Figures
Abstract
Transcription is controlled by multi-protein complexes binding to short non-coding regions of genomic DNA. These complexes interact combinatorially. A major goal of modern biology is to provide simple models that predict this complex behavior. The yeast gene RNR1 is transcribed periodically during the cell cycle. Here, we present a pilot study to demonstrate a new method of deciphering the logic behind transcriptional regulation. We took regular samples from cell cycle synchronized cultures of Saccharomyces cerevisiae and extracted nuclear protein. We tested these samples to measure the amount of protein that bound to seven different 16 base pair sequences of DNA that have been previously identified as protein binding locations in the promoter of the RNR1 gene. These tests were performed using surface plasmon resonance. We found that the surface plasmon resonance signals showed significant variation throughout the cell cycle. We correlated the protein binding data with previously published mRNA expression data and interpreted this to show that transcription requires protein bound to a particular site and either five different sites or one additional sites. We conclude that this demonstrates the feasibility of this approach to decipher the combinatorial logic of transcription.
Citation: Mao G, Brody JP (2010) Reverse Engineering the Yeast RNR1 Transcriptional Control System. PLoS ONE 5(11): e13895. https://doi.org/10.1371/journal.pone.0013895
Editor: Peter Butko, University of Maryland-Baltimore, United States of America
Received: May 4, 2010; Accepted: October 12, 2010; Published: November 16, 2010
Copyright: © 2010 Mao, Brody. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by the Defense Advanced Research Projects Agency (DARPA) Micro/Nano Fluidics Fundamentals Focus (MF3) Center, http://www.inrf.uci.edu/mf3/. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Proteins binding to short, specific DNA sequences can regulate gene expression. These proteins, called transcription factors, enhance or repress transcription. Transcription factor binding sites are generally short (less than 12 base pairs) in length and are usually located in the promoter region of the regulated gene. In the simplest case, the binding of a single protein to the gene's promoter can enhance or repress expression. In more complex cases, expression is regulated through a combination of multi-protein complexes binding to several distinct elements. The determination of the location and decoding of the combinatorial logic of all these regulatory elements would provide an important annotation to the complete genome sequence and could lead to a better understanding of development and evolution [1]–[4].
Deciphering the transcriptional regulatory code is a central challenge of modern biomedical research. Years of research have shown that cellular differentiation is mostly governed through regulatory control of transcription within each cell [2]. Thus deciphering this code will lead to a better understanding of cellular differentiation.
Several different assays have been applied to this problem. DNAse I protection mapping can be used to locate the binding sites of specific proteins on DNA or to identify locations where crude fractions of protein bind [5], [6]. Protein binding microarrays have produced comprehensive binding data for hundreds of different DNA binding proteins [7]–[10]. Chromatin immunoprecipitation is a powerful technique to identify, across the genome, sequences that are bound to specific transcription factors [11]–[16].
The different approaches to the problem have been synthesized into comprehensive identification of regulatory elements in the yeast genome [17] and for parts of the human genome by the NHGRI ENCODE project [18], [19]. These projects have led to mass identification of regulatory sites, but they do not provide any information on how these regulatory sites interact—the regulatory program.
Deciphering the regulatory program requires many measurements of binding between nuclear protein and specific DNA sequence. Neither protein binding microarrays nor chromosome immunoprecipitation can provide such measurements. The critical barrier to deciphering transcriptional control programs is the accumulation of data on nuclear protein binding to specific DNA sequences and resulting mRNA levels. Our approach to overcoming this barrier is to develop a surface plasmon resonance based assay [20]–[23].
Previously, we demonstrated that one could identify regulatory elements using surface plasmon resonance [24]. We did this by showing a significant change in SPR signal correlated with both nuclear protein binding to DNA sequence representing a particular regulatory element and an increased level of promoter activity. We also demonstrated that we can monitor dynamic changes in the occupancy of regulatory elements by monitoring yeast nuclear protein binding to a region of the RNR1 promoter as the cell cycle progresses [25].
Here we extend our previous work on one region of the RNR1 promoter to six other regions. These seven encompass most of the putative protein binding sites in the RNR1 promoter identified by a comprehensive, multi-pronged approach [17], as shown in Figure 1. Analysis of these seven regions allows for the determination of putative regulatory control systems.
The top shows the general region of Chromosome 5, while the bottom focuses specifically on the region between the coding sequences for ARG5,6 and RNR1. We labeled the seven sites monitored as Regions 0 through Region 6 or R0–R6. Each of the seven contained one or more sites previously identified as a putative protein binding site [17]. The image was generated by the UCSC genome browser [40].
Surface plasmon resonance sensors have previously been applied to nucleic acid/protein studies. Much of this work has focused on measuring kinetic rates between purified protein and short stretches of DNA [20], [22], [26], [27]. Surface plasmon resonance was used to characterize the interactions between human estrogen receptors and estrogen response elements [23]. A novel nanostructure based sensor was used to detect interactions between a nucleic acid aptamer and thrombin protein [28]. Aptamer/protein studies were performed with a novel PDMS microfluidic surface plasmon resonance imaging system [29]. A recent novel application used an SPR sensor to test whether specific transcription factors bind anywhere on an entire promoter (1,000–3,000 bp) [30].
Results
RNR1 lies on yeast chromosome 5, see Figure 1. About seven different regions have been identified as likely transcriptional regulatory sites [17]. Like most yeast genes, RNR1 has a compact and well characterized promoter. Thus it presents an ideal case for testing the ability of this assay to decipher its transcriptional regulatory program.
DNA (each 16 bp long) representing these seven regions were synthesized and separately attached to a surface plasmon resonant sensor. We treated a yeast culture to synchronize their cell cycles, took samples from the culture every 15 minutes, and purified nuclear protein from the samples. We measured the amount of nuclear protein binding to the seven different regions.
RNR1 mRNA levels were taken from published yeast microarray data [31]. These were collected following alpha factor arrest and are presented in Figure 2. Spellman et. al used DNA microarrays to comprehensively estimate relative mRNA levels of all yeast genes at 18 time points across the cell cycle [31]. They found that RNR1 mRNA levels reached two relative maximums, a first at about 21 minutes after synchronization, and a second at about 77 minutes [31].
This data was measured by others [31] with DNA microarrays.
Nuclear protein binding to different regions of the RNR1 promoter
After we established a synchronous yeast cell culture, we extracted nuclear protein at 15 minute intervals and measured the relative amount of nuclear protein that bound to the seven different 16 bp regions of the RNR1 promoter listed in Figure 1. Each measurement was repeated three times to provide error estimates. The results are presented in Figures 3 and 4.
Each measurement was repeated three times and the data points represent the mean value, while the error bars indicate three times the standard deviation in the mean.
Each measurement was repeated three times and the data points represent the mean value, while the error bars indicate three times the standard deviation in the mean. No significant differences were measured in nuclear protein binding to the control.
Control experiments
Two types of control experiments were performed. First, we extracted nuclear protein from unsynchronized yeast cells and measured the relative nuclear protein binding of this sample to each of the seven different regions of the RNR1 promoter. This provided a baseline with which to compare the binding fro synchronized cells.
Second, we immobilized degenerate 16 bp DNA (NNN NNN NNN NNN NNN N) (where N can be any of the four common nucleotides) to the sensor surface. This degenerate DNA was synthesized with equal molar concentrations of each base (A, C, G, and T) at each location. We measured how much nuclear protein, extracted at different time points, bound to this sequence. Since this degenerate DNA consists of many different ( billion) sequences, we expect it to bind to many different proteins, which will average out and not show any variation with the cell cycle. As expected, we found no significant change in the amount of protein bound to this degenerate DNA at each time point.
We collated this data into a “promoter profile”, as shown in Figure 5. This figure graphically depicts how proteins are binding and releasing from the targeted regions of RNR1's promoter. (Levels are depicted relative to unsynchronized cells, so in some cases these levels are negative.) It also summarizes the mRNA data (Figure 2) as either on/off or high/low. This digitization of the data, both the nuclear protein/DNA binding (input states) and the RNR1 mRNA levels (output states), allows us to suggest the regulatory program encoded into the DNA. Our suggestion is shown in Figure 6 as a digital circuit. Using standard notation, it could be equivalently written as,(1)indicating that RNR1 mRNA only results if protein complexes are bound to regions 2 and 3, or to regions 0, 1, 3, 4, 5, and 6.
The amount of nuclear protein binding to seven different regions of the RNR1 promoter is shown as black bars, with the height of the bars proportional to the difference between the binding measured at from nuclear protein at the given time point and binding measured with nuclear protein extracted from unsynchronized cells. On the right, the RNR1 mRNA levels are summarized. The left can be thought of as the input states, while the right represents the output.
High levels of expression occur if nuclear protein is bound to regions 0, 1, 3, 4, 5, and 6 or if nuclear protein is bound to regions 2 and 3.
This result suggests a hypothesis about how different regulatory elements interact to regulate transcription of RNR1. The hypothesis is generated from a correlation analysis of multiple observations. A rigorous test of the hypothesis could be performed by directly altering key regions of DNA, for instance deleting Region 3. The significance of this approach is that one can generate these hypotheses in a rapid, high throughput manner. Furthermore, the hypothesis could itself be tested through the accumulation of more data.
Discussion
This assay has some limits. It only identifies transcriptional regulation. Protein levels are regulated at many different points (e.g. transcription, translation, histones, ubiquitin-proteasome degradation). However, most regulation is thought to occur at the level of transcription [32].
The assay can not specifically identify protein complexes bound to the DNA. Three parameters could be measured to better identify these protein complexes: first, the kinetic binding constants between the protein and DNA [25], second, binding to specific antibodies [33], and third, the molecular weight of the complex, which is related to the increase in surface plasmon resonance signal. Each of these three factors should be dependent upon the identity of the protein complex. The measurement of all three, along with the knowledge that the protein binds to a specific DNA sequence should allow one to uniquely identify the protein complex.
This assay can be implemented as a higher throughput assay. The development of surface plasmon resonance imaging instruments [34]–[36] allow one to immobilize many different sequences of DNA onto a surface and simultaneously accumulate measurements of nuclear lysate binding to these different regions. This provides substantial improvement in throughput, when hundreds of targets must be tested. Surface plasmon resonance imagers can measure several hundred interactions simultaneously at the level of about
RIU. This is sensitive enough to measure the targeted interaction, in this work we measured the interaction as typically several times greater than
RIU.
In conclusion, we applied a novel method, surface plasmon resonance analysis of nuclear protein/DNA binding, to decipher how different regulatory elements interact to regulate transcription of a single gene, RNR1.
Materials and Methods
Yeast cell cultures and synchronization
As previously reported [25], we used the PY1 strain of Saccharomyces cerevisiae (yeast) [37]. Yeast cultures were grown in YPD medium (2% yeast extract, 4% peptone, and 4% dextrose) at C with a temperature controlled heater and shaker. Cultures were grown for approximately 20 hours until cells reached late-log phase. Following [31], yeast cells were synchronized by adding 12 ng/ml of
-factor to culture for 3 hours. This
-factor was subsequently removed by twice washing the cells, replacing the supernatant with fresh medium, and re-suspending the cells each time [25]. Samples were taken at regular intervals after establishing synchronous cultures and then processed to extract nuclear protein. Synchronization was confirmed by flow cytometry [25].
Nuclear protein extraction
We extracted nuclear protein from each sample of the synchronous culture, as previously described [25]. Briefly, yeast cells were converted to spheroplasts by digesting the cell walls. Spheroplasts were lysed, centrifugation (13,000 rpm for 30 min) separated the cellular lysate from the nuclear material. The nuclear material was further purified by gradient centrifugation and dialysis. Protease activity was inhibited by a cocktail of protease inhibitors added to the extraction buffer.
The concentration of nuclear protein was determined using the Bradford assay [38]. Nuclear extract from each sample was normalized to a concentration of 0.33 mg/ml in protein binding buffer (20 mM Hepes (pH 7.6), 10 mM MgSO, 1 mM EGTA, 20% glycerol, 75 mM ammonium sulfate).
Surface plasmon resonance measurements
Surface plasmon resonance measurements were made with the Spreeta SPR sensor. The experimental set up includes a data acquisition and control computer, a syringe pump, and a Spreeta evaluation kit. The three channel Spreeta evaluation kit consists of several Spreeta sensor modules, a three channel flow cell, an electronic controller with comprehensive software, and an integrated flow block. The sensor modules are made by Sensata; other components are made by Nomadics.
The flow block was used to connect the Spreeta sensor module with the control box and to secure the flow cell to the surface of the sensor. The flow cell provides three independent flow channels. Each channel is approximately 4.5 mm long and 0.1 mm wide. The flow cell confines solution to the narrow channels, which correspond to the sensor surface.
The sensor data was analyzed to determine relative protein binding by measuring the difference in steady-state refractive index level before and after the addition of nuclear protein. Each experiment was repeated three times to provide error estimates.
Attaching DNA to the gold surface
Double stranded DNA representing seven different regions of yeast chromosome 5, (see Table 1), was attached to the sensor surface using bovine serum albumin (BSA) as an intermediate. First, two complementary single-stranded DNA fragments, derivatized with biotin at the 5′ end, (see Table 1) each at 450 M concentration, were added together into a microtube. The microtube was placed into boiling water and allowed to slowly cool to room temperature. This annealing process produces double stranded DNA.
The immobilization scheme was implemented by flowing different solutions across the sensor surface. The sensor was monitored to confirm the appropriate surface modifications took place. The solutions contained (in order) biotin-BSA (0.67 mg/ml), streptavidin (0.33 mg/ml) and biotin-DNA (450 M) in PBS (1.37 mM NaCl, 2.7 mM KCl, 4.3 mM Na
HPO
, 1.4 mM KH
PO
at pH 7.3), which is also the running buffer. These were stored at
C, and thawed before use. Changes in refractive index adjacent to the sensing surface were monitored. Solutions remained in contact with the sensing surface until a stable refractive index value was reached, indicating the binding is at equilibrium. The running buffer was injected between each solution to remove non specifically bound molecules.
We previously measured the DNA surface density to be [39]. Using this surface density, we estimate an average spacing of about 30 nm between DNA molecules on the surface. This is much greater than the diameter of the DNA binding proteins (about 5 nm). Steric hindrance is not an issue.
Measuring nuclear extract binding to DNA
The nuclear extract (0.33 mg/ml of protein) in protein binding buffer (20 mM HEPES (pH 7.6), 10 mM MgSO, 1 mM EGTA, 20% glycerol, 75 mM ammonium sulfate) flowed across the sensor. The nuclear extract was stored at
C to prevent any degradation. Binding buffer was then injected to remove any non-specifically bound protein.
Cleaning
To restore the surface of the sensor to its original state, it was gently wiped with a Kimwipe wet by 6 N HCl and then flushed with water. This procedure was repeated three times. Then, 70% ethanol was used to wipe the surface followed by flushing with water; this was repeated three times. This cleaning procedure effectively removed all the immobilized layers. This was confirmed by measuring the refractive index of pure water as 1.3330. After each experiment was done, all syringes and tubes were rinsed thoroughly by water three times.
Author Contributions
Conceived and designed the experiments: JPB. Performed the experiments: GM. Analyzed the data: GM. Wrote the paper: JPB.
References
- 1. Kim HD, Shay T, O'Shea EK, Regev A (2009) Transcriptional regulatory circuits: predicting numbers from alphabets. Science 325: 429–432.
- 2.
Davidson EH (2001) Genomic Regulatory Systems. Academic Press.
- 3. Prud'homme B, Gompel N, Carroll SB (2007) Emerging principles of regulatory evolution. Proc Natl Acad Sci U S A 8605–8612.
- 4. Tirosh I, Reikhav S, Levy AA, Barkai N (2009) A yeast hybrid provides insight into the evolution of gene expression regulation. Science 324: 659–662.
- 5. Galas DJ, Schmitz A (1978) DNAse footprinting: a simple method for the detection of protein-DNA binding specificity. Nucleic Acids Res 5: 3157–3170.
- 6. Brenowitz M, Senear DF, Kingston RE (2001) DNase I footprint analysis of protein-DNA binding. Curr Protoc Mol Biol Chapter 12: Unit 12.4.
- 7. Bulyk ML, Gentalen E, Lockhart DJ, Church GM (1999) Quantifying DNA-protein interactions by double-stranded DNA arrays. Nat Biotechnol 17: 573–577.
- 8. Bulyk ML (2006) DNA microarray technologies for measuring protein-DNA interactions. Curr Opin Biotechnol 17: 422–430.
- 9. Vlieghe D, Sandelin A, Bleser PJD, Vleminckx K, Wasserman WW, et al. (2006) A new generation of jaspar, the open-access repository for transcription factor binding site profiles. Nucleic Acids Res 34: D95–D97.
- 10. Sandelin A, Alkema W, Engström P, Wasserman WW, Lenhard B (2004) Jaspar: an open-access database for eukaryotic transcription factor binding profiles. Nucleic Acids Res 32: D91–D94.
- 11. Solomon MJ, Larsen PL, Varshavsky A (1988) Mapping protein-DNA interactions in vivo with formaldehyde: evidence that histone H4 is retained on a highly transcribed gene. Cell 53: 937–947.
- 12. Buck MJ, Lieb JD (2004) Chip-chip: considerations for the design, analysis, and application of genome-wide chromatin immunoprecipitation experiments. Genomics 83: 349–360.
- 13. Lieb JD, Liu X, Botstein D, Brown PO (2001) Promoter-specific binding of Rap1 revealed by genome-wide maps of protein-DNA association. Nat Genet 28: 327–334.
- 14. Johnson DS, Mortazavi A, Myers RM, Wold B (2007) Genome-wide mapping of in vivo protein-DNA interactions. Science 316: 1497–1502.
- 15. Robertson G, Hirst M, Bainbridge M, Bilenky M, Zhao Y, et al. (2007) Genome-wide profiles of stat1 dna association using chromatin immunoprecipitation and massively parallel sequencing. Nat Methods 4: 651–657.
- 16. Schmidt D, Wilson MD, Spyrou C, Brown GD, Hadfield J, et al. (2009) Chip-seq: Using high-throughput sequencing to discover protein-dna interactions. Methods 240–248.
- 17. Harbison CT, Gordon DB, Lee TI, Rinaldi NJ, Macisaac KD, et al. (2004) Transcriptional regulatory code of a eukaryotic genome. Nature 431: 99–104.
- 18. E N C O D E Project Consortium (2007) Identification and analysis of functional elements in 1% of the human genome by the encode pilot project. Nature 447: 799–816.
- 19. Weinstock GM (2007) Encode: more genomic empowerment. Genome Res 17: 667–668.
- 20. Bondeson K, Frostell-Karlsson A, F?gerstam L, Magnusson G (1993) Lactose repressor-operator dna interactions: kinetic analysis by a surface plasmon resonance biosensor. Anal Biochem 214: 245–251.
- 21. Homola J, Yee SS, Gauglitz G (1999) Surface plasmon resonance sensors: review. Sensors and Actuators B-Chemical 54: 3–15.
- 22. Majka J, Speck C (2007) Analysis of protein-DNA interactions using surface plasmon resonance. Adv Biochem Eng Biotechnol 104: 13–36.
- 23. Teh HF, Peh WYX, Su X, Thomsen JS (2007) Characterization of protein–DNA interactions using surface plasmon resonance spectroscopy with various assay schemes. Biochemistry 46: 2127–2135.
- 24.
Lin L (2005) An SPR-Based Sensor to Measure DNA-Protein Interactions. Ph.D. thesis, University of California, Irvine.
- 25. Mao G, Brody JP (2007) Dynamic SPR monitoring of yeast nuclear protein binding to a cis-regulatory element. Biochem Biophys Res Commun 363: 153–158.
- 26. Kyo M, Yamamoto T, Motohashi H, Kamiya T, Kuroita T, et al. (2004) Evaluation of MafG interaction with Maf recognition element arrays by surface plasmon resonance imaging technique. Genes Cells 9: 153–164.
- 27. Neo SJ, Su X, Thomsen JS (2009) Surface plasmon resonance study of cooperative interactions of estrogen receptor alpha and transcriptional factor Sp1 with composite DNA elements. Anal Chem.
- 28. Kim DK, Kerman K, Hiep HM, Saito M, Yamamura S, et al. (2008) Label-free optical detection of aptamer-protein interactions using gold-capped oxide nanostructures. Anal Biochem 379: 1–7.
- 29. Wang Z, Wilkop T, Xu D, Dong Y, Ma G, et al. (2007) Surface plasmon resonance imaging for affinity analysis of aptamer-protein interactions with PDMS microfluidic chips. Anal Bioanal Chem 389: 819–825.
- 30. Moyroud E, Reymond MCA, Hamès C, Parcy F, Scutt CP (2009) The analysis of entire gene promoters by surface plasmon resonance. Plant J 59: 851–858.
- 31. Spellman PT, Sherlock G, Zhang MQ, Iyer VR, Anders K, et al. (1998) Comprehensive identification of cell cycle-regulated genes of the yeast saccharomyces cerevisiae by microarray hybridization. Mol Biol Cell 9: 3273–3297.
- 32. Maston GA, Evans SK, Green MR (2006) Transcriptional regulatory elements in the human genome. Annu Rev Genomics Hum Genet 7: 29–59.
- 33. Su X, Neo SJ, Peh WYX, Thomsen JS (2008) A two-step antibody strategy for surface plasmon resonance spectroscopy detection of protein-DNA interactions in nuclear extracts. Anal Biochem 376: 137–143.
- 34. Smith EA, Corn RM (2003) Surface plasmon resonance imaging as a tool to monitor biomolecular interactions in an array based format. Appl Spectrosc 57: 320A–332A.
- 35. Brockman JM, Nelson BP, Corn RM (2000) Surface plasmon resonance imaging measurements of ultrathin organic films. Annu Rev Phys Chem 51: 41–63.
- 36. Chinowsky TM, Grow MS, Johnston KS, Nelson K, Edwards T, et al. (2007) Compact, high performance surface plasmon resonance imaging system. Biosens Bioelectron 22: 2208–2215.
- 37. Kaiser P, Sia RA, Bardes EG, Lew DJ, Reed SI (1998) Cdc34 and the F-box protein Met30 are required for degradation of the Cdk-inhibitory kinase Swe1. Genes Dev 12: 2587–2597.
- 38. Bradford MM (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72: 248–254.
- 39. Lin L, Harris JW, Thompson HGR, Brody JP (2004) Surface plasmon resonance-based sensors to identify cis-regulatory elements. Anal Chem 76: 6555–6559.
- 40. Rhead B, Karolchik D, Kuhn RM, Hinrichs AS, Zweig AS, et al. (2010) The UCSC genome browser database: update 2010. Nucleic Acids Res 38: D613–D619.