High pressure-sensitive gene expression in Lactobacillus sanfranciscensis

Lactobacillus sanfranciscensis is a Gram-positive lactic acid bacterium used in food biotechnology. It is necessary to investigate many aspects of a model organism to elucidate mechanisms of stress response, to facilitate preparation, application and performance in food fermentation, to understand mechanisms of inactivation, and to identify novel tools for high pressure biotechnology. To investigate the mechanisms of the complex bacterial response to high pressure we have analyzed changes in the proteome and transcriptome by 2-D electrophoresis, and by microarrays and real time PCR, respectively. More than 16 proteins were found to be differentially expressed upon high pressure stress and were compared to those sensitive to other stresses. Except for one apparently high pressure-specific stress protein, no pressure-specific stress proteins were found, and the proteome response to pressure was found to differ from that induced by other stresses. Selected pressure-sensitive proteins were partially sequenced and their genes were identified by reverse genetics. In a transcriptome analysis of a redundancy cleared shot gun library, about 7% of the genes investigated were found to be affected. Most of them appeared to be up-regulated 2to 4-fold and these results were confirmed by real time PCR. Gene induction was shown for some genes up-regulated at the proteome level (clpL/groEL/rbsK), while the response of others to high hydrostatic pressure at the transcriptome level seemed to differ from that observed at the proteome level. The up-regulation of selected genes supports the view that the cell tries to compensate for pressure-induced impairment of translation and membrane transport. Correspondence


Introduction
High hydrostatic pressure (HHP) exerts manifold effects on cells and microorganisms, leading to adaptation, stress response and cell death.The possibility of HHP-inducible cell death has stimulated research by food scientists aiming at the improvement of the hygienic safety of minimally processed foods, with the retention of many of their natural properties such as color, vitamin content and "freshness".The behavior of vegetative cells and bacterial endospores has been studied in HHP treatment of material ranging from simple cell suspensions to complex food matrices.Whereas vegetative cells can be inactivated by HHP, the effective inactivation of bacterial endospores is only achieved in combined pressure/temperature treatments (1)(2)(3)(4)(5).
The inactivation kinetics of microorganisms by HHP follows a constantly declining curve which may end in a "resistant" fraction.It is currently unclear whether this should be referred to a sub-population.The survival of vegetative cells is strongly dependent on the food matrix (4,(6)(7)(8)(9).Sublethally injured cells may loose their resistance to adverse environmental conditions, e.g., low pH or the presence of osmolites (6,7,(10)(11)(12).
More recently, efforts have been made to elucidate the molecular mechanisms of HHPinduced cellular effects.It has been shown that HHP affects all levels of cellular physiology targeting cellular organization, transcription, translation, protein conformation, enzyme activity, and membrane function (12)(13)(14)(15).This knowledge has permitted preliminary interpretations of the mechanisms of HHP-induced cell death and stress response and has raised the possibility of using the high pressure response of cells and microorganisms as a novel approach to study biological systems and use cells as factories in HHP biotechnology.
In recent investigations, we have used lactic acid bacteria as Gram-positive model organisms applied to food biotechnology to study HHP-induced inactivation kinetics, mechanisms of stress response, and sublethal injury followed by cell death (16).
It was shown that above a threshold level the cell membrane changes its fluidity and permeability and most likely its hydration (9).The proton gradient is no longer maintained and membrane proteins lose their function (11).As a result, the intracellular pH adopts the extracellular level and cells may die as a result of adverse environmental conditions.Furthermore, solvents showed protective as well as antagonistic effects with respect to the barophysiology of the membrane.
In the present study, proteome and transcriptome analyses were used to gain insight into the global cellular mechanisms of the response to HHP stress.

Proteome analyses
For the proteome analyses Lactobacillus sanfranciscensis DSM 20451 T was grown at 30ºC in mMRS medium, pH 6.1 (17), and cells were harvested from the exponential growth phase.HHP treatment was carried out in fresh medium for 60 min at 30ºC and at pressures of 80 MPa at 30 MPa/min.Under these conditions, the maximal growth rate was reduced to 1/10.Protein extraction and 2-D electrophoresis were carried out as described by Drews et al. (18) using pH gradients of pH 3-10, 4-7, 4.5-5.5, and 5.5-6.5.
Protein analyses and sequencing were done by MALDI-MS and N-terminal sequencing by TopLab (Munich, Germany).Genes were identified by reversed genetics using degenerated primers deduced from the peptide fragments.

Transcriptome analyses and real-time PCR
L. sanfranciscensis genomic DNA was isolated by the method of Lewington et al. (19).All of the DNA manipulations in this study were performed according to standard procedures (20).For the transcriptome analyses shot-gun libraries from genomic DNA of L. sanfranciscensis DSM 20451 were prepared using E. coli DH5α, E. cloni™ and the plasmid psmart-HCAmp (Lucigen, Middleton, UK) with partially digested (BsuRI, AluI) or sonicated DNA of L. sanfranciscensis DSM 20451 T .Redundancy of the libraries was checked by random sequencing of 100 inserts and the inserts of 2000 clones were amplified by PCR and spotted onto silylated microscope slides to form the array.Array hybridizations were performed as described by Huang et al. (21) with cDNAs labeled with Alexa Flour ® 555 or Alexa Flour ® 647 from untreated and pressurized cells, respectively.The pressure treatment was carried out for 30 min at 45 MPa at 30ºC at 200 MPa/min.All experiments were done in triplicate.A more than two-fold change in signal intensity was considered to be significant.Selected pressure-sensitive genes were (partially) sequenced and their expression was verified by real-time PCR in the Light Cycler Instrument (Roche Molecular Biochemicals, Mannheim, Germany) with the QuantiTect™ SYBR ® Green kit (Qiagen, Hilden, Germany).Real-time PCRs were performed according to the instructions of the SYBR ® Green manufacturer.All expression ratios were normalized against the reference gene, phosphoketolase.

Proteome analysis
Five independent experiments repeatedly showed that the synthesis of 16 proteins (P1-P16) was affected by high pressure treatment (Figure 1).The expression of 9 of the 16 proteins showed an increase of 2-to 12fold while the expression of 7 proteins decreased by more than 50%.The sequences of four internal peptides of these proteins could be determined (data not shown).Comparison of these sequences with the BLAST databases at the National Center for Biotechnology Information revealed that the two most increased proteins P1 and P2 showed strong homology to ribokinase and an ATPdependent protease, ClpL. Figure 2 shows the differential expression of these proteins as determined by 2-D gel electrophoresis.The peptide sequences of proteins P9, P13 and P16 showed homologies to the general stress protein GroEL, also known as HSP60, to elongation factor Tu, and to glyceraldehyde-3-phosphate dehydrogenase, respec- combine the function of two enzymes in its two domains.

Transcriptome analysis
Changes in the transcriptional profile upon high pressure treatment were determined using DNA microarrays.Among the 750 spots that passed spot quality analysis, the mRNA level of genes or operons was affected in 48.About 6% of the genes evaluated were more than two-fold induced, while 1% were more than two-fold repressed.The gene encoding phosphoketolase (xpk) was not affected by high pressure and was therefore used as reference in the real-time verification of the transcriptome analyses.Realtime PCR results clearly supported the microarray data.A total of 48 high pressureaffected genes were classified with respect to their cellular function.As listed in Table 1, high pressure-affected genes of various cellular functions, namely protein synthesis and fate, stress response and cellular transport.Still, the function of 16 of the identified high pressure-sensitive genes remains unknown.
The most upregulated gene was an uncharacterized ORF with a conserved DEAD-ATP-dependent helicase motif (3.9-fold induced).DEAD helicases play overlapping roles in several processes including remodeling of RNA structures, biogenesis of the ribosome, and stabilization of mRNA (22,23).This gene was followed by groEL involved in heat shock, and a GTPase of unknown function (2.9-fold induced).An applicationbased array with 76 selected target genes including those involved in translation was designed to control the expression ratio in more detail.The translation elongation factors tuf and tsf, both responsible for binding of aminoacyl-tRNA to the ribosomal A site, were induced, while translation factors involved in later steps of protein synthesis were either downregulated (fusA and prfB) or not affected (prfC).In addition, transcription of genes whose products modify tRNAsynthetases (trmA, gidA and thd) and some ribosomal proteins was induced.tRNA-synthetases were hardly affected.

Discussion
In contrast to piezophilic microorganisms, bacteria adapted to atmospheric pressure do not have a specific HHP response.This is mainly due to the fact that HHP does not cause stress to piezophilic bacteria at the level present in their deep sea environment and in the experiments in which they have been studied so far.While HHP stress is probably unknown to lactobacilli, the cellular response appears to be targeted rather than confused, aiming at the compensation of pressure-induced defects.Some of these major defects appear to result from the general effects of high pressure, which at a level of up to 80 MPa promotes dissociation and conformation of macromolecules (24).The results of both the proteome and transcriptome analyses of the present study support this view.
The HHP response of L. sanfranciscensis differs from that observed at the proteome level but still overlaps with it.In the pres- ence of chloramphenicol no changes in the barotolerance and cross-resistance to other stresses were observed upon pressure treatment, indicating that de novo protein biosynthesis takes place in cells under sublethal pressure stress.The relative amount of mRNA of numerous genes is subject to change upon high pressure stress, which may result from selective transcription and also from mRNA stability.Typical HHP stress overlaps between proteome and transcriptome include induction of some general stress proteins (GroEL or ClpL) which may act as proteases or as molecular chaperones (25).
A more general interpretation of the cellular HHP response may be obtained from the transcriptome rather than the proteome analyses.This is due to the fact that the proteome analysis hardly includes low abundance regulatory proteins whose expression is rather determined by the highly sensitive real-time PCR.Assuming that the cell tries to compensate for pressure-induced inhibition or impairment of vital functions, differential expression of genes permits reaching conclusions about the affected functions in vivo.From these analyses it may be concluded that high pressure counteracts membrane damage and transport as well as translation to the ribosome.These conclusions are supported by the demonstration that translation is highly pressure sensitive in vitro and in vivo as a result of ribosome dissociation (26) and decreased binding of aminoacyl t-RNAs (27) due to changes in ribosome conformation (24,28), respectively.
These results support the idea that reduced ribosomal function may occupy a central position in the response to high pressure.A ribosomal sensor model implies that the physical state of the ribosome is a signal linking the stimulus (in this case high pres-sure) and the increased expression of a certain set of stress genes.In comparison, changes in translational capacity have also been reported to elicit cold and heat shock responses (29).Thus, such a model would explain the "links" between the response to various stresses.The strong response of a GTPase of unknown function further supports this model.GTPases are considered to be necessary for ribosome function and for transmission of information from the ribosome to specific targets to trigger specific cellular responses (30).
Apart from the effects triggered via changes in ribosomal structure and function, it has also been shown that pressure strongly affects membrane fluidity and transport (9).In the transcriptome this appears to be reflected by increased levels of proteins involved in transport.
In lactic acid bacteria the HHP stress response at first glance appears to be different from that observed with Escherichia coli, in which it is similar to heat shock (31) and may induce an SOS response (32).At the proteome level HHP stress was compared to other stresses including pH, temperature, cold shock and stationary phase, and overlaps were found with these stress responses (data not shown).On the basis of the results obtained with other organisms, it can be assumed that this is the case at the transcriptome level as well (33,34).This may be due to the lack in lactic acid bacteria of alternative sigma factors as they are found as global regulators in E. coli and some gram positives.While organisms employing global regulators are prone to react to stress in a stereotypical manner, the lack of these regulators may permit lactic acid bacteria to adapt to a variety of stresses in different environments.
The classification is based on the COGs functional annotation database available at (http://www.ncbi.nlm.nih.gov/COG/old/palox.cgi?fun=all).COGs = clusters of orthologous groups of proteins.