Chapter 3: Endogenous HSF Activity in Alfalfa

3.1 Introduction
Plant HSFs are controlled through several tiers of regulation encompassing transcription, trimer/dimer assembly, acquisition of DNA binding ability and transcriptional competence and the HSF proteins' movement within the cell. Unraveling the nature of how these transcription factors operate has been and continues to be a difficult task. The current model of plant HSF activation and attenuation (Chapter 1, 1.6) was established through analysis of transcript levels under different environmental conditions, immuno-localization of HSF proteins to identify cellular compartmentalization and GUS fusion expression constructs to identify organelle and tissue localization (Scharf et al., 1998).

Northern analyzes have revealed several important characteristics of HSF gene transcription: First, in response to heat, HSF family members are differentially expressed (Nover at al., 2001); second, members of both class A and B HSFs are either constitutively expressed or are heat induced (Czarnecka-Verner et al., 1994; Czarnecka-Verner et al., 1995); and third, some HSFs exhibit tissue specific expression (Nover at al., 2001).

An important clue to how the HSF proteins are regulated and transported within the cell has come from immuno-localization experiments and HSF-GUS fusion protein experiments. First, in non-stress conditions class A HSFs are found in the cytoplasm and class B HSFs are only found in the nucleus (Scharf et al., 1998a; Scharf et al.,1998b); second, upon heat stress class A HSFs are imported into the nucleus; third, the active HSFs function as homo-trimers and in some cases as hetero-trimers (Scharf et al., 1998a; Scharf et al.,1998b). These findings helped facilitate the development of the plant HSF regulation model (Chapter 1, 1.6) as well as exposing the complexity of the plant HSF system. The burning questions are, why do plant have so many HSFs, why form trimers when most transcription factors form dimers and how does the HSF system discern between heat and other stresses?

Expression of HSFs in response to heat stress has been well characterized by numerous different groups (Nover et al., 2001; Bharti et al., 2000; Czarnecka-Verner et al., 1995). However, little is known about HSF expression under other stresses in which HSPs are observed to accumulate. From the few studies that have been conducted, it is clear the HSF transcripts accumulate in response to low temperature stress (Friedberg et al., 2000; Choeng et al., 2002), pathogen attack, wounding and osmotic stress (Choeng et al., 2002). Among the class A4 HSFs, expression data is only available from Phaseolus acutifolius (tepary bean), Nicotiana tabacum and Arabidopsis thaliana. The Phaseolus class A4 HSF was isolated from the leaves of drought stressed plants, indicating that up-regulation was correlated with response to drought (Gaydos et al., 2001). The Nicotiana class A4 HSF was isolated from cultured cells (Shoji et al., 2000) and the Arabidopsis class A4 HSF was isolated from leaves and stems (Czarnecka-Verner et al., 2000). These results indicate that class A4 HSFs are involved in response to stresses other than heat, thus a more detailed investigation of HSF expression was warranted.

This chapter details an analysis of a class A4 HSF gene expression in alfalfa. The experiments described below were designed to determine whether this HSF (MsHSFA4) was a transcriptional activator, to observe its expression patterns under controlled laboratory temperature stress and to compare the observed responses to expression patterns during field temperature stress. Furthermore, the range of alfalfa fall dormancy groups, dormant to non-dormant was incorporated into these experiments to detect germplasm differences and to identify any links of MsHSFA4’s responses to temperature stress and cold acclimation (reviewed in Chapter 1, 1.8).

3.2 Materials & Methods

3.2.1 Tobacco Protoplast Expression
Constructs were built to determine the activation potential of the alfalfa HSF cDNA and C-terminal domain (CTD) . Expression constructs were designed to activate a ß-glucuronidase (GUS) reporter construct driven by a minimal CaMV 35S promoter fused to upstream optimal HSEs (see Chapter 1, 1.4) to simulate a typical heat shock gene promoter. Other constructs were also built to determine the ability of the CTD to function as a general activator of transcription.

The first set of HSF constructs (termed effector constructs) were built to determine HSF transcriptional competence by assaying for the ability of the produced protein to bind to a GUS reporter construct that contained nine optimal heat shock elements (HSE) up-stream from a minimal CaMV 35S promoter controlling the expression of a GUS reporter gene. In total, four effector constructs were built containing different HSF cDNAs and two control vectors were built to act as negative controls (Figure 3.1 and Table 3.1). The first effectors were as follows: the open reading frame (ORF) of AtHSFA1, the ORF of AtHSFA4a, the ORF of MsHSFA4 and complete cDNA of MsHSFA4 including its large 5’ UTR (Figure 3.1, C through F). The control vectors contained: 1) only the T7 tag and no insert and 2) only the Gal 4 leader sequence also with no insert (Figure 3.1, A and B).

The second set of constructs tested the general transcription activation domains of the aforementioned HSFs. These expression constructs involved fusing the CTD of the HSF to the DNA binding domain (DBD) of the yeast Gal 4 transcription factor. The Gal 4 DBD/HSF CTD fusion protein binds to a reporter construct containing 10 tandem Gal 4 DNA binding elements up-stream from a minimal CaMV 35S promoter driving the expression of a GUS reporter gene. Once bound, the CTD was free to activate transcription of the reporter gene. Fusion constructs were prepared for the AtHSFA4a CTD and the MsHAFA4 CTD (Figure 3.1, I and J). A negative control construct was built that contained only the Gal 4 DBD without any CTD (Figure 3.1, H). Vector Construction
HSF effector expression constructs were built using PCR cloning into a series of plasmid backbones (for vector summary see Table 3.1). The vector backbones contained a CaMV 35S promoter and a multiple cloning region subtended by a leader sequence (T7 tag or Gal 4 leader) and poly a nos terminator sequence. PCR products were produced under standard cycling conditions with an annealing temperature of 58°C. PCR products were checked for quality using agarose gels and digested with Sal I and Not I overnight at 37°C. Digested PCR products were purified through ethanol precipitation by adding 2.5 volumes of 100% ethanol and 1/10 volume of 3M sodium acetate and incubated at -20°C for 2 hrs. PCR products were centrifuged, resuspended in 20 ul of ddH2O and quantified on an agarose gel. Vector backbones were prepared in the same way. Ligation of digested PCR products into the digested backbone was accomplished using T4 DNA ligase (Invitrogen). Ligated vectors were then electroporated into competent DH5a E. coli and grown on ampicillin LB plates. Plasmids were extracted from positive colonies and successful cloning was confirmed by PCR and restriction digest. Control and reporter plasmids (GUS and LUC) were grown in E. coli and prepared as described above. Protoplast Transformation, Expression & Evaluation
Tobacco plants were grown from leaf cuttings under aseptic conditions on MS media in breathable culture containers (Fisher). Plants were grown to produce leaves of approximately 5 to 8 cm in diameter. Four leaves yielded approximately one million protoplasts per ml.

Day1 Leaf Digestion: Under aseptic conditions tobacco leaves were covered with 30 ml of enzyme solution (Appendix B) in a 15 cm culture plate. While in the solution the leaf surface was scored by gently tapping the surface with a scalpel blade on both sides of the leaf. The plates were sealed and incubated in the dark overnight at 28°C.

Day 2 Protoplast Isolation: The enzyme solution containing the protoplasts was gently filtered though a sieve into a 15 ml sterile glass centrifuge tube. The plate was rinsed with K3S media (Appendix B) to recover the remaining protoplasts. The tubes were covered and centrifuged in a swing bucket rotor at 40xg for 7 min. Due to the density of the suspension solution, live protoplasts floated to the top. The green upper ring of protoplasts was removed with a glass Pasteur pipette and transferred to a new tube containing 9 ml of W5 wash solution (Appendix B). The tubes were then centrifuged again at 77xg for 7 min (to pellet the protoplasts), and the supernatant was discarded. The protoplasts were then washed again in W5 wash solution, centrifuged as above and final protoplast pellets were resuspended in 1 ml of K3M media (Appendix B) by gentle mixing. Protoplast concentration was determined using a hemocytometer.

Each transformation tube contained, in a volume of 12.5 ul, 3 plasmids including: lulciferase standard (2.5 ug), effector construct (5 ug) and reporter construct (ug). To each plasmid-containing tube, 50 ul of resuspended protoplasts (approximately 100 000) were added. An equal volume of 25% PEG solution (Appendix B) was added to each tube, mixed gently and incubated at room temperature for 30 min. The transformation reaction was then stopped by adding 900 ul of K3M solution and gentle mixing. The tubes were allowed to recover by incubating overnight at 28°C.

Day 3 Harvest & Analysis: Tubes were centrifuged at 10 000 x g at 4°C for 10 min and the supernatant was discarded. The pellet was resuspended in 100 ul of GUS/Luc extraction buffer (Appendix B) and the cells were ruptured by sonication (16 pulses at 50% power). The cell debris was spun down and the supernatant was transferred to an new tube.
Luciferase Assay: In a new tube, 25 ul of luciferase substrate (luciferin) (Sigma) and 5 ul of protein extract were combined, briefly vortexed and kept on ice. Luciferase activity was then tested using a scintillation counter and observing direct photon emissions.
GUS Assay: In new tubes 75 ul aliquots of GUS/Luc buffer (Appendix B) were prewarmed in a 37°C water bath. To each tube, 50 ul of protein extract was added, mixed and incubated at 37°C. At 10 min and at 1hr 10 min, 50 ul of each reaction was removed, transferred to a new tube and the reaction was stopped by pipetting into 950 ul of stop solution (0.2M Na2CO3). A 200 ul sub-sample was diluted in 5 ml of stop solution and GUS activity was observed using a fluorometer (excitation l = 365nm and emission l = 445nm). The raw activity over the one hour period was calculated as the difference between the GUS value at 1hr 10 min and the value at 10 min. The normalized GUS activity was determined by dividing the raw GUS value by the corresponding luciferase value and multiplying by 1x106 to normalize variation of plasmid transformation.

3.2.2 Expression of Excised Alfalfa Apical Shoots
Approximately 500 mg of apical shoot tissue (top most part of the plant and the first expanding leaf) was harvested from alfalfa (cv N442) and placed directly in a 500 ml flask containing 100 ml of temperature acclimated SHK media (derived from SH media (Schenk and Hildebrandt, 1972) modified with 10 mM K2SO4). The experiment was arranged as a randomized complete block design with three blocks, two treatments and two controls. The first control (no treatment), involved freezing shoot tissue in liquid N2 directly harvested from the plant. The second control (liquid) involved harvesting shoot tissue, adding the tissue to a treatment flask containing 100 ml of SHK media pre-acclimated to 28°C and incubating it for 2 hrs. Heat shock treatments involved harvesting several shoot tissue and subjecting them to a treatment flask containing 100 ml of SHK media pre-acclimated to 41°C. Tissue was removed from the flask after 0.5 and 4 hrs, directly frozen in liquid N2 and stored at ­80°C. Similarly, cold shock treatment involved harvesting shoot tissue and subjecting them to a treatment flask containing 100 ml of SHK media pre-acclimated to 4°C. Tissue was removed from the flask at 0.5 and 4 hrs, directly frozen in liquid N2 and stored at -80°C. To aerate the SHK media, all treatment flasks throughout the experiments were kept on temperature controlled shakers rotating at 175 rpm.

3.2.3 Germplasm
The diurnal experiment, chamber experiment and field experiment were designed to sample tissues for RNA and protein analysis from among 10 alfalfa cultivars that span the spectrum of fall dormancy. The cultivars used, from dormant to non-dormant respectively include the following:

Cultivar : Dormancy Rating
Beaver 1 Dormant
Vernal 2 5246 3
Saranac 4
ABI 700 6
Sutter 7
Maricopa 8
CUF 101 9
UC 1887 10
Wadi Qurayat 11 Non-dormant

3.2.4 Diurnal Experiment
Plants from all 10 cultivars (section 3.2.3) were germinated from seed and grown in Turface (crushed clay) under a 16 hr light and 8 hr dark periods, at 21°C with 60% relative humidity. In the course of watering twice a day, at 9:00 am and 4:00 pm for 5 min, plants were supplemented with equal parts of dissolved nitrogen (N), phosphorus (P2O5) and potassium (K2O) at a concentration of 88 mg/L (Appendix F). The experiment was arranged as a randomized complete block design with three blocks where each block was conducted on a different day. Within a block, the first sampling occurred one hour before lights on (7:00 am). Subsequent samplings occurred at specific intervals throughout a 24 hour period (-1, -0.5, -0.25, 0, 0.25, 0.5, 0.75, 1.0, 1.5, 2.0, 4, 8, 12, 15. 16, 16.25, 16.5, 17, 19, 23) where 0 hrs equals lights on and 16 hrs equals lights off. At each sampling period apical shoot tissue was pooled across all ten of the cultivars. Total RNA was extracted from each sample and blotted onto macroarrays as described section 3.2.7.

3.2.5 Chamber Experiment - In vivo Expression
Plants were germinated from seed in Petri dishes and transferred to root trainers filled with wet turface (crushed clay medium). The plants were grown in a growth chamber with 60 % relative humidity and a 16 hour day at 21°C. Plants were watered twice a day at 9:00 am and 4:00 pm for 5 min and were supplemented with equal parts of dissolved nitrogen, phosphorus and potassium at a concentration of 88 mg/L. The experiment was arranged as a randomized complete block design with three blocks and three treatments where each treatment was conducted on a different day. Plants were grown for 6 weeks, cut back and allowed to regrow for another 3 weeks before the first sampling. Experimental units consisted of 12 plants of each cultivar. The plants were randomly arranged in root trainers with two empty rows at the end (Appendix I). There were three replications in time. The experiment consisted of 3 treatments; heat stress at 41°C, cold stress at 4°C and diurnal control (no stress 21°C). Within each treatment there were 6 sampling intervals; 0, 1, 3, 5, 8, & 12 hours. Sampling involved gently removing 10 random apical shoot tops (one per plant), from the 12 plants of each cultivar. The 10 shoot tops from each cultivar were pooled, bagged, frozen in liquid nitrogen and stored at -80°C. All sampling for a replication was completed inside the growth chamber within a 10 minute period. Total RNA was extracted from each sample and blotted onto macroarrays as described in section 3.2.7.

3.2.6 Field Experiment
The field experiment was arranged as a randomized complete block design (Appendix J), with three blocks containing all 10 cultivars described in section 3.2.3.. Experimental units were 1m x 1.5 m in size and contained a single cultivar consisting of 5 rows with 50 seeds per row. The field was seeded in May 2001 at the Elora Research Station in Elora, Ontario (43°39’ latitude, 80°25’ longitude, 376 m elevation). Herbage was removed on 30 June, 2001 and sampling began on 11 September, 2001 and concluded 11 December, 2001. Sampling occurred every Tuesday between 3:00 and 4:00 pm and involved only 5 cultivars (Beaver, Saranac, Sutter, CUF 101 and Wadi Qurayat). At a given sampling interval, 3 plants per experimental unit were excavated and the crown buds and roots (the upper 10 cm of tissue directly below the crown axial) were separated. The representative tissue from each cultivar was pooled over replicates. The tissue was frozen in liquid N2 on site and stored at -80°C. Total RNA was extracted from each sample and blotted onto macroarrays as described in section 3.2.7.

3.2.7 RNA Isolation
All solutions in the following protocol were prepared with DEPC (diethyl pyrocarbonate, 1 ml per L) treated water. Samples were ground with a mortar and pestle under liquid N2 into a fine powder. In a 12 ml centrifuge tube, 250 mg of ground sample were added to 1 ml of TRI-Reagent (BioCan Scientific cat# TS-120). The sample was homogenized using a Polytron disrupter for 30 sec and incubated at room temperature for 5 min. The cell debris was pelleted by spinning at 12 000xg for 15 min at 4°C. The supernatant was transferred to a 1.5 ml centrifuge tube, combined with 0.3 ml of chloroform, vortexed for 30 sec and incubated at room temperature for 15 min. Phase separation was induced by spinning at maximum speed for 15 min at 4°C. The upper aqueous layer was transferred to a new tube and the lower organic layer was frozen at -20°C for later protein extraction. To precipitate phenolics, 200 ul of Plant RNA Isolation Aid (Ambion cat # 9690) were added to each tube, vortexed and spun at maximum at room temperature for 2 min. The supernatant was transferred to a new tube and 0.4 ml of isopropanol, followed by 0.4 ml of precipitation solution (0.8 M sodium citrate; 1.2 M sodium chloride) were added, mixed by inversion and incubated at room temperature for 10 min. The RNA was pelleted by centrifugation at 12 000xg for 8 min at room temperature. The supernatant was removed and the pellet was washed with 1 ml of 75% ethanol and re-pelleted by spinning for 5 min at 12 000xg. The supernatant was discarded and the pellet was air dried for 5 min. To resuspend the RNA, 100 ul of H2O were added to each tube and incubated at 60°C for 20 min with intermittent vortexing. The tubes were spun at maximum to pellet any undissolved material and the supernatant was transfer to a new 1.5 ml centrifuge tube and stored at -80°C.

3.2.8 RNA Quantification
Total RNA was quantified using absorbance at 260 nm in a SpectraMax plus384 spectrophotometer (Molecular Devices, Sunnyvale, California). RNA samples were loaded in triplicate into a 96 well UV microplate. Quantification value was the average of the 3 independent readings.

3.2.9 RNA Quality
RNA quality was determined by running randomly selected samples on a denaturing agarose gel (1.2 % agarose; 5 % formaldehyde) and observing the presence of the prominent ribosomal bands and other prominent chloroplastic transcript bands.

3.2.10 DIG Probe Preparation
Using plasmid DNA as a template, the DIG PCR probe reaction was carried out using primers specific to the MsHSFA4 ORF (MsHSF-1a & MsHSF-1b) and the DIG PCR Probe Synthesis Kit (Roche cat # 1 636 090). Following the protocol supplied with the kit, successful PCR reaction were accomplished with an annealing temperature of 58°C, producing a 1 kb fragment. An aliquot of the probe reaction and the control reaction were separated on a 0.8 % agarose gel to observe a band shift indicating successful incorporation of the DIG labeled nucleotides. The probe was stored at -20°C until needed.

3.2.11 Northern Blot Hybridization
From each sample, 20 ug of total RNA were separated for 2 hours at 60 volts on a 1.2 % agarose gel prepared with 1x MOPS buffer and 5.4 % formaldehyde. The gel was then acclimated in 10x SSC while shaking in two, 15 min washes. The RNA was capillary transferred overnight in 10x SSC to a positively charged nylon membrane (Roche). Following the transfer, UV cross-linking of both sides of the membrane fixed the RNA to the membrane. The membrane was rinsed twice in ddH2O to remove excess SSC, placed in a glass hybridization tube containing 60 ml of pre-hybridization solution (Easyhybe™, Roche) and incubated for 2 hrs at 37°C. The pre-hybridization solution was then removed and stored at -20°C for future hybridizations. Finally, 30 ml of hybridization solution containing 50 ul of the DIG PCR probe reaction were filter sterilized, added to the hybridization tube and incubated overnight in a rotating hybridization at 50°C.

Post hybridization membranes were washed twice, 15 min each in 2x stringency wash solution (2x SSC, 1% SDS) at room temperature. A second set of two 15 min washes was preformed using 0.2x stringency wash solution (0.2x SSC, 1% SDS) at 65°C. The membrane was briefly washed in 1x wash solution (Roche) and placed in 1x blocking solution (Roche), gently shaking at room temperature for 1 hr. The blocking solution was then replaced with anti-DIG Fab solution (Roche) and incubated while gently shaking at room temperature for 30 min. The membrane was transferred to a new dish and excess antibody was removed by washing two times, 15 min each, in 1x wash solution (Roche). The membrane was then acclimated in 1x detection buffer for two min. The chemiluminescent substrate solution was prepared by diluting 10 ul of concentrated CDP-STAR (Roche) in 1 ml of 1x detection buffer. Prepared membranes were overlaid with CDP-Star chemiluminescent substrate solution and heat sealed between two pieces of clear acetate. The sealed membrane was exposed to Kodak Biomax film for two exposures of 10 and 20 min, and developed.

3.2.12 Preparation of RNA Macroarrays
All RNA samples were given a unique numeric code and randomized before blotting. Working in 96 well microplates, 10 ug of each RNA sample were diluted in RNA dilution buffer (6x SSC; 20 % formaldehyde) and randomly aliquoted into wells in a 96 well microplate. The microplates were then covered and incubated at 60°C for 20 min to denature the RNA. The samples in the microplates were transferred, using a multi-pipetting robot (Biomec 2000, Beckman Coulier), onto positively charged nylon membranes (Roche) in a 96 well Bio-Dot microfiltration device (BioRad, cat # 170-6545, Herculeus California). A vacuum applied to the Bio-Dot device pulled the samples through the wells and onto the membrane. Under vacuum, each well in the Bio-Dot device was rinsed with 10x SSC. The RNA was then fixed to the membranes by UV cross-linking. Each microplate and subsequent membrane was prepared and blotted in triplicate.

3.2.13 Screening of RNA Macroarrays
RNA macroarrays were probed using the aforementioned DIG DNA probe to the MsHSFA4 ORF (Chapter 3, 3.2.6). Membranes were stacked and separated using cut squares of nylon mesh and sealed in hybridization bags (Roche). The hybridization was performed as according to methods described in 3.2.7. Post hybridization, the macroarray membranes were developed as described in section 3.2.7. All membranes were processed at the same time to minimize experimental variation. Fluorograms of exposed membranes were scanned into digital images and values for relative video intensity was determined with the Array-Pro Analyzer software (version, Media Cybernetics L.P.). An example of a typical developed northern macroarray can be seen in Figure 3.8, A.

3.2.14 Statistical Analysis
MsHSFA4 expression exhibited a diurnal effect (see section 3.3.3). Thus in the chamber experiment, diurnal control values obtained under heat and cold stress were adjusted (normalized) to the values obtained in the diurnal control prior to statistical analysis. Furthermore, cultivars with the most similar fall dormancy were pooled to facilitate the comparison of cultivar means. For definition of cultivars and their dormancy group see Chapter 3, 3.2.3.

In the field experiment, the data for the CUF 101 cultivar at the Elora field location were removed from the analysis as there were not enough field plants to complete the experiment.
All statistical analyzes were performed using SAS (statistical analysis software, version 8.0, Cary California). All data were subjected to a variance analysis using the “proc mixed” procedure of SAS (Statistical Analysis Software). For this analysis the variance was partitioned into block, variety, time, block x variety and variety x time. The blocks and block x variety were assumed to be random effects, with time a repeated measure and varieties, time and variety x time considered to be fixed effects. A log transformation of the data was required to obtain homogeneity and normal distribution of error variance. The type I error rate was set at 0.05.

3.3 Results

3.3.1 Activity of MsHSFA4 in Tobacco Protoplasts
This experiment was designed to determine the potential of the alfalfa HSF cDNA to activate GUS reporter constructs driven by an HSE containing promoter. As well, to determine the transcriptional activation abilities of the isolated C-terminal domains (CTD) using fusion constructs involving the yeast Gal 4 DBD.

Activity of HSE/GUS Reporter Construct: In the presence of the HSE/GUS reporter construct the control T7 tag and control Gal4 leader constructs had negligible activity indicating that there was little to no background levels of endogenous GUS activity in the tobacco protoplasts (Figure 3.1, A and B). The AtHSFA1 open reading frame (ORF) construct produced very high levels of GUS activity (Figure 3.1, C). The AtHSFA4a ORF construct produced 20% of the level of activity as compared to the AtHSFA1 ORF (Figure 3.1, D), findings consistent with previous experiments (Czarnecka-Verner and Gurley, 2002; Czarnecka-Verner et al., 2000). The alfalfa MsHSFA4 ORF produced comparable levels of activity as compared to the AtHSFA4a homologue (Figure 3.1, E). The MsHSFA4 full cDNA containing the entire 5’ UTR, was unable to activate the HSE/GUS reporter construct, indicating that the 5’ UTR was acting as a negative regulator or was an incomplete sequence (Figure 3.1, F).

Activity of Gal4 BS/GUS Reporter Construct: The control T7 tag construct produced no transcriptional activity indicating the background levels of endogenous GUS activity in the tobacco protoplasts were very low (Figure 3.1, G). The Gal4 DBD construct exhibited a low level of activity (due to the presence of a cryptic activator), and was used as a control (Figure 3.1, H). The fusion construct of the AtHSFA4a CTD produced significantly higher levels of GUS activity, over 3 fold greater, as compared to the Gal4 DBD control (Figure 3.1, I). The fusion construct of the MsHSFA4 CTD produced significantly greater levels of GUS activity, over 2 fold greater than the AtHSFA4a CTD and 5 fold greater than the Gal4 DBD control (Figure 3.1, J).

3.3.2 Northern Blot of MsHSFA4 Temperature Expression
Alfalfa, cv. N442, apical shoot tissue was assessed for MsHSFA4 transcript activity under heat shock (41°C), cold shock (4°C) and field low temperature stress. Analysis of the experimental controls, revealed that MsHSFA4 had little to no expression under non-stress conditions (Figure 3.2, A and B). Under heat shock at 41°C, MsHSFA4 levels were detectable at 0.5 hrs and were still present after 4 hrs of exposure (Figure 3.2, C and D). Under cold shock at 4°C, MsHSFA4 levels were also detected after 0.5 hrs and still present after 4 hrs of exposure (Figure 3.2, E and F). Under field low temperature stress (approximately 5°C), specifically, tissue harvested from crown buds in November, high levels of MsHSFA4 transcripts were observed (Figure 3.2, G).

3.3.3 HSF Diurnal Effect
There were significant fluctuations in transcript levels of MsHSFA4 over the course of a 24 hr period (Figure 3.3). In this experiment, all environmental conditions with the exception of the light and dark periods and the two watering periods, were fixed in the growth chamber. Just before “lights on”, transcript levels spiked and then dropped immediately after “lights on” (8:00 am). Transcript levels then gradually climbed throughout the day reaching a maximum after 9 hrs of duration (4:00 pm). Levels then declined reaching a minimum by “lights off”, 17 hrs (12:00 am), followed by sharp spike an hour later (1:00 am). Transcript levels fell again through the dark period bottoming out at 20 hrs (3:00 am). Transcript levels then begin to climb again prior to lights on (Figure 3.3). The possibility that this diurnal effect is indicative of global RNA fluctuation in alfalfa was ruled out by using a probe to another transcript (DIG probe to an alfalfa sucrose-phosphate synthase cDNA - data not shown) which exhibited constant and unchanging transcript levels.

3.3.4 Chamber Experiment - Heat and Cold Stress Heat Stress
All cultivars exhibited increases in MsHSFA4 transcripts in response to heat stress but the nature of response differed significantly among all the cultivars (Figure 3.4). Among cultivars, the most striking differences were observed between dormant and non-dormant cultivars. Specifically, dormant cultivars exhibited an initial transcript induction peak (approximately 2 fold increase) after 1 hr of exposure and a second peak after 5 hrs of exposure, with the exception of cultivar 5246 (dormancy 3) which followed no trends (Figure 3.4, A). In contrast, non-dormant cultivars exhibited only a single peak (with the exception of dormancy group 9&10) after 3 to 5 hrs of exposure (Figure 3.4, B). Dormant cultivars also exhibited a more rapid onset of transcript accumulation in rate and magnitude (approximately 2 fold increase) as compared with non-dormant cultivars. The overall trend in MsHSFA4 expression in response to heat stress indicated that the more non-dormant the cultivar the more delayed was the expression of the transcript. Cold Stress
All cultivars exhibited increases in MsHSFA4 transcripts in response to cold stress but the nature of response differed significantly among all the cultivars (Figure 3.5). As was observed under heat stress dormant cultivars exhibited two expression peaks, one after 1 hr of exposure and the second after 5 hrs of exposure, with the exception of cultivar 5246 (dormancy 3) which followed no trends (Figure 3.5, A). Non-dormant groups exhibited only a single expression peak that occurred between 3 and 8 hrs of exposure (Figure 3.5, B). Overall trends were strikingly similar to those observed under heat stress signifying that in response to cold stress the more non-dormant the cultivar the more delayed the MsHSFA4 response.
There were also differences in expression between heat and cold stress among dormancy groups. Within the dormant cultivars, there was no change in the first response peak, however the secondary peak was significantly greater in magnitude under the heat stress as compared to the cold stress. Among the non-dormant cultivars, temperature stress had the reverse effect. In dormancy group 9&10, the amplitude of the response peak increased under cold stress. In dormancy group 11, the duration or amplitude of the response peak increased, from a peak at 3 hrs to a peak spanning 3 to 5 hrs under cold stress. In dormancy group 7, cold stress resulted in quicker response peak shifting from 5 hrs (heat stress) to 3 hrs (cold stress).

3.3.5 Field Experiment - Fall 2001
The MsHSFA4 transcript levels fluctuated throughout the fall and the patterns of fluctuation were significantly different between the four cultivars (Figure 3.6). MsHSFA4 responded to field conditions, but correlative conclusions can only be made in relationship to the effects of low temperature. Saranac and Sutter appear to have no significant trends and fluctuate around a horizontal axis. Beaver appears to strongly fluctuate up until the end of October and after that point holds a steady level of transcription. Wadi Qurayat appears to have a small general increasing trend throughout the fall, opposite to the temperature trends displayed in Figure 3.6, B. Analysis of root tissue revealed no significant differences among the four cultivars. MsHSFA4 transcript levels were generally constant throughout the fall (Figure 3.7); however, there may have been slight transcript increases in Sutter and Wadi Qurayat through November and December.

3.4 Discussion
As observed in Chapter 2, MsHSFA4 possesses an oddly large 5’ UTR, whose presence, in tobacco protoplasts, apparently knocks out its activity. Since the protein product produced from the open reading frame construct was functional, we can dismiss the possibility of post-translational regulation. The inhibitory effect of the 5’UTR may be occurring either at the level of transcription or translation but there is insufficient data to suggest its true mode of action. One possibility is that the inhibitory effect is a result of the deleted portion of the 5’ UTR observed in Chapter 2 (Chapter 2, and Figure 2.2). However, it is also possible that this effect is just a result of the experiment being performed in the context of tobacco as opposed to alfalfa. Additional studies of the same constructs but in alfalfa protoplasts are warranted.

The expression study of MsHSFA4 in tobacco protoplasts confirmed that the protein encoded within this cDNA sequence to indeed be a transcriptional activator and responsive to HSEs (Figure 3.4). Its level of activity in the construct containing only the open reading frame was comparable to the Arabidopsis thaliana A4 homologue. Furthermore, the Arabidopsis thaliana A4 HSF had lower but expected levels of expression compared to the Arabidopsis thaliana A1 HSF as is observed in previous studies (Czarnecka-Verner et al., 2000). There were, however, two notable differences in the activity of MsHSFA4. First, the expression construct containing the very large native 5’ UTR had no transcriptional activity in tobacco protoplasts; it was no different than the Gal4 leader control construct. Second, in the constructs where HSF CTDs were fused to the Gal 4 DBD, MsHSFA4 had over twice the transcriptional activity as compared to its Arabidopsis homologue AtHSFA4. This indicates that MsHSFA4 has the potential to be a much stronger activator as compared to AtHSFA4.

Upon sequence analysis of MsHSFA4 and AtHSFA4a, there were not any notable differences that could account for the significant boost in transcriptional activity in MsHSFA4 CTD as compared to the AtHSFA4a CTD. One small difference was that the AHA-2 motif of MsHSFA4 was shifted c-terminally by two amino acids. Czarnecka-Verner et al., (2000) have previously demonstrated that AHA-1 is redundant and the majority of transcriptional competence resides with AHA-2. Thus, this AHA shift may have effected activation ability through the nature of its extension. This greater reach theory is further supported from activity analysis of class A1 and class A3 Arabidopsis HSFs which have much greater levels of activity as compared to the class A4 HSFs. These subfamilies of HSFs possess either more AHA motifs (up to 4) or have a larger CTD that extend the reach of their AHA motifs. However, another explanation for these observed differences focuses on the c-terminal region of CTD ahead of AHA-2. Czarnecka-Verner et al., (2000) demonstrated that this c-terminal end of the sequence exhibits a repressive effect on transcriptional activation ability. Specifically, the removal the c-terminal end ahead of AHA-2 increases activity in a Gal 4 DBD/AtHSFA4a CTD fusion protein by two fold. It is possible that the observed difference in activity of MsHSFA4 and AtHSFA4a CTDs is a result of a change in some unidentified element in this c-terminal region. Thus the MSHSFA4 CTD’s activity does not represent an increase in activity but rather the lower level of activity observed in the AtHSFA4a CTD is a result of greater repression.

The controlled environment in the chamber experiment (Figure 3.4 and 3.5) confirmed the nature of MsHSFA4 as an HSF that is responsive to both heat and cold when all other environmental variables are held constant. This is the first report of a cold induced HSF and the first report of significant genotypic variations of HSF expression within a given species. The field experiment (Figure 3.6) confirmed that MsHSFA4 was active during the fall period and its activity varied between root an crown bud tissue.

Activation and attenuation of the heat shock genes by HSFs have been well documented under heat stress (Chapter 1), but activation of heat shock genes under low temperature is still a mystery. Furthermore, as only some of the full complement of genes involved in the heat shock response are active under low temperature, the mechanism of low temperature preferential selection is also unknown. The results from the field and chamber experiments strongly suggest that MsHSFA4 is active under low temperature stress to activate HSPs and protect against low temperature stress damage. MsHSFA4 was also observed to have diurnal fluctuations in transcript levels (Figure 3.3). This unexpected result suggests MsHSFA4 may be involved in the daily regulation of some HSPs, and in addition to stress protection, may be involved in other cellular activities.

Cultivar data from the chamber experiment revealed a relationship between dormancy classes and the response in amplitude and onset of MsHSFA4 in response to heat and cold; the less dormant the plants the more delayed the response. This cultivar dormancy effect suggests that the more adapted the cultivar is to warm climates the less it may require the heat shock response under heat stress. Similar responses were also observed under cold stress, indicating that the response is conserved under both heat and cold. These results imply that the heat shock response exists to supplement the adaptive physiology that evolves in a plant in a particular niche. Specifically, the breeding of cold tolerant cultivars also, and inadvertently, selected for a more rapid MsHSFA4 response and possibly a more rapid HSR. This theory further supports the idea that cold sensitive plants can be improved by modifying the HSR through breeding or genetic engineering (discussed in chapter 5).

3.5 Conclusion
The results detailed in this chapter describe a number of significant conclusions concerning MsHSFA4 activity in alfalfa. This HSF has the ability to bind HSEs in the promoters of HSP genes and activate gene transcription. It possesses a 5’ UTR that acts as a negative regulator in tobacco protoplasts. MsHSFA4 is inducible by heat stress, cold stress and fall field stress. Under the aforementioned stresses MsHSFA4 is active in shoot apical tissue, crown bud and root tissue. Under non-stress conditions, transcript levels of this HSF fluctuate over a 24 hr period.