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 1
Chromatin Immunoprecipitation

THIS CHAPTER IS COMPLETE
Brief Update Thu Jul 12 18:54:37 EDT 2007:   The ChIP verifications done in this chapter were published in the PLoS Biology paper: Large-Scale Mapping and Validation of Escherichia coli Transcriptional Regulation from a Compendium of Expression Profiles. Please see that paper for background, introduction, and a discussion of this work in the context of our work on network inference. For this work in particular, see the section of that paper entitled In vivo confirmation of new regulatory interactions. The detailed protocol that resulted from the work in this chapter is available in the appendix section on page with recipes available in section on page .
I'm trying to verify predicted interactions using ChIP.
Please see the pdf version for figures
Figure 1.1: Figure is drawn to scale. Big box is number of genes in regulon (genes only). Blue box is size of regulon. Space not covered by blue box is all non-regulon genes. Green box is number of TFs outsize regulon. Yellow is number of TFs in regulon.

1.1  Chip Target Primer testing

For each predicted target of the transcription factors to be detected we designed a primer pair to amplify the intergenic region and test if the region has been enriched by immunoprecipitation by an antibody directed at the transcription factor of interest. Primers were designed using Primer3 software with the following constaints: length 24bp+/-3, melting temp 60+/-3, %GC 50+/-8. Here we are verifying that all primers amplify only one region of the correct length.

1.1.1  Low-throughput testing

Wed Sep 28 13:57:27 EDT 2005 testing 8 genes (aroG, aroL, aroP, asd, cirA, cysC, dinD) also testing two efficiency parameters:
  1. can we run a 2-step rather than a 3-step reaction to speed simple primer design tests
  2. can we use half of qiagen's recommended 100ul reaction volume to save reagents
all eight genes are run in a 50ul reaction (half the qiagen recommended reaction):
PCR Reaction composition
H2O 23.5 ml
Qiagen Master Mix 25 ml
Forward and reverse primer1.5 ml
final primer concentrationN pM
Thermal cycler conditions
Initial denaturation5 min 95° C
2-Step cycling
Denature:30 sec 95° C
Anneal/Extend:45 sec 60° C
Number of Cycles:25
Final Extention: 7 min 72° C
aroG and aroL are included as a 100ul reaction (the recommended solution) the PCR parameters are:
PCR Reaction composition
H2O 47 ml
Qiagen Master Mix 50 ml
Forward and reverse primer3 ml
final primer concentrationN pM
Thermal cycler conditions
Initial denaturation5 min 95° C
2-Step cycling
Denature:30 sec 95° C
Anneal/Extend:45 sec 60° C
Number of Cycles:25
Final Extention: 7 min 72° C
additionally aroG, aroL, aroP, asd were run as in a 3-step 50ul reaction:
PCR Reaction composition
H2O 2 ml
Qiagen Master Mix 5 ml
Forward and reverse primer23 mM
Thermal cycler conditions
Initial denaturation5 min 95° C
3-Step cycling
Denature:30 sec 95° C
Anneal:30 sec 60° C
Extend:30 sec 72° C
Number of Cycles:30
Final Extention: 7 min 72° C
aroP and asd were run as 100ul reactions:
PCR Reaction composition
H2O 47 ml
Qiagen Master Mix 50 ml
Forward and reverse primer3 mM
Thermal cycler conditions
Initial denaturation5 min 95° C
3-Step cycling
Denature:30 sec 95° C
Anneal:30 sec 60° C
Extend:30 sec 72° C
Number of Cycles:25
Final Extention: 7 min 72° C
Please see the pdf version for figures
Figure 1.2: 80 ml, 1% agarose gel with 2.5 ul of 1% ethidium bromide run for 40 min at 100 volts. 10 ml of Fisher BioReagents exACTGene Low Range DNA Ladder, with bands:weights(ng) of 2000:105, 1500:87, 1000:68, 750:59, 500:94, 300:27, 150:34, 50:25 was used. Bold text indicates reaction was 100ul, italics represents a 3-step PCR. 6ul of each PCR was used. Product sizes for the 8 genes should be: aroG=173bp, aroL=161bp, aroP=122bp, asd=152bp, cirA=240bp, cysC=174bp, cysK=117, dinD=162bp. WARNING asd is probably designed for wrong region, need to check.
Brief Conclusions:   It appears that all primer sets tested showed amplification of a single band of the correct length. However, the gel itself is very difficult to read. It wasn't run long enough and the short bp fragments are very fuzzy. Also the far left of the figure shows that migration of etBr towards the opposite pole of the gel is a problem when running short fragments.

1.1.2  Improving gel of target genes

Thu Sep 29 14:30:28 EDT 2005
In Figure 1.2, it appears that all primers worked fine, but the image/gel is bad. Using the same PCR products, we'll try a more appropriate 2% gel run longer with a post-stain to prevent the effects of having etBr migrate the opposite direction. Also going to try a larger comb and different concentrations of PCR product. 1x TAE was made fresh from the premade 50x Fisher stock in the Collins lab for both the gel and the running buffer.

For large (6-well) comb: trying aroG 6ul, 4ul, 3ul, cirA 6ul, 3ul
For medium (10-well) comb: trying aroG 6ul, 4ul, 3ul, cirA 6ul, 4ul, 3ul
Please see the pdf version for figures
Figure 1.3: 70 ml, 2% agarose gel run for 40 min at 120 volts, poststained in 100 ml of H2O with 0.5 ug/ml ethidium bromide for 40 min followed by 15 min in water alone (both on orbital shaker at 50rpm) to reduce background. 10 ml of Fisher BioReagents exACTGene Low Range DNA Ladder, with bands:weights(ng) of 2000:105, 1500:87, 1000:68, 750:59, 500:94, 300:27, 150:34, 50:25 was used. Product sizes for the 2 genes should be: aroG=173bp, cirA=240bp.
Brief Conclusions:   The gel in Figure 1.3 is much better than the previous one (Figure 1.2). The amplifications are definitely the correct size and all the bands of the ladder are clear. Adding 5 minutes to the post-stain and the post-water wash might help even more.

1.1.3  Higher-throughput primer organization

Fri Sept 30, 2005
Previously primers were in 1.5ml eppy tubes premixed with a forward and reverse primer for an intergenic sequence of a particular gene. To speed up primer testing and more importantly improve accuracy when testing genes on a 384-well qPCR plate the primers have been moved to a 96-well Costar plate. The concentration of each is 4mM allowing use of a 800pM final concentration in a 10 or 15ml reaction using 2 or 3ml respectively (the bottom precision range of our multichannel pipettor).
The primer organization is shown in Table . Each transcription factor (TF) is tested by one or more rows (indicated in the far left column; the first three columns are negative controls (genes thought not to be regulated by the transcription factor); the next two columns are positive controls (known targets of the TF). For example the first row tests lexA with negative controls: serA, entC, and fliF; positive controls: recA and lexA; and potential new targets: dinD, dinG, dinI, dinP, ruvA, yceP, yebG.
96-well primer plate for ChIP-PCR
- 1 2 3 4 5 6 7 8 9 10 11 12
A: lexA serA entC fliF recA lexA dinD dinG dinI dinP ruvA yceP yebG
B: fliA serA entC recA fliF fliL nikA flgK flgM fliC - - -
C: fecI serA recA fliF fecA fecI cirA entC exbB fepA fhuA nrdH tonB
D: lrp recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB
E: lrp ilvC lysC metA metE metF dapB dapD - - - - -
F - - - - - - - - - - - -
G - - - - - - - - - - - -
H - - - - - - - - - - - -
Table 1.1: Primer organization. TFs tested by each row are indicated in the leftmost column. Hyphens indicate wells with H2O only.

1.1.4  Higher-throughput primer check

Mon Oct 3 14:01:38 EDT 2005
Primers from the plate in Table 1.1 are to be tested as described in Section 1.1. All of the genes on the plate will be tested using the following reaction:
PCR Reaction composition
H2O 15 ml
Qiagen Master Mix 25 ml
Forward and reverse primer10 ml
final primer concentration900 pM
Thermal cycler conditions
Initial denaturation5 min 95° C
2-Step cycling
Denature:30 sec 95° C
Anneal/Extend:45 sec 60° C
Number of Cycles:25
Final Extention: 7 min 72° C

Reactions are run in 4, 8-tube PCR strips (A5-A12, B2-B9, D5-D12, E1-E7) and one 12-tube strip (C1-C12) (actually one 8-tube and one 8-tube ripped in half).
Please see the pdf version for figures
Figure 1.4: 200 ml, 2% agarose gel with 4 ul of 1% ethidium bromide run for 45 min at 120 volts. Image is the inverse as the large amount of dye on the gel made the true image very difficult to see. 10 ml of Fisher BioReagents exACTGene Low Range DNA Ladder, with bands:weights(ng) of 2000:105, 1500:87, 1000:68, 750:59, 500:94, 300:27, 150:34, 50:25 was used.
Brief Conclusions:   A 250ml or 300ml gel would definitely be easier! The 200ml gel was very thin. Also, as can be seen, in 1.4 there is way too much loading dye and it migrates to the same region as most of the PCR fragments. Last the imaging system used is the one from the CAB. The picture is of so low resolution that it isn't possible to make out any details.

1.1.5  Higher-throughput primer check: trying for better gels

Tue Oct 4 12:57:55 EDT 2005
Going to try running the lower-throughput 10-well gels with the post stain as was optimized in section 1.1.2. Samples will be prepared as before using multichannel pipettors, but will be loaded with a single-channel pipettor, as the gel-wells don't line up.
Please see the pdf version for figures
Figure 1.5: 70 ml, 2% agarose gel run for 45 min at 120 volts, poststained in 100 ml of H2O with 0.5 ug/ml ethidium bromide for 45 min followed by 20 min in water alone (both on orbital shaker at 50rpm) to reduce background Product sizes for the genes should be: lexA: 197, dinD: 162, dinG: 128, dinI: 101, dinP: 201, ruvA: 106, yceP: 111, yebG: 141, entC, recA: 153, fliF: 156, fliL: 151, nikA: 101, flgK: 150, flgM: 165, fliC: 216. 10 ml of Fisher BioReagents exACTGene Low Range DNA Ladder, with bands:weights(ng) of 2000:105, 1500:87, 1000:68, 750:59, 500:94, 300:27, 150:34, 50:25 was used.
Please see the pdf version for figures
Figure 1.6: 70 ml, 2% agarose gel run for 38 min at 120 volts, poststained in 100 ml of H2O with 0.5 ug/ml ethidium bromide for 45 min followed by 20 min in water alone (both on orbital shaker at 50rpm) to reduce background. This gel sucks, wasn't poured smoothly or melted properly. 10 ml of Fisher BioReagents exACTGene Low Range DNA Ladder, with bands:weights(ng) of 2000:105, 1500:87, 1000:68, 750:59, 500:94, 300:27, 150:34, 50:25 was used.
Please see the pdf version for figures
Figure 1.7: 70 ml, 2% agarose gel run for 37 min at 120 volts, poststained in 100 ml of H2O with 0.5 ug/ml ethidium bromide for 45 min followed by 20 min in water alone (both on orbital shaker at 50rpm) to reduce background. Product sizes for the genes should be: fecI: 111, cirA: 240, entC: 175, exbB: 241, fepA: 195, fhuA: 199, nrdH: 162, tonB: 123, serA: 159, recA: 153, fliF: 156, fecA: 153. 10 ml of Fisher BioReagents exACTGene Low Range DNA Ladder, with bands:weights(ng) of 2000:105, 1500:87, 1000:68, 750:59, 500:94, 300:27, 150:34, 50:25 was used.
Please see the pdf version for figures
Figure 1.8: 70 ml, 2% agarose gel run for 34 min at 120 volts, poststained in 100 ml of H2O with 0.5 ug/ml ethidium bromide for 45 min followed by 20 min in water alone (both on orbital shaker at 50rpm) to reduce background. Product sizes for the genes should be: livK: 152, aroG: 173, aroL: 161, aroP: 122, codB: 206, cysC: 174, cysK: 117, dppB: 170, ilvC: 132, lysC: 180, metA: 104, metE: 158, metF: 187, dapB: 101, dapD: 117. 10 ml of Fisher BioReagents exACTGene Low Range DNA Ladder, with bands:weights(ng) of 2000:105, 1500:87, 1000:68, 750:59, 500:94, 300:27, 150:34, 50:25 was used.
Brief Conclusions:   All in all the results aren't too bad. Annoyingly, there is a small amount of signal around the 30-50bp region that looks like primers are amplifying nonspecific regions as well as the correct one. This could be the result of many things including:
  1. it's an artifact the 2-step PCR
  2. Tm needs to be raised because primer3 incorrectly estimates melting temperature
  3. bad primers
  4. amplification from whole cells is messy
Hopefully, the last is correct. For sure dumping a little overnite culture into your PCR master mix isn't as good as using pure DNA to amplify from. We should try a subset of the genes again on the cleaned up, sheared whole-cell extract before immunoprecipation to see if they are likely to influence our qPCR results (which uses a 2-step process). fhuA, flgK, fliC, and cysC should be repeated to see if their faint bands are the result of pipetting error or maybe those primers are no good. The aroP fragment looks too short and should be run again as well.

1.2  Testing the shearing range

It is desireable to have the range of the chromatin-fixed, sheared DNA to be between 200 and 1000bp. It is also important to know the approximate amount of pre-precipitation starting DNA. To test this we need to perform the initial steps ChIP reaction:
  1. from a 1:50 dilution grow cells in a flask to OD 0.6.
  2. fix cells with 1% formaldehyde solution
  3. wash cells 2x with PBS (to remove formaldehyde)
  4. lyse cells (with lysozyme or readylyse)
  5. shear DNA by sonication
Typically the above precedure is followed by immunoprecipitation. But we can take the lysate, reverse the crosslinks, and do a phenol:chloroform purification and ethanol precipitation to quantify the amount of starting DNA and the size range. It is probably also a good idea to preceed the phenol:choloroform reaction with an RNAase digestion or else the gel will contain RNA too making it difficult to determine the size range.

1.2.1  First shear test

Tue Oct 11, 2005
  1. Growing two strains (lexA:A, and lrp-myc:B) in a 50ml of LB in a 250ml culture flask and a 250ml baffled flask (4 samples total). Cultures started with a 1:100 dilution from an overnite culture. All cultures contain 50 ml Ampicillin (100mg/ml stock solution).
  2. After 1hr, 25 ml IPTG (1M stock solution) was added to each culture to induce production of the transcription factor
  3. After an additional 2hr 15min, two 10ml samples were taken from each flask and placed into 15ml Falcon centrifuge tubes for crosslinking (8 samples total named lexN1, lexN2, etc using the basename below). Also, 300 ml was taken from each culture to take an OD reading:
    Strain Flask OD 600 basename
    lexA:A culture 0.26 lexN
    lexA:A baffled 0.27 lexB
    lrp-myc:B culture 0.22 lrpN
    lrp-myc:B baffled 0.18 lrpB
  4. 280 ml of 37% formaldehyde was added to each 10ml sample and mixed by inversion. Sample was incubated at RT for 10min
  5. cells were pelleted at 2900g for 8 min and washed 2x with PBS
  6. washed pellets were lysed for 30min at 37C in 500 ml of Pallson lysis buffer (no-shaking)
  7. 500 ml of 2x Pallson IP buffer with 1 ml of RNAaseA (1 mg/ml stock [Ambion])was added and the tubes were placed in a shaking incubator (300rpm) for 10min yielding a very clear lysate
  8. the 1ml lysate was transferred to a 1.5ml eppy tube
  9. lysed samples were sonicated with a Branson Sonicator at 3 x 20 sec with power 1.5; samples lexB2 and lrpB2 (the second samples from the baffled flask) were sonicated a 4th time
  10. sonicated lysates were spun at top speed (13,200 rpm) for 5 min at 4C to pellet cellular debris (there was very little)
  11. supernants were transferred to new eppy tubes
  12. 2.5 ml of Proteinase K (20 mg/ml stock [Ambion]) was added to each tube and they were put in a 65C heat-block overnite to remove cross-links
ERROR: Palsson 2x IP buffer was made with 8% Triton X-100 not the correct 4%
More precise times:
Tue Oct 11 11:57:33 EDT 2005 in incubator
Tue Oct 11 12:50 EDT 2005 added 25mM IPTG
Tue Oct 11 2:15 EDT 2005 out incubator
Tue Oct 11 19:11:20 EDT 2005 put sheared chromatin-DNA in 65C heat-block to remove cross-links.
Wed Oct 12 10:42:47 EDT 2005 removed from heat-block
Wed Oct 12 14:28:55 EDT 2005 removed from -85C freezer to centrifuge
Wed Oct 12 18:43:51 EDT 2005 running gel with sheared DNA to check size range
DNA was phenol:chloroform extracted and ethanol precipitated using the method of Barker. The Ethanol/DNA 2:1 mix was placed in the -85C for 60min followed by centrifugation for 20 min at 4C.

DNA quantification

Thu Oct 12, 2005
Each of the eight samples was quantified using the NanoDrop. Different amounts of lysate (200ml or 100ml of the total 1ml) were used to see if things scaled linearly. The sample nomenclature is as follows [samplename : amount of lysate used for extraction].
Sample DNA (ng/ml ) 260/280 260/230 total DNA (ug) 1
lexN1:200 364.0 182.0
lexN2:200 327.5 163.8
lexN2:100 134.9 134.9
lexB1:200 283.1 141.6
lexB2:200 285.1 142.6
lrpN1:200 203.4 101.7
lrpN2:200 205.3 102.7
lrpB1:200 135.6 67.8
lrpB2:200 156.9 78.5
lrpB2:100 74.5 74.5
Figure 1.9: amount of sheared DNA obtained from lysate as a function of the OD of the culture it was extracted from.
Brief Conclusions:   It looks like the sheared DNA yield is a function of the cell density of the culture it was taken from (see Figure 1.9) [not surprising] and better yields are obtained when extracting larger fractions of culture (probably because it's easier to get more of the aqueous from the phenol:aqueous interface).

DNA shearing range

1ml and 2ml of each samples lexN2:200, lexN2:100, lrpB2:200, and lrpB2:100 were run on an agarose gel.
Please see the pdf version for figures
Figure 1.10: 70 ml, 1.5% agarose gel run for 48 min at 120 volts, poststained in 100 ml of H2O with 0.5 ug/ml ethidium bromide for 32 min followed by 10 min in water alone (both on orbital shaker at 50rpm) to reduce background. 10 ml of Fisher BioReagents exACTGene 1KB Plus DNA Ladder, with bands:weights(ng) of 10000:43, 8000:31, 6000:28, 5000:41, 4000:36, 3000:31, 2500:26, 2000:23, 1500:22, 1000:35, 900:26, 800:52, 700:31, 600:18, 500:35, 400:12, 300:17, 200:12, 100:7 was used. Desired shearing range is 200-300bp.
Brief Conclusions:   It looks like the shearing range is too high. This result could be due to one of two problems: 1 we didn't shear long enough or on high enough power, 2 there is a lot of RNA masking the DNA signal. We can do an RNAse digest on the DNA to see if that cleans it up to test hypothesis 2.

RNAse digesting sheared DNA

Thu Oct 13 18:37:19 EDT 2005 lexAN2:200 and lrpB2:200 will be RNAse digested to remove any contaminating RNA.
The RNAseA / DNA mixture was purified using a Qiagen PCR cleanup kit. DNA was resuspended in 50ml , roughly half the original volume. Cleaned up yields were:
Sample DNA (ng/ml ) 260/280 260/230
lexAN2:100 55.8
lrpB2:200 49.3
Please see the pdf version for figures
Figure 1.11: 70 ml, 1.5% agarose gel run for 48 min at 120 volts, poststained in 100 ml of H2O with 0.5 ug/ml ethidium bromide for 32 min followed by 10 min in water alone (both on orbital shaker at 50rpm) to reduce background. 10 ml of Fisher BioReagents exACTGene 1KB Plus DNA Ladder, with bands:weights(ng) of 10000:43, 8000:31, 6000:28, 5000:41, 4000:36, 3000:31, 2500:26, 2000:23, 1500:22, 1000:35, 900:26, 800:52, 700:31, 600:18, 500:35, 400:12, 300:17, 200:12, 100:7 was used. Desired shearing range is 200-300bp.
Brief Conclusions:   The Nanodrop spec shows a cleaner DNA as the 260/280 is closer to the more ideal 1.8, but the yield is about half its pre-RNAse digested value (so either there was a LOT of RNA or the Qiagen columns are very inefficient). Also, the gel from the RNAse digested samples looks cleaner, but the sheared range is still a little on the high side. Perhaps the lrpB2, which was sheared 20 extra seconds is a little smaller average size but not that much. Next round we should do at least that much shearing. Also, the RNAse digest was done with an Ambion RNAse A enzyme. I have purchase an Ambion RNAse cocktail that should chop up the RNA to a smaller size range next time in case there was a little left. Not sure why the ladder looks bad. I changed out the buffer for the next gel in case it was that.

Primer testing on sheared DNA

PCR: Thur Oct 13, 2005
Doing a quick positive control / sanity check to make sure the primers tested earlier can amplify the sheared DNA. Sheared DNA was from RNAse digested lrpB2:200. 1.3mg was used to amplify all 44 tested targets (I just pipetted it into the multichannel reservior).
Gel: Fri Oct 14 19:05:27 EDT 2005
Please see the pdf version for figures
Figure 1.12: 70 ml, 2% agarose gel run for 38 min at 120 volts, poststained in 100 ml of H2O with 0.5 ug/ml ethidium bromide for 40 min followed by 20 min in water alone (both on orbital shaker at 50rpm) to reduce background. 10 ml of Fisher BioReagents exACTGene Low Range DNA Ladder, with bands:weights(ng) of 2000:105, 1500:87, 1000:68, 750:59, 500:94, 300:27, 150:34, 50:25 was used. Product sizes for the genes should be: livK: 152, aroG: 173, aroL: 161, aroP: 122, codB: 206, cysC: 174, cysK: 117, dppB: 170.
Brief Conclusions:   It is clear when comparing Figure 1.12 to those from section 1.1.3 that the PCR from cleaned up sheared DNA is much more effecient and cleaner than when I dumped in 10ml of overnite to amplify from. The only question this figure raises is: What happened with codB? Earlier cysC was tentative, but here this gene amplifies cleanly. Now codB doesn't show up at all. The best part of this figure is there are no fuzzy bands around the 50bp point as was consistently found when amplifying from whole cells. Perhaps sheared DNA should be used when cloning in the any future TOPO constructs? Also, it should be noted that this DNA was cleaned 2x, once with phenol:chloroform extraction and once by a Qiagen PCR purification kit.

1.3  Checking the cloned TOPO constructs

Two Invitrogen TOPO vectors were used for cloning transcription factors to add tags to the C terminal or N terminal. The pTrcHis vector adds N-terminal 6xHis and Xpress epitopes. While the pTrcHis2 add C-terminal c-myc and 6xHis epitopes. We must sequence the vectors to confirm there are no PCR errors and that the gene is inserted in frame. All of the vectors were previously verified by restriction digest, but the digest will be repeated prior to sequencing to make certain all freezer stocks were labeled properly. Sequences will be send with primer and template mixed.

1.3.1  DNA Sequencing

Sequencing is to be done using the Agencourt QuickLane Sequencing service. Required concentrations are:

Amount of Template Amount of Primer Total Volume
0.6-1.0ug 20pmol 40ml
The concentration of primer roughly corresponds to 1ml of the primers supplied with the Invitrogen kit. For the pTrcHis vector we are using the Xpress forward primer. For the pTrcHis2 sequences we are using the pTrcHis Reverse priming site.
Minipreps were performed using an eppendorf kit on overnite cultures from the following plasmids: fec-myc, lrp-myc, lrp, fliA, lexA. Yields were determined with the Nanodrop and are as follows (additionally the amount of plasmid needed for a 600ng DNA sequencing reaction are included):
Gene Amount of Vector (ng/ml) 260/280 260/230 ml for 600 ng
fecI-myc 32.3 18.57
lrp-myc 31.3 19.17
lrp 36.7 16.35
fliA 26.3 22.81
lexA 39.9 15.04
Restriction digests to reconfirm insert size are being run before sending the constructs for sequencing. The following reactions are being performed:
lrp, fliA, lexA
Restriction Digest
vector/DNA 9 ml
EcoRI buffer 2 ml
BSA 2 ml
EcoRI enzyme 0.5 ml
BamHI enzyme 0.5 ml
H2O 6 ml
lrp-myc, fecI-myc
Restriction Digest
vector/DNA 9 ml
EcoRI buffer 2 ml
BSA 0 ml
EcoRI enzyme 0.5 ml
NcoI enzyme 0.5 ml
H2O 8 ml
Digests placed in 37C incubator at: Mon Oct 10 15:07:14 EDT 2005
out at: Mon Oct 10 16:07 EDT 2005
As seen in Figure , the inserts are all the appropriate sizes and will be sent for sequencing.
Please see the pdf version for figures
Figure 1.13: 70 ml, 2% agarose gel with 0.75 ul of 1% ethidium bromide run for 40 min at 120 volts. 10 ml of Fisher BioReagents exACTGene 1KB Plus DNA Ladder, with bands:weights(ng) of 10000:43, 8000:31, 6000:28, 5000:41, 4000:36, 3000:31, 2500:26, 2000:23, 1500:22, 1000:35, 900:26, 800:52, 700:31, 600:18, 500:35, 400:12, 300:17, 200:12, 100:7 was used. 20ul (all) of each digest was used. Product sizes for the 8 genes should be: fecI-myc=521bp, fliA=719bp, lexA=608bp, lrp=494bp, lrp-myc=494bp.
Vectors were picked up by Agencourt courier for sequencing on: Tue Oct 11 11:27:49 EDT 2005 with project name: ChIP TF clones
4 of 5 sequences passed the Agencourt QC (800bp Phred 20), fecI didn't and is being resequenced
Please see the pdf version for figures
Figure 1.14: Where the finger points should be a T.







Notice the error at bp 758!
Brief Conclusions:   As can be seen from the unfortunate sequences above, there are MANY errors. It looks like I need to use a proofreading Taq, I ordered Easy-A High-Fidelity PCR Cloning Enzyme and Master Mix designed specially for TOPO vectors because it adds the extra A on the end and uses a Pfu proofreading Taq. Supposedly has an error rate of less than 1 in 10 vectors. Hopefully, that will fix this problem.
One sequence lrp on the pTrcHis vector is ok. The fliA vector was inserted backwards. The other lrp sequence on the pTrcHis2 strain only has one mutation which converts a lysine to an arginine. These are similar amino acids both with positively charged R-groups so it might not matter, and it'd be interesting to see if it does.
The lexA vector is almost fine; there are no PCR errors, but unfortunately there is a human error. I removed the LAST THREE CODONS from the primer AHHHHHHHHHH!!!!! I hate human errors.
Last, something is screwy with the MiniPrep kit from eppendorf, I've been getting very low yields ( 30ng/ml ). I will try the Qiagen kit, but last time I used this kit the results had a white precipitate that bothered me.
This weekend I need to fix the miniprep problem (trying both eppy and qiagen) so I can send all the vectors I've cloned out to see if there is anything useful in there before I start cloning.

1.3.2  Do any strains not have errors?

Sun Oct 16 14:51:03 EDT 2005
Only one in five clones from the previous round was error-free (though lexA was caused by a human error and one of the lrp's only have a single mutation. I have many previously cloned genes left, so I'm minipreping all I can find to try and see if any are usable and how much more cloning I need to perform.
gadX:A, gadX-myc:A, fecI:A, pdhR:A, pdhR-myc:A, flhC:A, yheO:A, sfsA:A, yhiF:A, hyaC:A, rhaR:A, nac:A, cbl:A, bolA:A were all grown overnite from freezer stocks in 6ml of LB. 3 genes I thought I previously cloned were not in the freezer box (crp, csgD, dctR) 2 One clone, gadX-myc:A didn't grow. This left 13 cultures to be minipreped and sequenced.
All 13 genes were prepped with an eppendorf kit. In addition, 4 genes (bolA, gadX, hyaC, nac) were prepped with the QIAprep spin kit to see how the kits compare (these are shown in italics below).
Gene Amount of Vector (ng/ml) 260/280 260/230 ml for 600 ng
bolA 27.8 21.6
bolA 52.6 11.4
cbl 33.4 18.0
fecI 28.3 21.2
flhC 39.1 15.3
gadX 35.2 17.0
gadX 28.3 21.2
hyaC 33.3 18.0
hyaC 46.8 12.8
nac 30.7 19.5
nac 44.5 13.5
pdhR 23.1 25.8
pdhR-myc 21.0 28.6
rhaR 33.9 17.7
sfsA 33.3 18.0
yheO 34.7 17.3
yhiF 28.5 21.0
Fourteen samples were sent to Agencourt for sequencing. The bolA, hyaC, and nac plasmids were from the Qiagen kit. bolA was also sent from the Eppy miniprep for comparison (and to check out Agencourt's consistency).
Mon Oct 17 15:51:28 EDT 2005
All samples were digested to reverify insert size (this is mainly because I lost the original image). hyaC was mistakenly taken from the eppy clean up, but everything else is exactly the same as for the sequencing with regards to Eppy and Qiagen. The following protocol was used:
pdhR-myc
Restriction Digest
vector/DNA 8 ml
EcoRI buffer 2 ml
BSA 2 ml
EcoRI enzyme 0.4 ml
BamHI enzyme 0.4 ml
H2O 7 ml
all other vectors
Restriction Digest
vector/DNA 8 ml
EcoRI buffer 2 ml
BSA 0 ml
EcoRI enzyme 0.4 ml
NcoI enzyme 0.4 ml
H2O 9 ml
Please see the pdf version for figures
Figure 1.15: 70 ml, 1.5% agarose gel run for 40 min at 120 volts, poststained in 100 ml of H2O with 0.5 ug/ml ethidium bromide for 45 min followed by 20 min in water alone (both on orbital shaker at 50rpm) to reduce background. 10 ml of Fisher BioReagents exACTGene 1KB Plus DNA Ladder, with bands:weights(ng) of 10000:43, 8000:31, 6000:28, 5000:41, 4000:36, 3000:31, 2500:26, 2000:23, 1500:22, 1000:35, 900:26, 800:52, 700:31, 600:18, 500:35, 400:12, 300:17, 200:12, 100:7 was used. 20ul (all) of each digest was used in 40:1 6x fisher dye. Product sizes for the genes should be: bolA=351bp, cbl=951bp, fec=521bp, flhC=579bp, gadX=825bp, hyaC=708bp, nac=917bp, pdhR=764bp, rhaR=939bp, sfsA=705bp, yheO=735bp, yhiF(dctR)=531bp.
Vectors were picked up by Agencourt courier for sequencing on: Mon Oct 17, 2005
Sequences:





















Brief Conclusions:   The Qiagen prep definitely outperformed the Eppy kit. The yield was higher and the purity was better. Also, the Qiagen columns can use up to 5ml of culture, while the Eppy kit recommends 3ml max, so Qiagen's yield should be higher if I use more than 3ml of culture next time.
The sequencing results are horrible of the 19 sequencing reactions on 14 vectors, only 1 came back error-free. One error was my fault because I designed the primer wrong. Hopefully the next round will see improved results.

1.3.3  Recloning TOPO constructs

Out of the 15 or so sequences sent to Agencourt, all but one had errors. I'm going to reclone them using Easy-A High-Fidelity PCR Master Mix. Hopefully, the error rate will go down considerably. I'm going to prepare genomic DNA to PCR from this time, which should give a cleaner PCR.
Plated colonies from freezer stock: Mon Oct 31 10:45AM Need to make 40 LB:glucose:agar plates with amp according to the TOPO kit instructions.

Checking the imaging system on large gels

Mon Oct 27, 2005 In order to clone in higher-throughput I need to get the larger gel system working. With the CAB imaging system the resolution it was impossible to get high enough resolution for a good image. Plus, my loading dye was too concentrated and created a huge visual obstruction on the gel (see Fig 1.4).
Here I ran 2.5ml of each PCR product using 1/10 fisher dye in a 10ml total volume. Gel was increased to 300ml to make it thicker and more loadable.
Please see the pdf version for figures
Figure 1.16: 300 ml, 1% agarose gel with 1.5 ul of 1% ethidium bromide run for 38 min at 120 volts. 2.5ml of each PCR sample was used.
Brief Conclusions:   I'm still not totally satisfied with the look of the gel. For the TOPO system the framents will be longer which will make identification easier, but the resolution is still a bit crappy. I need to try using the zoom lense to see if this helps, as the bands themselves really are quite small. Perhaps optics can help. Next time I should move too the 1/40 fisher dye as this 1/10 is still slightly in the way. Also, I made samples on parafilm. This is tricky with the multichannel.

Transfering oligos to 96-well plate

Wed Nov 2 11:08:18 EST 2005
Cloning will be done from plates. This sample size 36 is right on the border where I could do it by hand or by plate. Since every other step is on plates, I'm going to put in the extra effort to move the oligos to plates. This will make things easier to scale up later should I need to. The contents of the plate are shown in Table
TOPO TF cloning oligo plate
- 1 2 3 4 5 6 7 8 9 10 11 12
A bolA cbl fecI flhC gadX hyaC lexA* nac pdhR rhaR sfsA yheO
B yhiF glcC nusA yhiW ydaK gfp* - bolA-m cbl-m fecI-m flhC-m gadX-m
C hyaC-m lexA-m* lrp-m nac-m pdhR-m rhaR-m sfsA-m yheO-m yhiF-m gfp-m fliA fliA-m
D - - - - - - - - - - - -
E - - - - - - - - - - - -
F - - - - - - - - - - - -
G - - - - - - - - - - - -
H - - - - - - - - - - - -
Table 1.2: Oligo concentration is 4mM. Wells contain both a forward and reverse primer. lexA contains the corrected lexA reverse primer. gfp requires a different template (since it is not in genome it must be amplified from plasmid). Oligo names with -m indicate the myc version of the primer that removes the stop codon (necessary for proper production of the myc tag). bolA-m didn't make it into the final plate because I forgot to order the primer.

Preparing Genomic DNA

Wed Nov 2 11:14:22 EST 2005
A plate of MG1655 was grown 9hrs and placed in the refridgerator for 1 day (because I had jury duty on Nov 1, 2005). Two colonies were inoculated at 10AM, Nov 2 into 4ml of LB. Genomic DNA will be prepared as described in using 4ml of sample. Yields are:
Gene Amount of Vector (ng/ml) 260/280 260/230 ml for 75 ng
genomicA 678.2 0.1106
genomicB 444.7 0.1687
Brief Conclusions:   It should be noted that no RNAse digestion was performed, some of the yield is RNA.

PCR

PCRs will be performed on all 33 genes in Table 1.2 using the following reaction:
PCR Reaction composition
H2O 20 ml
Easy-A Master Mix 25 ml
Forward and reverse primer0.4 mM
template DNA 75 ng
Thermal cycler conditions
Initial denaturation5 min 95° C
3-Step cycling
Denature:30 sec 95° C
Anneal:30 sec 58° C
Extend:30 sec 72° C
Number of Cycles:28
Final Extention: 25 min 72° C
Please see the pdf version for figures
Figure 1.17: 300 ml, 1.5% agarose gel with 1.5 ul of 1% ethidium bromide run for 45 min at 120 volts. 5ml of each PCR sample was used. Sizes for the genes should be: bolA=351bp, cbl=951bp, fecI=521bp, fliA=719bp, flhC=579bp, gadX=825bp, hyaC=708bp, lexA=608bp, lrp=494bp, nac=917bp, pdhR=764bp, rhaR=939bp, sfsA=705bp, yheO=735bp, yhiF(dctR)=531bp, gfp=717bp, glcC=765bp, nusA=1488bp, yhiW=729bp, ydaK=909bp
Enough master mix for 39 reactions (3 extra) will be prepared, gfp will have additional template added by hand but will also contain genomic DNA. genomicA DNA was sticky (probably because of the high DNA conc?), so I'm going to use genomicB.
Brief Conclusions:   There was just enough master mix in reservoir using 3 extra. These giant gels should really be run a little longer. All the PCRs looked great except the lexA which has a second band for some reason. Also, yheO-myc doesn't appear to have amplified.

Cloning

Thur Nov 3, 2005
0.75ml of each PCR product will be cloned into the appropriate pTrsHis TOPO vector according to the manual. Two concentrations of TOPO/insert mix are plated onto LB:agar:glucose:amp plates.
I didn't have enough pTrcHis2 vector for all of the genes I PCR'd, so yhiE and fliA were not cloned into the myc-tagged vector.

Plasmid prep

Fri Nov 4 17:34:17 EST 2005
3 colonies are chosen from each 96 total for miniprep (only two were chosen for yheO-myc. The matrix of picked colonies is:
Picked TOPO Colonies for Minipreping
- 1 2 3 4 5 6 7 8 9 10 11 12
A pdhR Laf pdhR Lb pdhR Ha gadX La gadX LbfgadX Ha rhaR La rhaR Ha rhaR Hb fliA La fliA Lb fliA Ha
B sfsA La sfsA Lb sfsA Ha yheO La yheO Lb yheO Haf gfp Laf gfp Lb gfp Ha yhiF La yhiF Lbf yhiF Ha
C hyaC La hyaC Lb hyaC Haf bolA La bolA Lb bolA Ha cbl La cbl Lbf cbl Haf fecI Ha fecI Hb fecI Hg
D flhC La flhC Lb flhC Haf yhiW Laf yhiW Lb yhiW Ha nac La nac Lb nac Ha lexA La lexA Haf lexA Hb
E glcC La glcC Lb glcC Haf ydaK Laf ydaK Lb ydaK Ha nusA La nusA Lb nusA Ha cbl-m La cbl-m Lb cbl-m Haf
F hyaC-m La hyaC-m Lb hyaC-m Ha fecI-m La fecI-m Lb fecI-m Haf lexA-m La lexA-m Lbf lexA-m Haf pdhR-m La pdhR-m Lb pdhR-m Ha
G sfsA-m La sfsA-m Lb sfsA-m Ha nac-m La nac-m Lb nac-m Ha rhaR-m La rhaR-m Lb rhaR-m Ha lrp-m La lrp-m Lb lrp-m Ha
H gfp-m Laf gfp-m Lb gfp-m Ha gadX-m La gadX-m Lb gadX-m Ha flhC-m La flhC-m Lb flhC-m Ha yheO-m Haf yheO-m Hbf -
Table 1.3: L = low conc. plate (plated 10ml of ligation reaction), H = high conc. plate (35ml of ligation), f = incorrect/no insert in digest (see Figure )
It took 3 hrs to pick all those colonies, and if I had had half a brain (as I will if I ever repeat this). I would have taken a ml of each 1200ml sample and added it to another plate to make a freezer stock from. Alas, I'm too stupid this time; next time I won't be. Random samples were taken to determine how much of each sample to use for digests to check insert size. H12 was a negative control containing only LB, and its yield is the lowest of the bunch in the table below. When I spun down the plate prior to lysing, it was the only well with no colonies. The samples yields are:
Sample DNA (ng/ml ) 260/280 260/230
C10 30.2 1.84 1.38
G1 32.3 1.77 1.40
E6 20.7 1.62 1.43
H10 20.5 1.79 1.38
B9 15.4 1.56 1.00
H5 14.9 1.67 1.47
H12 7.7 1.46 0.65
Brief Conclusions:   I noticed later that, either my samples evaporated or spilled out randomly from the overnight or I wasn't consistent with my pipetting of the 1.2ml sample in each well. I'd really prefer something that would allow 2ml of sample. I should use the air permeable tape next time and the maximum 1.3ml sample if not a different plate that allows 2ml. Also, I didn't have the right attachment to the vacuum to allow me to use the Qiagen TurboFilter plate properly. I've ordered one.

Insert check

Mon Nov 7, 2005
8 ml of each vector was digested to check for an insert of the correct length using the following reactions:
pTrcHis
Restriction Digest
vector/DNA 8 ml
EcoRI buffer 2 ml
BSA 2 ml
EcoRI enzyme 0.4 ml
NcoI enzyme 0.4 ml
H2O 7 ml
pTrcHis2
Restriction Digest
vector/DNA 8 ml
EcoRI buffer 2 ml
BSA 2 ml
EcoRI enzyme 0.4 ml
BamHI enzyme 0.4 ml
H2O 7 ml
I got these parameters from an earlier digest (see section 1.3.2). Unfortunately they are backwards. The NcoI enzyme works on pTrcHis, it is just expensive and makes the insert 100bp longer. The BamHI doesn't have a cutter site in pTrcHis2, so none of those insert checks worked, and I need to redo them.
Please see the pdf version for figures
Figure 1.18: 300 ml, 1.8% agarose gel with 1.5 ul of 1% ethidium bromide run for 48 min at 120 volts. 20ml of each PCR sample was used.
Tue Nov 8 20:06:40 EST 2005
I reran the 39 genes with a myc tag using the proper enzymes (NcoI and EcoRI); things are much better this time (see Figure ).
Please see the pdf version for figures
Figure 1.19: 300 ml, 1.8% agarose gel with 1.5 ul of 1% ethidium bromide run for 48 min at 120 volts. 20ml of each PCR sample was used.
Brief Conclusions:   In Figure 1.18, it is clear that none of the myc tagged genes are possible to check, because I used the wrong enzymes. However, most of the pTrcHis genes have the correct insert (see Table 1.3 for which ones). Also, 8ml seems sufficient even though the miniprep yield was poor.

Sequencing

Tue Nov 8 12:55:47 EST 2005
Twenty plasmids are being sent for initial sequencing. Others will be done if the first twenty look good. A few genes are being sequenced 2x (two different vectors, not he same vector 2x).
Sample DNA (ng/ml ) 260/280 260/230 ml for 600ng
A2 28.0 21.42
A3 23.6 25.42
A4 50.9 11.79
A6 33.9 17.70
A7 18.0 33.33
A10 13.9 43.17
A11 70.2 8.54
B1 31.3 19.17
B4 18.8 31.91
B8 13.6 44.11
B9 15.4 38.96
B10 17.1 35.09
C1 29.1 20.62
C4 15.2 39.47
C7 22.5 26.67
C10 30.2 19.87
C11 32.1 18.69
D1 25.7 23.35
E1 32.1 18.69
E5 22.8 26.32
The sequences with alignment to the known MG1655 sequence are:













Please see the pdf version for figures
Figure 1.20: hello
Brief Conclusions:   Nine of the twenty sequences made it through the first round of Agencourt sequencing. Seems pretty low, perhaps because I'm not good at the 96-well miniprep right and I didn't have the vacuum regulator yet. Two vectors were ok: pdhR Ha and hyaC La. Two in nine is pretty crappy, but much better than one in twenty or so before. Most of the problems are due to fragments being in backwards. There are very few PCR errors this time, so at least one problem is solved. Now I need a way to screen for those reverse guys or to sequence more plasmids per gene.
The beginning sequences of the two correct vectors appeared to be screwed up, containing a deletion each in the 5' upstream TOPO section (not the part I PCR'd). This seemed odd, so I checked it out for a lot of sequences and they all have similar problems (see alignment below). When I looked in more detail at the chromatagraphs however (see Figure 1.20) it was clear that the sequence was ok, it was just a bad phred call.
If I had this to do over, I'd sequence 2-3 of each. Now that I know the pfu Taq improved the error rate.
By gene summary:
fecI            --CTCTCATCCGCCAAAACAGCCAAGCTTCGAATTCGCCCTT- 40
gfp             TTCTCTCATCCGCCAAAACAGCCAAGCTTCGAATTCGCCCTT- 42
fliA1           ----------------------CAAGCTTCGAATTCGCCCTT- 20
fliA            -----------------------AAGCTTCGAATTCGCCCTT- 19
lrp             -------------------------------------CCCTT- 5
topo            -------GATCTGTACGACGATGACGATAAGGATCCAACCCTT 36
hyaC            --------------------ATGACGATAGGA--TCAACCCTT 21
pdhR            ---------------------TGACGATAAGGATCCACCCTT- 21

More Sequencing

Tue Nov 15 12:17:21 EST 2005
The low success rate from previous runs and the time-constraint to get this paper out means I gotta focus my effort on fewer TFs. I'm going to pick 8 colonies for the 4 most relevant genes from regulon given our current dataset (lexA, pdhR, fecI, fliA). LexA because we have so many chips run with antibiotics; fliA because the flagellar network has a huge response in many chips; pdhR because of Josh's experimental work on its interaction with the fec genes; fecI because all the algorithms find it as the main iron regulator. In addition two other genes outside regulon are being focused on: abgR (i.e. ydaK) and nusA.
From these eight colonies all found to have inserts will be sent for sequencing (up to five total). Let's hope this works. I'm using the Qiaprep 96-well miniprep again. With the new vacuum, I hope for better results...
Forty-four colonies were picked for the following genes: fecI, ydaK, lexA, nusA, fliA, pdhR, gfp (2-5 colonies per gene). Forty-two colonies grew. Twenty-four were prepped with a Qiagen centrifuge miniprep. The remaining were done in a 96-well plate, the plate contains the following:
96-well primer plate for ChIP-PCR
- 1 2 3 4 5 6 7 8 9 10 11 12
A lexA H4 lexA H5 y1 pdhR 1 nus3 pdh4 gfp2 pdh6 fli3 gfp3 gfp5 nus2
B gfp6 gfp4 pdh2 ydaK1 pdh7 gfp1 - - - - - -
C - - - - - - - - - - - -
D - - - - - - - - - - - -
E - - - - - - - - - - - -
F - - - - - - - - - - - -
G - - - - - - - - - - - -
H - - - - - - - - - - - -
Since the agencourt courier arrived faster than I could run my digest on a gel, I didn't get to select genes with an insert. I grabbed twenty-eight sequences and hoped for the best. In addition rhaR La from earlier was cloned into MG1655 and miniprepped. I included it for sequencing with a forward and reverse primer, so we can see if this long sequence is error-free.
The following sequences were sent (big thanks to Ilaria or this would never have been completed on time).
Sample DNA (ng/ml ) 260/280 260/230
D4 rhaR1 (F) 81.7
D5 rhaR1 (R) 81.7
D6 fec1 (F) 238.4
D7 fec2 (F) 57.4 - -
D8 fec3 (F) 214.6
D9 lexA1 (R) 87.4
D10 lexA2 (R) 94.6
D11 lexA3 (R) 46.9
D12 lexA4 (R) 120.3
E1 lexA5 (R) 89.1
E2 fliA1 (R) 96.5 - -
E3 fliA2 (R) 63.4
E4 fliA3 (R) 71.3
E5 fliA4 (R) 88.6
E6 fliA5 (R) 71.9
E7 pdhR1 (R) 76.0
E8 pdhR2 (R) 59.1
E9 pdhR4 (R) 66.1
E10 pdhR5 (R) 104.3
E11 pdhR7 (R) 41.0
E12 gfp1 (R) 88.3
F1 gfp2 (R) 80.2
F2 gfp3 (R) 81.7
F3 gfp4 (R) 76.1
F4 ydaK1 (R) 59.9
F5 ydaK2 (R) 117.3
F6 ydaK3 (R) 117.3
R = Reverse primer
F = forward primer
The digest (that I didn't use to pick the vectors for sequencing) can be seen in Figure .
Please see the pdf version for figures
Figure 1.21: 300 ml, 1.5% agarose gel with 4.5 ul of 1% ethidium bromide run for 40 min at 100 volts. 10 ml of Fisher BioReagents exACTGene 1KB Plus DNA Ladder, with bands:weights(ng) of 10000:43, 8000:31, 6000:28, 5000:41, 4000:36, 3000:31, 2500:26, 2000:23, 1500:22, 1000:35, 900:26, 800:52, 700:31, 600:18, 500:35, 400:12, 300:17, 200:12, 100:7 was used. 4.5ml of the centrifuge minipreps was used. 6.0ml of the 96-well plate minipreps was used. y=ydaK, nus=nusA, lex=lexA, fec=fecI

















Brief Conclusions:   Finally some good news on the cloning front!!!!! definitely correctly cloned genes: fecI, fliA, lexA, gfp, ydaK. genes the are most likely correct but need another resequencing effort to make sure: nusA and pdhR.

Brief Conclusions:   Wed Dec 14, 2005
Finally resequenced the pdhR from above that it looked like it was alright but the read wasn't very good.  It looks fine.

1.4  Cloning TOPO constructs into MG1655

Vectors for lrp, rhaR were originally cloned into TOP10 and will be transferred to MG1655 ATCC cells.
Competent cells were made according to the protocol on page section . They were digested to check for proper size post-tranformation (see Figure ).
Please see the pdf version for figures
Figure 1.22: 80 ml, 1.5% agarose gel with 0.4 ul of 1% ethidium bromide run for 40 min at 100 volts. 10 ml of Fisher BioReagents exACTGene 1KB Plus DNA Ladder, with bands:weights(ng) of 10000:43, 8000:31, 6000:28, 5000:41, 4000:36, 3000:31, 2500:26, 2000:23, 1500:22, 1000:35, 900:26, 800:52, 700:31, 600:18, 500:35, 400:12, 300:17, 200:12, 100:7 was used. 5ml of the centrifuge minipreps was used.
Mon Dec 5 18:23:28 EST 2005 Additional genes to be transformed into MG1655 are: fecI3, fliA2, nusA1, and ydaK4. 0.5ml of each was added to compentent cells on ice for appx 15min heat-shocked for 30 sec, added 450ml of LB, incubated with shaking at 37 for 45 min. 30ml of each was spread onto an ampicillin plate. nusA1 had no colonies. Others were grown and minipreped to make sure the insert was there.
Please see the pdf version for figures
Figure 1.23: 80 ml, 1.5% agarose gel with 0.4 ul of 1% ethidium bromide run for 40 min at 100 volts. 10 ml of Fisher BioReagents exACTGene 1KB Plus DNA Ladder, with bands:weights(ng) of 10000:43, 8000:31, 6000:28, 5000:41, 4000:36, 3000:31, 2500:26, 2000:23, 1500:22, 1000:35, 900:26, 800:52, 700:31, 600:18, 500:35, 400:12, 300:17, 200:12, 100:7 was used. 5ml of the centrifuge minipreps was used.
Also gfp4, lexA5, pdhR7 had been previously miniprepped and eluted into a 96-well tube format provided by qiagen and they all dried out in the fridge. So they were regrown from freezer stocks (prepared right before I did the minipreps) and miniprepped and transformed into MG1655.

1.5  ChIP optimizing parameters

1.5.1  Round1 with lrp:A only

Tue Oct 18 20:37:53 EDT 2005
Trying to figure out best OD to choose and further clarify the link between OD and DNA yield. Also trying to lessen the RNA contamination by using an RNAse cocktail and a pre-phenol:chloroform RNAse digestion for 30min at 37C. All strains are lrp:A, which is the only one that was verified to be error free by sequencing.
Running 4 samples in duplicate (8 total). Procedure begins as in 1.2.1. Two samples R3:A, R3:B were grown for 5hr 15min (R3). The other two samples, R2:A, R2:B, were grown for 3hr 45min (R4).
sample times:
Round3 (R3) in 1PM
Round2 (R2) in 2:30PM
Round1 in 4PM but I broke the flask (there was only going to be one sample of this)
all samples out at 6:15PM
Strain growth time OD 600 (-bkgrnd) avg(OD 600)
R3:A 5hr 15min 0.63, 0.66 0.64
R3:B 5hr 15min 0.59, 0.60 0.59
R2:A 3hr 45min 0.31, 0.32 0.32
R2:B 3hr 45min 0.32, 0.34 0.33
Two 10ml samples were taken from each 50ml sample to make 8 samples: R3:A1, R3:A2, R3:B1, R3:B2, R2:A1, R2:A2, R2:B1, R2:B2. All samples were crosslinked and sheared using the crosslink/shearing protocol in section . Sheared lysates were visualized on an agarose and can be seen in Figure .
Please see the pdf version for figures
Figure 1.24: 80 ml, 1% agarose gel with 0.5 ul of 1% ethidium bromide run for 40 min at 100 volts. 10 ml of Fisher BioReagents exACTGene 1KB Plus DNA Ladder, with bands:weights(ng) of 10000:43, 8000:31, 6000:28, 5000:41, 4000:36, 3000:31, 2500:26, 2000:23, 1500:22, 1000:35, 900:26, 800:52, 700:31, 600:18, 500:35, 400:12, 300:17, 200:12, 100:7 was used. 3ml of each sheared DNA sample was used. Samples were sonicated using a Branson 250 digital sonifier. R3A1 and R3A2 were sonicated 3x 30sec at 20% power. R3B1 and R3B2 were sonicated 4x 30sec at 20% power.
Brief Conclusions:   Sonication with the digital Branson is much cleaner looking. Some of this might be due to the ability of the newer gel imaging system to make prettier gels. The smears are quite consistent though.
qPCR   To reduce cost I ran only one of the samples (plus one sample with 4 replicates takes an entire plate). The following was for sample R3:A1. The plate was set up to include 4 pcr Replicates.
Reactions were performed in a 384-well Abi qPCR plate. For details see . The phenol:chloroform extracted DNA was resuspended in 100ml of 10mM Tris. This was diluted with 302ml of water to ease pipetting into the plate. 40ml was alliquoted into 10 pcr tubes to facilate multichanneling. 4ml of the 400ml total target DNA solution was used per well. This is equivalent to using 1ml of the 100ml resuspension per reaction.
The final qPCR reaction and order of addition to the plate qPCR reaction was:
  1. add 12 ml of master mix + water (10ml sybr green master mix, 2ml water) from reservoir
  2. add 4 ml of primer mixes (final conc. 800 pM)
  3. add 4 ml of dilute target DNA
lrp qPCR plate setup
- 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
A recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD blank - - - -
B recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD blank - - - -
C recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD blank - - - -
D recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD blank - - - -
E recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD blank - - - -
F recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD blank - - - -
G recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD blank - - - -
H recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD blank - - - -
I recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD blank - - - -
J recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD blank - - - -
K recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD blank - - - -
L recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD blank - - - -
M recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD blank - - - -
N recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD blank - - - -
O recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD blank - - - -
P recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD blank - - - -
Table 1.4: lrp:A1 qPCR plate setup. Yellow rows are + control template (sheared lysate DNA). Cyan rows are immunoprecipitated DNA with a specific antibody. Red are immunoprecipitated with an unrelated antibody. White are immunoprecipitated with no antibody.
Out of curiousity, I took a picture of the plate when the qPCR plate finished (see Figure ).
Please see the pdf version for figures
Figure 1.25: qPCR plate from the first round of chipping with lrp. Last 4 columns (black) were water. The little circles are water droplets on the tranparent cover formed when I jiggled the plate a little too much on my way to the imager.
The raw data (Ct values round to the nearest thousandth) can be seen in Table .
lrp R3:A3 ChIP results
- pCnt1 pCnt2 pCnt3 pCnt4 sAnti1 sAnti2 sAnti3 uAnti1 uAnti2 uAnti3 uAnti4 nAnti1 nAnti2 nAnti3 nAnti4
recA 14.006 14.039 14.011 13.987 25.139 24.945 24.897 25.314 25.479 25.293 25.593 32.077 30.309 30.347 30.390
cysK 14.173 14.283 14.241 15.692 26.478 26.254 26.131 26.918 26.998 26.815 25.585 32.440 32.793 32.148 33.056
entC 13.848 14.019 13.924 13.846 24.245 24.379 24.273 24.421 24.644 24.609 24.761 33.460 35.060 36.558 34.014
dppB 14.511 14.734 14.611 15.306 23.902 23.441 23.211 25.477 25.665 25.383 24.708 - - 38.019 -
fliF 14.520 15.029 14.640 14.556 26.447 26.713 27.126 26.885 27.448 27.433 27.598 - - 38.182 -
ilvC 13.609 13.696 15.696 13.637 23.847 23.983 23.791 24.172 24.304 24.213 23.375 36.641 31.985 38.575 35.788
serA 13.792 13.879 13.821 13.813 23.114 23.009 23.279 24.714 24.547 24.578 24.560 38.939 - 38.414 -
lysC 14.000 14.022 14.215 13.984 24.887 24.538 24.662 25.280 25.270 25.250 25.184 32.378 30.427 32.859 -
livK 13.704 13.956 13.768 13.770 20.212 20.159 20.233 23.743 23.963 23.509 23.861 36.320 38.059 39.246 -
metA 13.651 13.651 13.668 13.654 23.905 23.863 23.861 24.278 24.273 23.775 26.687 28.233 30.776 29.501 -
aroG 13.931 14.055 13.969 14.198 24.281 23.223 23.529 24.573 24.722 25.003 24.657 - - - -
metE 14.510 14.440 14.444 14.501 29.197 29.213 29.461 29.382 29.610 29.448 30.338 37.004 - - -
aroL 13.941 14.046 13.957 13.937 25.218 24.312 24.833 24.690 24.789 24.582 24.915 - - 38.248 -
metF 13.968 14.219 13.943 13.941 24.622 24.448 24.799 24.915 24.951 25.106 25.130 - - 39.012 -
aroP 13.939 14.037 13.972 14.007 24.156 23.905 24.146 24.433 24.417 24.314 24.547 35.690 36.375 - -
dapB 13.790 13.732 13.754 13.870 23.999 23.869 23.777 24.618 24.518 24.671 23.924 31.301 32.195 33.531 38.918
codB - - - - - 39.360 - - - - 39.936 - - - -
dapD 14.076 14.102 14.130 14.184 24.637 24.561 24.245 24.575 24.694 24.579 24.863 38.938 - - 36.849
cysC 14.662 14.803 14.821 15.070 16.640 16.873 16.227 16.573 17.005 16.727 16.780 19.960 20.215 19.621 19.222
Table 1.5: qPCR Ct values rounded to the nearest 1000th. pCnt = positive control, sheared lysate not precipitated; sAnti = target DNA, immunoprecipitated with the antibody specific to the TF of interest; uAnti = target DNA, immunoprecipitated with an unrelated antibody (myc for an XPress tagged protein), nAnti = beads only, no antibody control. All samples except the sAnti have four replicates. One of the sAnti replicates was lost by pipetting into the wrong hole.
The Ct values must be compared to determine a p-value for the enrichment and a a-value must be chosen to determine at what level genes are accepted as having had their promoters positively enriched when using the correct antibody relative to the unrelated antibody. A quick inspection of the no-antibody/beads-only columns reveals that all genes are significant relative to the no-antibody control so it is likely I'll remove this control from my experiments in the future to save time and cash.
Currently, I'm using a one-sided unpaired t-test or a one-sided paired t-test to determine p-value. In the end I should have more replicates than are presented here, so the t-tests will have a little more meaning than a 4 against 4 t-test. I'm thinking 3 sample replicates and 3 qPCR replicates = 9 total replicates will be sufficient.
gene paired t-test unpaired t-test
recA 0.0586 0.0047
cysK 0.3537 0.2513
entC 0.0449 0.0092
dppB 0.0033 9.3641×10-4
fliF 0.0129 0.0335
ilvC 0.4308 0.3038
serA 0.0023 4.4390×10-6
lysC 0.0140 8.0723×10-4
livK 2.6981×10-4 3.5336×10-7
metA 0.1852 0.1548
aroG 0.0658 0.0068
metE 0.0959 0.0970
aroL 0.5845 0.5697
metF 0.0332 0.0064
aroP 0.0068 0.0050
dapB 0.0577 0.0238
codB NA NA
dapD 0.4970 0.0928
cysC 0.3302 0.1809

From the above results it looks like with an unpaired t-test serA, livK, lysC, dppB, and aroP are significantly enriched. For the paired it appears that aroP would not be included. This interpretation is taken by looking at genes included because they were not likely lrp targets (i.e. negative control genes): recA, entC, fliF and looking at their p-values. You can see that there is often some enrichment even for these negative control genes, so we must be careful with our thresholds. Hopefully, adding more samples will tighten the p-values. The positive controls (known targets serA and livK) are very enriched for both types of t-test which is promising.
Going to repeat the qPCR reaction using R3:A2
Brief Conclusions:   Used 2ul of mg/ml glycogen and non-siliconized tube: much easier to see pellet. For the master mix from the reservoir, I added enough for around 5 extra reactions to account for the lose in the reservoir, this was definitely not enough, I was about 10 wells short. Next time I should probably dilute the target DNA less (e.g. use 3ml instead of 4ml ). To allow a little more water to be added to the reservoir. Also, should add extra for around 10 reactions (sucks cause this is expensive).

OD vs DNA yield

One monoclonal vs two monoclonal

To preclear or not to preclear

This step was not tested.

1.5.2  Round1 with lrp:A2

Thur Nov 10 17:47EST, 2005
A quick note, these tables showing where the genes are on the chip are actually wrong. They derive from the table on page pageref that shows the primer layout. Using that table you see all of the qPCR tables thus far presented show rows D and E concatenated to make 24-wells. But in a qPCR 384-well plate, the multichannel pipettor output spacing is such that only every-other well is filled by the pipettor. Thus two pipettes worth, slightly shifted make the entire 24-wells fill up. The tables should show D1,E1,D2,E2,etc... Staggering the two columns.
lrp qPCR plate setup: round 2
- 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
A recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD 5kbUP 1kbUP pntA gltB stpA
B recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD 5kbUP 1kbUP pntA gltB stpA
C recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD 5kbUP 1kbUP pntA gltB stpA
D recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD 5kbUP 1kbUP pntA gltB stpA
E recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD 5kbUP 1kbUP pntA gltB stpA
F recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD 5kbUP 1kbUP pntA gltB stpA
G recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD 5kbUP 1kbUP pntA gltB stpA
H recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD 5kbUP 1kbUP pntA gltB stpA
I recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD 5kbUP 1kbUP pntA gltB stpA
J recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD 5kbUP 1kbUP pntA gltB stpA
K recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD 5kbUP 1kbUP pntA gltB stpA
L recA entC fliF serA livK aroG aroL aroP codB cysC cysK dppB ilvC lysC metA metE metF dapB dapD 5kbUP 1kbUP pntA gltB stpA
M blank blank blank blank empty
N empty
O empty
P empty
Table 1.6: lrp:A1 qPCR plate setup. Yellow rows are + control template (sheared lysate DNA). Cyan rows are immunoprecipitated DNA with a specific antibody. Red are immunoprecipitated with an unrelated antibody. White are immunoprecipitated with no antibody. Blanks are template + master mix only, no primers. 5kUP and 1kbUP were added to see how good the shearing is. pntA is a new potential target from Boris' algorithm. gltB and stpA are new extra positive controls.

1.5.3  Round1 with lrp:B1

Fri Nov 11 17:53:52 EST 2005
This is one of the two samples that used both His and Xpress antibody. In addition these samples underwent a four shearings instead of three. The layout is identical to that from lrp:A2 except only 1.5ml of primer was used (300nM) instead of the normal 800nM, and there were not empty wells with primer only.

1.6  lrp ChIP in Minimal and Rich media in MG1655

In the first round of ChIP their was definitely enrichment for the positive controls, but not that many of the predictions were verified (around 5-9 I think). It was also noisy as everything enriched including the negative controls, so determining the true interactions was done by looking at things enriched compared with the negative controls. It appears that just too much lrp is binding everywhere.
This round I'm moving from 1mM IPTG to 100 mM. In addition all cultures have 0.5% glucose which should slow expression from this lac repressed promoter (according to the pTrcHis Invitrogen manual). The minimal media is also the condition where lrp is doing stuff, and hopefully a more diverse expression of different transcription factors will lower the background. Last, the plasmids were cloned in to MG1655 rather than the original TOP10 cloning strain. All samples were grown as 50ml cultures in 250ml baffled flasks with shaking at 300rpm at 37C.
Davis minimal media was used because it is easy to prepare and very minimal.

1.6.1  growing lysing shearing

First round on Saturday failed as I didn't realize how slow the culture would grow in minimal media when they have a vector.
Sat Nov 19 around 4PM inoculated two cultures lrp:J and lrp:K freezer stock into Davis minimal with 1% glucose
Sun Nov 20 19:40:00 EST 2005 inoculated 3 250ml baffled flasks with 1:100 dilution of lrp:K (OD 1.04)
Mon Nov 21 09:15ish overnite was too long :(.....
Mon Nov 21 09:30ish started growing from 1:100 dilution into another baffled flask
Mon Nov 21 15:24:44 EST 2005 incubated the LB/1% glucose 1:100 from overnite will grow both. At this time I also added the IPTG.
Mon Nov 21 5:29PM 2005 removed samples from incubator because minimal samples were getting overgrown
Two samples were taken from the initial 6 samples (lrpMa, lrpMb, lrpMc, lrpOa, lrpOb, lrpOc) and OD600 absorbances were take:
Strain OD 600
lrpMa 0.693
lrpMb 0.726
lrpMc 0.736
lrpOa 0.290
lrpOb 0.283
lrpOc 0.297
It's clear that the minimal strains (lrpM*) were grown more than the LB cultures (lrpO*), but I didn't want one of the cultures to have a lot more time in the prescence of IPTG, and thus much higher Lrp-tagged protein concentration. The twelve total samples were lysed and sheared according to the protocol starting on page .
The concentration of each sample was checked on the nanodrop:
Sample DNA (ng/ml ) 260/280 260/230 ml for 25 mg ml dilution buffer
lrpMa1 217.0 115.2 1037
lrpMa2 240.0 104.2 938
lrpMb1 241.0 103.7 933
lrpMb2 228.0 109.6 986
lrpMc1 238.7 104.7 942
lrpMc2 255.2 98.0 882
lrpOa1 202.6 123.4 1111
lrpOa2 180.8 138.3 1225
lrpOb1 188.2 132.8 1195
lrpOb2 211.7 118.1 1063
lrpOc1 216.4 115.5 1040
lrpOc2 164.2 152.3 1371
Approximately 600ng of all twelve samples (3ml ) was loaded onto an agarose gel (see Figure ).
Please see the pdf version for figures
Figure 1.26: 80 ml, 1.5% agarose gel run for 40 min at 120 volts, poststained in 100 ml of H2O with 0.5 ug/ml ethidium bromide for 40 min followed by 20 min in water alone (both on orbital shaker at 50rpm) to reduce background. Shearing range appears between 1000 and 100 basepairs.
Brief Conclusions:   Shearing range is beautiful; DNA yields are a little smaller than I'd like. I'm surprised how similar the LB and minimal media yields are given the OD was 2x higher for the minimal media. However, when spinning down the minimal not all cells were pelleted (the pellet was still much larger than for the LB cultures). I'll proceed to immunoprecipitation after the Turkey break...

1.6.2  Misc.

Negative control primers are going to be rechoosen randomly this time rather than the previous 5 that were chosen by picking targets of other proteins I'm testing. I choose the numbers by picking 20 random genes from the 4345 genes in the microarray and chosing the first 12 in order that were more than 5000bp from a predicted target of all of the TFs I'm testing and that were not hypothetical genes.

1.6.3  Bead washing

Thu Dec 1 21:55:48 EST 2005 precleared samples set to rotate Fri Dec 2 21:52:02 EST 2005 put washed complexes on heat block overnite to reverse crosslinks
I ran the 36 samples in two batches of 18. This is a LOT of work takes 5 hrs a batch. Makes for a very unpleasant day. During the breaks I did more research on how I can make this work in a 96-well format (written in section
The 12 original samples (each with A, B, C for correct antibody, incorrect antibody, and bead only control) were split into the following random blocks of six (with the A,B,C kept together for 18 samples in each block. The samples in the two sets where pick randomly to be: 4, 9, 5, 7, 3, 8; 6, 2, 10, 11, 1, 12. The numbers correspond to their placement in the DNA yield table above (i.e. 4 = lrpMb2, 9 = lrpOb1, etc...).

1.6.4  Ideas for higher-throughput and automation

Here's what I learned (most of this I already planned earlier).
The initial growth part is tricky. I grow 50 ml, take a 15 ml sample and end up with around 200mg of DNA, I need 50mg if I assume only one of the negative controls is important. Really 55mg would be nice so I can have a good positive control. With roughly 13.33mg DNA per ml I need around 5ml to make this work. Then I'd need to scale down the lysing procedure accordingly. 2 ml is the biggest 96-well block I know of. Perhaps they go up to 5 ml, but my BioHit pipettor doesn't....
(Small addition: Tue Jan 17 12:00:20 EST 2006 there are 48-well plates that use 5ml of samples; that's probably the perfect size. 48-cultures can be split into two correct antibody/incorrect antibody. The resulting 96 (i.e. 48 sonicated lysates split into two) will go into the 96-well plates. If that's not enough culture, there are 24-well 10ml plates, but that would involve setting up and distributing from many different plates.
Also would be very useful to buy a thermo-sealer AB-0384 $1500; need this too AB0724 a completely automated one costs 30K ouch! AB-0950)
For the shearing, it could be done with one of these bioruptors if this machine worked. http://www.diagenode.com/Research/Optimisation.php Costs around 12 grand. If it worked it would really cut down on the worst step. Machine is not set up for 96-well, but even as it is doing 12 samples in falcon tubes in 10 minutes with out my having to open and hold each tube brings tears to my eyes.
bead washing with 96-well costar filtrex plates: much faster! especially if I had a robot to load the 96-wells!
qPCR robots Corbett is cool but only has 96-well
interesting unrelated robot thing: genomic solutions has cool way to grow LOTS of cells biotek robot doesn't do what I want neither does this one ttp labtech
Lissy 2002 and bu has one cmld.bu.edu/instrumentation/lissy.html problem is it doesn't go that low in volume
the hummingbird looks just as good as the deerac looks good to but it is slow
deerac has by far the best one, but it only accommadates one extra plate to transfer to the production plate. Not sure if it can do things like A12 sample to H23 production. Wouldn't be too bad if it can handle 384 in both places. Equator HTS or GX pipettes down to 50nl!!!!! Smaller qPCR reactions could save a lot of dough. Plus doesn't need tips.
(Thu Jan 19 13:40:06 EST 2006 labchip 90 would be nice for operon prediction, we could run normal pcr's in 384 well plates, only need 150nl. Gives list of bands, so we could multiplex at different band sizes and get a moderately quantitative reading (certainly good enough for a yes, no answer). 384-well plates can resolve between a 140bp and a 210bp fragment. Also between a 400 and a 420bp fragment. So we might get 5-10 rxns per well. Total pcr rxns is 8690 (17380 primers $$$), could be done in 3, 384-well plates. Would take 2hrs for the PCR and 12 hrs for the subsequent gels. One gel can run 1800 rxns, so you could check the genomes operons 2x per gel.

1.6.5  qPCR

new primers were designed this time we have 12 random genes (rather than just genes from other targets that I guessed wouldn't be bound by the TF of interest). The random genes were picked using a few rules.
  1. generate a random list of genes (did in matlab from all 4345 genes I pick 150 or so).
  2. throw away genes that are putative (to be far we want to compare apples and apples, so we should make sure the protein is most likely a real one
  3. throw away things that are in the list of the top 40 targets or are within 5kb of one of those genes (by looking at the ecocyc genome browser)
Also, primers are now ordered in 96-well plates from IDT prenormalized to 100 mM. Primer pairs are placed consecutively by column (e.g. if the forward primer is in well C3 the reverse will be in D3), to make for creation of primer mixes. It's also cheaper and MUCH faster for me not to have to keep track of all that stuff now.
From now until further notice, all TFs will have their top 36 interactions tested (minus a few positive controls). Combined with the 12 random genes that's 48 targets per TF times the number of replicates (4-6 haven't decided yet) times the number of controls (4) for a total of 768-1152 qPCR rxns per TF. Yikes that's a lot. If I ever do this again I need to drop one of those controls as it'll reduce the number of rxns by a forth.

New list of lrp targets

The previous 36 targets have been altered. To be scientifically honest, I felt I shouldn't just remove a bunch of stuff to fill it with Boris' algorithms predictions and try again, especially since only two genes appeared to be new targets. I only removed one predicted target codB from the previous plate because those primers never worked, so I can say nothing about that gene plus or minus. Also, one new positive control serC was added.
The major change however was for the negative control genes. In order to get a better estimate of the background noise a true set of random genes was chosen. This time we have 12 genes rather than the previous 3-5 and they aren't just chosen from the list of primers I already had. All of the previous negative control primers have been removed. The 1kb and 5kb controls are still there, but will likely not be used for other TFs I check.
Specifically the genes removed were codB, recA, fliF, and entC. New targets are shown in italics in the new primer table (see Table ).
updated lrp target plate
- 1 2 3 4 5 6 7 8 9 10 11 12
A metN thrL yagU cvpA serC yhjE leuLABCD gdhA purC argI argA yebR
B argG nlp ompT purM ilvC metE metF metA serA livK lysC pntA
C dppB stpA aroG cysK aroL cysC aroP gltB dapD dapB 5kb 1kb
D gcl mog pinO idnD yhaF nhaA amiA goaG kdtB yagG citC fruK
Testing the primer plate  
The new primer plate needs to be tested to make sure all the primers work. Unfortunately, I'm having issues with this simple task.
Yesterday, Thur Dec 8 2005, I tried running the 48 PCRs in a plate rather than in PCR strips. No luck. I used a Costar Thermowell 96-well plate (part number 6551), which has the nice feature that it's easy to break it into sections (e.g. to use in our PCR machine which only has 48-well blocks. I sealed the plate with Costart Thermowell Sealers (part number 6524). Unfortunately, getting a good seal was not easy, especially since the plate was broken in half, leaving a proper edge on only one side. The bad seal led to the evaporation of most of the solution in the plate toward the bottom right. As you went towards the top left, it didn't look as bad. The results show that the bottom of the plate didn't work well. But also for some reason (which I don't know) the top of the plate didn't work either. Only the middle really worked (see Figure .




Please see the pdf version for figures
Figure 1.27:
Then this morning, I ran a new PCR dropping the volume down to 25 ml (previous was my typical 50 ml ). However, before loading the gel, I realized I forgot to add the template DNA to the mix. Now I'm rerunning the 25 ml reaction (I lowered the volume for these reactions for two reasons: 1) why not, I don't need much DNA and this will be cheaper; 2) I'm very low on Taq, in fact I ran out of the cheap Qiagen Taq, and this round I'm using the Easy-A proofreading Taq.
Please see the pdf version for figures
Figure 1.28: 300 ml, 1.5% agarose gel with 4.5 ul of 1% ethidium bromide run for 40 min at 100 volts. 10 ml of Fisher BioReagents exACTGene Low Range DNA Ladder, with bands:weights(ng) of 2000:105, 1500:87, 1000:68, 750:59, 500:94, 300:27, 150:34, 50:25 was used. 8ml of the PCR/dye was loaded per lane.
Brief Conclusions:   It looks like the primers work fine when I don't screw up the PCR (see Figure 1.28. The only problem is that three reactions didn't work. I'll know better when I run the qPCR if they were just random failed PCRs or bad primers. Since all previously verified primer pairs worked I think it's probably bad primers. I don't mind some failures, as I'm planning to test 36 per gene and dropping a couple doesn't hurt too much. Replacing them is unweildy given the way I pipette these by hand with a multchannel, with a robot, it'd be easy to replace them, but that's another day. The only problem is that one of the failures is a random gene, which knocks my number of negative control genes for every TF I test down to eleven, which is an odd number to have and weakens our error model. All in all I'm happy with the new plate oligo format from IDT and I need to hurry up and order the other plates.
The layout for the 384-well qPCR plate derives from the primer place and is shown in Table .
Lrp minimal vs rich qPCR setup
- 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
A metN argG thrL nlp yagU ompT cvpA purM serC ilvC yhjE metE leuLABCD metF gdhA metA purC serA argI livK argA lysC yebR pntA
B dppB gcl stpA mog aroG pinO cysK idnD aroL yhaF cysC nhaA aroP amiA gltB goaG dapD kdtB dapB yagG 5kb citC 1kb fruK
C metN argG thrL nlp yagU ompT cvpA purM serC ilvC yhjE metE leuLABCD metF gdhA metA purC serA argI livK argA lysC yebR pntA
D dppB gcl stpA mog aroG pinO cysK idnD aroL yhaF cysC nhaA aroP amiA gltB goaG dapD kdtB dapB yagG 5kb citC 1kb fruK
E metN argG thrL nlp yagU ompT cvpA purM serC ilvC yhjE metE leuLABCD metF gdhA metA purC serA argI livK argA lysC yebR pntA
F dppB gcl stpA mog aroG pinO cysK idnD aroL yhaF cysC nhaA aroP amiA gltB goaG dapD kdtB dapB yagG 5kb citC 1kb fruK
G metN argG thrL nlp yagU ompT cvpA purM serC ilvC yhjE metE leuLABCD metF gdhA metA purC serA argI livK argA lysC yebR pntA
H dppB gcl stpA mog aroG pinO cysK idnD aroL yhaF cysC nhaA aroP amiA gltB goaG dapD kdtB dapB yagG 5kb citC 1kb fruK
I metN argG thrL nlp yagU ompT cvpA purM serC ilvC yhjE metE leuLABCD metF gdhA metA purC serA argI livK argA lysC yebR pntA
J dppB gcl stpA mog aroG pinO cysK idnD aroL yhaF cysC nhaA aroP amiA gltB goaG dapD kdtB dapB yagG 5kb citC 1kb fruK
K metN argG thrL nlp yagU ompT cvpA purM serC ilvC yhjE metE leuLABCD metF gdhA metA purC serA argI livK argA lysC yebR pntA
L dppB gcl stpA mog aroG pinO cysK idnD aroL yhaF cysC nhaA aroP amiA gltB goaG dapD kdtB dapB yagG 5kb citC 1kb fruK
M metN argG thrL nlp yagU ompT cvpA purM serC ilvC yhjE metE leuLABCD metF gdhA metA purC serA argI livK argA lysC yebR pntA
N dppB gcl stpA mog aroG pinO cysK idnD aroL yhaF cysC nhaA aroP amiA gltB goaG dapD kdtB dapB yagG 5kb citC 1kb fruK
O metN argG thrL nlp yagU ompT cvpA purM serC ilvC yhjE metE leuLABCD metF gdhA metA purC serA argI livK argA lysC yebR pntA
P dppB gcl stpA mog aroG pinO cysK idnD aroL yhaF cysC nhaA aroP amiA gltB goaG dapD kdtB dapB yagG 5kb citC 1kb fruK
Table 1.7: yellow is positive control, cyan = A (correct antibody), magenta = B (incorrect antibody), white = C (beads only)
The schema in Table 1.7 allows two replicates per plate (hence the duplicated 1970's rainbow pattern). Takes six plates to run 12 chIP samples. Yikes, I'd better buy more qPCR master mix!
Each well will be filled as follows:
  1. add 15.5 ml master mix + water
  2. add 1.5 ml primer mix (150nM)
  3. 3 ml of template
Master mixes are prepared for 8 extra reactions. qPCR master mix (3940ml ) + water (2167). For template add 100 ml of water to each tube (200ml total), allow for appx 65 rxn, we're running 48.
I draw a random number to choose which sample (O or M) should occupy the top half of the plate (in case there are differences in the machine across the plate).
First batch plate A302JWOB   Sun Dec 10, 2005
M first then O
Ran short by 7 wells of master-mix + water. Pipetted those by hand (P12, P14, P16, etc). One other slight problem, when I spin down the plate prior to walking it over to the LSEB building where the 7900HT is, the centrifuge was at 4C. Normally this step puts all the sample on the bottom of the wells, but the temperature difference made a little condensation form on the top of each well (a very small amount but still annoying). I let the lid of the centrifuge stay open for a while and it pretty much fixed the problem.




No other mishaps occured to my knowledge. Rows A-H contain sample lrpMa1. Rows I-P contain sample lrpOa1. Just a reminder M = Davis + 0.5% glucose, O = LB + 0.5% glucose. The batch took a little less than 1hr 30min to set up. Remaining templates were put back in freezer.
Second batch plate A302JWOA   Sun Dec 11 16:00 EST 2005
O first then M
Ran way short of master mix + water. Had to make more for entire last row.
Third batch plate A302JWO9   Sun Dec 11 18:00 EST 2005
M first then O
Figured out how to stop running out of master mix + water. When pipetting the 15.5 ml from the multichannel I would press until the pipette hit the firm part and the press a little further to get the last little bit out. This has the negative affect of creating bubbles. It seems it also makes you run out of material faster. I wouldn't have guessed that. I need to repeat this and see if that's true, but I didn't run out of reagent this time.




This round was the most error prone so far. It was a little from fatigue (768 PCR rxns in one day is a little much), but mostly because the silicon grease must have worn off in the two previous batches. From now on I need to grease the low volume multichannel before every qPCR plate. There were a number of times the tips didn't stick to the channel. One time I ended up pipetting 300nM into 12 wells (F1,F3,F5,...,F23), because the first time not all the tips sucked up the 1.5ml and I lost track of which ones didn't. Hopefully, primer concentration doesn't have a huge impact, we'll have to take a closer look at that row.




Fourth batch plate A302JWO8   Mon Dec 12 20:54:08 EST 2005
M first then O (I seem to be good at generating a random number less than 0.5 (which means I do M first)
No problems that I saw; I'm becoming a good robot.
Fifth batch plate A302JWNS   Mon Dec 19 13:28:22 EST 2005
M first then O.
No problems that I saw.
Sixth and final batch plate A302JWNT   Mon Dec 19 21:44:22 EST 2005
O first then M.
No problems that I saw.

1.7  ChIPrndC: pdhR, fecI, lexA, ydaK

Wed Dec 7 11:23:33 EST 2005 cells in incubator in davis with 0.5% glucose for overnite growth (which takes 24hrs to hit stationary with minimal)
I thought they didn't grow, but they were just growing very slow.
ydaK is getting the can. While designing the primers, I noticed that it's predicted targets are all right next to each other. Typically this is a good sign as neighboring regions of bacterial genomes are often regulated by the same TF. Then I noticed fnr which seemed fishy and was also a neighbor, and I saw the problem. Almost half of the chips were run on the strain Jamey got from Tim Gardner, which I noticed had an fnr deletion (Jamey then found a paper describing this deletion in the Yale stock that is NOT in the stock used by Blattner to sequence the genome). These neigboring genes probably appear corregulated because they are all correlated due to the knockout...
Tue Dec 13 22:36:33 EST 2005 Finished with pdhR and lexA; fecI didn't make it, grew too slow. I would've guessed pdhR would've been the slow one. Trying fecI again tomorrow. Did minipreps to make sure tubes were labeled correctly (they were; see Figure . Also saved 1ml of each sample to try SDS-page with his-stain to check protein expression.
Please see the pdf version for figures
Figure 1.29:
Wed Dec 14, 2005 Finished growing lysing, shearing fecI. Minipreps from previous day checked out ok; seems to be no problems with mislabeling fecI just grows slow. This time fecI was started without; it seemed to grow much faster. 100mM IPTG was added 2hrs prior to adding formaldehyde (unfortunately an hour less than the other two samples). Also the cell grew a bit more dense than I would've liked (but I was trying to increase exposure to IPTG).
Sample OD600 growth time DNA (ng/ul) 260/280 260/230 ml in 25 mg ml dilution buffer
fecI F 0.680 5 hr 2 min 376.6 66.4 597
pdhR B 0.592 3 hr 20 min 407.3 61.4 552
fecE E 0.727 5 hr 2 min 459.9 54.4 489
pdhR D 0.658 3 hr 20 min 468.4 53.4 480
pdhR A 0.657 3 hr 20 min 422.5 59.2 532
fecI A 0.710 3 hr 20 min 382.0 65.4 589
pdhR C 0.658 3 hr 20 min 444.4 56.3 506
lexA D 0.317 3 hr 20 min 173.3 144.3 1298
pdhR F 0.635 3 hr 20 min 413.1 60.5 545
lexA C 0.273 3 hr 20 min 137.9 181.3 1632
lexA B 0.260 3 hr 20 min 123.7 202.1 1819 (1738)*
lexA E 0.272 3 hr 20 min 137.4 182.0 1638
lexA A 0.269 3 hr 20 min 111.5 224.2 2018 (1716)*
fecI B 0.701 5 hr 2 min 462.8 54.0 486
fecI D 0.694 5 hr 2 min 434.5 57.5 518
fecI C 0.689 5 hr 2 min 392.2 63.7 574
pdhR E 0.671 3 hr 20 min 403.0 62.0 558
lexA F 0.271 3 hr 20 min 128.5 194.6 1751
The shearing range was checked by running appx 600ng on a 1.5% agarose gel. Unfortunately the gel pour wasn't the best, I ran the gel too long and my samples evaporated a little on the parafilm (I know when one thing goes wrong, they all do!), so it's not the most beautiful gel (see Figure ). But if you look close, you see the shearing range is the typical one or perhaps a little shorter (I think the shorter is just because there is less DNA in the gel than I typcically run), but overall looks good as is typical for this step now.
Please see the pdf version for figures
Figure 1.30: Sheared DNA samples were run on a 1.5% gel to check the size range, which appears to be in the range of 100-800bp.
Thu Dec 15 23:13:46 EST 2005 Just added the antibodies. ChIPing sucks some times. I gotta find a way to shorten this thing into 24-48 samples a day.

1.7.1  primer design

Wed Dec 7 16:53:58 EST 2005
Primers are now designed automatically with a couple scripts I wrote that are now in the CVS (numerical/primer_design) and the publically available primer3 software. Primer order is randomized except all 12 random genes need to be kept together to simplify pipetting. However, the row which the random genes are placed in (A-D) is chosen randomly.
lexA primer plate, random genes are in row A
- 1 2 3 4 5 6 7 8 9 10 11 12
A gcl mog pinO idnD yhaF nhaA amiA goaG kdtB yagG citC fruK
B lasT yfiR galE araJ cspG ydjM araA dcrB araE sufA ydaN fnr
C aceE fldB cytR nhaR polB recN b1141 dinB mdh ydaM dinD b1458
D cspI envY aceA galP dinG fliA oraA dinF hemC dinI dnaA rfaB
fecI primer plate, random genes are in row D
- 1 2 3 4 5 6 7 8 9 10 11 12
A yedF yebG wzxE wecD sieB yrdC yncE cspG mutT ddlB entF map
B bfd fecA amiB exbB fdhE feoA rffG fecI fhuF wecG ynaI mntH
C entA sodA proV iscA oraA fhuA nrdE apaG dinF sulA ydiE ybaN
D gcl mog pinO idnD yhaF nhaA amiA goaG kdtB yagG citC fruK
pdhR primer plate, random genes are in row D
- 1 2 3 4 5 6 7 8 9 10 11 12
A b3171 fdhE cspG yjaG map yrdD yjgP yfiR rrmB ndh cdsA ybeB
B yjeQ aceE ispB ubiA ybjM dinF oraA rrnG ubiB ubiD alr yhbE
C lpxA rfaB dinG ybaP ddlB yebG yrbF b3790 cspB yhhF fecI fecA
D gcl mog pinO idnD yhaF nhaA amiA goaG kdtB yagG citC fruK
Primers were tested by PCR.
Please see the pdf version for figures
Figure 1.31: 300 ml, 1.5% agarose gel with 1.8 ul of 1% ethidium bromide run for 40 min at 100 volts. 10 ml of Fisher BioReagents exACTGene Low Range DNA Ladder, with bands:weights(ng) of 2000:105, 1500:87, 1000:68, 750:59, 500:94, 300:27, 150:34, 50:25 was used. 8ml of the PCR/dye was loaded per lane. The strange thing on the fourth row a kind of fuzzy aberation in the image is a piece of scotch tape that feel into the gel. Fortunately it didn't seem to affect the way gel ran. The first two rows are fecI; the last two are pdhR.
Please see the pdf version for figures
Figure 1.32: 300 ml, 1.5% agarose gel with 1.8 ul of 1% ethidium bromide run for 40 min at 100 volts. 10 ml of Fisher BioReagents exACTGene Low Range DNA Ladder, with bands:weights(ng) of 2000:105, 1500:87, 1000:68, 750:59, 500:94, 300:27, 150:34, 50:25 was used. These are the lexA targets. 8ml of the PCR/dye was loaded per lane.

1.7.2  qPCR

It will take 9 full 384-well plates to complete this experiment. Two samples will be run in each plate as in Table 1.7, except the samples will be run in the order of the initial randomization (the order is the one used in the table above showing the sheared DNA yields for each sample). For the previous ChIP experiment each plate contained one lrpMinimal sample and one lrpLB sample. These plates may contain samples from different TFs as I'll just go down the random list consecutively.
The target ordering can be inferred be looking at the individual 96-well oligo plates detailed in the tables above and combining two rows with alternating columns (e.g. row1 = A1, B1, A2, B3, ..., A12, B12; row2 = C1, D1, C2, D2, ..., C12, D12). This is the same way the lrp plates were handled. This round of chip is labeled ChIPrndC on all of the eppendorf tubes containing samples and on the plates containing oligo mixes.
First round of CHIPrndC A302JWNU   Tue Dec 20 17:48:04 EST 2005
fecIf (1) first then pdhRb (2)
No problems that I saw.
Second round of CHIPrndC A302JWNV   Tue Dec 20 19:50 EST 2005
fecIe (3) first then pdhRd (4)
No problems that I saw.
Third round of CHIPrndC A302JWNW   Wed Dec 21 17:49:49 EST 2005
pdhRa (5) first then fecIa (6)
No problems that I saw. I switched from the Marsh reservoirs to corning costar 4871 50ml reservoirs, and it is even easier to pipette the entire plate without running out of mastermix. I could probably use less extra master mix (but I'm not going to, better safe than sorry).
Fourth round of CHIPrndC A302JWNX   Wed Dec 21 19:51:46 EST 2005
pdhRc (7) first then lexAd (8)
Fifth round of CHIPrndC A302JWNY   Wed Dec 21 22:13:58 EST 2005
pdhRf (9) first then lexAc (10)
Sixth round of CHIPrndC A302JWO1   Thu Dec 22 00:15:48 EST 2005
lexA (11) first then lexA (12)
1536 qPCR reactions in one day is too much. I don't think I'll do this again anytime soon. Maybe I'll be puking SYBR green for xmas...
Seventh, Eighth, Ninth rounds of CHIPrndC   Thu Dec 22, 2005 performed at 4PM, 6PM, 8PM.
A302JWO0 lexA (13) first then fecI (14)
A302JWNZ fecI (15) first then fecI (16)
A302JWNI pdhR (17) first then lexA (18)
Round Eight was run on the other ABI 7900HT machine.
It is finished Thu Dec 22 22:50:59 EST 2005
Brief Conclusions:   Tue Jan 17 12:08:56 EST 2006
During the last round, lubing up the pipettor with silicon grease really got annoying, because there is a very fine balance between too much and too little grease. If the pipettor is not in that sweet spot, it is very frustrating to pipette. Getting uniform volumes requires throwing out reagent (I always find the sweet spot when I'm pipetting primers, since I have a very large supply of them) and adding or removing silicon until you hit the magic spot where the tips are all accurate, even and pipetting is sooooo much faster. It is stupid to have to deal with this unnessary problem. The problem to a large extent seems to be dependent on how warped the tip box becomes after autoclaving.
For the results above I used almost entirely Fisher tips (21-277-2B). They work well for standard pipetting in the lab with a Rainin P2 or P10. The last round I used boxes of finntip tips from Thermo Electron (the company making the multichannel). It was a dream come true, silicon woes were over and they stuck on the pipettor channels like flies on stink. No problems, always accurate, a little expensive. However, I did a little research today. By using the manufactors refill packs I can save the environment (a little) and the cost is almost the same as the fisher tips (3 one-hundredths of a penny more per tip; 0.0426 per tip (thermo) vs 0.0423 (fisher)).
One slight worry is the tip boxes containg 2x as many tips (192 instead of 96). Having 96 per box made it easy to keep track of location on the plate and not pipette into the wrong set of wells. 192 will speed things up, but make it more difficult to have a quick glance and immediately know where you are if someone stops by to talk to you. The are numbers are (9400327 for a refill and 9400326 for a starter kit that contains the 192 hole, tip boxes).

1.8  Sharing the Transcription Factor TOPO cds cloning primers

Thu Jul 13 14:20:01 EDT 2006
Hemali Patel (in our lab) needed to be able to amplify the full cds of many genes for some microarray spike-in studies. I gave here a 5 uM sample of 25 genes I thought I might use for the ChIP studies. I made and 200 ml plate for myself (in case I want to amplify these in the future) and gave 30 ml to here. The layout of the plate is below.
Full cds primer plate #7136
- 1 2 3 4 5 6 7 8 9 10 11 12
A nus F gadW F glcC F flhC F sfsA F cbl F pdhR F nac F yheO F yrb F fliA F rhaR F
B nus R gadW R glcC R flhC R sfsA R cbl R pdhR R nac R yheO R yrb R fliA R rhaR R
C gadX F lrp F hyaC F lexA F bolA F yhiF F abgR F fecI F ymfL F sspA F cytR F yidP F
D gadX R lrp R hyaC R lexA R bolA R yhiF R abgR R fecI R ymfL R sspA R cytR R yidP R
E ymfN F - - - - - - - - - - -
F ymfN R - - - - - - - - - - -
G - - - - - - - - - - - -
H - - - - - - - - - - - -
F indicates the forward primer; R indicates reverse primer.