I used to attach a long cannula needle to some narrow flexible tubing. To get the needle to the bottom of the membrane band, I would put the gradient tube into a rack on a lab jack, and gently raise it until the tip of the needle was where I wanted it. Then I would use suction at the end of the tubing to start draining the gradient. By lowering the end of the tubing below the height of the gradient tube, I would use it as a siphon.

There is some automated equipment that can be used both to form and drain gradients, if you can afford it.

If you use tubes for SW41.Ti tubes, a long 1ml tip fit to a pipette would help. I would cut the sharp end of the tip to make the opening more wide, then push the piston of pipette to the 1ml max (to avoid bubbling after inserting through the gradient), and then insert the tip from the top of the tube till the membrane band (upper gradients will not enter tip unless you release pipette piston) and then collect by releasing the pipette piston slowly at the same time hovering on the membrane band(neither above nor below band) until I get all thick membrane band inside tip. Then take out and add it to eppendorf tubes. Note: if you need multiple bands (ER/Golgi/PM band at different densities), collect from top bands and then bottom ones.

There are alternative ways where a syringe can be inserted from sides at the band region or puncturing the bottom of tube and collect 1ml fractions. But I prefer pipette method which is much easy. Good luck.

Hi, Adam, 

It is very interesting and quite innovative method. one concern is will you collect a large volume of sucrose if you want to harvest the membrane fraction thoroughly? I mean if the band is thin, sucrose solution surrounding the tube will be sucked into the tube easier than (or at least at the same time as)  the membrane fractions?

Hi, Goodwin,

Thanks. I used the 1ml pipette tip last time, and find we collected too much sucrose and the band wasnot thoroughly collected as well. But we didnot cut the end of the tip last time, and maybe I can try it again to see how it works.

Hi Bin,

Even if the band is collected in a larger volume than is strictly necessary, it isn't a problem because the collected sample can be diluted with sucrose-free buffer. Then the diluted membranes can be pelleted in the ultracentrifuge with a fixed angle rotor in an hour and resuspended in a small volume of buffer.

Hi, Adam,

Thanks. i was thinking to collect as less sucrose as we can, so we can wash it out more easily.  but yes, as you said, it seems not affecting the further steps too much.

Best option will be to use oil and fraction collector along with equipment which make a small hole in the bottom of the tube followed by slow upward push by specific oil, and collect the fractions of 0.5ml or 1mL. Than by using a small fraction from each tube and testing specific marker for raft and not raft membrane and pool later these fractions if interested. You will always get messed up by other methods. I am no more active researcher that is why can not write exact name of  the oil. Optiprep was for gradient and may be if I find instrument name as well as name of the oil, I write. This oil does not get mixed with sample and help in fraction collections. 

Thanks, Dev. I am using a recyclable ultracentrifuge tube to minimize the cost, and therefore, it is not quite compatible to drill a hole on it. But it is an interesting method and will be applicable for the case of using single-use tube.

Now, I just cut the sharp end of the pipette tips, and generally it works fine but only need to do it by concentrating the focus.

In our lab, we always used disposable centrifuge tubes and specifically made to have hole at the bottom after use to collect the fractions. These tubes are from Backman Coulter