Note from Andi: Theresa Culley used ISSR primers to assess the genetic diversity within and among populations of Viola pubescens. Theresa kept meticulous notes on her primer optimizations and the trouble-shooting she had to go through to collect her data. The concentration of the MgCl stock was 50 mM. The other students in my lab referred to in the text below are Shannon Datwyler and Chris Randle. These gel pictures will give you a good idea of the optimization protocol used in my lab.

Theresa's ISSR Primer Optimization


The purpose of my project is to compare genetic variation within and among five different populations of the yellow violet, Viola pubescens. Three of the populations (Stump, Hill, and Etter) are located in Crawford County, Ohio. The fourth population (Bohannan) is located two counties south of Crawford Co., in Delaware County, Ohio. The last population (MI) serves as an "outgroup" and can be found at the northern tip of the lower Michigan peninsula, in Emmet County.


primer1b 1.

On my first gel, I screened three different primers (844, 814, 17898) using 2-3 samples from each population for each primer. I chose these primers because they were recommended to me by Harvey Ballard. I used an annealing temperature of 45C, as suggested by Andi. Unfortunately, the gel did not develop too well because I had used old dNTP's that Lawrence had made months ago. I did obtain some faint bands for primer 17898 and 844, but nothing for 814 (thus, primer 814 was discarded).

eluquik 2.

Andi suggested that I should try the Elu-Quik clean-up method with my samples. I cleaned 1-2 samples from each population using instructions found on the website and suggestions from Chris (as I worked alongside him). I then ran the cleaned samples out on a gel next to identical uncleaned DNA for both primers 17898 and 844. For primer 17898, the Elu-Quik samples sometimes had more distinct bands, but overall they were fainter and at least one of the samples was blank. Thus, I decided not to use the cleaned DNA for primer 17898. For primer 844, the Elu-Quik samples looked very similar to the uncleaned DNA so I thought I should try it once again (see #5 below).

Lanes 1 -2: negative controls; 3 - 8: Elu-Quik treatment; 9 - ladder: uncleaned DNA; 17 - 22: Elu-Quik treatment; 23 - 28: uncleaned DNA.

eluquik 3.

I showed Shannon my Elu-Quik gel (as Andi was gone) and she suggested that I should try different dilutions of the DNA to try to reduce the smearing and brightness of the bands. For primer 844, I ran 50% diluted DNA next to identical undiluted DNA and also compared different MgCl2 concentrations - 1.0 Ál versus 1.5 Ál . I had a much better band resolution with the undiluted DNA combined with 1.0Ál MgCl2, so this is what I decided to use for primer 844.

eluquik 4.

Since the Elu-Quik samples were not the best with primer 17898, I chose to ignore them and to focus next on the Taq and MgCl2 concentrations. On this gel, I ran identical samples from all populations next to one another using the following combinations:

1: 0.05 Ál Taq / 1.5 Ál MgCl2
2: 0.05 Ál Taq / 1.0 Ál MgCl2
3: 0.10 Ál Taq / 1.5 Ál MgCl2
4: 0.10 Ál Taq / 1.0 Ál MgCl2

Andi said that the last combination yielded the most distinct bands, but that I should try it at a higher temperature to increase resolution.

eluquik 5.

For primer 844, I wanted to re-test the Elu-Quik samples and to also verify that the 1.0 Ál MgCl2 was better than 1.5 Ál MgCl2. I also wanted to test a higher annealing temperature (46íC) to see if that would increase band resolution. Again, the best results were obtained using the uncleaned DNA with 1.0 Ál MgCl2 and it appeared that there was a definite improvement using the higher annealing temperature.

1: 1.5 Ál MgCl, uncleaned DNA
2: 1.5 Ál MgCl, Elu-Quik treatment
3: 1.0 Ál MgCl, uncleaned DNA
4: 1.0 Ál MgCl, Elu-Quik treatment

eluquik 6.

Moving back to primer 17898, I wanted to test the different annealing temperatures and different MgCl2 concentrations (1.0 Ál MgCl2 versus 1.5 Ál MgCl2). I used the gradient robocycler, set to 45-47C. The gel was not the best, but did show that the 47C annealing temperature was not ideal (only one sample appeared). The best band resolution for primer 17898 seemed to be with 1.5 Ál MgCl2 (contrary to previous run) and a temperature of 45C.

1: 1.0 Ál MgCl, 45C
2: 1.0 Ál MgCl, 46C
3: 1.0 Ál MgCl, 47C
4: 1.5 Ál MgCl, 45C
5: 1.5 Ál MgCl, 46C

eluquik 7.

For primer 844, I wanted to test the effects of the different annealing temperatures and Taq concentrations (0.05 Ál and 0.10 Ál). Again, the 46C annealing temperature yielded the best results and 0.05 Ál Taq appears to be best.

1: 0.05 Ál Taq, 45C
2: 0.05 Ál Taq, 46C
3: 0.05 Ál Taq, 47C
4: 0.1 Ál Taq, 45C
5: 0.1 Ál Taq, 46C
6: 0.1 Ál Taq, 47C


Next, I wanted to begin screening several more primers. Shannon suggested using the lab's trinucleotide primers. I screened both Omar and Terry, and Mao and Manny. Because Mao had such a good band resolution and showed some population band differences, I chose Mao as one of my three primers.

eluquik eluquik
Test of Omar and Terry primersTest of Mao and Manny primers


Later on, I ran into problems while running my Mao gels. Previously, I had been getting some pretty nice gels (e.g., gel on left) when suddenly I began to get blank gels (e.g., gel on right). I tried various things (re-calibrated pipettors, mixed cocktail solution up extra well, etc.) to no avail.

mao4 eluquik
Mao gel that worked; Numbers on photo refer to different populations tested (C = positive and negative controls).Mao gel that failed

cocktail 10.

After talking with Chris and Shannon, I decided to try a variety of things on one gel instead of separately on different gels. I took one sample from each population (5 total) and used different cocktail solutions on each set of samples to find what the problem was:

1. all new solutions except for old dNTP.
2. " " " for old PCR buffer
3. " " " for old Taq
4. all old solutions

mao4b 11.

As expected, the cocktail made of old solutions gave blank samples, but I also got faint bands using the all new solution/old buffer combination. Because the other two cocktail solutions gave solid bands, I deducted that the problem must have been primarily with the MgCl2 solution and secondarily with the PCR buffer. Upon thinking back, I remembered being in a rush once and taking out some solution from these two tubes while there was still a bit of ice present. This probably changed the concentration of these two solutions and rendered them inaccurate for all subsequent gels. So, just to double-check my conclusion, I made up a new cocktail solution with all new components and the gel looked pretty good.

Numbers on photo refer to different populations tested (C = positive and negative controls).


One last difficulty I ran into was with primer 17898. After I had run all of the first set of gels (e.g., gel on left) and was on the replicates, I noticed that the bands were looking progressively fainter. I suspected the primer was the problem because the same primer stock solution had been used for all gels, whereas I had been constantly using new solutions of the other cocktail components. I obtained a new stock solution of the primer and immediately saw an improvement (gel on right). However, now there were several new bands which were absent from previous gels. I eventually had to re-run all of the gels using the new stock solution of primer.

17898 eluquik
Early gel with 17898 primer. Numbers on photo refer to different populations tested (C = positive and negative controls).Gel with degraded 17898 primer. Numbers on photo refer to different populations tested (C = positive and negative controls).

Return to:

ISSR Resource Website

Go to:

Wolfe Homepage

Wolfe Lab Homepage

Penstemon Website

Please send your suggestions, comments, corrections to
Last updated 17 January 2000.