucleic acids at all.  While they have been around for a while now, recently they seem to have taken on a new popularity.  In our lab I have used streptavidin-coated beads to pull out DNA from a genome that hybridize to certain probes (SSR hunting), and carboxylated beads for size selection (NGS prep) or for Sanger sequencing cleanup.

A couple of months ago, I was talking to a visiting sales rep and mentioned that I had found a cheaper alternative to AMPure beads that seemed to work just as well.  The rep asked what I was using, and then told me that someone else she met was making their own from scratch and had posted their own blog about it.  As soon as the rep left, I googled the concept and quickly found this post by Brant Faircloth at UCLA.  The recipe is to be found here, but I will reproduce it below for your convenience.

After a brief wait for the beads, I started mixing up recipes.  I first made up the SeraMag recipe from the Genome Research paper (Rohland & Reich, 2012).  I was doing a little extra reading on the process and came across a few other useful ideas, so I made up two more solutions for doing Sanger cleanups (TEG - Elikin et al, 2002- and BioMag - Mijatovic-Rustempasic et al, 2012).  First, I wanted to see if I could reproduce Dr. Faircloth's results.  His gels look really nice, so I followed a similar approach, although I had a ladder with fewer bands (KAPA Express DNA Ladder).  I followed this protocol:

• Recipe (from Rohland & Reich, 2012)
• 0.1% carboxyl-modified Sera-Mag magnetic Speed-beads (Fisher, 09-981-123)
• 18% PEG-8000 (w/v) (Fisher, NC0107553)
• 1M NaCL
• 10mM Tris-HCl, pH 8.0
• 1mM EDTA, pH 8.0
• Optional: 0.05% Tween-20

1. Add well-mixed beads at the desired volume:volume ratio.
2. Pipet up and down a few times to mix, and wait 10 min.
3. Place plate (384 well in this case) onto SPRI magnet plate and wait about 2 min for binding.
4. Withdraw and dispose of liquid once beads have formed nice pellets on the well sidewalls.
5. Add 20uL 70% EtOH, wait ~30 sec, withdraw and dispose, and repeat once for a total of two 70% EtOH washes.
6. Let stand 10 min to air dry.
7. Resuspend in 10uL 10mM Tris-Cl pH 8.8.
8. Run on 1% gel (LB buffer, 250V, 150mA, 15 min).

Here is the gel from my attempt:

Hey, this stuff works!  Unfortunately, I was distracted during this exercise and the binding step took closer to 25 min than the desired 10 min.  You can see some difference, but clearly I wasn't 100% effective at removing the small (100bp) fragment.  Immediately, I figured that the binding step is time-dependent since you are waiting for the DNA to precipitate in the presence of PEG, and then to bind to the carboxylated beads.  I decided to try this again, but this time choose a constant volume:volume ratio (1:1), and instead add beads at specific time intervals (1min, 2min, 5min, 10min, 15min).

Again I ran  a gel of my results:

I relegated myself to the bottom corner of someone else's gel, so the resolution isn't great.  In fact, it looks terrible!  Not satisfied with the ambiguity of my 1% gel, I ran the results on a Bioanalyzer DNA1000 chip.  Here the results were much more clear.

No treatment Bioanalyzer lane (0 min):

Overlays from each of the time treatments (1min, 2min, 5min, 10min, 15min):

So there is some time-dependence, but for fragments ~400bp or greater, maximum binding is achieved within 10 min.  If you are in a hurry, 5 min may even suffice.  No matter what, some small fragments carry through which means we must continue doing multiple bead cleanups to remove adapter-dimers and other pesky products that interfere with our NGS applications for the time being.  Fortunately, the recovery of larger fragments is very good (~100%?), so unlike with column cleanups, you can count on recovering most of your sample.

Magnets: The SPRI magnet plate I linked to above is quite expensive.  Our lab uses such things since we do high-throughput work in 96 and 384 well plates.  However, the average lab user needn't fork over  an astronomical sum just to do magnetic separations because some company glued a $5 Nd magnet into a tiny tube rack.  You could get creative and glue your own magnets to your favorite rack for a fraction of the cost, or your could be lazy like me and simply hold a button magnet to your microcentrifuge tube while the beads migrate to the wall, and use your other hand to do the pipeting steps.  I did this for years before we got the magnet plates.

Magnetic field: For plate separations, we have two different types of plates.  A 384well post magnet array, and a 96well ring magnet array.  Initially I thought the beads should go directly to the point of contact with the magnet, and for the 384well plate this is true.  However, for the 96well ring magnet it turns out not to be the case.  This has important implications for minimum sample volumes that can be cleaned up using a bead assay.  For the 384 well plate, you need about 10uL total volume to get the beads to go to the magnet, so you can clean up 5uL sample in a 1:1 ratio.  For the 96 well plate, you need a larger volume and this has to do with the shape of the magnetic field.  An engineer from Alpaqua sent me the following image depicting a well sitting in a ring magnet and the region of maximum magnetic field strength is above the contact with the magnet, and that is where the beads actually travel to:

Final notes: Some users concern themselves with carryover of beads from step to step.  I haven't found a situation where this is a problem.  You could resuspend your beads, and use the resulting slurry in PCR, for a ligaton, or even to load onto a Sanger instrument for sequencing.  If you are really concerned about a little carryover, I recommend rather than applying a magnet again, just throw your sample in the centrifuge at high speed for a minute.  The beads will pellet, and really won't want to come apart again.  Now you can pipet and just avoid the beads if you prefer.

Also, don't overdry your beads or they won't want to break up and go back into suspension, which does seem to affect recovery.  10 min at room temperature seems OK, but lately I've been using the vacuum centrifuge in non-centrifuge mode for 3 min at 60C to dry samples in a 384 well plate with good results.  I had been doing 10 min, and then 5 min, but the longer times cause me to lose some samples that presumably wouldn't elute very well once so dessicated.

Other applications for carboxylated beads include cleanup of Sanger sequencing products (Elkin et al, 2002 - works well; Mijatovic-Rustempasic et al, 2012 - tried once, didn't work for me...yet), and preparation of transcriptome products (Stranneheim et al, 2011 - looks promising).

Enjoy your cleanups!!