Chapter 1: History & Background

Stress is an integral component of the forces that drive the course of evolution. Changes in the physical environment generate stress, which in turn affects homeostasis. Without stress there would be little change outside of random mutation or genetic drift and the primordial soup would have remained just a soup. It was rapid environmental change that led to the massive diversification of life on Earth. Thus, the study of biological stress encompasses the nature of a changing environment, the mechanisms living organisms employ to survive and the resulting evolutionary effect.

1.1 Temperature Stress
Temperature stress exists in many forms both within and between the different climates of the world. In particular, plants that exist in regions of higher latitudes must endure dramatic seasonal temperature change, which can span from -40 to +40°C. At one extreme, the effects of freezing can induce dehydration, ice nucleation, protein inactivation and in extreme cases, cell death (Levitt, 1971). Alternately, plants exposed to excessive heat can also succumb to dehydration, protein degradation and cell death (Nover, 1991). Plants have evolved mechanisms to cope with stresses in a particular niche. Typically, this specific regional adaptation serves plants well, but can leave them susceptible if they are relocated to regions with different stresses. Thus, in new environments, the effects of temperature stress become more pronounced. This translates into significant limitations for agriculture as it expands onto marginal lands or into different climates. Understanding the mechanisms used to counteract temperature stress will allow researchers to breed and engineer improved plants, thus extending the current climatic limits of agriculture.

Research on low temperature stress has revealed several different protective mechanisms; some of which are common to both plants and animals and others that are unique to plants. Unlike animals, plants are essentially fixed in their location, and cannot move to avoid the effects of low temperature. Instead, plants exhibit unique patterns of gene expression that cause changes in cellular and organismal physiology to endure the effects of low temperature stress (Nover 1991; Levitt, 1971). This response to low temperature is not a single physiological pathway but a conglomerate of many pathways. On further observation, low temperature responses are intertwined with responses from other stresses such as drought, salt, oxidation and heat

Research into heat stress has revealed an evolutionary mechanism which is conserved among a diverse range of organisms including bacteria, animals, fungi and plants (Feder and Hofmann, 1999). During this “heat shock response” (HSR) organisms respond to heat stress by inducing specific patterns of transcription and translation that confer basal and acquired thermotolerance (Nover, 1991). More specifically, the HSR helps maintain cellular homeostasis during heat stress by preventing proteins from losing their functions. In eukaryotes, the HSR has a variety of physiological functions during both stress- and non-stress conditions, indicating the importance of the HSR to cell function (Nover, 1991). The expression of a partial HSR can be induced upon exposure to environmental stresses other than heat, including drought, oxidation, heavy metals and cold (Nover 1991). Furthermore, the HSR carries the characteristic of cross adaptation, that is, induced resistance to one stress can provide resistance to another stress.
The apparent relationship of the responses to high and to low temperatures signified the need to unify the two areas of research. Presently, temperature stress investigations are uncovering control mechanisms that govern numerous response pathways. Identifying the communication nodes among these pathways may reveal the complete network of cellular stress response mechanisms.

1.2 The Heat Shock Response
In the early 1960s, the heat shock response was serendipitously discovered in the cells of Drosophila melanogaster salivary glands (Ritossa, 1962). Heating these cells induced puffs to form at various regions of the polytene chromosomes. On closer inspection, these puffs were found to be areas of localized transcription that correlated with the increase of several large families of proteins, referred to as heat shock proteins (HSPs). This heat induced transcription and protein accumulation was then termed the heat shock response (HSR). To date, the body of research on the HSR comprises over 12,000 references and the data continue to grow.

Soon after its discovery, researchers began using this response as a model system for studies of gene expression. Over the last few decades, researchers collectively studied the HSR in a variety of organisms from bacteria to higher eukaryotes and made several important observations: 1) Heat shock genes and heat shock proteins exhibit a high degree of conservation across a variety of organisms (Lindquist, 1986; Schlesinger et al., 1982); 2) The HSR can be induced to different degrees by many agents, including elevated temperature, heavy metal ions, drought, oxidative stress, low water potential, anoxia and low temperature stress (Cabané et al., 1993; Lindquist, 1986; Czarnecka et al., 1984); 3) The organism’s ability to endure these stresses may be the result of HSP chaperoning capabilities (Parsell and Lindquist, 1993); 4) Some heat shock genes are known to be up-regulated during meiosis, constitutively expressed under non-stress conditions and up-regulated during other non-stress cellular activities (Atkinson et al., 1993; Bouchard et al., 1993). These characteristics indicate the essential role that heat shock genes play in adaptation and development and suggest that they probably developed early in the evolution of animals and plants.

Presently, HSR research continues to reveal the diversity of HSP functions. For example, it is now known that in addition to environmental stress, HSPs play a considerable role in plant and animal development (Atkinson et al., 1993; Bouchard et al., 1993; Heikkila, 1993), the mammalian immune response (Feige and Mollenhauer, 1992) and many aspects of mammalian disease (Morimoto et al., 1994). Animal HSPs are up-regulated under motional stress, mechanical stresses (Sironen et al., 2002), and both high blood pressure and cardiac failure (Schafler et al., 2002). Parasites and symbionts also require and augment host HSPs to cope with cellular stress (Feder and Hofmann, 1999). The induction of HSPs, specifically HSP 70, can extend the lifespan of Drosophila (Tatar, 1997). In Homo sapiens, mental stress has also been shown to induce HSPs (Lewthwaite et al., 2002). HSPs have also acquired a role in the controversial prion debate, which implicates damaged HSPs (which fail to properly refold proteins) as the indirect source of infectious prion proteins (Kryndushkin et al., 2002). HSPs are induced in both plants and animals by trace amounts of many different toxins, which make HSPs an excellent indicator of environmental pollutants. Moreover, observations of the HSR are currently utilized as a bioindicator to detect water and soil toxins (Krasko et al., 1997; de Pomerai, 1996).

When considered at the molecular level, HSPs have been implicated in the regulation of gene expression, the regulation of translation (Pal, 1998), inhibition of apoptosis (Samali and Orrenius, 1998) and, in some cases, altering chromatin structure by binding DNA (Soti and Csermely, 1998). Additionally, these proteins are now being employed to engineer personalized vaccines against cancer and infectious diseases (Srivastava, 1998).
We have only begun to understand the full nature of the HSR, specifically the functions and abilities of HSPs. However, based on what is already known, this response is essentially universal among organisms and thus fundamental to life.

1.3 Heat Shock Protein Families
In eukaryotes, heat shock genes are classified into a number of families that are named principally for the average molecular weight (kDa) of the protein products they encode. The families primarily include HSP 100, HSP 90, HSP 70, HSP 60, HSP 40 and a large family of small (or low molecular weight) heat shock proteins (sHSPs) that range from 12-30 kDa (Table 1.1). Many family members have counterparts, referred to as heat shock cognates (HSCs), that are expressed under normal non-stress conditions. New HSPs and HSP isoforms are discovered regularly and to date there are 11 known HSP subfamilies (Nover and Scharf, 1997). The function and structure of HSPs varies dramatically between and within different families, but phylogenetically, they are highly conserved.

The majority of HSPs are functionally linked to cellular chaperone activities but they are also involved in numerous cellular activities. HSPs are expressed during heat stress and at specific developmental stages, and some are localized within specific organelles. HSPs are implicated in a vast number of cellular processes, including stress damage protection, protein maturation, trafficking of newly synthesized proteins, the assembly of protein complexes and the functioning of protein degradation machinery. It has been demonstrated that members from the HSP 70, HSP 60 and HSP 40 families are required at different stages within protein folding systems (Frydman and Hartl, 1996; Frydman et al., 1994). Other HSPs are essential for the recovery of stress-damaged proteins that have precipitated out of the cytosol (Parsell et al., 1994), involved in signal transduction (Kimura et al., 1996), and have been found to interact with other HSPs (Schneider et al., 1997).

The HSP 70 family has been the most widely studied of all the heat shock proteins. In addition to protein folding, HSP 70s are involved in several cellular activities including protein degradation, reorganization of cytoskeletal components, translation initiation, nuclear protein import and export, ribosome assembly, chromatin structure and DNA synthesis (Forreiter and Nover, 1994 and references within). HSP 70’s have also been implicated in the cytosolic regulation of HSF monomers (discussed below) (Schöffl et al., 1998). During the recovery of denatured proteins, HSP 70 is believed to function in an ATP-dependent complex involving HSP 40 and BAG-1 (Nover and Forreiter, 1998). This complex binds, contorts and releases non-native proteins in a cyclic process that continues until the non-native proteins have regained their functional conformation. However, the HSP 70 chaperone machine (HSP 70/HSP 40/BAG 1) has no measurable effect on protein renaturation on its own. It requires another HSP protein complex, namely HSP 20, to recover protein function (Forreiter and Nover, 1998).

Members of other HSP families (HSP 60, HSP 90 and HSP100) also form protein complexes. HSP 60s can form barrel-like structures for the purpose of protein refolding. The barrel resembles a truncated tube that pumps denatured proteins through its interior in a unidirectional ATP-dependent manner (Forreiter and Nover 1998). HSP 90s prevent protein precipitation and are involved in hormone signal transduction (Forreiter and Nover 1998). Reminiscent of the HSP 70s, HSP 100 complexes can also refold denatured proteins and resolubilize precipitated proteins (Nover and Forreiter, 1998; Parsell et al., 1994).
In plants, the sHSPs are the most abundant, the most diverse and the least characterized of the HSP families. Plant sHSPs have been organized into four classes based on amino acid similarity (Vierling, 1991) (Table 1.1). Classes I and II include cytoplasmic sHSPs, class III includes sHSPs that are localized to the chloroplast and class IV includes endomembrane sHSPs. As with other HSP families, sHSPs are capable of forming large aggregates during their cellular activities.

HSP 20s function in conjunction with HSP 70s to prevent heat induced protein aggregation (Jinn et al., 1995). Under stress, HSP 20 oligomers coalesce to form complexes ranging from 200-800 kDa that are visible as cytosolic granules of approximately 40 nm in diameter (Nover and Forreiter, 1998). It is believed that these complexes form a net that traps denatured proteins and guides them to the HSP 70 chaperone machine for refolding. In vivo experiments involving heat denatured luciferase revealed that HSP 17.6 and HSP 70 are both required to recover heat inhibited luciferase activity (Forreiter et al., 1997).

In addition to stress related activities, some sHSPs are up-regulated during meiosis and development (Atkinson et al., 1993). They also exhibit tissue-specific expression as seen in the heat shocked meristematic tissue of maize (Zea mays L.) seedlings (Greyson et al., 1996). Furthermore, some sHSPs (specifically HSP 18) are expressed in rapidly cycling undifferentiated callus cells. This indicates that HSP function is universally required during stress in physiologically active cells, regardless of tissue differentiation (Friedberg and Walden, 1999). It also demonstrates a strong correlation between HSPs and mitosis, signifying that HSPs are necessary to protect mitotically active cells from heat damage (Friedberg et al., 1998).

1.4 Heat Shock Genes
Heat shock genes (HSGs) are classified into two main groups based on the nature of their transcriptional activation. The first group is transcribed in response to heat shock, and its products are referred to as HSPs. Those in the second group are constitutively transcribed under non-stress conditions and are referred to as heat shock cognates (HSCs). HSGs are further divided into sub-classes according to their cellular targeting (Czarnecka-Verner et al., 1994).

Promoters of HSGs possess several regulatory elements that are required for transcriptional activation. The ability of heat shock promoters to respond to heat is credited to the presence of a consensus sequence, which is located in the proximal and distal up-stream promoter region and is bound by heat shock transcription factors (HSFs). This DNA recognition sequence is referred to as the heat shock element (HSE) and consists of pentanucleotide repeats with a core sequence of 5'-nGAAn-3' (head module) or its inverse complement, 5'-nTTCn-3' (tail module) (Pelham, 1982). These repeat core elements are arranged in alternating orientations and can be found in multiple copies upstream from the TATA box element (Barros et al., 1992). In plants, the most optimal HSE consensus sequence was found to be 5’-aGAAg-3’ or 5’-cTTCt-3’. Deletion analysis has revealed that HSEs in the distal promoter have redundant function, but in the proximal promoter, at least two HSE sites are required for full heat inducible activity (Czarnecka et al, 1989). Although the HSEs with highest degree of conservation to the consensus sequence provide the strongest HSF binding, numerous variations of the HSE core sequence exist that allow significant levels of HSF binding (Nover et al., 2001). Several other response elements have been identified in HSG promoters and may be required for its activation. Some of these sequences include the GAGA sequence which is bound by the GAGA binding factor, scaffold attachment regions (SARs), CCAAT boxes and AT-rich sequences. The presence of these elements and their interacting proteins indicate that the modification of chromatin structure may also be a requirement for activation. It has been hypothesized that the GAGA sequence and the SARs alter the chromatin structure so that the TATA binding proteins (TBPs) can gain access to the TATA box to recruit the transcription preinitiation complex (PIC) (Tsukijama et al., 1994). The presence of the CCAAT box may elicit a degree of heat inducibility but is also required for HSC basal transcription. The AT-rich sequences found upstream of the TATA box frequently exist in TAAT or ATTA repeats and they may function in a similar manner to intergenic AT-rich sequences. These AT-rich sequences are bound by high mobility group proteins (HMGs) and AT binding factors (ATBFs) which are believed to increase transcription through interaction with nuclear scaffold proteins (Czarnecka-Verner et al., 1994).

1.5 The Heat Shock Transcription Factors
The heat shock transcription factor (HSF) regulates the expression of the majority of HSGs (Nover et al., 2001; Schöffl et al., 1998). HSFs have remarkable structural similarities among animals and plants, but there are significant differences in the complement and activity of members of the HSF gene family. Some organisms seem to have only one HSF, such as yeast (Wiederrecht et al., 1988) and Drosophila (Clos et al., 1990), but most have more. Animal HSF genes have been isolated from a number of species including Drosophila melanogaster (Clos et al., 1990), Xenopus laevis (Stump et al., 1995), Gallus domesticus (Nakai and Morimoto, 1993), Mus musculus (Sarge et al., 1991) and Homo sapiens (Rabindran et al., 1991). From all the animals surveyed and with the aid of the complete sequence of the human and mouse genomes, animals appear to contain up to four members in the HSF family designated HSF 1 to 4 (Pirkkala et al., 2001; Scharf et al., 1998) (Table 1.2). Among these four classes, uncharacterized isoforms, generated by alternative splicing (denoted a and ß) have been identified. In some species, a number of these isoforms exhibit tissue specific expression. For example, in some vertebrates, HSF 2 appears to be developmentally regulated, while HSF 3 appears to be avian specific and HSF 1 is universally induced by heat stress (Pirkkala et al., 2001).

In contrast to animals, plant HSF gene families are more complex. For clarity, a different nomenclature has been employed (Nover et al., 2001; Nover et al., 1996) (Table 1.2). Plant HSF gene family members have been characterized from Lycopersicon peruvianum (Scharf et al., 1990), Arabidopsis thaliana (Hübel and Schöffl, 1994), Zea mays (Gagliardi, et al., 1995), Glycine max (Czarnecka-Verner et al., 1995), Pisum sativum (Aranda et al, 1999), Phaseolus acutifolius (GenBank direct submission), Nicotiana tabacum (Shoji et al., 2000), Oryza sativa (GenBank direct submission) and Medicago sativa (Friedberg et al., 2003). Genome sequencing projects have revealed the size of plant HSF gene families, with up to 21 currently known members observed in Arabidopsis (Nover et al., 2001). Plant HSF genes have been divided into classes A, B and C based on homology analyses of several highly conserved regions such as the DNA binding domain and oligomerization domains (Nover et al., 2001; Czarnecka-Verner et al., 2000). Within these classes, plant HSF genes can be further organized into subclasses based on differences in amino acid sequences. Among these classes HSFs can be identified by expression patterns involving genes that are constitutively expressed, exhibit tissue specific expression or are heat induced. The presence of class B HSFs is particularly intriguing as they are not found in animal systems and this constitutes a striking difference between the two kingdoms.

In both animals and plants, protein alignments of known HSFs have revealed regions of high conservation (Figure 1.1) involving a DNA binding domain (DBD) at the N-terminus , an oligomerization domain (OD) consisting of two hydrophobic heptapeptide repeat regions denoted HR A and HR B, and a C-terminal domain (CTD). Of these regions, the DBD has the greatest conservation and has become the hallmark of HSFs, distinguishing them from other families of transcription factors.

The DBD of the HSF protein monomer consists of an antiparallel four stranded ß-sheet (a1-a4) and three alpha helices (a1-a3). The DBD forms a helix-turn-helix (HTH) structure (Harrison et al., 1994) that is essential for HSF binding of HSEs in HSG promoters. In addition, HSFs have a sole intron of variable size located within the HTH motif of the DBD in the C-terminal region of the third helix (a3).

The oligomerization domain (OD), located C-terminally from the DBD (Figure 1.1), is composed of HR A and HR B, and is the region that connects HSF monomers into a trimer configuration (Czarnecka-Verner et al., 1994). The OD is crucial for trimerization and only trimerized HSFs can be imported into the nucleus (Westewood, 1991). However, in contrast, there is evidence that class B HSFs, specifically subclass B1 HSFs, function as dimers (Nover et al., 2001). The OD is believed to form a parallel triple stranded a-helical coiled coil in which the HR A trimerizes through intermolecular interactions between three a helices. HR B is thought to buttress up against the HR A a helix and actually stabilize the trimeric structure. This domain is essential for HSF activity, as HSF DNA binding ability is increased by three orders of magnitude in the trimer configuration (Hardy et al., 2000). Oligomerization is thought to be regulated by a third hydrophobic region (HR C, located in the CTD) and also by specific HSPs. Under non-stress conditions, HSF monomers maintain a folded conformation through intermolecular interactions between the HR C and the HR A/HR B region. It is believed that this folded state is maintained by HSPs (predominantly HSP 90 and perhaps HSP 70) (Kim and Schöffl, 2002). Upon stress, HSP 70 and HSP 90 are sequestered to other cellular activities, freeing (i.e. derepressing) the folded HSF monomers and thereby allowing trimerization (reviewed by: Scharf et al., 1998; Schöffl et al., 1998; Czarnecka-Verner et al., 1994).
Variation among HSFs is primarily localized to the C-terminal activation domain, which includes a nuclear localization signal (NLS), a region of hydrophobic repeats (HR C) and one or more regions of conserved amino acids residues. These amino acids consist of aromatic, bulky hydrophobic and acidic residues, referred to as the AHA motif (reviewed by Nover and Scharf, 1997). The AHA motif has been shown to be critical in the transcriptional activation of class A HSFs. In addition, it is thought to physically interact with the TATA-binding protein and the TFIIB complex, which is crucial for forming the transcription pre-initiation complex (Czarnecka-Verner et al., 2003; Czarnecka-Verner et al., 2000; Döring et al., 2000). A fundamental difference between class A and B HSFs resides in the structure of the CTD. Class B HSFs are largely truncated in the CTD. They still possess an NLS, but are missing most of the CTD, including the HR C and the AHA motifs (Figure 1.1).

Class A and B HSFs exhibit a spectrum of activities at the RNA level, but their abilities to act as transcriptional activators seems to be specific to the Class A HSFs (Czarnecka-Verner and Gurley, 2002). Although class B HSFs can still bind to the HSE in the HSG promoter region, they seem unable to functionally activate the transcription and are therefore thought to act as attenuators of active HSGs (Czarnecka-Verner et al., 2000; Czarnecka-Verner et al., 1997).

1.6 Regulation of the Heat Shock Response
The HSR is a remarkably powerful genetic system. Under heat stress, normal gene activity is inhibited and HSGs are preferentially up-regulated. It is believed to be an emergency response mechanism that has evolved to rapidly collect and divert all cellular resources to protect and sustain the cell. This is exemplified in the general structure of HSGs and the eloquent system of activator and or repressor HSFs that are employed to activate and attenuate the complete response.

There are a number of key features that have evolved to increase the speed and accuracy of this response. First, HSGs generally have very few introns, and those that are present are quite small. This shortens the time required to transcribe and process pre-mRNA into a mature mRNA. Second, although only two HSEs are required to activate transcription, HSGs tend to posses many HSEs to provide a strong binding site for HSFs and ensure activation of the response. Third, the cooperative activity of class A and B HSFs, through concentration based competitive binding to HSEs allows for rapid activation and attenuation of transcription. Fourth, plants use class B HSFs to bind HSEs and repress HSGs, which keeps the localized chromatin unwound and thereby exposing the promoter. Essentially, this keeps the promoter primed and ready for activation (discussed below). Collectively, The evolution of these genetic features have produced a response so rapid that HSG transcripts are easily detectable within three minutes of exposure to heat stress (Key et al., 1985; Schöffl and Key, 1982).

The regulation of the HSF and its control of the HSR are still not fully understood. However, cumulative research has resulted in two working models to describe the activation and attenuation of the HSR in animals and plants (Schöffl et al., 1998; Wu 1995; Czarnecka-Verner et al., 1994). The animal system is based on a negative feedback inhibition model, whereas the plant system involves a negative feedback and competitive regulation model (Figure 1.2). In animal systems, the regulatory model has three stages:

Stage 1: Repression
Under non-stress conditions, constitutively expressed HSF monomers are repressed in the cytoplasm by HSPs. It is believed that the HSF is folded such that HR C interacts with HR A and HR B, which essentially masks the oligomerization domain and prevents trimerization.

Stage 2: Activation
With the onset of stress, the repressing HSPs are sequestered to attend to stress denatured proteins, thereby freeing up (derepressing) the HSF monomers. HSF monomers gain the ability to bind DNA by trimerizing, which allows them to be imported into the nucleus and subsequently find their target HSEs. The acquisition of transcriptional competence occurs through one or a number of mechanisms, but it may involve a phosphorylation to further unfold or modify the CTD, or to modify associated proteins in the transcription preinitiation complex.

Stage 3: Attenuation:
Post stress HSPs complete their stress related activities and begin to bind, fold and repress HSF monomers. Consequently, levels of active trimerized HSFs fall and monomers accumulate. Once below critical HSF trimer concentrations, the HSG transcription shuts off, returning the system to a ready (control) state.

The current model for the activation and attenuation of the HSR in plants is similar to the animal model, but involves three stages situated in two cellular locations (Figure 1.2).

Stage 1: Repression (Figure 1.2, A)
Under non-stress conditions, constitutively expressed class B HSFs (repressors) are present in the nucleus and are bound to HSEs in the promoters of HSGs, thus inhibiting gene transcription. Concurrently, constitutively expressed class A HSF are folded and repressed by HSPs in the cytoplasm. Immunolocalization experiments have revealed that under non-stress conditions class B HSFs are only present in the nucleus and class A HSFs are present only in the cytoplasm (Nover and Forreiter, 1998). Further support for this model comes from the structural analysis of class B HSFs. They lack much of the CTD regulatory region, indicating that trimerization cannot be repressed due to the lack of HR C, and that transcriptional activity is eliminated due to the lack of AHA motifs.

Stage 2: Activation (Figure 1.2, B, C & D)
During stress, constitutively expressed class A HSFs are derepressed, trimerized, imported into the nucleus and begin to out-compete the class B HSFs that are bound to HSEs. Once a critical level of trimerized class A HSF binding is achieved compared with the class B HSFs, HSG transcription begins. It is hypothesized that that class A HSFs out-compete class B HSFs through the advantage of varying affinities built into their DBDs and the use of cooperativity through concentration based binding (Czarnecka-Verner et al., 1997). To further enhance the activation cascade, some stress induced class A HSFs are transcribed which may be further utilized to greatly increase HSP gene transcription.

Stage 3: Attenuation (Figure 1.2, E)
To attenuate the HS response, stress induced class B HSFs are imported into the nucleus and out-compete the class A HSFs bound to the HSEs. This initiates the shutdown of HSG promoters. In the cytoplasm, there is an excess of newly synthesized stress-induced HSPs that bind class A HSFs, repressing trimerization and returning class A HSFs to their monomeric state (Kim and Schöffl, 2002; Schöffl et al., 1998). Accumulations of constitutively expressed class A and B HSFs return to normal non-stress levels (Scharf et al., 1998) leaving the system primed and ready for the next round stress.

In both the animal and plant models, HSP 70 and HSP 90 facilitate the cytosolic repression of constitutively expressed HSF monomers. Much of the support for this model comes from immunohistochemical experiments observing HSP and HSF localizations during heat stress and recovery (Schöffl et al., 1998).

Plant and animal systems exhibit significant differences in their regulation, specifically the involvement of phosphorylation (which may activate HSFs in the animals system), the lack of class B and C HSFs in the animal system, and the limited number of HSFs in animals (4 HSFs) as compared with plants (up to 21 HSFs) (Satyal and Morimoto, 1998; Nover and Scharf, 1997). It remains a mystery as to why plants and animals evolved divergent regulatory systems in response to the same stresses. A possible explanation may lie in the fact that animals are mobile, while plants are often anchored. Hence, animals can regulate body temperature while plants must endure unregulated temperature change. Moreover, in plants the HSR is likely integrated with other stress response systems (oxidative stress, etc. see below), or portions of the HSP complement may be needed for cold or drought protection. Thus, of the 21 potential HSFs in plants, there may be a high degree of specialization (i.e. some may be required for specific HSG activation and others may only be required for specific stresses).

1.7 The Arabidopsis HSP & HSF Families
The complement of HSPs varies between kingdoms. For example, the smaller HSP families are the most abundant in plants but are not in animals and yeast. The susceptibility of plants to environmental extremes, due to their sessile and poikilothermic existence, probably contributed to the differences in the plant HSP complement as compared to animals. The sequencing of the Arabidopsis thaliana genome has illuminated a nearly complete complement of plant HSPs Described below are important highlights from the complement of HSPs identified within the Arabidopsis thaliana genome.

The HSP 100 family contains eight genes distributed among all five chromosomes and has corresponding proteins ranging from 92.7 kDa to 108.7 kDa (reviewed by Agarwal et al., 2001) (Table 1.1). Previously, Arabidopsis was thought to have only one HSP 100 family member, but organelle targeting sequences indicate that three members exist in the cytosol and the other five are targeted to plastids. All eight family members of HSP 100 are expressed under heat stress.

In the HSP 90 family, seven members have been identified, six of which were previously known (reviewed by Krishna and Gloor, 2001) (Table 1.1). They range from 80 to 90 kDa and all have a basal level of expression that greatly increases during heat shock.

The HSP 70 family (also known as DnaK), the most studied of all the families, contains 18 members and the genes are distributed among all five chromosomes. Members are targeted to the cytosol, plastid, mitochondria and endoplasmic reticulum. Most family members are heat inducible, but exhibit a wide range of tissue and developmentally specific expression patterns (reviewed by Lin et al., 2001).

The HSP 60 family, also known as chaperonins, has 29 members (Table 1.1). They are targeted to the mitochondria (type I), the chloroplast (type I), the cytosol (type II) and the genes are found among all five chromosomes. Most of the family members are 60 kDa, except for one member that is 21 kDa and four others that are 10 kDa (reviewed by Hill and Hemmingsen, 2001).

Analysis of the HSP 40 family, also known as J-domain proteins, has revealed a surprising 89 members spread throughout the genome (Miernyk, 2001) (Table 1.1). Most of the members are approximately 40 kDa but a few obscure members range from 17 to 277 kDa. Members localize to the cytosol, nucleus, mitochondria and throughout the secretory pathway. The genes have a wide range of expression patterns including constitutive expression, heat responsive expression and circadian or diurnally regulated expression (reviewed by Schaffer et al., 2001).
Compared to animals, the sHSPs in plants are the most abundant HSPs by weight. They are an ubiquitous family of proteins that range from 12 to 30 kDa (reviewed by Scharf et al., 2001) (Table 1.1). Most are heat induced and the genes can be found among all five chromosomes. Previously, only eight sHSPs were identified, but now 35 open reading frames (ORFs) have been found in Arabidopsis.

In the HSF family, 21 members have been identified (reviewed by Nover et al., 2001). Class A contains 15 members, class B contains 5 members and class C has only 1 member (Table 1.1; Table 1.2). Family members are found throughout the genome and exhibit a variety of expression levels. The HSF proteins vary in size from 28.3 to 55.7 kDa.

Through the current survey of Arabidopsis genome it appears that the HSR encompasses at least 207 genes (Table 1.1), but due to technical limitations, some probably have yet to be identified. Knowing the complete complement of genes in the HSR provides a unique perspective with which to study the dynamic activity of the HSGs and overall stress response in plants.

1.8 Low Temperature Stress in Plants
In the past century, research on low temperature stress in plants has evolved into two main areas of study: 1) cold acclimation: a change in physiology conferring protection against sustained low temperature stress (reviewed by Thomashow, 1999); and 2) cold shock: the rapid response to decreases in ambient temperature in non-acclimated plants (reviewed by Guy et al., 1997). Research has tended to focus on either the changes in physiology or the changes in gene expression and researchers have discovered some key features about low temperature stress. First, low temperature stress is composed of several co-stresses, including freezing, chilling, dehydration, and disease susceptibility. Second, some low temperature co-stresses overlap with other non-low temperature stresses; for example, cold shock and heat shock both exhibit dehydration stress. Third, several different genetic pathways are active during cold shock and cold acclimation, including the CBF/DREB (C repeat binding factor / dehydration responsive element) gene cascades, oxidative stress response, HSR, etc. Fourth, observable changes in physiology include; oxidative damage, membrane damage due to crystallization, cold labile enzymes, denatured proteins and cytoplasmic dehydration.

The current challenge is to determine the link between pathways and unite the genetics with the physiology. Specifically, to resolve the hierarchy, where low temperature stress causes a change in gene expression. This results in the production of stress related proteins that alter the cellular physiology, thus providing protection. At the top of this hierarchy, a number of different families of genes have been observed to be active during both cold shock and in cold acclimated plants (reviewed by Thomoshow, 1999). These genes include fatty acid desaturases, molecular chaperones, mitogen activated (MAP) kinases, calmodulin-related proteins, late embryogenesis proteins (LEA), COR (cold regulated) proteins, absisic acid (ABA) responsive genes and a number of transcription factors.

The discovery of cold activated transcription factors was critical in determining the genetic starting point for cold responsive pathways. The CBF/DREB family of transcription factors is only induced during drought and / or low temperature stress. The duality of this family’s activity is indicative of the similarities in the physical strain caused by both drought and cold (i.e. loss of water from the cytosol). Transgenic Arabidopsis plants have been created that over-express CBF1 and preferentially up-regulate the COR genes (which play a role in preventing membrane damage). These plants show a small increase in freezing tolerance as compared to the non-transgenic control, indicating that CBF1 is, at least in part, is involved in the acquisition of freezing stress tolerance (Thomashow et al., 1998; Thomashow et al., 1997).

Most of the physiological damage from low temperature stress occurs in organelle and plasma membranes, or is related to dehydration (Thomashow, 1999; Maurel 1997; Nishida and Murata 1996). Under normal conditions, plasma membranes are composed of a protein-studded lipid bi-layer which acts as a lamellar fluid mosaic. As temperature decreases, phospholipids become more tightly packed and eventually shift from lamellar to a hexagonal II phase which is significantly more rigid. In a hexagonal II phase configuration, plasma membranes are prone to fracture and the membrane-bound proteins become retarded and lose function. Fracturing results in water loss which in turn, may cause further fracturing. The effect on membrane-bound proteins is most notable in the thylakoid photosystems. In this case, certain proteins must be able to move through the membranes to carry electrons from the p680 photosystem to the p700 photosystem. Malfunctioning photosystems can not keep up with the electrons coming in, and excess electrons are subsequently lost or converted to free radicals. The accumulation of free radicals can have devastating effects on the cell by oxidizing or reducing proteins and fatty acids, further compounding the damage (Moller, 2001). This explains the observed concurrent accumulations of enzymes such as superoxide dismutase, ascorbate peroxidase and glutathione reductase, which specifically absorb and remove free radicals from the cell (Noctor and Foyer, 1998).

The maintenance of membrane integrity during stress involves a number of genes. COR proteins are believed to prevent hexagonal II phases from forming by binding to membranes, allowing them to maintain their natural curved structure (Steponkus et al., 1998). The activity of desaturases adds double bonds to the fatty acid chains of phospholipids which increase the bulkiness of the phospholipids. This increased bulk decreases the temperature at which a hexagonal II phase will form. Certain HSPs help stabilizes cytoskeletal proteins that in turn support the membrane. In addition to the aforementioned proteins many other proteins are directly or indirectly involved in membrane stabilization (Steponkus et al., 1998). As these proteins are encoded by genes from different systems in the plant, it is clear that several genetic pathways are involved. Thus, researchers continue to look for other cold induced transcription factors to identify all the master switches activated during low temperature stress.
In order to cope with water loss, plants that can cold acclimate exhibit one of two strategies during extended periods of prolonged low temperature: supercooling and alterations in cell physiology. Super cooling is observed in many woody plants and is essentially the physical separation of any sensitive tissue from the rest of the plant. During the acclimation period, the plant sets up a heavily lignified and suberinified wall to isolate bud primordia from the rest of the plant. This prevents extracellular ice from penetrating the bud and acting as a nucleation site for continued dehydration and ice formation. The result is that any water trapped in the bud fluctuates with the ambient temperature, cooling to sub-zero temperatures without crystallization.

The second approach is to change the cellular physiology throughout the tissue (usually in the root and crown of the plant). This serves to prevent water loss due to extra cellular ice formation, which would lead to the formation of intracellular ice. In this case, plasma membrane and cell wall structures are altered to eliminate possible ice nucleation sites. Specifically, cell walls become heavily lignified and accumulations of proline and other metabolites are observed between the plasma membrane and the cell wall. The composition of soluble carbohydrates changes in the cytoplasm to shift the osmotic potential so that water will be forced to stay in the cell.

1.9 Low Temperature Stress & HSPs
As temperature decreases a number of plant proteins have been identified as cold labile (reviewed by Guy et al., 1998a). Specifically, some active proteins lose their structures and some newly synthesized proteins are unable to obtain their native conformation. These proteins include enzymes, structural proteins and multi-protein complexes. Logically, wherever there are denatured proteins, one expects to find HSPs and, thus, the activation of the HSR. Some elegant experiments have found HSPs to be up-regulated during low temperature stress and during cold acclimation (Guy and Haskell, 1987; Neven et al., 1992; Anderson et al., 1993). However, only selected members of HSG families are expressed, raising a question about the control mechanism under low temperature (discussed below).

The connection between HSPs and low temperature stress was first established from investigations into the cross adaptation of the HSR. These experiments revealed that inducing the protection of the HSPs under one stress can also protect during other stresses. Specifically, the increase of HSPs has been correlated with increased resistance to several diverse stresses (Collins et al., 1993; Lurie and Klein, 1991; Chen et al., 1982; Palta et al., 1981). Numerous groups continue to report induction of both HSP transcripts and proteins during low temperature stress (Guy et al., 1998; Sabehat et al., 1998 Krishna et al., 1995), drought (Wehmeyer and Vierling, 2000), high salinity (Bhagwat and Apte, 1989) and heavy metal toxicity (Wollgiehn and Neumann, 1999). The connection was solidified through experiments confirming HSP expression under chilling (0º - 10ºC) stress (Cabané et al., 1993; Collins et al., 1993) as well as HSP expression during cold acclimation (Ukaji et al., 1999; Neven et al., 1992). A heat shock pretreatment, causing HSPs to accumulate, was shown to reduce the effects of chilling stress damage in developing Lycopersicon esculentum fruits (Sabehat et al., 1998; Kadyrhanova et al., 1998) and Cucumis sativus cotyledons (Lafuente et al., 1991). The HSP mRNA expression and HSP accumulation during low temperatures clearly implicate the necessity for the presence of HSPs in cells in order for plants to cope with low temperature stress.

On a molecular level, HSPs appear to have three modes of action: 1) re-solubilizing precipitated proteins, 2) folding proteins and assembling protein complexes, and 3) stabilizing structural proteins, specifically those that interact with plasma membranes, thus preventing solute leakage. These concepts are based on expression levels and correlations, but definitive data has yet to be obtained. However, HSPs,are clearly active and required during low temperature stress.

The question remains; what is the control mechanism for the induction of HSGs under low temperature? It stands to reason that HSFs play a role, but only selected HSPs are expressed under low temperature (specifically, members from the HSP 90, 60, 70 and sHSP families). It is possible that the low temperature regulation of HSPs is connected through another pathway or, that the HSF regulatory system has the ability to differentiate between heat and cold and alter HSG expression accordingly. The latter possibility would explain the large number of genes in the HSF family in plants. Furthermore, some HSFs may be specifically adapted to work under low temperature stress conditions. One potential candidate is the Glycine max GmHSFB4 (formerly GmHSF5) whose RNA is expressed at lower temperatures and is actually degraded upon heat stress (Czarnecka-Verner et al., 1995).

1.10 Temperature Stress Medicago sativa L.
Alfalfa, endearingly known as the “Queen of the Forages,” has a long history of cultivation and use as a primary forage. Cultivated types are believed to have originated in the areas of Asia Minor, Transcaucasia, Iran and the Highlands of Turkmenistanis - regions that exhibit cool winters and hot dry summers. Cultivation is thought to have begun around 8000 to 9000 years ago (Ivanov, 1977) and since then, cultivars have been transported all over the world. Alfalfa is a backbone of modern agriculture and in Ontario it is the third most widely grown plant by acreage. Its primary uses include feed for animals and due to its symbiotic nitrogen fixing abilities, a plow-down source of nitrogen in crop rotations.

Alfalfa is perennial and a tetraploid (2n = 4x = 32) which is both advantageous and disadvantageous in alfalfa production. As a perennial, alfalfa remains productive for several years, but when grown in northern latitudes it must endure hot summers and harsh winters. Breeding programs have adapted alfalfa to live in northern regions, but during severe winters, alfalfa can exhibit freezing damage and severe winter kill, resulting in significant yield loss. As a tetraploid, the genetic redundancy of having four homologues allows for the retention of deleterious mutations that result in severe inbreeding depression, as well as the inability to generate homozygous or hybrid lines. However, tetraploid alfalfa offers breeders a wide variety of allelic combinations and the ability to constantly maintain a completely heterozygous population - a distinct advantage over diploid breeding programs.

Alfalfa, like most annual and perennial field crops, is susceptible to heat stress. The primary effect of heat stress is water loss, which translates into reduced yield. Breeding programs have produced alfalfa cultivars that are well adapted to hot climates, but there is limited ability to increase yields under arid conditions. Most cultivars can withstand short periods of heat stress under which vegetative growth halts, but quickly reverses under normal conditions. In regions of hot climates and low rainfall alfalfa productions systems are frequently supported by irrigation.
The effects of winter damage have profound effects on perennial production systems in the agricultural community. The trait of winter hardiness, defined as a plant’s ability to survive the entire complement of winter stresses (McKenzie et al., 1988) has been the quest of breeding programs for over a century. However, due to the complexity of the genetics behind winter hardiness, little progress has been made. From the research amassed on expressed genes during winter stress, researchers have incorporated transgenic technology in to breeding programs to vastly augment alfalfa’s traits.

The spectrum of alfalfa population genetics ranges from dormant (dormancy rating 0) to non-dormant (dormancy rating 11) genotypes. Fall dormancy ratings are based on fall regrowth (Barnes et al., 1978). As winter approaches plants that can become dormant will slow their vegetative growth. Dormancy itself refers to the physiological state induced to cope with winter stresses, or is the end result of the process known as cold acclimation (reviewed by Thomashow, 1999). To clarify, dormancy is a physiological state and fall dormancy refers only to the slowing of vegetative regrowth during the fall. The dormancy and fall dormancy need not be dependent on each other and thus, some researchers have suggested that the two traits can be bred independently (Brummer at al., 1999).

Cold acclimation provides plants with the characteristic of winter hardiness. In the last century, considerable amounts of research have been devoted to understanding low temperature stress, what stimulates cold acclimation and how to improve winter hardiness. Based on this research, several important discoveries have been made: 1) Cold acclimation is triggered by environmental factors, such as temperature and photoperiod (Palva, 1998); 2) Survival is directly related to root size, crown size and carbohydrate abundance (Schwab et al., 1996), and 3) Increased winter hardiness can be achieved by engineering plants to increase the expression of enzymes including superoxide dismutase (SOD) and alcohol dehydrogenase (ADH) which reduce free radicals (McKersie et al., 2000; McKersie et al., 1993).

Building on our existing knowledge, more detailed investigations into the mechanisms with which alfalfa responds to low temperature and high temperature will fuel new approaches for developing stress resistant plants. The hope for this collective effort being to improve alfalfa production, utility and extend its’ climatic growth limits.

1.11 Research Hypothesis
The expression of HSPs is an integral part of a multi-pathway system that protects plants during exposure to various forms of low temperature stress. From the full HSP complement induced during the HSR, only a subset of those HSPs are expressed under low temperature, which suggests one of two hypothesizes for low temperature regulation of HSPs: 1) HSPs are regulated by another cold induced pathway independently of the HSFs or 2) Some HSFs have been adapted to act specifically during low temperature stress and have specificity built into their DNA binding domains for the promoters of cold induced HSGs. Identifying low temperature regulated HSFs will aid in the understanding of how the HSR fits into the multi-pathway low temperature protective system. My research was to identify and characterize a potential transcriptional regulator for HSP induction under low temperature stress and observe the deferential response to heat and cold stresses.

My objectives were as follows: 1) To clone and characterize a native alfalfa heat shock transcription factor 2) To analyze levels of HSF transcripts, HSP transcripts and to test HSF and HSP activities under laboratory heat stress, cold stress and field low temperature stress, and 3) To assess the utility and potential of a genetically engineered HSF over-expressed in transgenic alfalfa plants to increase low temperature stress tolerance