Note: this is a search engine friendly version of my lab notebook, please see the pdf version of this document for a more human friendly, printer friendly version.

Chapter 10
Miscellaneous Experiments

This chapter is for things that are too big to include in other chapters, but too little for their own chapters. This includes optimization of protocols in the appendix (where it isn't desirable to have all of the optimization info, I just want the best protocol there).

10.1  New color proteins from the Tsien lab

I requested 3 fluorscent proteins from the Tsien lab after reading their Nature Methods paper: A guide to choosing fluorescent proteins. Two of the colors, mCherry and mOrange, come from the paper: Improved monomeric red, orange and yellow fluorescent proteins derived from Discosoma sp. red fluorescent protein. The other protein mCitrine was published in Reducing the Environmental Sensitivity of Yellow Fluorescent Protein. This set really fills out our possibilities (before we only had GFP in the lab, then recently I got DsRed). I didn't request the proteins for an particular project. I thought I might attach the Cherry to the other side of the toggle41. But I have lots of other projects where multiple colors will be very useful. For the moment we have Red, Orange, Yellow, and Green. The only colors we're missing are a Cyan (Tsien lab doesn't have a good one) and a dark red / plum (Tsien lab has a decent, but dim one however I already requested my limit of 3 proteins; mCherry is pretty close to plum in excitation anyways).
I choose these three because they were the brightest and most separated, based on the info in A guide to choosing fluorescent proteins. GFP the other color we have in the lab is too close to Citrine to really add it as a fourth color. However, CyPet and Cerulean are cyan proteins that are still pretty bright (all of the blues are currently way too dim, not photostabile, and require UV excitation which damages DNA) and are pretty distant from mCitrine. Getting either of those two would allow 4 color imaging. If you are interested in three color, that's possible now using GFP, mOrange, and mCherry. However, you shouldn't use the wild-type GFP because it has a bimodal absorption pattern (395 and 475). EGFP with a single peak at 484 is a better option. A better three protein option is Cerulean (433), mCitrine (516), mCherry (587).
Figure shows a couple images I took on the blue light transilluminator using the cheap lab digital camera. The cells are in Davis minimal media42. The second photo is the same as the first except I pelleted the cells to concentrate the proteins to one location and make them easier to see. The order of the second photo is Citrine, mCherry, mOrange.

10.1.1  Filter Selection

Below I list the filters recommended in the Tsien lab paper A guide to choosing fluorescent proteins. I also list the closest filters to the recommended ones that we have in our plate reader (which are the ones I actually use for the moment). The Single column refers to a sample with a single-protein. The multiple column refers to samples with multiple fluorescent proteins mixed together.
Protein Best available in our plate reader Tsien Single Tsien Multiple
mCitrine 485/20, 528/20 490/30, 550/50 495/10, 525/20
mOrange 530/25, 590/35 525/20, 595/80 545/10, 575/25
mCherry 590/20, 645/40 560/20,640/100 585/20, 675/130

10.1.2  Archiving the colors

Oct 26, 2006
The samples came via standard US post on filter paper. I added 50-100 ml of TE (enough so that I could see some liquid at the bottom of the tube and let the filter paper sit in the TE for 30-60 minutes. I transformed 1 ml of the TE into DH5a cells.

10.1.3  Putting the colors in a strain where they'll express

Tue Oct 31 18:07:52 EST 2006
The proteins all have T7 promoters, so you get very little expression in DH5a cells (though I could still clearly see the cherry cells by eye). I'm putting the plasmids containing the proteins into BL21(DE3)pLyseE cells [Invitrogen] which have the T7 polymerase, so that I can mess around with the cells a little with the light microscope and plate reader to get a feel for them. The BL21(DE3)pLyseE has T7 lysozyme on a chloramphenicol maintained plasmid. The lysozyme serves to lower the basal expression from the T7 promoter (which is supposed to be extremely high). The cells also have the nice property that if you freeze thaw them a few times on ice, they lysozyme causes them to lyse themselves. This might have some biotechnology applications.

miniprepping the plasmids

Tue Oct 31 18:35:00 EST 2006 I grew the cells up last night and miniprepped them this morning:
Sample DNA (ng/ul) 260/280 260/230 total yield
mCherry 352.4 17.6 mg
Citrine 302.4 15.1 mg
mOrange 49.6 2.5 mg
The mOrange is low because I messed up the elution of the miniprep. Instead of doing the final spin-down to remove excess ethanol, I added the EB elution buffer. I didn't want all of the ethanol in my sample. I added 750 ml ethanol in a hope that it would change the pH or at least precipitate the DNA a bit so that it wouldn't flow through the column. I then ran the EtOH through, and did a second spin to dry the residual etOH and finally added 50 ml EB buffer to elute. This killed my yield, but it's still enough DNA to do what I want. After I was finished, I realized it would've been much easier to elute the DNA + EB + ethanol and then EtOH precipitate the sample. Next time....

Cloning into BL21(DE3)pLyseE

Tue Oct 31 18:35:00 EST 2006
I cloned the cells according the Invitrogen protocol. I added chloramphenicol to some amp plates I already had and I plated 70 ml of the transformation.

Inducing expression

Wed Nov 1, 2006
IPTG induces the expression of T7 in the BL21(DE3)pLyseE cells. They say to use 0.5 mM in the Invitrogen manual. I picked one colony from each BL21(DE3)pLyseE transformation. I picked the same colony 2x, the first went into a falcon tube with LB and the appropriate antibiotics, the second went into a falcon tube with LB, appriopriate antibiotics, and 500 uM IPTG.
I was surprised to find that the samples with IPTG didn't really grow any noticable amount by the time the non-IPTG samples had reached stationary phase. I made a freezer stock of the stationary phase cultures (strains are in row 2 of the main box). I then reinoculated the IPTG samples using the left over stationary phase culture (1/7 dilution).
I saw a fairly immediate jump in the concentration of mCherry and especially Citrine. mOrange, which is supposed to be the most highly expressed protein, was not really detectable (Figure ). However, the next morning, after overnite growth in IPTG, it was clear that the orange worked because the sample was very orange. The plate reader agreed with this observation . I read that the mOrange is very pH sensitive. It could be that the acidic conditions of LB was lowering the amount of fluorescent mOrange protein?
The time samples and the final overnite reading are available here in excel format. Notice also how the readings improve when using the autosensitivity adjustment (see Figure ).
Please see the pdf version for figures
Figure 10.1: The fluorescence of the colors proteins from the Tsien lab was measured using the plate reader. The measurements after 1/7 incubation into 500 uM IPTG were done using a sensitivity of 50. The next day, I realized that using autoadjust was a way to get a nicer scaling. Initially (A), the more towards red you got, the lower the numbers were, even for the - control. The color separation is pretty good for the Citrine and mOrange. The color separation is very good for the mCherry.
Please see the pdf version for figures Please see the pdf version for figures Please see the pdf version for figures
Figure 10.2: I washed the cells, from an overnight in the BL21(DE3)pLyseE strain with 500 uM IPTG, with Davis minimal media. Image (a) shows proteins mCherry, Citrine, and mOrange (left to right) in suspension. Image (b) shows the same tubes but with the cells pelleted to concentrate the proteins/cells to one location (left to right is Citrine, mCherry, mOrange). Image (c) shows pellets of cultures expressing mCitrine, mOrange, and mCherry in white (ambient) light.
Brief Conclusions:   Thu Nov 2 17:45:57 EST 2006
The plasmids certainly produce the correct colors (see Figure 10.2). The only thing that still worries me a little is why didn't the mOrange show up initial post-induction with IPTG? I learned that the autoscale feature can be very useful to make the detection readings more comparable with each other (Figure 10.1). Choosing the same sensitivity for every protein yielded much lower results for the high wavelength (red) readings - even for the LB blank. By autoscaling each well so that the LB blank is at value 100 the colors become much more comparable. For davis media, scaling the blank to 40 would probably be more appropriate.

10.2  Sephacryl column separation of DNA

I found this protocol: http://www.genome.ou.edu/protocol_book/protocol_partI.html Fragment purification on Sephacryl S-500 spin columns that described using a spin column stuff with S-500 Sephacryl to size-fractionate a sheared sample. Column size-fractionation is the method preferred for generating cDNA libraries. I guess it produces more ligatable DNA than gel-purification. I decided to give this protocol a whirl.
Here is the original protocol (just cut-and-paste from their website): DNA fragments larger than a few hundred base pairs can be separated from smaller fragments by chromatography on a size exclusion column such as Sephacryl S-500. To simplify this procedure, the following mini-spin column method has been developed.
  1. Thoroughly mix a fresh, new bottle of Sephacryl S-500, distribute in 10 ml portions, and store in screw cap bottles or centrifuge tubes in the cold room.
  2. Prior to use, briefly vortex the matrix and without allowing to settle, add 500 ul of this slurry to a mini-spin column (Millipore) which has been inserted into a 1.5 ml microcentrifuge tube.
  3. Following centrifugation at 2K RPM in a table top centrifuge, carefully add 200 ul of 100 mM Tris-HCl (pH 8.0) to the top of the Sephacryl matrix and centrifuge for 2 min. at 2K RPM. Repeat this step twice more. Place the Sephacryl matrix-containing spin column in a new microcentrifuge tube.
  4. Then, carefully add 40 ul of nebulized cosmid, plasmid or P1 DNA which has been end repaired to the Sephacryl matrix (saving 2 ul for later agarose gel analysis) and centrifuge at 2K RPM for 5 minutes. Remove the column, save the solution containing the eluted, large DNA fragments (fraction 1). Apply 40 ul of 1xTM buffer and recentrifuge for 2 minutes at 2K RPM to obtain fraction 2 and repeat this 1xTM rinse step twice more to obtain fractions 3 and 4.
  5. To check the DNA fragment sizes, load 3-5 ul of each eluant fraction onto a 0.7% agarose gel that includes as controls, 1-2 ul of a PhiX174-HaeIII digest and 2 ul of unfractionated, nebulized DNA saved from step 4 above.
  6. The fractions containing the nebulized DNA in the desired size ranges (typically fractions 1 and 2) are separately phenol extracted and concentrated by ethanol precipitation prior to the kinase reaction.
Bruce A. Roe, Department of Chemistry and Biochemistry, The University of Oklahoma, Norman, Oklahoma 73019 broe@ou.edu
TM buffer shows up as many different things on many different websites, protocol books. The TM recipe that the people which posted the sephacryl recipe have on the website is: 50 mM Tris [pH 8], 15 mM MgCl2.

10.2.1  Testing the old protocol

Wed Oct 25, 2006
I want to test the original protocol I got off the web first. I used 2 sources of genomic DNA and a cut plasmid in the size-selection process. The genomics sources were samples 1 and 5 (Figure 5.5) from section 5.1 on page pageref. I choose the two genomic samples because they provide a nice smear across two different ranges (large for genomic 1 and medium/small for genomic 5 see Figure 5.5). I added the cut plasmid so that I would have a strong reference band at one particular size. Cut pUC19 is around 2700bp.
I tried fractionating each sample individually and all three combined. The all three combined was the only sample that I ran on a gel. It consisted of 500 ng cut pUC19 and 2 mg of each of the genomic samples. The samples were combined into a total volume of 40 ml in TE.
Please see the pdf version for figures         Please see the pdf version for figures
Figure 10.3: 500 ml of Sephacryl S500 (a) was washed 3x with TE. The centrifugal force distributed the sephacryl on a hard angle inside the Spin-X column (b).
I added the 500 ml of Sephacryl S500 [Amersham] to a Spin-X filter column [Costar] and washed it as described in the section above. Notice that the centrifugation pulls the sephacryl onto a hard slant 10.3, so that at it's thickest the sephacryl extends to the top of the tube, but on the other exterme it is only 0.25 cm or so thick.
I always loaded the samples at the thickest part of the sephacryl. I ran the initial fraction followed by 3 more fractions with 40 ml TM buffer. I ran all 4 appx 40 ml fractions out on a gel along with a non-fractioned sample with the original composition as my fractionated sample .
Please see the pdf version for figures
Figure 10.4: 1% agarose gel and 4 sephacryl fractions. an unfractionated sample is included for comparison.
Brief Conclusions:   It is vaguely clear that each fraction contains progressively more short fragments and progresively fewer long pieces with each subsequent fraction (Figure 10.4). However, the different between the fractions is pretty miniscule here, and I'm not sure it is useful except to get rid of the smallest of the small fragments. I want to try again with more sephacryl in the hopes that having the minimum thickness of sephacryl increased (see Figure 10.3) will also increase the separation achieved by the fractionation.

10.2.2  Improving the sephacryl protocol

Oct 26, 2006
I want to determine if increasing the amount of sephacryl or changing the MgCl2 concentration (both of which I assume would slow the DNA's migration through the sephacryl) would better the fractionation.
Please see the pdf version for figures
Figure 10.5: Using 850 ml of sephacryl S500 created a much thicker layer at the bottom of the centrifuge created slant.
I used the concentration of all of the DNA components I used last time (genomic samples 1 and 5, plus cut pUC19). This time I ran two samples, one with 750 ml of sephacryl (50% more than last time) and 15 mM MgCl2 (the same as last time) and another with 850 ml of sephacryl (70% more than last time) and 7.5 mM MgCl2 (50% less than last time).
Please see the pdf version for figures
Figure 10.6: 1% agarose gel stained with EtBr. The two different MgCl2 and Sephacryl concentrations are indicated above the gel.
Like last time, I ran each 40 ml fraction in a separate lane on a gel (Figure 10.6). I only ran one non-fractionated sample, since both protocol variants I tried had the same starting material.
Brief Conclusions:   The increased sephacryl amount certainly increased the minimum thickness at the bottom of the centrifuge generated slant (compare new Figure 10.5 with the old Figure 10.3). And the resulting size-fractionation was much more marked (Figure 10.6). The 850 ml with the lower MgCl2 concentration seemed to work better. I really would like to try one more time with even more sephacryl, though it's getting to the point that it'll be pretty cramped in there. Perhaps there's something other than Spin-X which will hold more volume at the top (besides the really long standard chromatography columns which I'd rather not mess with). I should also consider buying the ones from invitrogen and comparing with these home-made ones.
I'd really like to get that initial cut off for the second column to be greater than 500 bp. Right now, it is pretty strong still at 500 bp and weakens to low at around 300 bp and to nothing at around 200 bp.
To Do!!!  Try to run with 1 ml or 1.25 ml of sephacrl. Will likely need to spin down 800 ml sephacryl, add 200-400 more, spin down again and THEN wash 3x. Buy fractionation columns from Invitrogen?

10.2.3  Improving the sephacryl protocol, part 2

Thu Nov 9 17:45:53 EST 2006
I'm going to try and fractionate with 1 ml or 1.2 ml of sephacrl. I spun down 800 ml sephacryl, added 200-400 more, and spun down again. The 1 ml column was almost full. The 1.2 ml column was completely full, any more and the lid wouldn't really close properly. Then I washed the column 3x with TE. I still used 2 mg of genomic 1 and 2 mg of genomic 5. However, I didn't have any pUC19. Instead I used 1 ml of cut pNEB193. I'm not sure how much it was? Maybe 200-400 ng total.
One last problem. The original protocol says to spin 2K rpm for 2 minutes between each TE wash. I only used 1 min for the first two. I thought it wouldn't matter, however when I ran the first fraction through and spun for 5 minutes there was way more than my 40 ml starting material (probably 60-80 ml ).
I ran the fractions on a 80 ml, 1.5mm comb, 1% agarose gel for 50 minutes at 90V 43 (Figure ).
Please see the pdf version for figures
Figure 10.7: 1.0% agarose gel of sheared genomic DNA and plasmid run fractionated through columns with 1 ml and 1.2 ml sephacryl S500.
Brief Conclusions:   Fri Nov 10 10:42:09 EST 2006
Next time during the TE wash steps, I need to make sure to centrifuge for 2 minutes. Perhaps use 3 or 4minutes for the last wash centrifugation. I'm still not entirely pleased with the size separation this gives me (Figure 10.7), but I'm kinda running outta things to try. It might be good enough to just use the first two fractions. I just hope the cloning efficiency of the 300mers doesn't overwhelm the more abundant longer pieces. Another thing to consider is using the invitrogen premade columns. I should buy some and give them a try. But their expense make them unsuitable for everyday use.
To Do!!!  buy invitrogen sephacryl columns; test the ability of the microcon columns to remove large numbers of short adaptors and blunt ligated adaptors.

10.3  Comparison of short DNA fragment removal with Microcon 30, 50, or Qiagen PCR purification kit

Whenever I use adaptors in a ligation reaction, the huge excess necessary to prevent incorrect blunt ligations makes it dang hard to run on a gel. For an agarose gel you get a giant band where the primers are, for a polyacrylamide gel you get a giant black spot and a huge amount of DNA noise that pretty much makes the rest of the gel unreadable. How can I get rid of these pieces? When adaptoring cDNA I used a Qiagen PCR column to remove them, but there was still a lot left (cite figure). I'm thinking maybe one of the microcon super-small filter devices from millipore might provide better exclusion. Or perhaps some combination of the two. For sure one nice feature of the Qiagen column is that it does effectively remove very large fragments (see the cDNA cloning chapter).
Ilaria gave me primers to amplify at 80mer and 120mer fragment of pLtet. These sizes are right at the boundary of what the different filtering and concentrating can keep/reject. My hope is to keep these bands strong while virtually eliminating the shorter primer bands.

10.3.1  preparing the 80mer and 120mer PCR products

Nov 15, 2006
I used Phusion Taq [Fermentas] master mix to produce blunt-end products. The reaction was 2 ml 10 mM primer (400 nM), 1 ml plasmid (11 ng), 22 ml H2O , 25 ml Phusion master mix. I ran two reactions for the 80mer and two reactions for the 120mer. The two reactions for each were combined and cleaned up using a Qiagen PCR purification kit. Yields were:
Sample DNA (ng/ul) 260/280 260/230 total yield
80 mer pLtetO 55.0 1.65 mg
120 mer pLtetO 45.0 1.35 mg
Please see the pdf version for figures
Figure 10.8: 2.0% agarose gel of Phusion PCR amplification of pLtet. Looks like an upside-down cross - yikes!
Brief Conclusions:   Minus the wierd upside-down cross thing, so far so good.

10.3.2  preparing the test DNA

Nov 16, 2006
I'm going to use a similar composition to what I used with the sephacryl but with the addition of some shorter stuff. Per rxn I'm using 2 mg genomic 1, 2 mg genomic 5, 200 ng 80 mer, 200 ng 120 mer (appx 4 ml of each), and 25 fold excess of each small primer pair (see section 6.7.1.1 page pageref), which corresponds to 1.8 mg of the 30mer and 900 ng of the 15mer and 18 ml from the 10 mM primer stock. In total, this is around 7.1 mg of DNA.
I made 4x this amount to try with a Qiagen cleanup, a YM30, YM50, and no column. I need to make sure to run the no column lane far from the others to prevent saturation problems.
Fri Nov 17 15:16:43 EST 2006
The Qiagen PCR prep DNA was eluted in 30 ml . The microcon were spun at 14,000g for 12 minutes and then eluted. The elution was too small so I added around 30 ml to the top of each and weakly vortexed for 30 seconds.
The yields for all of the different methods were:
Sample volume DNA (ng/ul) 260/280 260/230 total yield
non-processed 80 ml 99.0 7.92 mg
microcon YM30 17 ml 338.0 5.75 mg
microcon YM50 15 ml 376.9 5.65 mg
Qiagen PCR 30 ml 104 3.12 mg
I ran 1 mg of each sample in a total volume of 20 ml (including dye) on a 15% TBE Urea gel for 50 minutes at 190V. The gel was stained 20 minutes in EtBr and destained 10 minutes in H2O (Figure ).
Please see the pdf version for figures
Figure 10.9: 15% TBE Urea gel
Versadoc calibrated bands from Figure 10.9
Lane Base Pairs Peak Int Average Int Trace (Int x mm) Relative Qty Calibrated (ng) Normalized (Qty)
1 163 1478 1006 2717 5 38
1 102 1531 1055 2742 5 38
1 31 4095 2052 9027 16 206
1 22 1074 889 2134 4 41
4 178 2003 1314 4337 7 73
4 108 2239 1410 3667 6 59
4 32 4095 1891 10969 17 261
4 22 834 751 2402 4 34
5 181 1562 1069 2886 5 40
5 110 1671 1119 2798 5 39
5 32 4095 2083 9583 16 221
5 22 1103 932 2330 4 33
6 184 1616 1097 3073 5 42
6 113 1691 1160 2900 5 40
6 32 4095 2070 10352 16 243
6 22 1024 902 2165 3 40
Brief Conclusions:   Fri Nov 17 15:16:43 EST 2006
Well, these things don't do a terribly good job of removing what they said they remove. The microcon columns to a great job of concentrating and not lossing much DNA (at least relative to the Qiagen column), but the primers weren't really removed (Figure 10.9).

10.4  Comparison of short DNA fragment removal with Qiagen PCR purification kit and ChargeSwitch with 1x, 2x, 3x, 4x N5 buffer

Mon May 14, 2007
Once again, I'm continuing my thus far unsatisfactory search for a non gel-purification based method to remove short DNA fragments. Removing the short DNA pieces without the inefficiencies of gel-purification would really speed things up and improve my ability to do several of the techniques I'm working on.
This time I'm comparing my default Qiagen PCR purification with a newer magnetic bead based method call ChargeSwitch from Invitrogen. I says in the Invitrogen manual that it works for purification of 25 ml - 50 ml PCR reactions. It also mentions that by altering the concentration of the N5 buffer, you could change the minimum DNA size retained by the purification kit (higher N5 concentration = fewer short pieces).
I tested their statement with the 50 bp ladder from NEB, so that I could look for DNA retainment at a varieaty of sizes. I made 25 ml samples (2 ml NEB 50 bp ladder and 23 ml TE). I tested the Qiagen kit and the ChargeSwitch kit with 1x, 2x, 3x, and 4x N5 buffer. The DNA was then run on a 2% agarose gel .
Please see the pdf version for figures
Figure 10.10: 2% agarose gel showing short DNA size removal (or rather the lack thereof) of Qiagen PCR purification kit and ChargeSwitch with different concentrations of N5 buffer.
Brief Conclusions:   To really say something more precise about this result I could use the versadoc to estimate the concentration of the bands of interest (in particular 50bp and 100bp). But generally, this looks like another size-selection disappointment. The ChargeSwitch yield is much worse than the Qiagen, and the removal of even the 50bp piece by any of the kits is negligable. Also notice that the main influence of increased N5 concentration is decreased yield.

10.5  Comparison of Qiagen Column Gel cleanup and QiaexII gel cleanup

Thur Jul 5, 2007
I think there is a need to really understand these kits that we use all of the time in the lab. We need an understanding that goes beyond just the printed stuff that comes with the kit. I want to be able to gel purify stuff, obtaining the highest yield with the cleanest DNA.
I'm testing the two kits from Qiagen: the column based method and the QiaexII bead based method. To test these methods I ran 1x NEB 2-log ladder and 0.5x NEB ladder on a 1% TAE agarose gel for 45 minutes. For each ladder, I cut and purified 2 bands (one big and one little) for each kit (so 4 total for each kit). I also cut two bands from lanes with no DNA as a negative control. I purified them according to each kit's instructions. I used 550 ml of buffer QG in all of the column methods to solubilize the gel. And I used 550 ml of buffer QX1 in all Qiaex methods to solubilize the gel. I eluted all of the purifications into 30 ml EB buffer. To quantify the yield of each purification, I used 20 ml of the 30 ml elution with the HS dsDNA Qubit fluorescence quantification platform.
The results were:
kit ladder band (bp) starting DNA (ng) cleaned up DNA (ng) recovered
column none blank 0 0.543 -
column 0.5x 100 30.5 16.5 0.54
column 0.5x 500 62 39.6 0.64
column 1x 200 32 11.79 0.37
column 1x 3000 120 69.0 0.58
qiaexII none blank 0 0.759 -
qiaexII 1x 100 61 24.87 0.41
qiaexII 1x 500 124 78.6 0.63
qiaexII 0.5x 200 16 9.12 0.57
qiaexII 0.5x 3000 60 26.79 0.45
mean recovery from Qiagen column 0.53
mean recovery from QiaexII 0.51
raw qubit readings for this data
Brief Conclusions:   Sat Jul 14 19:44:55 EDT 2007
Both of these Qiagen kits performed quite similarly. And to my eye, I don't notice any great differences in recovery via gel extraction with differing sizes and amounts of DNA. I'm not too surprised by the average recovery of 50%; it's actually a little higher than I expected. But these recoveries are much lower than the numbers posted on Qiagen's websites. For the column kit, Qiagen claims up to 95% recovery (I only got up to 64% recovery) with a typical recovery being around 80%. For their Qiaex II kit, they claim a recoveries of 60-95%. Maybe the difference is that they started with way more DNA than I did for each size. They used 2 mg vs my 16-124 ng of DNA. Perhaps, I'll try again some day with more DNA, but it's not very common that I have 2 mg of a single size of DNA that I want to purify on a gel - that's too much for my everyday needs.

10.6  Comparison of short DNA fragment removal with PCR Purelink kit from Invitrogen

Sat Jan 19, 2008
I still haven't found a non-gel based way to remove short DNA efficiently. I noticed in the Invitrogen Molecular Biology catalogue that they have a PCR purification kit that comes with a special buffer for removing dsDNA less than 300bp. So I bought the kit and decided to try yet another method to remove short DNA fragments.
I ran 5 mg of 50 bp ladder [NEB] through:
  1. Purelink PCR column with standard buffer
  2. Purelink PCR column with HC buffer
  3. Qiagen PCR purification column
I eluted all three samples into 50 ml of the elution buffers that came with the respective kits. The yields were:
Sample DNA (ng/ul) 260/280 260/230 total yield loss
1 61.2 3.06 mg 38.8%
2 70.1 3.51 mg 29.9%
3 82.4 4.12 mg 17.6%
I ran 1 mg of each sample on a TAE agarose gel. I also ran 1 mg of the ladder that had not been run through the column (Figure ).
Please see the pdf version for figures
Figure 10.11: It appears that the HC buffer did remove a part (but certainly not all) of the DNA below 300bp.
Brief Conclusions:   The HC buffer with Purelink certainly did a better job than the Qiagen column at removing the short fragments. Although the manual stated 300 bp, the real cutoff seems to be more like 200 bp with it being more efficient at removal the smaller the fragment is. I'd like to try again, but this time to run the wash buffer and/or the binding buffer across the column multiple times to see if I can wash the shorter fragments through a little better. As it is, it appears that this kit might to a better job of removing the excessive amount of adaptors (compared with the Qiagen column) that I have after adaptoring my cDNA.

10.7  diffusion of DNA when loading agarose gels

Tues Jul 10, 2007
When I load agarose gels, particularly with DNA that has been cleaned with a Qiagen PCR purification kit, every once-in-a-while it happens that the stupid sample diffuses really fast right after I pipette it into the well of the gel. Needless-to-say this results in a faint band with no chance for quantitive comparisons across samples, but why does this happen sometimes? It seems to happen more with TAE than TBE, but I don't have data to confirm this hunch. My guesses as to whats happening are: 1) eluting into EB buffer contains no salt so EB DNA diffuses away; 2) EB contains no EDTA; 3) trace EtOH from the purification is causing the diffusion.
To test this, I took 1.5 ml of 25 bp ladder [invitrogen] added it into a 10 ml total volume of EB, EB + 15% EtOH, TE, STE. I loaded each sample onto a 2% TAE EtBr agarose gel (Figure ). When I loaded the samples, they all seemed to sink pretty well - no fast diffusers. Which kinda sucked, because it did really relect what I figured would happen, which is that some of the samples would diffuse out quickly and others would not.
Please see the pdf version for figures
Figure 10.12: 2% agarose gel run at 110 V for 80 minutes
Brief Conclusions:   Err, make sure you don't have too much EtOH in your sample. Clearly this makes your sample disappear. As far as the other three go, it is less clear. It seems that EB is a little fainter and fuzzier then the other 2 samples. I'm not sure if the smeary TE lane was a fluke or not, but the 2-log ladders were both in TE and they ran fine. I don't know if the DNA will come off the PCR purification column with a salty buffer, since a high salt buffer is used to bind the DNA in the first place (not sure if it is the pH the salt or both that matter for the binding). But it would be easy to elute into TE and add the appropriate amount of salt to bring the NaCl concentration to 50 mM prior to running the gel.