Team:Johns Hopkins/Project

From 2010.igem.org

(Difference between revisions)

Latest revision as of 03:53, 28 October 2010

Abstract

If the goal of iGEM and the Parts Registry is to take the messy world of genetic engineering and transform it into something like the standardized world of electrical engineering, it may be useful if electronic systems could directly interface with biological systems. Past iGEM projects have used chemical or optical stimuli to actuate transcriptional responses. Our project, however, seeks to add voltage sensitivity to Saccharomyces cerevisiae. Baker’s yeast was chosen because in some sense yeast have a system that responds to voltage input. With a voltage stimulus one can open the voltage-gated calcium channels of yeast, causing calcium ions to rush into the cytoplasm. This causes calcineurin to dephosphorylate Crz1, which enters the nucleus and binds various promoters. Our group presents a library of characterized Crz1-sensitive binding sites of both naturally-occurring and synthetic varieties that can be integrated into promoters. Genes downstream of these promoters are thus voltage-regulated in media containing calcium.

The Plan, Visualized: In this image, the pathway we are utilizing is schematically represented in an extremely accurate fashion by blobs and arrows. The calcium channels (mid1 and cch1) are shown in blue. With electrostimulation, they admit calcium into the cell. This calcium activates the the orange protein phosphatase calcineurin which dephosphorylates the purple transcription factor Crz1p. The dephosphorylated version of Crz1p, Crz1, is rapidly transported into the nucleus where it acts on the purple and blue promoters, transcribing GFP and/or RFP. Ultimately, our system did not get as far as having two coexisting promoters, but that was the plan.

Aims

Show that voltage can be used to stimulate a transcriptional response in S. cerevisiae.
Develop a library of voltage-inducible elements with differing voltage response curves.
Determine the functional range of and optimized values for our system with respect to the following variables:
- Voltage applied
- Duration of voltage application
- Presence or absence of vesicular calcium pumps, yeast’s natural mode of intracellular calcium control
Develop an effective experimental apparatus to apply voltage and measure response.

Methods

Visualizing the Crz1 Transcription Factor

We wanted to visualize the behavior of Crz1 under electrostimulation, and confirm localization to the nucleus upon stimulation. The cells were induced, then mounted glass plate where they were imaged with a laser-scanning confocal microscope. Crz1-GFP was observed to move in and out of the nucleus just 5 minutes after a voltage stimulus. In a given image, some cells displayed Crz1-GFP densely packed in the nucleus while other cells displayed Crz1-GFP in the cytoplasm, but excluded from the nucleus. Here we see a confirmation of the oscillating behavior described by Cai et al.

In the above video we can see the oscillatory behavior of Crz1-GFP after electrostimulation. Note that Crz1-GFP's oscillating movement into and out of the nucleus is not in phase throughout the cell population. Only a few cells actually exhibit a change in Crz1 localization during the course of the video.

Optimizing Parameters for the CDRE from the FKS2 Promoter

We wanted to find an optimal voltage amplitude and electrostimulation duration for yeast containing the CDRE-RFP plasmid. Shocking at 10V caused widespread cell death, so it was necessary to find a balance between transcription and cellular damage. CDRE-RFP were grown and passed into 96 well plate. The electroporator was used to shock the cells with voltages from 2-10V for 0-80 seconds. We found that optimal transcription occurs when electrostimulation voltage is 8V. We also needed at least 40 seconds of electrostimulation to see expression.

	2s	5s	10s	20s	40s	80s
10V	None	None	None	None	None	N/A
8V	None	None	None	None	High	High
6V	None	None	None	None	Moderate	High
4V	None	None	None	None	Low	Low

Cells were analyzed using the above described confocal microscopy. As such, our data are purely quantitative.

Qualitative Characterization of the FKS2 CDRE

Having decided on an optimal voltage, we then wanted to optimize the duration of electrostimulation. Additionally, these experiments provided the opportunity to compare yeast with and without functional PMC1 and VCX1 (vesicular calcium pumps). The data support a linear increase in expression of our RFP reporter with increased stimulation time. We also observe much higher expression in cells with vesicular pumps than in those without.

The graph to the left shows a increasing, linear relationship between electrostimulation time and expression of the RFP reporter.

FKS2 CDRE with vesicular pumps, 0 seconds electrostimulation

FKS2 CDRE with vesicular pumps, 90 seconds electrostimulation

FKS2 CDRE with vesicular pumps, 110 seconds electrostimulation

In the images above we observe some constitutive expression of RFP in the FKS2 CDRE strain with the vesicular pumps. However, there is a linear increase in expression of the RFP reporter with increase in electrostimulation time up to 110 seconds.

FKS2 CDRE without vesicular pumps, 80 seconds electrostimulation

FKS2 CDRE without vesicular pumps, 90 seconds electrostimulation

FKS2 CDRE without vesicular pumps, 110 seconds electrostimulation

FKS2 CDRE without vesicular pumps, 130 seconds shocking

Again, there is a linear relationship between stimulation duration and RFP reporter expression. However, in this case, the cells lacking vesicular pumps expressed less than those with vesicular pumps, and were less likely to be viable under electrostimulatory conditions.

Quantitative Characterization of FKS2 CDRE and PMC1 CDRE in Vacuole Positive and Negative Yeast

At high time points, we observed that the unstimulated control often expressed as much as experimental samples. We postulated that this was a result of widespread cell death. While we normalized against OD, this can only account for cell density, not cell viability. We also noted that the data point at 40s (and later in one case) is consistently higher than many of the data with longer electrostimulation duration. We believe that this is the same effect. If this is so, it implies that the cell death that we observe is a steplike function of duration.

In this experiment, cells with the FKS2 CDRE and the PMC1 CDRE were grown and passed into two rows of a 96 well plate and shocked at 8V with an exposure time of 0-80 seconds to quantitate the relationship between electrostimulation time and expression of a fluorescent protein. We again investigated the effect of knocking out the PMC1 and VCX1 vesicular calcium pumps. The data for both pump positive and pump negative cells are shown below.

This plot shows the fluorescence intensity normalized by OD 8 hours after induction by electrostimulation in cells positive for vesicular calcium pumps and using the FKS2 UAS in the reporter's promoter

This plot shows the fluorescence intensity normalized by OD 8 hours after induction by electrostimulation in cells negative for vesicular calcium pumps and using the FKS2 UAS in the reporter's promoter

This plot shows the fluorescence intensity normalized by OD 8 hours after induction by electrostimulation in cells positive for vesicular calcium pumps and using the PMC1 UAS in the reporter's promoter

This plot shows the fluorescence intensity normalized by OD 8 hours after induction by electrostimulation in cells negative for vesicular calcium pumps and using the PMC1 UAS in the reporter's promoter

Future Research Plans

Finish categorizing the synthetic CDREs using our high throughput electrostimulation device.
Apply our optimized high throughput experimentation technique to a library of synthetic CDRE sequences we have developed to characterize their voltage activation domains. We hope this library might allow others to have a more fine-grained control over the voltage response in their cells.
Develop a genetic switch utilizing the repressible operator already present on FKS2.
Develop an iGEM-compatible plasmid for yeast with a yeast-optimized origin of replication and markers. A shuttle vector would also be useful, enabling DNA to be stored or copied in bacteria and quickly transfered for testing in yeast. Our project would not be possible in a prokaryotic chassis. We feel that iGEM's apparent focus on E. coli may hinder the development of parts and systems that require eukaryotic organisms.

References

Cyert MS. Calcineurin signaling in Saccharomyces cerevisiae: how yeast go crazy in response to stress. Biochem Biophys Res Commun (2003)
Long Cai, Chiraj K. Dalal & Michael B. Elowitz. Frequency-modulated nuclear localization bursts coordinate gene regulation. Nature (2008).
Stathopoulos-Gerontides, et al. Yeast calcineurin regulates nuclear localization of the Crz1p transcription factor through dephosphorylation. Genes & Development (1999).
Yoshimoto et al. Genome-wide analysis of gene expression regulated by the calcineurin/Crz1p signaling pathway in Saccharomyces cerevisiae. J Biol. Chem. (2002).

@@ Line 17: / Line 17: @@
 {
 display:block;
-width:129px;
+width:112px;
 font-weight:bold;
 color:#FFFFFF;
@@ Line 39: / Line 39: @@
 <li><a href="https://2010.igem.org/Team:Johns_Hopkins/Modeling">Modeling</a></li>
 <li><a href="https://2010.igem.org/Team:Johns_Hopkins/Parts">Parts</a></li>
+<li><a href="https://2010.igem.org/Team:Johns_Hopkins/Device">Device</a></li>
 <li><a href="https://2010.igem.org/Team:Johns_Hopkins/Notebook">Notebook</a></li>
 <li><a href="https://2010.igem.org/Team:Johns_Hopkins/Safety">Safety</a></li>
@@ Line 48: / Line 49: @@
 <br>
-[[Image:plan.jpg]]
+[[Image:plan.jpg|1000px|center|frame|'''The Plan, Visualized:''' In this image, the pathway we are utilizing is schematically represented in an extremely accurate fashion by blobs and arrows.  The calcium channels (mid1 and cch1) are shown in blue.  With electrostimulation, they admit calcium into the cell.  This calcium activates the the orange protein phosphatase calcineurin which dephosphorylates the purple transcription factor Crz1p.  The dephosphorylated version of Crz1p, Crz1, is rapidly transported into the nucleus where it acts on the purple and blue promoters, transcribing GFP and/or RFP.  Ultimately, our system did not get as far as having two coexisting promoters, but that was the plan.]]
 <br>
 =='''Aims'''==
 *Show that voltage can be used to stimulate a transcriptional response in ''S. cerevisiae''.
-*Develop a library of voltage-inducible promoters with differing voltage response curves.
+*Develop a library of voltage-inducible elements with differing voltage response curves.
 *Determine the functional range of and optimized values for our system with respect to the following variables:
 **Voltage applied
 **Duration of voltage application
-**Presence of vacuoles, yeast’s natural mode of intracellular Calcium control
+**Presence or absence of vesicular calcium pumps, yeast’s natural mode of intracellular calcium control
 *Develop an effective experimental apparatus to apply voltage and measure response.
 =='''Methods'''==
 ===Visualizing the Crz1 Transcription Factor===
-We want to characterize the electro stimulation voltage domain in which we see movement of the GFP tagged CRZ1 transcription factor into the nucleus. Crz1-GFP was grown and passed into two rows of a 96 well plate. The cells were then shocked every two hours at 10 Volts. The cells were passed to a glass plate where they were fixed and then imaged. GFP was observed to move in and out of the nucleus just 5 minutes after voltage stimulus. In a given image, some cells displayed GFP densely packed in the nucleus while other cells displayed GFP in the cytoplasm, but excluded from the nucleus. Here we see a confirmation of the oscillating behavior described by Elowitz et al (Elowitz 2008).<br>
+We wanted to visualize the behavior of Crz1 under electrostimulation, and confirm localization to the nucleus upon stimulation. The cells were induced, then mounted glass plate where they were imaged with a laser-scanning confocal microscope. Crz1-GFP was observed to move in and out of the nucleus just 5 minutes after a voltage stimulus. In a given image, some cells displayed Crz1-GFP densely packed in the nucleus while other cells displayed Crz1-GFP in the cytoplasm, but excluded from the nucleus. Here we see a confirmation of the oscillating behavior described by Cai ''et al''.<br>
 <html>
 <object width="640" height="385"><param name="movie" value="http://www.youtube.com/v/HivMIxt6cAA?fs=1&amp;hl=en_US"></param><param name="allowFullScreen" value="true"></param><param name="allowscriptaccess" value="always"></param><embed src="http://www.youtube.com/v/HivMIxt6cAA?fs=1&amp;hl=en_US" type="application/x-shockwave-flash" allowscriptaccess="always" allowfullscreen="true" width="640" height="385"></embed></object>
 </html><br>
-In this video we can see the GFP tagged CRZ1 cells displaying nuclear localization after shocking. Note that Crz1's oscillating movement into and out of the nucleus is not in phase throughout the cell population. GFP appears focussed in the nucleus in some cells, but in other cells GFP fills a donut-shaped area, leaving a void (the nucleus). Only a few cells actually exhibit a change in Crz1 localization during the course of the video.
+In the above video we can see the oscillatory behavior of Crz1-GFP after electrostimulation. Note that Crz1-GFP's oscillating movement into and out of the nucleus is not in phase throughout the cell population. Only a few cells actually exhibit a change in Crz1 localization during the course of the video.
 ===Optimizing Parameters for the CDRE from the FKS2 Promoter===
-We want to find an optimum voltage amplitude and electro stimulation time for yeast containing the CDRE-RFP plasmid. Shocking at 10 Volts caused large amount of cell death, as such, it was necessary to find an optimal voltage to cause transcription, but not to damage our cells. CDRE-mCherry were grown and passed into 96 well plate. The electroporator was used to shock the cell with voltages from 2-10 Volts with an exposure time from 0-80 seconds. We find optimum electro stimulation voltage is 8V and  we need at least 40 seconds of electrostimulation to see expression.
+We wanted to find an optimal voltage amplitude and electrostimulation duration for yeast containing the CDRE-RFP plasmid. Shocking at 10V caused widespread cell death, so it was necessary to find a balance between transcription and cellular damage. CDRE-RFP were grown and passed into 96 well plate. The electroporator was used to shock the cells with voltages from 2-10V for 0-80 seconds. We found that optimal transcription occurs when electrostimulation voltage is 8V.  We also needed at least 40 seconds of electrostimulation to see expression.
 {| border="1" cellpadding="5" cellspacing="0" style="background: #ffffff;"
@@ Line 85: / Line 86: @@
 |}
-===Further Honing FKS2 CDRE===
+Cells were analyzed using the above described confocal microscopy.  As such, our data are purely quantitative.
-We want to hone in on the FSK2-CDRE systems region of activation in the domain of time of shocking. Also to compare yeast without PMC1 and VCX1 knocked out versus yeast with it not knocked out to determine effect of the knockout on the CDRE systems expression levels. From here we see a through characterization of the system that shows increase in expression of our reporter RFP with increase in electro stimulation time. We also see that there is much higher expression of the cells without vacuoles knocked out than in cells with vacuoles knocked out.
-[[Image:FSK2-CDRE.jpeg|550px|center]]
+===Qualitative Characterization of the FKS2 CDRE===
-This graph shows a very clear increasing linear relationship between electrostimulation time and expression of the RFP tagged FSK2-CDRE expression.
 {|
-|[[Image:Hone_in_0_Sec_KI.jpg|200px|thumb|FSK2 CDRE  with vesicular pumps, 0 seconds shocking]]
+|[[Image:FSK2-CDRE.jpeg|left|frame]]
-|[[Image:Hone_in_90_Sec_KI.jpg|200px|thumb|FSK2 CDRE  with vesicular pumps, 90 seconds shocking]]
+||Having decided on an optimal voltage, we then wanted to optimize the duration of electrostimulation. Additionally, these experiments provided the opportunity to compare yeast with and without functional PMC1 and VCX1 (vesicular calcium pumps). The data support a linear increase in expression of our RFP reporter with increased stimulation time. We also observe much higher expression in cells with vesicular pumps than in those without.
-|[[Image:Hone_in_110_Sec_KI.jpg|200px|thumb|FSK2 CDRE  with vesicular pumps, 110 seconds shocking]]
+The graph to the left shows a increasing, linear relationship between electrostimulation time and expression of the RFP reporter.
 |}
-From the images above we can see that there is some constitutive expression of RFP in the FSK2 CDRE  with vesicular pumps, but there is a very clear increase in expression of RFP with increase of electro stimulating time to 90sec. In the picture taken of the cells that were stimulated for 110 seconds, we see a large vesicle, which makes sense as the was a huge calcium influx which was relocated to the vesicle. This also explain why the cells are a lot healthier looking than the cells in which the vesicular pumps were knocked out.
+{|
+|[[Image:Hone_in_0_Sec_KI.jpg|200px|thumb|FKS2 CDRE  with vesicular pumps, 0 seconds electrostimulation]]
+|[[Image:Hone_in_90_Sec_KI.jpg|200px|thumb|FKS2 CDRE  with vesicular pumps, 90 seconds electrostimulation]]
+|[[Image:Hone_in_110_Sec_KI.jpg|200px|thumb|FKS2 CDRE  with vesicular pumps, 110 seconds electrostimulation]]
+|}
+In the images above we observe some constitutive expression of RFP in the FKS2 CDRE strain with the vesicular pumps. However, there is a linear increase in expression of the RFP reporter with increase in electrostimulation time up to 110 seconds.
 <br>
 {|
-|[[Image:Hone_in_80_Sec_KO.jpg|200px|thumb|FSK2 CDRE  without vesicular pumps, 80 seconds shocking]]
+|[[Image:Hone_in_80_Sec_KO.jpg|200px|thumb|FKS2 CDRE  without vesicular pumps, 80 seconds electrostimulation]]
-|[[Image:Hone_in_90_Sec_KO.jpg|200px|thumb|FSK2 CDRE  without vesicular pumps, 90 seconds shocking]]
+|[[Image:Hone_in_90_Sec_KO.jpg|200px|thumb|FKS2 CDRE  without vesicular pumps, 90 seconds electrostimulation]]
-|[[Image:Hone_in_110_Sec_KO.jpg|200px|thumb|FSK2 CDRE  without vesicular pumps, 110 seconds shocking]]
+|[[Image:Hone_in_110_Sec_KO.jpg|200px|thumb|FKS2 CDRE  without vesicular pumps, 110 seconds electrostimulation]]
-|[[Image:Hone_in_130_Sec_KO.jpg|200px|thumb|FSK2 CDRE  without vesicular pumps, 130 seconds shocking]]
+|[[Image:Hone_in_130_Sec_KO.jpg|200px|thumb|FKS2 CDRE  without vesicular pumps, 130 seconds shocking]]
 |}
-From the above images we can see that there is a significant increase in the expression of RFP, which is liked to our FSK2-CDRE UAS, with the increase of time of electrostimulation, with little to no expression in the 80 second case, medium expression in the 90 second case, and high expression in the 130 second case.
+Again, there is a linear relationship between stimulation duration and RFP reporter expression.  However, in this case, the cells lacking vesicular pumps expressed less than those with vesicular pumps, and were less likely to be viable under electrostimulatory conditions.
-===FKS2 CDRE and PMC1 CDRE in Vacuole Positive and Negative Yeast===
+===Quantitative Characterization of FKS2 CDRE and PMC1 CDRE in Vacuole Positive and Negative Yeast===
-Many cells were dying from the stress of voltage shocking, so FKS2 CDRE-mCherry was inserted into yeast that still had their calcium vacuoles. These cells were grown and passed into two rows of a 96 well plate and shocked at 2-10 volts with an exposure time of 0-80 seconds. We wanted to characterize the relationship between electrostimulation time and expression of florescent protein of the synthetic CDRE promoter sequence and the PMC1 promoter sequence. We also want to demonstrate the effect of knocking out the PMC1 and VCX1 genes in yeast.  The data for both vacuole positive and vaculoe negative cells are shown below.All results are normalized against OD.<br>
+At high time points, we observed that the unstimulated control often expressed as much as experimental samples.  We postulated that this was a result of widespread cell death.  While we normalized against OD, this can only account for cell density, not cell viability.  We also noted that the data point at 40s (and later in one case) is consistently higher than many of the data with longer electrostimulation duration.  We believe that this is the same effect.  If this is so, it implies that the cell death that we observe is a steplike function of duration.
-[[Image:fsk2 +.jpg|550px|center]]
-[[Image:fsk2 -.jpg|550px|center]]
+In this experiment, cells with the FKS2 CDRE and the PMC1 CDRE were grown and passed into two rows of a 96 well plate and shocked at 8V with an exposure time of 0-80 seconds to quantitate the relationship between electrostimulation time and expression of a fluorescent protein.  We again investigated the effect of knocking out the PMC1 and VCX1 vesicular calcium pumps.  The data for both pump positive and pump negative cells are shown below.<br>
+{|
+|[[Image:fsk2 +.jpg|500px|thumb|This plot shows the fluorescence intensity normalized by OD 8 hours after induction by electrostimulation in cells positive for vesicular calcium pumps and using the FKS2 UAS in the reporter's promoter]]
+|[[Image:fsk2 -.jpg|500px|thumb|This plot shows the fluorescence intensity normalized by OD 8 hours after induction by electrostimulation in cells negative for vesicular calcium pumps and using the FKS2 UAS in the reporter's promoter]]
+|}
-[[Image:pmc1 +.jpg|550px|center]]
+{|
+|[[Image:pmc1 +.jpg|550px|thumb|This plot shows the fluorescence intensity normalized by OD 8 hours after induction by electrostimulation in cells positive for vesicular calcium pumps and using the PMC1 UAS in the reporter's promoter]]
-[[Image:pmc1 -.jpg|550px|center]]
+|[[Image:Pmc1 final.jpg|550px|thumb|This plot shows the fluorescence intensity normalized by OD 8 hours after induction by electrostimulation in cells negative for vesicular calcium pumps and using the PMC1 UAS in the reporter's promoter]]
+|}
-<br>
-From these graphs we can see a clear linear relationship between the period of stimulation and the expression of RFP or GFP in the domain of stimulation time in which we tested. We also observed that there were significantly higher percentages of live cells in the cells in which the vacuoles were not knocked out, and the presence of the vacuoles did not have significant impact on the expression rates, aside from some constitutive expression being seen in the vacuole plus samples. The controls on all our experiments seem to indicate very high constitutive expression, but that is due to a combination of cell death due to stimulation, only a small fraction of cells being stimulated and background fluorescence and so it can be discounted.
-=='''Instrumentation'''==
-We tried several approaches to efficiently and accurately electrostimulate our cells without
-killing them:
-*Microfluidic Chip
-**Consists of 300 micron-wide channels with titanium and gold electrodes to electrically stimulate the cells.
-**We have completed the fabrication of this device and are currently perfecting the fabrication technique to improve device quality and performance.
-**High-throughput and high-precision possible.
-**See our [[Team:Johns_Hopkins/Notebook#Device Design|fabrication protocol]].
-[[Image:Mask.jpeg|300px|center|]]
-*A 8 Well plate with aluminum electrodes which had a brass backing. This did show us our first results, but due to the inconsistency and inflexibility of the instrument we abandoned its use.
-*Finally, we have started using a gold plated coaxial 8 well electroporater to electrically stimulate the cells. It has shown excellent results and allows us to do large scale experimentation using 96 well plates. In the image below, you can see our complete setup.
-**In the bottom left, the 96 well plates where we kept our cells.  These 96 well plates also fit into the confocal microscope, flourometer and spectrophotometer we used to quantify and image our cells' response.
-**Also in the bottom left, our 8 well gold-plated coaxial electroporator.  The electroporator fits neatly into a vertical column of our 96 well plate, allowing us to easily subject multiple conditions (i.e. with and without vesicles) to identical conditions, or to run a particular condition in up to 8-fold multiplicity.
-**On the lower right, the oscilloscope that shows the alternating current produced by the function generator
-**In the middle right, the function generator we used to produce the voltage that induced our cells.  In this image, it is set to 19 Hz alternating current.
-<br>
-[[Image:Hightp.jpeg|550px|center|]]
 =='''Future Research Plans'''==
 * Finish categorizing the synthetic CDREs using our high throughput electrostimulation device.
-* We also plan to apply our optimized high throughput experimentation technique to a library of synthetic CDRE sequences we have developed to characterize their voltage activation domains. We hope this library might allow others to have a more fine-grained control over the voltage response in their cells.
+* Apply our optimized high throughput experimentation technique to a library of synthetic CDRE sequences we have developed to characterize their voltage activation domains. We hope this library might allow others to have a more fine-grained control over the voltage response in their cells.
-* We hope to develop a genetic switch utilizing the repressible operator already present on FKS2.
+* Develop a genetic switch utilizing the repressible operator already present on FKS2.
-* We noticed a lack of iGEM compatibility with the officially supported yeast chassis. We plan to develop an iGEM-compatible plasmid for yeast with a yeast-optimized origin of replication and markers. A shuttle vector would also be useful, enabling DNA to be stored or copied in bacteria and quickly transfered for testing in yeast. Our project would not be possible in a prokaryotic chassis. We feel that iGEM's apparent focus on E. coli may hinder the development of parts and systems that require eukaryotic organisms.
+* Develop an iGEM-compatible plasmid for yeast with a yeast-optimized origin of replication and markers. A shuttle vector would also be useful, enabling DNA to be stored or copied in bacteria and quickly transfered for testing in yeast. Our project would not be possible in a prokaryotic chassis. We feel that iGEM's apparent focus on ''E. coli'' may hinder the development of parts and systems that require eukaryotic organisms.
 ==References==
+*Cyert MS. Calcineurin signaling in ''Saccharomyces cerevisiae'': how yeast go crazy in response to stress. Biochem Biophys Res Commun (2003)
 *Long Cai, Chiraj K. Dalal & Michael B. Elowitz. Frequency-modulated nuclear localization bursts coordinate gene regulation. Nature (2008).
+*Stathopoulos-Gerontides, ''et al''.  Yeast calcineurin regulates nuclear localization of the Crz1p transcription factor through dephosphorylation.  Genes & Development (1999).
 *Yoshimoto ''et al''. Genome-wide analysis of gene expression regulated by the calcineurin/Crz1p signaling pathway in ''Saccharomyces cerevisiae''. J Biol. Chem. (2002).