Team:Harvard/Template:NotebookData

June 7th

Miniprep of pKD42 (lambda red)

The lambda red plasmid is needed to enable the recombination used to insert the selection/expression systems into our E. coli cultures.

Procedure: followed Qiagen Kit instructions, each student (8) using 1 mL cell suspension

Results: DNA reasonably pure (260/280 between 1.8 and 2) and between 25 and 50 ng/µL

June 8th

PCR to connect ultramers into OZ052 (Zif268 F2 triplicate, GCCGATGTC)and OZ123 (Zif268 F2 triplicate, GAGTGGTTA):

OZ052:

• 3µL OZ052_F (10µM stock)
• 3µL OZ052_R (10µM stock)
• 5µL 10x Pfx amplification buffer
• 1.5µL dNTPs
• 1µL MgSO4
• 0.4µL DNA polymerase
• 36.1µL ddH2O

OZ123:

• 3µL OZ123_F (10µM stock)
• 3µL OZ123_R (10µM stock)
• 5µL 10x Pfx amplification buffer
• 1.5µL dNTPs
• 1µL MgSO4
• 0.4µL DNA polymerase
• 36.1µL ddH2O

Parameters:

• 1) 94⁰C for 5 min
• 2) 94⁰C for 15 sec
• 3) 60⁰C for 30 sec
• 4) 68⁰C for 1 min
• 5) Repeat 2-4 for 25 cycles
• 6) 68⁰C for 5 min
• 7) 4⁰C forever

June 9th

• Created cell culture with selection construct (contains ZFB, His3, pyrF on plasmid) and reporter RFP (this will be used to test positive control ZFs, cells fluoresce green when ZF binds)
  • Picked colonies, grew in LB/amp liquid media until mid-log
    • 3 mL of LB, 1.5 µL of 2000x amp
  • Once mid-log reached, created glycerol stock, stored stock at -80⁰C.
    • 300 µL bacteria, 1200 µL 80% glycerol This should have been 1200 µL bacteria media, 300 µL 80% glycerol (Corrected 6/14/2011) (80% pure glycerol, 20% molecular grade water)
  • Spiked new tubes of media with 25 µL bacteria from the mid-log tube to leave overnight

NOTE: reporter RFP did not grow to mid-log by end of day, will let grow overnight to saturation and continue creating glycerol stock tomorrow.

• Plated selection strain from gel stab onto tet plate.

• Began primer design for creating the kan/selection construct fusion (see our primer spreadsheet for details).

June 9th - Bioinformatics

Today we focused on reacquainting and familiarizing ourselves with Python. We completed the parsing (reading in) of the sequence and amino acid data so that it is easy to work with: by substituting each amino acid abbreviation (ex. A, N) with its numeric equivalent (ex. 1, 14), we can use a lot of nice math comparisons instead of messy letter/"string" comparisons.

After that, we worked on counting the number of times each amino acid appears in each of the 7 positions (unfortunately given by -1,1,2,3,5,6,7), and counting the number of times amino acids are next to each other (ex. ACTQRNF has AC, CT, TQ, etc pairings). Taken overall, we found that L is overwhelmingly in position 5.

For pairings, we found patterns, but none as obvious as the L-in-position-5. Read this like a multiplication table: the intersection of L row and M column is how often that pairing was observed.

Follow up work on this will be to convert this table to frequencies instead of values: values are less meaningful.

June 10th

• What we learned today: don't put E. coli plates in the -20C freezer!

• Observed a well populated selection strain plate and placed it in the 4C refrigerator

• Took the selection construct culture and extracted the plasmid through miniprep
  • Observed 260/280 ratio of 1.90 and 1.88 through Nanodrop
  • Observed concentrations of 87.7 and 100.6 ng/µL through Nanodrop

• Made 10 new agar plates with LB and amp

June 10th - Bioinformatics

Visualizations

We spent the first few hours today making cool visualizations and graphs of the data we found on the 9th: heatmaps turned out to be an annoying limitation of Python, so a Python/R hybrid was used, and bar charts were made exclusively in Python. See the dropbox for our pretty (and hopefully informative compared to spreadsheets) charts/graphs.

We then started work on TNN and GNN properties specifically (essentially repeating the June 9th work, but confined to smaller data sets). There are some differences between TNN and GNN: see graphs in dropbox. We decided that there was not enough data for fingers that bind to ANN and CNN triplets to perform significant analysis on it.

• Overall, similar color clusters are found in the heatmaps. In all cases, L and N are often placed consecutively on the helix. There are fewer clusters of high frequency when looking at TNN binders.

We then, using the theorized framework from a paper (2011 Persikov [http://iopscience.iop.org.ezp-prod1.hul.harvard.edu/1478-3975/8/3/035010/]), tried to match amino acid binding to each base pair to see if there was a pattern. See dropbox document .../bioinformatics/Binding Frequency for that data. There's a lot of it.

Properties of amino acids

We then worked on finding properties of the each position (hydrophobic/phillic, non/polar):

Hydrophilic vs Hydrophobic

Polar vs Nonpolar

Follow up work here is to check more properties, and maybe try individual pairings (ex. phobic-philic, polar-phillic).

June 13th

The control zinc fingers OZ052 and OZ123 were amplified with overhanging primers to allow its insertion into the Wolfe plasmid:

Overhang PCR for ultramers: the template was the product of the ultramer PCR (see 6/8/11), and several concentrations were used

In all the tubes:

• 5 µL Pfx amplification buffer
• 1.5 µL dNTPs
• 1 µL MgSO4
• 0.4 µL polymerase
• 38.1 µL ddH2O
• 1.5 µL OZ052_up and 1.5 µL OZ052_down OR 1.5 µL OZ123_up and 1.5 µL OZ123_down

In OZ052 (1) and OZ123 (1):

• 1 µL of ultramer PCR product

In OZ052 (1:10) and OZ123 (1:10):

• 1 µL of a 1 in 10 dilution of ultramer PCR product

In OZ052 (1:100) and OZ123 (1:100):

• 1 µL of a 1 in 100 dilution of ultramer PCR product

Parameters:

• 94⁰C for 5 min
• 94⁰C for 15 sec
• 55⁰C for 30 sec
• 68⁰C for 30 sec
• Repeat steps 2-4 for 25 cycles
• 68⁰C for 5 min
• 4⁰C forever

Gel to verify proper amplification (1% agarose gel, 10 µL 1 kb ladder, 120 V):

The OZ052 lanes (1-3) had bands at the proper length (328 bp) at all three concentrations, although there were several fainter bands likely from side products. Only the undiluted OZ123 lane showed any bands, and from the faint band at 328 and the stronger band around 250 it appears that the PCR did not work well, and the majority of the product was the ultramer from the first PCR.

PCR around vector: the template used was the Wolfe selection construct plasmid miniprepped 6/10/11 (100.6 ng/µL stock) Reagents the same as above except:

• 1.5 µL of Wolfe_F and 1.5 µL of Wolfe_R primers to each tube
• plasmid tube (1 ng) given 1 ng of template (1 µL of a 1 in 100 dilution)
• plasmid tube (10 ng) given 10 ng of template (1 µL of a 1 in 10 dilution)

Parameters same as above except:

• elongation (step 4) 5 minutes (vector approximately 5 kb)

Gel to verify proper amplification (1% agarose, 10 µL 1 kb ladder, 170V)

There were no bands of the correct size in the lanes. The only band that appeared was a faint, short band in one lane that likely was a primer. Since the DNA ladder worked, the problem likely was not with the electrophoresis but with the PCR reaction, perhaps due to issues with the primers.

Gel images

June 13th - Bioinformatics

Today we started work on a program to statistically generate possible sequences.

The four functions needed to do this are:

• generate(matrix, pseudocounts (lambda), dependency tuples)

takes a matrix of zinc-finger AA position counts, a list of dependent amino acid pairs, and a pseudocount multiplier and generates a list of potential amino acid sequences weighted by independent and dependent probabilities

• add_pseudo(dependent matrix row,independent matrix row)

given a matrix row of dependent counts (i.e. how many times 'a' occurs at position n when 'b' is set to some AA at position m) and a row of independent matrix counts (how many times 'a' occurs at n regardless of b's AA) return an adjusted matrix row, based on the dependent matrix row, that has pseudocounts added to the values that are empty in the dependent matrix row but filled in the independent matrix row.

• generate_indep(matrix)

randomly pick an amino acid, given a matrix row, from a weighted random distribution based on the values in the row

• generate_dep(indep_row, dep_row, lambda)

add pseudo counts (call add_pseudo) and generate a dependent random call for a position (using generate_indep on the adjusted matrix)

We finished generate_indep, generate_dep, and add_pseudo today, along with creating a 140x140 matrix of needed values.

June 14

• Made four LB-based media solutions, and later created glycerol stocks from these and placed in -80⁰C freezer
  • Selection strain (ΔHis3ΔPyrFΔrpoZ) in 3 mL LB and 3µL of 1000x Tet solution stock
  • Selection strain (ΔHis3ΔPyrFΔrpoZ) in 3 mL of LB only solution stock (control)
  • Kan cassette (pZE22G) in 3 mL of LB solution stock and 3 µL of kanamycin solution stock
  • Lambda Red (pKD42) in 3 mL of LB and 3 µL Amp solution stock
    • For all of these stocks, we tried to grow all to mid-log and then place them in 1,200 µL of culture and 300 µL of 80 % glycerol solution (This is the correct protocol for creating a glycerol stock; refer to June 9th)
    • We were only able to get the kan cassette to mid-log and created glycerol stock of the kan cassette
    • Observations included contamination of a pKD46 liquid culture, and we are leaving Lambda Red and both solutions with the selection strain for overnight growth
• Ran 1% gel (150V) with the rest of the OZ123 and OZ052 overhang PCR samples
  • Used better ladder today, less the 1 kb ladder
  • Bands followed the same pattern as the gel run on 6/13/11
• Used gel extraction to obtain the correct OZ123 and OZ052 PCR product from the gel
  • Used Qiagen quick gel extraction kit
  • OZ052 (from undiluted lane): 7.0 ng/µL, 260/280=2.42 (Note: this sample had a strange yellow substance in the column--may have been contaminated)
  • OZ052 (from undiluted lane): 12.0 ng/µL, 260/280=2.02
  • OZ052 (from 1:10 dilution): 10.8 ng/µL, 260/280=2.04
  • OZ052 (from 1:10 dilution): 15.2 ng/µL, 260/280=1.82
  • OZ052 (from 1:100 dilution): 20.6 ng/µL, 260/280=2.17
  • OZ123 (from undiluted lane): 6.3 ng/µL, 260/280=2.17
• PCR the backbone fragment of the plasmid using Wolfe_R and Wolfe_L primers and a lower annealing temperature than before due to the lower melting point of Wolfe_L
  • Reagents:
    • 22µL Invitrogen Platinum PCR supermix
    • 1µL template from a 1 in 100 dilution (1 ng)
    • 1µL Wolfe_F and 1µL Wolfe_R
  • 94°C for 30 s
  • 94°C for 30 s
  • 53°C for 30 s
  • 70°C for 5 min
    • Previous three steps repeat 30 times
  • 70°C for 5 min
  • 4°C forever
• Performed a restriction enzyme digestion on the selection construct plasmid using EcoRI to test for presence/absence of inserted selection construct
  • 1µL EcoR1
  • 1µL buffer 4
  • 2µL backbone plasmid
  • 6µL ddH2O
  • Incubate at 37 degrees for 90 min
  • There is only one EcoR1 site (GAATTC) in the plasmid, so we should see 1 band at about 5kb
• Ran a gel (1%, 170 V) on the backbone fragment (Wolfe primers) PCR product and restriction digestion result
  • Observations: the EcoR1 digest produced the expected band of around 5kb. The backbone did produce a 5kb band but also had a secondary smaller product, perhaps due to one of the primers annealing to a sequence that is a close match to its target.
• Began gradient PCR on the selection strain backbone with Wolfe_R and Wolfe_F primers because of the large difference in melting temperatures between the two
  • Set the annealing temperatures within the PCR to go from 50-57 C and ran with 5 minute extension phases at 70 C
  • Ran overnight

Today's Gel Images

June 14 - Bioinformatics

We finished writing the generate function, and now have a working sequence generator. We also began more in-depth research into the 2011 work by Persikov which deals with how the zinc finger binds to DNA. He predicts several relations which we should be able to test.

Persikov sent us his SVM code (used to calculate the probability of a sequence binding to given DNA), so we also worked on adapting this to use when narrowing our sequences to those most likely to work.

• There are four canonical amino acid-base interactions involved in zinc finger interactions.
  • Amino acids in positions -1, 3, and 6 on the helix are known to interact with the 3 bases of the triplet. In addition, the amino acid in position 2 interacts with the upstream base of the complementary strand (Klug 2010).
  • In addition, Persikov (2011) has proposed novel interactions between these four amino acids based on his analysis of zinc finger binding data upto 2005.
  • Persikov uses the information between these four amino acid-base interactions in his SVM, to determine whether a finger would be a good binder to a particular DNA sequence.
• In order to use his program, we need to convert the helices that are created by the generator into a format the SVM accepts.
  • The SVM only considers the four canonical interactions. It assigns a numerical value to each possible amino acid-base combination. The program that converts our data into a format the SVM accepts creates a string with these numerical values based on Persikov's key. (See this page for more details on this program.)

Brandon learned basic Python today. Justin created a JavaScript program that recognizes potential binding sites from a given sequence.

June 15th

Gradient PCR: 5µL of each PCR product were run on a 1% gel. No bands appeared: the PCR appears to not have worked.

Selection construct: bacteria containing the selection construct (plasmid containing ZF, omega subunit, ZFB, His3, URA3, etc.) was made into a glycerol stock (see 6/14 and 6/9), miniprepped, and used for PCR:

• Miniprep: used Qiagen kit
  • 82.5ng/µL, 260/280=1.99
  • 91.1ng/µL, 260/280=2.01
  • 77.9ng/µL, 260/280=1.98
• PCR:
  • PCR used to amplify section of plasmid containing zinc finger binding site, weak promoter, His3, and URA3 (with homology to join it to kan cassette)
  • Reagents
  • zinc finger binding site and weak promoter, selection construct plasmid:
    • 1µL ZFB-wp-f (5µM) (made by 1:20 dilution of 100µM stock)
    • 1µL ZFB-wp-hisura-r (5µM) (made by 1:20 dilution of 100µM stock)
    • 1µL selection construct (1:100 dilution of overnight culture)
    • 22.5µL of invitrogen's Platinum PCR SuperMix
  • PCR used to amplify the kan cassette
  • Reagents
  • KAN cassette, pZE22g plasmid:
    • 1µL hisura-kan-f (5µM) (made by 1:20 dilution of 100µM stock)
    • 1µL kan-r (5µM) (made by 1:20 dilution of 100µM stock)
    • 1µL pZE22g (1:100 dilution of glycerol stock culture)
    • 22.5µL of invitrogen's Platinum PCR SuperMix
  • Parameters:
• 1) 94°C for 2 min (denature template, activate enzyme)
• 2) 94°C for 30 sec (denature)
• 3) 53°C for 30 sec (anneal)
• 4) 72°C for 2 min (extend)
• 5) Repeat 2-4 for 25 cycles total
• 6) 72⁰C for 5 min
• 7) 4°C forever

PCR Purification:

• Used the Qiagen PCR purification kit and instructions in order to purify the Kan cassette and selections construct PCR products
  • Nanodrop the purification results and observed 3 ng/µL for Kan cassette and 29.8 ng/µL for ZFB-wp-His3: purification did not work well, especially for the kan

Selection strain (ΔHis3ΔPyrFΔrpoZ):

• saturated overnight culture was inoculated again: 3mL LB, 3µL tet, 30µL of overnight culture, at 37C until mid-log
• glycerol stock
• For transformation tomorrow we grew up pKD42 in 3 mL of LB, 1.5 µL of ampicillin(2000x) and one colony at 30 C
  • Also grew up more of the selection strain so it will be ready for electroporation transformation

Gel

• Ran gel with Kan cassette and selection construct (Binding site, His3, and URA3)
  • Observations successful and image below
  • Used 1 kb plus ladder

PCR Overlap

• since the purification was not very successful, we used 3µL saved from the original PCR product
• Procedure
  • 25µL of 2x Phusion Master Mix
  • 1 µL of ZFB-wp-HisURA-R (100µM)
  • 1µL of HisURA-Kan-F (100µM)
  • 21 µL of water
  • 1 µL Kan template and 1 µL of ZFB-wp-His3-URA3
    • 4 tubes
      • Both undiluted
      • Both 1:10 dilution
      • Both 1:100 dilution
      • Both 1:1000 dilution
• Protocol
  • 98 C for 30 s
  • 98 C for 10 s
  • 53 C for 30 s
  • 72 C for 3 min
  • Repeat steps 2-4 for 24 more cycles
  • 72 C for 5 min
  • 4 C 4EVA!!!

June 15th - Bioinformatics

• We continued research into Persikov's and others' work on binding.

• We worked on using the OPEN data to test Persikov's binding-predicting program: SVM

• Justin continued work on a sequence-finding program, the most up to date version can be found in the Dropbox under code/zfsitefinder.html.
• Justin and Will found 10 candidate sequences across 4 diseases that hopefully should encompass a good amount of diversity in terms of expanding the ZF library. These sequences can be found in the table below, with more details here. The most up to date version can be found in the Dropbox under sequences/Target Loci Sequences.

• We collected 15 alternative zinc finger backbones (different from zif268 backbone) and their corresponding base sequences. Many of these were from Persikov 2011 and all binding sequences were confirmed on the Protein Data Bank website at http://www.pdb.org/pdb/home/home.do. The zinc finger PDB ID's and related links are:

June 16th

• So there was totally a crazy bee hive outside today!!

Glycerol Stock pKD42

• Grew up pKD42 in 30 C and once reached mid-log created glycerol stock and placed in -80 refrigerator

Overlap PCR gel

• Ran gel to test if overlap PCR that ran through the night worked, and it did not
  • Used PCR product without purification which gives good explanation for why it didn't work

• PCR: Since the PCR done the previous day (6/15) we made a back up PCR using phusion mastermix (Finnzyme)
  • PCR used to amplify section of plasmid containing zinc finger binding site, weak promoter, His3, and URA3 (with homology to join it to kan cassette)
  • Reagents
  • zinc finger binding site and weak promoter, selection construct plasmid:
    • 1µL ZFB-wp-f (100µM) (taken directly from the primer tube)
    • 1µL ZFB-wp-hisura-r (100µM) (taken directly from the primer tube)
    • 2µL selection construct (1:100 dilution of overnight culture)
    • 25µL Phusion High-Fidelity PCR Master Mix
    • 21µL distilled water (for total volume of 50µL)
  • PCR used to amplify the kan cassette
  • Reagents
  • KAN cassette, pZE22g plasmid:
    • 1µL hisura-kan-f (100µM) (taken directly from the primer tube)
    • 1µL kan-r (100µM) (taken directly from the primer tube)
    • 2µL pZE22g (1:100 dilution of glycerol stock culture)
    • 25µL of Phusion High-Fidelity PCR Master Mix
    • 21µL distilled water (for total volume of 50µL)
  • Parameters:
    • 1) 94°C for 2 min (denature template, activate enzyme)
    • 2) 94°C for 30 sec (denature)
    • 3) 53°C for 30 sec (anneal)
    • 4) 72°C for 2 min (extend)
    • 5) Repeat 2-4 for 25 cycles total
    • 6) 72⁰C for 5 min
    • 7) 4°C forever
• Ran PCR product on a gel: bands of the correct size were observed, though the kan band was much fainter than the ZFB-wp-his3

• Repeat PCR (to get a higher concentration of the Kan cassette and ZFB-wp-his3 constructs)
  • same as the above backup PCR (since it was successful), but to a 4x total volume of 200µL, compared to 50µL
  • PCR products were run on a gel: the correct bands were observed--see image below ("second gel")
  • PCR product purification: followed Qiagen kit instructions. Strangely, the conc. was 64.6 ng/µL for Kan, purity 2.09 (260/280) and for ZFB-wp-his3, the conc. was 23.7ng/µL and the purity was 1.92 (260/280).

Overlap PCR: used the kan cassette (64.6 ng/µL) and ZFB-wp-his3-ura3 (23.7 ng/µL) purified above

• 1 µL of kan and 1 µL of ZFB-wp-his3-ura in each tube according to the following conditions:
  • two tubes: undiluted
  • two tubes: both diluted 1 in 10
  • two tubes: both diluted 1 in 100
• 12.5µL Phusion master mix
• 8 µL ddH2O
• primers: 1.25 µL ZFB-wp-hisura_r (10 µM) and 1.25 µL hisura-kan_f (10 µM)
  • we tried two different reaction types: one added the primers as usual before starting the PCR reaction, the other added the primers after 10 PCR cycles (allowing the polymerase to first use the overlapping kan and ZFB to elongate, and then the primers)
• Parameters for PCR starting with primers:
  • (PCR machine 5, program name EXT3KB in IGEM folder)
  • 1) 98°C for 1 min (denature template, activate enzyme)
  • 2) 98°C for 15 sec (denature)
  • 3) 65°C for 15 sec (anneal)
  • 4) 72°C for 2 min (extend)
  • 5) Repeat 2-4 for 30 cycles total
  • 6) 72⁰C for 5 min
  • 7) 4°C forever
• Parameters for PCR starting without primers:
  • 1) 98C for 1 min
  • 2) 98C for 15 sec
  • 3) 65C for 15 sec
  • 4) 72C for 1 min
  • 5) back to step 2 for 10 cycles (PCR paused after 10 and primers added)
  • 6) 98C for 15 sec
  • 7) 65C for 15 sec
  • 8) 72C for 2 min
  • 9) back to step 6 for 20 cycles
  • 10) 72C for 5 min
  • 11) 4C forever

Transformation

• Used the selection strain (ΔHis3ΔPyrFΔrpoZ) cells at mid-log and attempted to place lambda red (pKD42) plasmid into the cell
• Procedure
  • Keep on ice through out whole procedure before use of the electroporation machine
  • Spin 1.5 mL of mid-log cells at 4 C for 1 minute at 18000 rcf (we created two tubes through the following steps)
  • Discard supernatant and resuspend with 1 mL of cold water
  • Spin again and repeat for a second water wash
    • With each wash, try to get as much supernatant out as possible(even use pipette) because don't want salts to interfere with the electrical pulse
  • Resuspend pellet with 50 µL of cold water
  • Add 1 ng of pKD42 to one of the tubes and 45 ng of pKD 42 to the other
  • Take all of the liquid in each tube and place in two separate cuvettes for electroporation
    • Make sure the electroporation machine is on the right setting (for the cuvettes we used today it was "Ec2")
    • Wipe off all water on the side of the cuvette
  • Have 1 mL of LB in hand and after pulsing, immediately put LB in cuvette
  • Transfer to culture tube and place in 30 C for 2 hours
  • Make 4 LB/amp plates and spread E. coli using glass beads:
    • Plate 1: 10 µL of 1 ng culture
    • Plate 2: take 700 µL of 1 ng culture, spin down and remove supernatant, resuspend in about 30 µL of LB and plate
    • Plate 3: 10 µL of 45 ng culture
    • Plate 4: take 700 µL of 45 ng culture, spin down and remove supernatant, resuspend in about 30 µL of LB and plate
  • Grow overnight at 30 C

PCR to confirm knockouts of selection strain

• this PCR was to confirm that the ΔHis3ΔPyrFΔrpoZ was indeed a knockout for the His3, PyrF, and rpoZ genes
• each primer set was used for two conditions: wild-type (we used a pKD42 culture) and knockout (ΔHis3ΔPyrFΔrpoZ culture, left over from transformation)
• 1 µL of either wt or ko template, diluted 1:20
• 12.5 µL Phusion master mix
• 1.25 µL of each 10µM primer:
  • test for His3:
    • 1)His3_F, His3_R
    • 2)His3_F, His3_internalR
  • test for PyrF:
    • 3)PyrF_F, PyrF_R
    • 4)PyrF_F, PyrF_internalR
  • test for rpoZ:
    • 5)rpoZ_F, rpoZ_R
    • 6)rpoZ_F, rpoZ_internalR
  • test for Zeocin (there are two primer sets because we don't know what orientation the Zeocin gene is in)
    • 7)Zeocin_R, rpoZ_F
    • 8)Zeocin_R, rpoZ_R
• ddH2O up to 25 µL
• Parameters:
  • 98 C for 5 min
  • 98 C for 10 sec
  • 65 C for 25 sec
  • 72 C for 45 sec
  • cycle 30 times
  • 72 C for 5 min
  • 4 C forever
• Results: ran PCR products out on 1% gel (see below). There were some nonspecific bands, but the PyrF and rpoZ genes do appear to be knocked out in the selection strain. His3, however, looks like it's still present--we'll test again to confirm.

June 16 - Bioinformatics

• Research Targets

1. Clinically relevant targets
2. Existing ZFs that bind under-represented triplets

Updating our programs

• Many of our current programs currently look at overall data or data based on specific DNA triplets (for example: 'GAT' or 'AAA'). However, in order to more easily understand some of the patterns that occur in the datasets, we want to examine broader subsets of data. For example, do different patterns appear when looking at fingers that bind to 'GNN' triplets versus 'NGN' triplets (where 'N' represents any of the 4 bases)?
  • We added the capability for our programs to accept inputs with the variable 'N' by using regular expressions.
    • We can now create lists of the zinc fingers that bind to any triplet, and create interaction matrices and frequency tables for any triplet input.

June 17th

Update on selection strain knockout status: We are trying to reach Addgene to check how His3 was knocked out---instead of deleting the gene, they may have simply introduced an early stop codon. If that's the case, our gel would have the correct bands because the primers we designed can only show whether a deletion or insertion was in that locus.

Transformation results: successful!!

• The only plate with colonies was the one plated with 700 µL of cells transformed with 45 ng of pKD42
• Chose a colony to grow in 3mL LB, 1.5µL amp, 30C; make glycerol stock with mid-log cells
• Plate with colonies at 4C

Miniprep of pZE22G: (to have the plasmid containing the kan cassette on hand)

• used 2 tubes of 1.5mL overnight culture, followed Qiagen kit instructions
• 38.0 ng/µL, 260/280=1.99
• 27.8ng/µL, 260/280=2.02

Overlap PCR gel and extraction: 1%, 150V

• Results: adding the primers in after 10 cycles was much more successful than adding them at the beginning, and all three dilutions showed the expected product band (about 2.5kb). The rest of the 1:10 dilution will be run on a gel and extracted.
• 11.2ng/µL, 260/280=2.10

June 17 - Bioinformatics

Goals

1. Make BB Database in program-readable format ✓
2. Edit out BB with incomplete helices ✓
3. GNN, TNN, CNN, ANN frequencies

• Targets (5-10; 8) x Backbones (???) x Helices (≥500)=55,000
  • Backbones: similar, but not too similar to zif268; more than 1-2 aa changes, but <10
  • Helices fixed based on our program-- eventually saturates and levels out
    • Graph: # of var (# of tries by the computer) vs. % space covered

Options for Target DNA Sequences / ZF Helices

1. F3(known) / F2(known) / F1(novel)
2. F3(known) / F2(SNP in b₁ position) / F1(known)
3. F3(unknown) / F2(unknown) / F2(unknown)

• Excluded Rare Codons (for E. coli)CodonUsage OpenWetWareCodonUsage NIHRareCodonCalculator:
  • CTA
  • ATA
  • CCC
  • CGA
  • CGG
  • AGA
  • AGG
  • GGA
  • GGG

References

<biblio>

1. Persikov2011 pmid=21572177
2. CodonUsage http://www.sci.sdsu.edu/~smaloy/MicrobialGenetics/topics/in-vitro-genetics/codon-usage.html
3. OpenWetWareCodonUsage http://openwetware.org/wiki/Escherichia_coli/Codon_usage
4. NIHRareCodonCalculator http://nihserver.mbi.ucla.edu/RACC/

</biblio>

June 20th

• Grew up colony of the selection strain with pKD46 in an attempt to reach mid-log and create glycerol stock
  • Unable to reach mid-log, so going to leave growing over night and use saturated culture tomorrow
• Determined primers in order to piece together the omega subunit and ZFP genes into the pZE21G plasmid (spec cassette)
• Ran PCR on His3 locus and sent to GENEWIZ to be sequenced
  • used the same procedure as the earler WT/KO PCR, but with 1µL undiluted template and only His_F and His_R primers
  • ran 3 reactions and sent in three primers (His_F, His_R, His_internalR)

June 20th - Bioinformatics

Goals for the week

• Finish designing the chip, by Wednesday hopefully
  • Need chip order out, takes 4 weeks
  • Need all sequences by Friday!!!
• FIRST PRIORITY: If we can get Persikov to work, good!
  • Step one: get results he’s published, get the web app to "work" with his data, then OPEN data, and finally our data
• Brainstorming session (tomorrow?) to decide how many targets/sequences
• Determine the importance of the first/second/third nucleotide positions
  • Look at NGN, NTN, NAN, NCN (Not just GNN, etc.)
  • Pick a particular GNN, plot vs. TNN- is there a pronounced difference in position 1, or -1?

Today

• Testing Persikov's Data for validation
  • Persikov v. himself ✓
  • Persikov v. OPEN
  • Persikov v. our sequences

Probability data

• The following are graphs of the probability of finding each amino acid at each position on the alpha helix.

June 21st

His3 sequencing results:

The sequencing results showed that the His3 (HisB) gene is still present in the strain and without any early stop codons. There is a 2 aa deletion in the middle of the protein, but its purpose is unknown and the gene likely is still fully functional.

• Restreak selection strain on plate from glycerol stock--tomorrow we will PCR the His3 locus and sequence again just to be sure.
• Made oligos for MAGE to insert stop codons and make a frame shift in the endogenous His3 gene, so that if necessary we can knockout His3 ourselves.

Selection strain with lambda red:

• Reinoculated and made glycerol stock
• prepared for MAGE tomorrow

June 21st - Bioinformatics

Persikov Statistics - Graphs

• FQCRICMRNFS_{zif268 F2 Backbone}/Helix F1/TGEKP_linker

• The Persikov data shows weak predictive power for OPEN amino acid sequences. Our conclusion is that Persikov's program is not well-suited for incorporation into our helix generator. Testing Persikov's helices in his program yeilded mostly accurate results (approximately 24/25 matched known binding information). This is an important test because it proved that we are using the program correctly and that the program is in fact working properly. However, testing the OPEN sequences in Persikov's program resulted in numerous false negative values which informed our decision not to use Persikov's program to check our own hellix-generating program.

Phone Call with Dan

• How conservative/risky should we be in terms of using other backbones?
  • Conservative
    • Possible Pros:
      • More likely to get something that will work
      • Depending on how "smart" our probabilities are (from our ZF generation algorithm), we could cover a lot of novel space without straying too far from zif268
      • Worst Case:Something we can show for iGEM (we covered the same ground OPEN did, and found many of the same ZFs, but with a targeted approach, a "smarter" method-- not throwing random things at it; Chip is not ours, but the program is "smarter")
    • Possible Cons:
      • Might end up covering the same ground as OPEN, but doing a "worse" job than they did
      • Less likely to discover new/groundbreaking things (i.e., TNN triplets)
  • Less Conservative
    • Have 3-6 target sequences (we're currently going for 8)
    • More backbones from non-zif268 than zif268
    • Pros:
      • We could get luck and find something no one has ever seen before (TNN, ANN). If we throw enough things at it, we're more likely to get luck.
    • Cons:
      • Risk: Many of these backbones (from entire ZF world)may NOT bind DNA (i.e., may bind proteins)
      • Risk: May not find anything that binds, then the whole project is a dud
• What is the more important variable, helices or backbones?
  • Helices seem to be more important, backbones of secondary importance
  • Backbones: ZF's unravel DNA, open the major groove-- backbone is important here, changes the bond angle, etc. (Brandon's paper-??)
• Balance needed between low and high risk
  • If we find backbones that we know bind DNA, greatly lowers our risk
  • Limited spaces on chip: zero-sum game
  • With a middle of the road approach, we diminish both benefits and risk (diminishes the benefits of the high risk approach much more than it diminishes the benefits of the conservative approach; i.e., if you're playing the lottery, you're more likely to win if you buy many more tickets)
• We need to compare probabilities of randomly-generated OPEN sequences vs. probabilities of sequences randomly generated by our program
  • OPEN tries to cover all space: smaller probability
  • If we have a "smarter" algorithm, we can produce fewer
  • However, the idea is not to repeat OPEN, but to go somewhere else, non-GNN sequences
  • Remember: OPEN is a Cell paper; the point of the project is not to compare ourselves to them
• If we find binders for 1-2 of our sequences, that would be awesome
  • Probably we'll have some that find none, some have 10, our last one might have 1,000 hits (then, we do bioinformatics to figure out why/what those hits were)
  • Point: to learn and do high-level bioinformatics, and high-tech cloning techniques in the lab
  • If you do find binders, you can write a paper about it!
• We have all the resources we need right now to build our chip
  • We need to pick out targets
  • Need to decide exactly what we want for:
    • No. of target sequences/which ones
    • No. of helices/ which ones
    • Ratio of zif268 backbones: non-zif268 backbones
  • Avoid switching Leucine out of position 4, then change other positions based on our frequencies

Chip Design

• No. of sequences will be more than we can put on the chip
  • Helices: essentially unlimited
    • Put more-likely-to-bind helices into the risky backbones
    • Put less-likely-to-bind helices into a zif268 backbone
• Backbones
  • Maybe revert to a more targeted approach: pick backbones that we know are transcription factors (TFs), that we know bind to DNA
  • OR research the ZFs from the phylogenetic tree
    • Pick clades to research, see if one looks better than the other
  • Why did OPEN cover so many helices, without changing the backbone, but still yield predominantly GNNs?
  • If we have an idea of how the backbone might affect binding, maybe we could look into some sort of low-level modeling, etc. so that we wouldn't be grasping? Could Vatsan help with this?
    • See 2000 Wolfe paper [http://www.ncbi.nlm.nih.gov/pubmed/10940247]
  • Backbones could affect interactions between fingers
  • Theory: energy penalty to ZF binding-- unravels DNA when binds to it
• We have 12 target sequences
  • 2 per 4 diseases, 4 for the 5th disease
  • If we want to be more conservative, we could throw out Type III, but it could be something cool
  • We should have mostly Type I (CoDA argument, if this is an F2)
  • Proposed: 3 diseases, 6 sequences
    • 4 Type I (F3 and F2 known, F1 novel)
    • 1 Type II (GNN, ANN, GNN)
    • 1 Type III (All unknown, e.g., TNN, ANN, TNN;max 1)

Or, for 3 diseases:

1. Type I's
2. Type I, Type II
3. Type I, Type III

• Clinical Targets

1. Colorblindness (Type I's)
2. Familial Hypercholesterolemia (FH) (1 in 500)
3. Cystic Fibrosis (CF)
4. Tay Sachs
5. KRAS- (oncogene/cancer)

• Main goal of project: to build outside of what is already known
  • If we wanted to cure a disease only, we could just use existing ZFs (i.e., find GNN binding locations)
  • Also, we lend a level of specificity for insertion/deletion
  • There is the possibility that there might be some area where specificity might demand ANN codons

Current decision on chip design:

• We will have 6 target sequences, 2 each from colorblindness, FH, and KRAS. All are "Type I" targets (only F1 is novel) with the middle finger chosen from the CODA paper (either GNN or TNN)
  • N.B.: the CB and FH sequences make up full ZF nuclease cut sites. The KRAS sites, due to the small number of GNNTNN F3F2 combos available in CODA, are separate, with the flanking ZF nuclease site added afterwards in parentheses

1. GGTGGTAAG (CB)
2. GGAGTCCTG (FH)
3. GGCTGATGC (KRAS) (CTGAAAATT)
4. GGCTGACAC (FH)
5. GGCTGGAAT (KRAS) (GACAAGAGC)
6. GTCGCCTCC (CB)

• Targets 3, 4, and 6 are similar to sequences Zif268 variants successfully bind to, so the backbones will be weighted accordingly:
  • Zif268_F2 backbone: 6000 helices (per target)
  • 10 backbones more closely related to Zif268: 300 helices each
• Targets 1, 2, and 5 will have equal distributions of backbones:
  • Zif268_F2: 3000 helices
  • 10 backbones closely related to Zif268: 300 each
  • 10 backbones more distantly related to Zif268: 300 each

Identifying dependencies

• We looked at the probability graphs to determine which amino acid positions on the finger's helix interact with which bases.
  • Some interactions are fairly well estabilished, while others have been more recently proposed (See interaction map (Persikov 2011))
  • To identify these interactions in our own data we looked at which helix positions varied most when you changed the bases. A more rigorous way to do this is to calculate the entropy change as you change the amino acids in each position.
    • xNN(Vary base 1): Amino acid 6 changes
    • NxN(Vary base 2): Amino acid 3 changes
    • NNx(Vary base 3): Amino acid -1 and 2(?) changes
  • Our program looks at dependencies between amino acids when generating sequences.
    • We decided on these amino acid dependencies, using both established data and patterns we saw in the OPEN data:
      • -1 and 2
      • 2 and 1
      • 6 and 5
  • Because there is not much data for 'CNN' and 'ANN' sequences (with 16 and 29 known fingers that bind to each triplet, respectively), we should use pseudocounts for these sequences, so that our frequency generator is not too biased toward probabilities that may not be significant.