Thellungiella has been proposed as an alternative model species to Arabidopsis to investigate plant abiotic stress tolerance mechanisms. Thellungiella shares many features with Arabidopsis that make it an attractive candidate for both physiological and molecular studies [14, 21, 29]. The main argument in favor of Thellungiella, however, is that it is considered an “extremophile” that is much more tolerant to various stresses than Arabidopsis. On the other hand, it has been shown that there is considerable natural variation between different accessions of Arabidopsis that results in different levels of tolerance under various environmental growth and stress conditions (see e.g.  for a recent review). This natural variation has been investigated most extensively for cold acclimation and freezing tolerance [7, 8, 10, 12, 13]. Since natural accessions are also available for Thellungiella this opens the unique possibility to directly compare the range of stress tolerance and possible differences in adaptive mechanisms between these species.
In the present study, we have for the first time compared the range of natural variation in the freezing tolerance of Arabidopsis and Thellungiella. We conclude from the wide overlap in the freezing tolerance that at least with regard to this trait Thellungiella should not be considered an extremophile. Its range of freezing tolerance, however, extends to lower temperatures than that of Arabidopsis with about one-third of the available Thellungiella accessions more freezing tolerant than any Arabidopsis accession. The acclimated freezing tolerance of Thellungiella was positively correlated with the average minimum habitat temperature recorded during the coldest month of the growth season, consistent with previous results for Arabidopsis[7, 12].
Only the freezing tolerance of the Yukon accession of Thellungiella has previously been reported in the literature . LT50 values of −13°C for nonacclimated and −18.5°C for cold acclimated plants were recorded when whole-plant survival was evaluated. These temperatures are substantially lower than the −6.4°C (NA) and −11.7°C (ACC) obtained from our electrolyte leakage measurements. However, corresponding electrolyte leakage data in  suggest a similar temperature range to our results although no LT50 values were given. In addition, since no direct comparison with Arabidopsis was presented, any comparison between the species remained speculative in this paper.
From the comparison presented here we suggest that although Thellungiella may not be an extremophile with regard to freezing tolerance, its range of freezing tolerance after cold acclimation clearly extends beyond Arabidopsis. We therefore consider Thellungiella a useful additional model species to identify superior or alternative freezing tolerance mechanisms.
During cold acclimation in Arabidopsis, the composition of the metabolome is strongly changed (see  for a review). The pool sizes of several metabolites are increased and there are significant differences in the cold-responsive metabolomes of different Arabidopsis accessions [7, 31, 32]. Significantly, the leaf contents of the four sugars Glc, Fru, Suc and Raf were linearly correlated with leaf freezing tolerance [8, 11, 13] and these sugars were also found among a small group of metabolites that could be used to predict the freezing tolerance of several Arabidopsis genotypes with high accuracy . In addition, although the Pro contents of the leaves also increased during cold acclimation, there was no correlation with freezing tolerance among the 54 accessions investigated previously  and Pro was also not among the predictive metabolites .
The present data suggest that the role of these five compatible solutes may be significantly different between Arabidopsis and Thellungiella. Among the sugars, a positive correlation with acclimated freezing tolerance was only observed for Suc, while there was actually a negative correlation for Fru. In addition, the Thellungiella accessions did not accumulate Raf to the same extent as Arabidopsis. Instead, Thellungiella accumulated much higher amounts of Pro during cold acclimation and we found a significant correlation with acclimated freezing tolerance. The accumulation of compatible solutes, particularly Suc and Pro, was not only found in Thellungiella plants during cold acclimation. Especially Pro contents also increased much more than in Arabidopsis when plants were challenged with high NaCl concentrations [15, 33, 34] suggesting a different metabolic adaptation strategy between the species under abiotic stress conditions. Obviously, this hypothesis has to be tested in the future by metabolomic approaches using appropriate collections of accessions from both species.
We would like to stress at this point that it is highly unlikely that the differences in compatible solute content are the only reason for the observed differences in freezing tolerance. Although the constitutively freezing tolerant esk1 mutant in Arabidopsis shows a high accumulation of Pro under nonacclimated conditions , it also shows hundreds of changes in gene expression, making it impossible to attribute the higher freezing tolerance to a single factor . Similarly, although freezing tolerance in Arabidopsis is strongly correlated with Raf content, a knock-out mutant of the raffinose synthase gene in Col-0 resulted in the absence of Raf in the cold acclimated leaves without an impairment of freezing tolerance . All these findings emphasize the well-known fact that plant freezing tolerance is a multigenic, quantitative trait. In addition, the present data indicate that even in closely related species, different metabolites may be important.
One additional class of metabolites that has frequently been implicated in plant freezing tolerance are polyamines . They are thought to be involved in many aspects of plant growth, development and stress tolerance (see [38–40] for reviews). Their exact functions in these processes have not been completely elucidated, but it was demonstrated that Put is an essential component of the cold acclimation process in Arabidopsis. This is at least in part mediated through a role in the regulation of ABA biosynthesis.
The measurement of free polyamine levels in several accessions of both Arabidopsis and Thellungiella revealed that not all accessions showed an increase in the content of Put or Spd during cold acclimation. Also, the levels of free Put and Spd were not correlated with leaf freezing tolerance. In fact, the most freezing tolerant Arabidopsis accession in this study (Te-0) showed no increase in the pool size of either polyamine. In addition, the overall amounts of Put and Spd were very similar in all studied plants. Only the contents of free Spm showed higher levels in Thellungiella under nonacclimating conditions than in Arabidopsis. This was, however, strongly decreased during cold acclimation, leading to similar pool sizes between the species in the acclimated state. In Thellungiella we found a negative correlation between Spm contents and LT50 ACC, indicating that low levels of Spm may be a requirement for efficient cold acclimation. A similar reduction of Spm levels was previously already observed in the Arabidopsis accession Col-0  and in wheat  in response to cold exposure. However, the functional relevance of this reduction of free Spm levels is currently unknown. The natural variation in Spm content revealed in this study may offer an interesting possibility to elucidate the molecular basis and functional significance of this phenomenon.