during exponential growth of Salmonella Typhimuriummmi_69291169..1186
The challenge of compacting the chromosome into the limited size of the bacterial cell is met by a number of mechanisms including macromolecular crowding and DNA supercoiling . 
The chromosomal DNA is further condensed by small nucleoid-associated proteins that influ-ence local DNA topology ( Luijsterburg et al. , 2006 ) . 
The architectural role of these proteins also plays a major role in the transcriptional regulation of genes that respond to environmental changes ( McLeod and Johnson , 2001 ; Fang and Rimsky , 2008 ) . 
The protein composition of the bacterial nucleoid is not static and reflects varying cellular requirements as it adapts to changing environmental conditions or growth phase ( Azam and Ishihama , 1999 ) . 
The 15.5 kDa proteins StpA ( suppressor of td mutant phenotype A ) and H-NS ( histone-like nucleoid structuring protein ) are major components of the Escherichia coli and Salmonella enterica serovar Typhimurium nucleoid ( Luijsterburg et al. , 2006 ) . 
The StpA and H-NS proteins are closely related and share 52 % identity . 
Conserved amino acids are predominantly located in the dimerization and DNA-binding domains . 
The close relatedness of the dimerization domains allows StpA and H-NS to interact to form heteromers ( Williams et al. , 1996 ; Johansson and Uhlin , 1999 ; Leonard et al. , 2009 ) . 
Both StpA and H-NS bind DNA nonspecifically . 
They display a slight preference for curved DNA , as found in the AT-rich segments found close to bacterial promoters ( Yamada et al. , 1990 ; Owen-Hughes et al. , 1992 ; Sonnenfield et al. , 2001 ) . 
Consequently , both StpA and H-NS can repress transcription from a synthetic gal promoter containing an upstream curved sequence in bacterial cells ( Zhang et al. , 1996 ) . 
More recently , an AT-rich ( 78 % ) consensus sequence has been proposed as a high-affinity H-NS binding site in proteobacteria ( Lang et al. , 2007 ; Sette et al. , 2009 ) . 
Mutations in the hns structural gene cause pleiotropic phenotypes in both Salmonella and E. coli , many of which are linked to adaptation to environmental changes such as increased resistance to osmotic and cold shock in Salmonella ( Hinton et al. , 1992 ) , carbon source utilization ( Atlung and Ingmer , 1997 ) , homologous recombination and genome stability in E. coli ( Lejeune and Danchin , 1990 ; Dri et al. , 1992 ) . 
In contrast , mutations in stpA have 
StpA is a paralogue of the nucleoid-associated protein H-NS that is conserved in a range of enteric bacteria and had no known function in Salmonella Typhimurium . 
We show that 5 % of the Salmonella genome is regulated by StpA , which contrasts with the situation in Escherichia coli where deletion of stpA only had minor effects on gene expression . 
The StpA-dependent genes of S. Typhimurium are a specific subset of the H-NS regulon that are predominantly under the positive control of s38 ( RpoS ) , CRP-cAMP and PhoP . 
Regulation by StpA varied s38 with growth phase ; StpA controlled levels at mid-exponential phase by preventing inappropriate s38 activation of during rapid bacterial growth . 
In contrast , StpA only activated the CRP-cAMP regulon during late exponential phase . 
ChIP-chip analysis revealed that StpA binds to PhoP-dependent genes s38 but not to most genes of the CRP-cAMP and s38 regulons . 
In fact , StpA indirectly regulates-s38 dependent genes by enhancing turnover by repressing the anti-adaptor protein rssC . 
We discovered that StpA is essential for the dynamic regula-s38 tion of in response to increased glucose levels . 
Our findings identify StpA as a novel growth phase-specific regulator that plays an important physi-s38 ological role by linking levels to nutrient availability . 
Accepted 11 October , 2009 . 
* For correspondence . 
E-mail sacha.lucchini@bbsrc.ac.uk; Tel. ( +44 ) 1603 255000 ; Fax ( +44 ) 1603 255288 . 
no obvious phenotype in E. coli , although StpA often has a ( minor ) role in processes in which H-NS is involved ( Bertin et al. , 2001 ) . 
This minor role was confirmed by the fact that deletion of stpA has no effects at the proteomic or transcriptomic levels in E. coli , as defined by two-dimensional protein gel electrophoresis and global gene expression analysis ( Zhang et al. , 1996 ; Mueller et al. , 2006 ) . 
In fact , inactivation of stpA only has phenotypic effects in the absence of hns , indicating that the deletion of stpA is fully compensated by H-NS in E. coli . 
The observed phenotypic differences between stpA and hns deletions might be partially explained by the fact that intracellular H-NS levels are higher than those of StpA ( Zhang et al. , 1996 ; Sonnenfield et al. , 2001 ) . 
This would be supported by the observation that multicopy expression of stpA rescues several hns mutant phenotypes ( Bertin et al. , 2001 ) . 
One mechanism that maintains the imbalance in levels of the two proteins is mediated by the negative cross-regulation that both proteins exert on each other ; H-NS represses stpA transcription more strongly than StpA controls hns ( Zhang et al. , 1996 ) . 
Additionally , StpA is susceptible to degradation by the Lon protease , while H-NS is not . 
StpA degradation by Lon increases in the absence of H-NS , suggesting that the formation of H-NS-StpA hetero-oligomers protects StpA from Lon and that only low levels of StpA Homo-oligomers are found in bacterial cells ( Johansson and Uhlin , 1999 ) . 
Although it is clear that StpA can mediate multicopy suppression of many H-NS-dependent genes , both proteins are not functionally identical . 
StpA was discovered by virtue of its ability to suppress the splicing defect of a mutant phage T4 thymidylate-synthase gene , and the protein was shown to promote RNA splicing in vitro ( Zhang et al. , 1995 ) . 
In fact , StpA has a stronger effect on splicing than H-NS in vivo ( Zhang et al. , 1996 ) . 
The involvement of StpA in RNA splicing probably reflects its ability to promote RNA annealing , strand displacement and folding of RNA ( Zhang et al. , 1996 ; Cusick and Belfort , 1998 ; Mayer et al. , 2007 ) . 
In addition to its RNA splicing role , StpA induces OmpF expression in E. coli by destabilizing the micF antisense RNA . 
In contrast , H-NS positively regulates OmpF levels by transcriptionally repressing micF ( Deighan et al. , 2000 ) . 
This suggests that under specific conditions StpA has a distinct function as an RNA chaperone that can not be fulfilled by H-NS . 
More recently , the H-NS regulon of S. Typhimurium was defined by transcriptomic analysis . 
Consistent with the pleiotropic effects of hns mutations , 1439 genes were regulated by H-NS ( twofold cut-off , FDR 0.05 ) in S. Typhimurium ( Ono et al. , 2005 ; Lucchini et al. , 2006 ; Navarre et al. , 2006 ) . 
H-NS is responsible for silencing the expression of horizontally acquired DNA ( Lucchini et al. , 2006 ; Navarre et al. , 2006 ) . 
We have now determined the regulatory role of StpA by assessing the impact of deleting stpA upon S. Typhimurium global gene expression at different stages of growth . 
We show that , unlike in E. coli , StpA regulates a large number of genes in S. Typhimurium . 
StpA plays an important role in the growth phase-specific regulation of gene expression ; during mid-exponential phase , StpA prevents the premature expression of the s38 regulon whereas , during late-exponential growth , StpA is required for full expression of the CRP-cAMP regulon . 
Results
Growth phase-dependent expression of StpA The transcription of stpA is known to be restricted to a short period of the exponential growth phase during growth of E. coli in rich medium ( Free and Dorman , 1997 ) . 
This suggested that the cellular requirement for StpA could change in Salmonella during batch culture . 
To determine whether the levels of StpA were growth phase-dependent in S. Typhimurium SL1344 , we monitored stpA transcription and protein levels throughout growth in rich medium . 
+ A stpA : : gfp transcriptional fusion showed that expression of stpA peaks during the early stages of exponential growth , as observed in E. coli ( Fig. 1A ) . 
As StpA is regulated at the post-translational level by Lon protease in E. coli ( Johansson and Uhlin , 1999 ) , we monitored StpA protein levels . 
To achieve this we first constructed strain JH3573 that expresses a version of StpA with a C-terminal 3 ¥ FLAG epitope from the native stpA locus . 
3 ¥ FLAG Transcriptomic analysis of the stpA strain JH3573 showed that no genes were differentially expressed during exponential growth compared with SL1344 wild type , confirming that the addition of the epitope did not compromise StpA function ( data not shown ) . 
This strain was used to detect StpA protein at all growth stages by Western blotting , with levels peaking at the entry to exponential growth ( Fig. 1B ) . 
This confirms that , unlike in E. coli , stpA continues to be expressed at a reduced level throughout exponential and stationary growth in S. Typhimurium . 
Identification of the StpA regulon
The growth phase-dependent expression of StpA prompted us to use a transcriptomic approach to define the StpA regulon at four stages of growth , corresponding to different levels of StpA expression ( Fig. 1A ) : 60 ′ ( early exponential phase , EEP ) , 160 ′ ( mid-exponential phase , MEP ) , 250 ′ ( late exponential phase , LEP ) and 22 h pos-tinoculation ( late stationary phase , LSP ) . 
Bacterial RNA was extracted at each of the four time points , labelled and hybridized against SALSA microarrays ( see Experimental procedures ) . 
The transcriptome of the SL1344 parent and a strain lacking stpA ( JH3003 ) was compared under identical growth conditions , and revealed that the StpA regulon varied with growth phase ( Table S1 ) . 
The absence of StpA did not affect gene expression in EEP or LSP ( 1 and 22 h ) . 
In contrast , expression of up to 5 % of the S. Typhimurium genome was significantly altered in JH3003 compared with SL1344 during exponential growth ( MEP and LEP ) , and we define these genes as being StpA-dependent . 
The number of StpA-dependent genes ( twofold change and a false discovery rate ( FDR ) 0.05 ) at any time point was 183 . 
There were 96 StpA-dependent genes at MEP and 129 at LEP . 
Most of the StpA-dependent genes were derepressed , with 92 % ( MEP ) and 81 % ( LEP ) genes being upregulated in the absence of StpA ( Table S1 ) . 
It has been reported that StpA can negatively regulate expression of a reporter gene under the control of the hns promoter in E. coli ( Zhang et al. , 1996 ) ; although this could only be observed in the absence of hns , this raised the possibility that StpA-dependent gene expression could simply reflect changes in the level of H-NS . 
Because StpA potentially represses hns , we expected such genes to be upregulated in an hns mutant and downregulated in a DstpA strain . 
After comparing the list of StpA-dependent genes with the previously published H-NS regulon of S. Typhimurium ( Ono et al. , 2005 ) , only one gene was identified that displayed such behav-iour ( STM3138 ) . 
This indicates that under the conditions tested , StpA does not regulate gene expression by modulation of H-NS levels , and is consistent with the transcriptomic data , which do not show significant StpA-dependent changes in hns gene expression in S. Typhimurium JH3003 at any time point . 
We used transcriptional gene fusions and RT-PCR to confirm the transcriptomic data for selected genes ( Fig . 
S1 ) . 
StpA represses genes required for cell envelope modification and resistance to cationic peptides
Among the genes that were derepressed in the stpA deletion mutant were mig-14 , virK and ugtL , which are required for resistance to the cationic antimicrobial peptide polymyxin B ( PmB ) ( Brodsky et al. , 2002 ; Det-weiler et al. , 2003 ; Shi et al. , 2004 ) . 
Interestingly , more than half ( 54 % ) of the genes that were repressed at any time point by StpA were inducible by PmB ( Bader et al. , 2003 ) , suggesting that StpA could be involved in the resistance of Salmonella to antimicrobial peptides . 
To test the phenotypic consequences of StpA-dependent gene regulation , we analysed the effects of deleting or overexpressing stpA on survival of PmB challenge . 
Moderate overexpression of stpA was achieved by cloning the stpA gene into the low-copy-number vector pWKS30 [ six to eight copies per cell ; ( Wang and Kushner , 1991 ) ] to generate pMDH20 . 
A significant increase in resistance to PmB was seen in JH3003 ( DstpA ) , and we discovered that overexpression of StpA caused a strong PmBsensitive phenotype ( Fig. 2A ) . 
Increased resistance of Salmonella to PmB involves alteration of the cell envelope via modification of the lipid 
A component of LPS . 
It has been demonstrated that ugtL encodes a protein that mediates the formation of mono-phosphorylated lipid A ( Shi et al. , 2004 ) and reduces the negative charge of the lipid A . 
This decreases the interaction of the cationic PmB , leading to reduced plasma membrane destabilization and increased PmB resistance ( Ernst et al. , 2001 ) . 
Our data show that ugtL is StpA-dependent , as are other genes involved in LPS modification including pagC and pagP . 
These genes are under the control of the PhoPQ two-component system ( Belden and Miller , 1994 ; Soncini et al. , 1996 ; Zwir et al. , 2005 ) and have been shown to control the permeability of the outer membrane to a variety of molecules , including the bile component deoxycholate ( Murata et al. , 2007 ) . 
We tested the effect of deoxycholate on Salmonella survival in response to stpA deletion and overexpression , and discovered that deoxycholate and PmB showed similar sensitivity patterns . 
Deletion of stpA resulted in a mar-ginal effect , and overexpression of StpA caused a deoxycholate-sensitive phenotype ( data not shown ) , supporting a role of StpA in the regulation of LPS modification . 
Analysis of the StpA regulon ( Table S1 ) suggests that the role of StpA in altering the composition of cell envelope is not restricted to modulation of LPS components . 
StpA also represses a variety of genes thought to be involved in the synthesis of cell surface structures such as pili ( stdAB , safA , fimA ) , capsule ( wca operon ) and cell membrane components ( ybhO ) . 
StpA represses s38-activated genes
More than one third of the StpA-dependent genes that are derepressed in JH3003 ( DstpA ) during MEP are activated by the alternative sigma factor s38 in Salmonella ( 39 % , Table S1 ) . 
s38 is encoded by rpoS and is required for the cellular reprogramming associated with entry into stationary phase or adverse growth conditions known as the general stress response ( Ishihama , 2000 ) . 
The transcription of a large number of stationary phase-associated genes is activated by s38 , and confers increased resistance to a number of physical and chemical stresses , including acid , heat and oxidative agents ( Klauch et al. , 2007 ) . 
The list of s38-dependent genes that are derepressed in JH3003 included genes involved in resistance to salt shock ( otsAB ) and oxidative stress ( katN and katE ) . 
Two of the most highly upregulated s38-dependent genes in JH3003 , osmY and yciE , code for proteins that are induced by acid in E. coli ( Audia et al. , 2001 ) . 
This suggested novel phenotypes that could be regulated by StpA in S. Typhimurium . 
We determined the level of resistance of the stpA mutants to stationary phase-relevant stresses , namely salt , acid and peroxide killing . 
The absence of StpA caused increased stress resistance at MEP ( Fig. 2B , D and E ) . 
This role was con-firmed by overexpression of StpA , which increased the sensitivity to salt and peroxide , compared with the wildtype strain ( Fig. 2B and D ) . 
We discovered that the deletion of stpA caused a far less pronounced effect on the transcription of s38-dependent genes at LEP compared with MEP ( Table S1 ) . 
This is exemplified by the osmY and otsAB genes , which were repressed by StpA at the MEP stage of growth , but not at LEP . 
This effect had phenotypic consequences as the stpA deletion mutant JH3003 displayed increased resistance to salt at MEP , but not at LEP ( Fig. 2C ) . 
StpA prevents s38-dependent gene expression during mid-exponential growth
During the early stages of exponential growth in rich medium , several mechanisms ensure that s38 levels are maintained at low levels ( Klauch et al. , 2007 ) . 
Consequently , s38 inactivation has a mild effect on gene expression during exponential growth in E. coli ( Dong et al. , 2008 ) . 
Consistent with this , comparison of the transcriptional profile of S. Typhimurium SL1344 with that of an rpoS-deletion mutant revealed few differences at MEP . 
Only eight genes showed minor changes in gene expression ( Table S3 ) , leading us to consider the possibility that the activation of s38-dependent genes in the absence of StpA could be mediated by other factors . 
It may be relevant that the s38-dependent genes osmC and katE have been shown to possess a second s38-independent promoter that is activated by the RcsC-YojN-RcsB phosphorelay ( Tanaka et al. , 1997 ; Bouvier et al. , 1998 ; DavalosGarcia et al. , 2001 ; Majdalani and Gottesman , 2005 ) . 
We needed to know if the increased gene expression observed in JH3003 in the absence of the repressive activity of StpA was dependent on s38 or not . 
We used a genetic strategy to determine whether deletion of stpA caused upregulation of gene expression in the absence of s38 at MEP . 
We compared the expression profiles of rpoS and stpA/rpoS mutant strains , and found that the majority of the 88 genes derepressed in JH3003 , including osmC and katE , were no longer derepressed in response to stpA deletion in the absence of s38 at MEP ( Fig. 3 ; Table S1 ) . 
The data show that s38-activation is responsible for the derepression observed in JH3003 . 
We conclude that StpA prevents s38-dependent gene expression during exponential growth of S. Typhimurium . 
StpA is essential for linking s38 expression to glucose availability
As well as the large number of genes that are directly or indirectly repressed by StpA , 32 genes were downregulated in the absence of StpA . 
Many of these genes , including mglAB , aspA and glpF , are involved in the catabolism of carbohydrates and are activated by CRP-cAMP in E. coli ( Gosset et al. , 2004 ) ( Table S1 ) . 
This raised the possibility that StpA could directly induce the expression of crp at LEP , as had been observed in E. coli ( Johansson et al. , 1998 ; 2000 ) . 
Indeed , a StpA-dependent decrease of crp expression was observed in our experiment ( 1.6-fold ) . 
This change was less than twofold , and therefore did not meet one of the criteria for detecting differentially expressed genes . 
However , the StpA-dependent reduction of the crp mRNA was highly statistically significant ( FDR 0.001 ) . 
A similar low-level StpA-dependent decrease in crp was seen in E. coli ( Johansson et al. , 2000 ) . 
Interestingly , the expression of the CRP-cAMP regulon was only StpA-dependent at LEP ( Fig. 4 ) . 
In E. coli , CRP-cAMP represses rpoS transcription during logarithmic growth ( Lange and Hengge-Aronis , 1994 ) . 
However , there are indications that CRP-cAMP also plays a different role as an activator of rpoS expression during the late stages of growth ( Hengge-Aronis , 2002 ) . 
As StpA regulates both s38 and CRP , we hypothesized that StpA might link the expression of s38 with sugar availability . 
As CRP requires cAMP for its activity , we determined the effects of levels of glucose that lower intracellular cAMP levels upon s38 ( Deutscher et al. , 2006 ) . 
Growth in the presence of glucose caused a decrease of s38 in wild-type S. Typhimurium , and these changes in s38 levels were StpA-dependent ( Fig. 5 ) . 
Importantly , the non-PTS sugar maltose had a reduced effect on s38 levels in wild-type S. Typhimurium . 
To rule out an indirect effect on s38 through a change in medium osmolarity , we looked at the effect of the addition of lactose , a sugar that can not be utilized by Salmonella . 
The addition of lactose did not alter s38 levels . 
While this does not prove a mechanistic link between CRP-cAMP and StpA in the regulation of s38 , these results do show that the modulation of s38 by glucose requires functional StpA . 
StpA-regulated genes are also H-NS-dependent To determine whether StpA and H-NS regulate the expression of similar genes , we compared the regulons of StpA and H-NS . 
During exponential growth , 1439 genes are regulated by H-NS in S. Typhimurium ( twofold , FDR 0.05 ) ( Ono et al. , 2005 ) . 
We found that 90 % of the 162 StpA-dependent genes are regulated , to varying extents , by H-NS . 
Further analysis revealed that StpA regulates a defined subset of the H-NS regulon , largely 38 consisting of the PhoP and s regulons ( Fig. 6A ) . 
These findings show a limited but distinct role for StpA in the control of global gene expression . 
Identification of StpA-binding sites on the Salmonella chromosome
Because StpA regulates a subset of the H-NS regulon , and we and others have shown that H-NS directly silences expression of ~ 250 S. Typhimurium genes ( Lucchini et al. , 2006 ; Navarre et al. , 2006 ) , it was important to compare the binding of H-NS and StpA to the chromosome in S. Typhimurium . 
As StpA and H-NS form hetero-oligomeric complexes ( Williams et al. , 1996 ; Leonard et al. , 2009 ) , we determined whether StpA and H-NS colocalize in vivo and could cooperate to regulate gene expression . 
We used chromatin-immunoprecipitation ( ChIP-on-chip ) to identify StpA and H-NS DNA-binding sites on the genome of S. Typhimurium SL1344 . 
ChIP-on-chip was performed on strain JH3573 that expresses a 3 ¥ FLAG-tagged version of StpA . 
The parental SL1344 strain was used as negative control . 
We identified 285 chromosomal regions that bind StpA , ranging in size from 400 to 3500 base pairs and containing 572 genes . 
Comparison of the ChIP data revealed that StpA displayed an identical binding profile to H-NS ; all genomic regions associated with StpA were also bound by H-NS ( Fig. 6B and C ) . 
As previously seen for H-NS ( Lucchini et al. , 
2006 ; Navarre et al. , 2006 ) , StpA displayed a slight preference for AT-rich DNA ( Fig . 
S2 ) . 
Of the 88 StpA-dependent genes that were derepressed in JH3003 during mid-exponential growth , only 25 were bound by StpA in the wild-type strain ( Table S1 ) . 
This interesting finding shows that StpA generally regulates gene expression indirectly . 
We investigated the s38 regulon and found that only a small minority of StpA-dependent genes that were activated by s38 were bound by StpA ( 6/35 ) . 
In contrast , the majority of PhoP-dependent genes were bound by StpA ( 8/9 ) ( Table S1 ) . 
StpA modulates s38 stability
Our observation that the regulation of s38-dependent genes by StpA was largely indirect was intriguing . 
It was possible that StpA controlled the s38 regulon by modulating levels of s38 at the transcriptional level . 
However , transcriptomic data showed that the rpoS mRNA levels did not vary in JH3003 at MEP , compared with wild type ( Table S1 ) , and this was confirmed by real-time PCR ( Fig. 7C ) . 
As post-transcriptional and post-translational regulation of s38 play an important role in E. coli ( Hengge-Aronis , 2002 ) , we used Western blotting to analyse s38 levels and discovered a clear increase in s38 in the stpA deletion strain at MEP ( Fig. 7A ) . 
This suggested that StpA regulates s38 by a post-transcriptional or post-translational mechanism at MEP . 
In contrast , StpA did not significantly modulate s38 protein levels at LEP ( Fig. 7A ) . 
Consistent with this observation , deletion of stpA had little effect on the transcription of s38-dependent genes at LEP compared with MEP ( Table S1 ) . 
This was exemplified by the expression profile of the s38-dependent gene osmY ; the expression of osmY is StpA-dependent ( i.e. altered in the stpA deletion strain ) during mid-exponential growth ( Fig. 7C ) . 
However , despite the fact that the deletion of stpA did not affect s38 protein levels at LEP , our data showed that rpoS mRNA levels were significantly lower in the absence of StpA ( in strain JH3003 ) at LEP compared with wild type ( Table S1 , Fig. 7C ) . 
This suggests that during the later stages of exponential growth , the StpA dependence of the rpoS mRNA is compensated by another mechanism that increases the level of s38 protein . 
During exponential growth in LB , s38 is maintained at low levels by a series of post-translational mechanisms , including ClpXP-mediated protein degradation ( Hengge-Aronis , 2002 ) . 
We tested a role for StpA in the posttranslational control of s38 by monitoring s38 stability after blocking protein synthesis with spectinomycin . 
The LEP growth phase was chosen because s38 levels are similar at this time point in wild type and JH3003 , and any observed difference in stability would not simply reflect the concentration of s38 . 
The results show that deletion of stpA leads to the stabilization of s38 , while overexpression of StpA reduces the half-life of s38 ( Fig. 7D ) , revealing that StpA is involved in the control of s38 protein turnover . 
StpA represses expression of the RssC anti-adaptor protein
Degradation of s38 requires the binding of the adapter protein RssB to target s38 to the ClpXP proteolytic system . 
Anti-adapter proteins are an important component of this pathway and interfere with the binding of RssB to the sigma factor to stabilize s38 . 
The RssB inhibitor IraP has been shown to be dependent on PhoP ( Bougdour et al. , 2006 ; Tu et al. , 2006 ) . 
As StpA repressed several PhoP-dependent genes , we speculated that StpA might regulate s38 stability via modulation of IraP activity . 
However , deletion of stpA in a DiraP background resulted in the same fivefold increase in s38 half-life seen in a DstpA iraP + background ( data not shown ) showing that StpA does not directly modulate s38 stability via IraP . 
A screen for genes that increased s38 activity in S. Typhimurium identified the rssC gene ( STM1110 ) . 
Overexpression of rssC was reported to cause a 10-fold increase in the stability of s38 ( Bougdour et al. , 2008 ) . 
Based upon its similarity to the E. coli IraM protein , RssC is predicted to be a RssB anti-adaptor ( Bougdour et al. , 2008 ) . 
Our data show that the rssC gene is highly upregulated in the stpA deletion mutant at LEP ( Table S1 ) , raising the possibility that StpA could indirectly modulate s38 by controlling the transcription of rssC . 
To confirm the role of RssC , we conducted an experiment at LEP and discovered that the stabilization of s38 caused by deletion of stpA was RssC-dependent ( Fig. 7E ) . 
It should be noted that StpA also regulates s38 stability via another mechanism , because s38 levels still show a partial StpA-dependence in the absence of rssC . 
We also determined whether rssC was necessary for the StpA-dependent modulation of s38 levels at MEP . 
The 
Western blot data clearly showed that the stabilization of s38 caused by deletion of stpA is also RssC-dependent at MEP ( Fig. 7B ) . 
In summary , StpA controls s38 stability during MEP and LEP by modulating the level of RssCmediated degradation of s38 . 
Discussion
We characterized the role of an H-NS paralogue in S. Typhimurium during growth in rich medium , and discovered that StpA is a true global regulator of gene expression that directly and indirectly regulates 183 genes . 
The majority of these genes are repressed by StpA . 
Many of the StpA-repressed genes were involved in the synthesis/modi fication of the cell envelope and response to various stresses such as osmotic shock , oxi-dative stress and resistance to cationic antimicrobial peptides . 
A large proportion of the StpA-dependent genes belong to the s38 , CRP and PhoP regulons . 
To validate the transcriptomic data at the phenotypic level , we showed that the stpA mutant exhibited increased resistance to various environmental stresses and antimicrobial peptides . 
The majority of the StpA-dependent genes belong to a well-defined subset of the H-NS regulon . 
Most of the overlap between the two regulons involved s38 and PhoP-dependent genes . 
However , StpA does not regulate the Salmonella pathogenicity island 2 or the motility system , which are strongly repressed by H-NS . 
Therefore , StpA does not simply share the same biological properties of H-NS , but has a distinct and specific role . 
Importantly , gene regulation by StpA was found to be growth phase-dependent . 
Repression of the s38 regulon by StpA was restricted to the exponential phase of growth , while activation of CRP-dependent genes was limited to LEP . 
This identifies StpA as an important factor in the growth phase-dependent control of gene expression in Salmonella . 
StpA prevents premature expression of s38-dependent genes and potentiates the expression of genes involved in nutrient acquisition during the late stages of exponential growth , when nutrient availability becomes a limiting factor for cellular replication . 
To differentiate between direct and indirect regulation of gene expression by StpA , we used chromatin immunoprecipitation ( ChIP-chip ) to identify StpA binding sites on the Salmonella genome in living bacterial cells . 
We found StpA coimmunoprecipitated with the majority of PhoP-dependent genes . 
Two StpA-repressed PhoP-dependent genes bound by StpA , pagC and ugtL , are also directly repressed by H-NS ( Perez et al. , 2008 ) . 
Interestingly , these genes require SlyA to alleviate H-NS repression . 
As SlyA is not able to directly activate transcription in vitro , its role appeared to be limited to counteracting the repressive action of H-NS on genes that were recently acquired by horizontal gene transfer ( Perez et al. , 2008 ) . 
Our finding that a large proportion of SlyA-activated genes are repressed and bound by StpA ( Table S1 ) suggests that StpA regulates these genes in a similar manner to H-NS , and that SlyA can counteract the effect of both H-NS and StpA mediated repression . 
In contrast to the binding of StpA to members of the PhoP regulon , only a minority of s38-dependent genes were bound by StpA . 
This led us to determine whether StpA regulated s38-dependent genes indirectly by modulation of s38 levels . 
We discovered that s38 protein levels were much higher in the stpA mutant than in the parental strain during exponential growth , and that StpA modulates s38 stability via repression of rssC . 
H-NS has also been shown to modulate s38 levels in E. coli by regulating translation initiation and s38 stability through mechanisms that still require clarification ( Hengge-Aronis , 2002 ; Zhou and Gottesman , 2006 ) . 
No regulatory function has been attrib-uted to StpA in E. coli ( Hengge-Aronis , 2002 ) . 
Strikingly , we observed that deletion of stpA does not affect s38 levels during late exponential growth , even though StpA does modulate s38 stability during this growth phase . 
The stpA deletion mutant displayed lower rpoS mRNA levels at LEP , indicating that the negative effect of StpA on s38 stability may be compensated at the transcriptional level under conditions when higher s38 amounts are required ( Fig. 8 ) . 
We speculate that the dual role played by StpA in both the transcriptional and post-transcriptional control of s38 makes the RpoS system particularly sensitive to environmental conditions . 
The expression of genes required for the resistance against environmental challenges is costly for the bacterial cell . 
This is exemplified by the fact that rpoS mutants of Salmonella display a fitness advantage in the absence of stress ( Robbe-Saule et al. , 2003 ) . 
Moreover , natural isolates of E. coli that displayed higher levels of s38 activity were shown to metabolize fewer carbon sources and to be less able to compete for low nutrient concentrations ( King et al. , 2004 ) . 
These findings led to the hypothesis that bacteria need to balance self-preservation and nutritional capacity ( SPANC ; Ferenci , 2005 ) . 
The fitness advantage derived from reduced s38 levels under low stress conditions has been linked to sigma factor competition ( Farewell et al. , 1998 ) . 
It is postulated that there is a limiting amount of RNA polymerase ( RNAP ) and that s38 and s70 compete for binding to the RNAP . 
High s38 levels therefore allow less s70 binding to RNAP leading to lower expression of s70-dependent genes , many of which are linked to metabolism and cellular growth ( Gruber and Gross , 2003 ) . 
Consequently , bacterial cells must tightly regulate the expression levels of key global regulators and this is exemplified by the complex regulatory network that determines s38 levels . 
Because StpA represses the s38 regulon while stimulating expression of the CRP-cAMP regulon , we propose that StpA plays a role in promoting an appropriate SPANC balance by linking stress response to nutrient availability in S. Typhimurium . 
This hypothesis is supported by the observation that the glucose-mediated decrease of s38 levels is StpA-dependent . 
Analysis of the ChIP-chip data revealed that StpA colocalizes with H-NS on the Salmonella genome . 
Together with our finding that all StpA-dependent genes are also regulated by H-NS , the localization results suggest that StpA and H-NS cooperate to regulate gene expression by forming hetero-oligomers . 
This is reminiscent of the H-NS-like proteins MvaT and MvaU of Pseudomonas aeruginosa that regulate gene expression in a synergistic fashion ( Castang et al. , 2008 ) . 
It is important to note that although H-NS and StpA bind to the same genes , their effect on gene expression is not equivalent . 
For example , the SPI2 genes , which are strongly derepressed in an Dhns mutant ( Lucchini et al. , 2006 ) , do not respond to the deletion of stpA despite the fact that StpA and H-NS share a similar binding profile over the entire locus ( Fig. 6C ) . 
Our data fit with other observations that binding of StpA to a particular locus is not sufficient to exert repression . 
At the bgl operon in E. coli , StpA can compensate for the deletion of the H-NS DNA-binding domain , but StpA alone is not capable of significant repression ( Free et al. , 1998 ) . 
The fact that StpA can compensate for the lack of the H-NS DNA-binding domain implies that the functional differences between these two proteins are associated with the dimerization and oligomerization domains . 
Consistent with this , recent work in S. Typhimurium and E. coli shows that StpA and H-NS have different dimerization and oligo-merization properties ; StpA homodimers are more thermostable than H-NS homodimers ( Leonard et al. , 2009 ) . 
In addition , H-NS-StpA hetero-oligomers are more stable than homo-oligomers of either StpA or H-NS . 
There is growing evidence that interactions between H-NS and other regulatory proteins are an essential component of H-NS-mediated repression . 
Hha and YdgT are homologues of the H-NS dimerization and oligomerization domains that can interact with both H-NS and StpA ( Nieto et al. , 2002 ; Paytubi et al. , 2004 ) . 
Deletion of Hha and YdgT induces derepression of many more H-NS-dependent genes than we observed in the absence of StpA . 
These findings suggest that the complexes that H-NS forms with Hha/YdgT have distinct properties compared with H-NS homo-oligomers or H-NS-StpA hetero-oligomers , possibly due to a modulation of the activity of H-NS ( Vivero et al. , 2008 ) . 
Distinctions between the regulatory roles of H-NS and StpA may therefore lie in their differing abilities to form homo-oligomers , and to interact with other proteins , with concomitant effects upon local DNA conformation and gene expression . 
Overall , our data reveal a novel role for StpA in the transcriptional regulatory network of S. Typhimurium as a node that connects the CRP-cAMP , PhoP and s38 regulons ( Fig. 8 ) . 
This contrasts with the situation described for E. coli , where inactivation of stpA only significantly affects gene or protein expression in the absence of hns ( Zhang et al. , 1996 ; Mueller et al. , 2006 ) . 
The fact that StpA plays a more limited role in gene regulation in E. coli was also observed at the phenotypic level , as deletion of stpA in E. coli does not show the clear effects we observed in Salmonella ( Atlung and Ingmer , 1997 ; Bertin et al. , 2001 ) . 
Our findings correlates with the fact that hns is an essential gene in S. Typhimurium ( Navarre et al. , 2006 ) , while deletion of hns only has a moderate effect on growth rate in E. coli ( Zhang et al. , 1996 ) . 
The differing roles of StpA in the closely related organisms S. Typhimurium and E. coli could reflect the different requirements and the evolution of an intracellular pathogen . 
Analysis of the conservation of the E. coli transcriptional network in 175 prokaryotic genomes revealed that transcription factors evolve faster than their target genes ( Babu et al. , 2006 ) . 
It has also been shown that bacterial adaptation can not only be achieved by gene acquisition and/or loss but also requires changes at the level of gene regulation ( Winfield and Groisman , 2004 ) . 
Our discovery that StpA represses the s38 regulon during exponential growth of S. Typhimurium , but not in E. coli , is consistent with the function of orthologous regulatory proteins rapidly diverging to allow adaptation to new environmental niches . 
Experimental procedures
Bacterial strains and growth conditions
The S. enterica serovar Typhimurium strain SL1344 was provided by Catherine Lee ( Hoiseth and Stocker , 1981 ) and is the same isolate used in previous transcriptomic studies from the Hinton laboratory ( Eriksson et al. , 2003 ; Mangan et al. , 2006 ; Nagy et al. , 2006 ; Hautefort et al. , 2008 ) . 
Where necessary , antibiotics were used at the following concentrations : ampicillin ( 100 mg ml-1 ) , chloramphenicol ( 12.5 mg ml-1 ) , kanamycin ( 35 mg ml-1 ) . 
Cultures were grown in LB broth ( Sambrook and Russell , 2001 ) under aeration ( 250 r.p.m. ) at 37 °C and harvested at four different growth phases : EEP ( A600 = 0.005 -- 0.010 ) , MEP ( A600 = 0.12 -- 0.15 ) , LEP ( A600 = 1.0 -- 1.2 ) and LSP ( A600 = 3.7 -- 3.8 ) . 
Bacterial strains and plasmids used in this study are shown in Table 1 . 
Strain construction and DNA manipulation
To obtain a strain overexpressing StpA , we used the low-copy plasmid pWKS30 ( Wang and Kushner , 1991 ) . 
First , we PCR-amplified the stpA gene including 779 bp upstream of the structural gene using primers stpA-FO2 and stpA-RO2 , which carry the restriction sites XbaI and HindIII respectively . 
The resulting PCR product and plasmid pWKS30 were digested with XbaI and HindIII , purified by agarose gel electrophoresis and ligated to generate plasmid pMDH20 . 
This plasmid was then transferred into S. Typhimurium SL1344 to generate strain JH3750 . 
Overexpression of StpA from pMDH20 was confirmed at both RNA and protein levels ( data not shown ) . 
To generate a stpA deletion derivative of S. Typhimurium SL1344 we first constructed a plasmid derivative of pMDH20 , where the stpA orf was replaced by the Campylobacter coli cat gene from the plasmid pAV35 ( van Vliet et al. , 1998 ) . 
To do so , we used primers StpA-Ri and StpA-Fi2 carrying a BamHI site to perform an inverse PCR on pMDH20 . 
These primers prime outward from the stpA structural gene and do not amplify it . 
The PCR product was then ligated to generate plasmid pMDH21 . 
The pAV35 cat gene was then inserted into the BamHI site of pMDH21 . 
This generated a new plasmid , designated pMDH22 , which contains the cat gene flanked by approximately 780 and 450 base pairs of the regions upstream and downstream of the stpA gene respectively . 
The DNA fragment containing the cat flanked by the stpA flanking regions was excised from pMDH22 by a XbaI and HindIII restriction digestion . 
This DNA fragment was then electroporated into S. Typhimurium SL1344 expressing the l Red recombinase ( Datsenko and Wanner , 2000 ) . 
The other deletion derivatives of S. Typhimurium SL1344 generated for this study ( Table 1 ) were obtained using the un-modified l Red method ( Datsenko and Wanner , 2000 ) . 
After mutagenesis , all mutations were P22-transduced to a clean background using phage P22 HT105/1 int-201 . 
EBU plates were used to select for nonlysogens ( Bochner , 1984 ) . 
P22-transduction was also used to combine mutations for the generation of double mutants . 
All mutant strains were verified by PCR and DNA sequencing . 
Strains carrying the deletion of rpoS and/or stpA were additionally validated by Western blot . 
The C-terminal tagging of StpA with the 3 ¥ FLAG epitope was performed using a modified l Red method ( Uzzau et al. , 2001 ) . 
The DNA fragment to be recombined into the chromosome was PCR-amplified from the pSUB11 plasmid . 
The successful fusion of StpA with the 3 ¥ FLAG epitope was confirmed by sequencing and Western blot . 
To obtain a stpA : : gfp + transcriptional fusion we used a derivative of the l Red system ( Hautefort et al. , 2003 ) ; we used primers stpA-gfp_F and stpA-gfp_R to amplify from plasmid pZEP07 a DNA fragment containing a promoterless gfp + gene followed by a chloramphenicol resistance cassette . 
The primers were designed to direct integration of the gfp + and cat genes in the Salmonella chromosome 20 bp downstream of the stpA ORF . 
All primers used to generate the different constructs are shown in Table S2 . 
Flow cytometric analysis
For measurement of GFP + expression , samples were immediately fixed for 1 min at room temperature in 3.7 % ( w/v ) formalin , diluted in 1 -- 2 ml of PBS to obtain a maximum of approximately 106 particles ml-1 and analysed with a FACS-calibur flow cytometer ( Becton Dickinson , Franklin Lakes , N.J. ) equipped with a 15 mW air-cooled argon ion laser as the excitation light source ( 488 nm ) as previously described ( Hautefort et al. , 2003 ) . 
RNA extraction for transcriptomic experiments
Overnight cultures were diluted 1000-fold into 250 ml flasks containing 25 ml of LB-broth ( Sambrook and Russell , 2001 ) . 
These cultures were incubated at 37 °C for 60 min , 160 min , 250 min or 22 h in a water bath ( New Brunswick Innova 4000 ) under agitation ( 250 r.p.m. ) . 
Synthesis and degradation of RNA were then blocked by adding 1/5 volume of stop-solution ( 90 % ethanol/10 % phenol ) ( Tedin and Blasi , 1996 ) . 
RNA was prepared from each culture flask using the Promega SV total RNA purification kit according to the manufacturer 's instructions . 
The quality of RNA was checked using the RNA nanochip ( Labchip , on an Agilent 2100 Bioanalyser ) and quantified by measuring the absorbance at 260 nm on a Nanodrop 1000 spectrophotometer . 
Template labelling and hybridization
The ` Common reference ' experimental design used S. Typhimurium genomic DNA as the cohybridized control for one channel on all microarrays . 
This method has the advantage of allowing the direct comparison between multiple samples , and is ideal for time-course experiments ( Eriksson et al. , 2003 ; Mangan et al. , 2006 ) . 
Total RNA and chromosomal DNA were labelled by random priming according to the protocols described at http://www.ifr.ac.uk/safety/ microarrays / #protocols . 
Briefly , a total of 16 mg RNA was reverse-transcribed and labelled with Cy3-conjugated dCTP ( Pharmacia ) using 200 U of Stratascript ( Stratagene ) and random primers ( Invitrogen ) . 
Chromosomal DNA ( 400 ng ) was labelled with Cy5-dCTP using the Klenow fragment . 
After labelling , each Cy5-labelled cDNA sample was combined with Cy3-labelled chromosomal DNA and hybridized to a microarray overnight at 65 °C . 
Data acquisition and transcriptomic data analysis
After hybridization , slides were washed and scanned using a GenePix 4000A scanner ( Axon Instruments ) . 
Fluorescent spots and the local background intensities were quantified using GenePix 5.0 software ( Axon ) . 
To compensate for unequal dye incorporation , data centring was performed bringing the median Ln ( Red/Green ) for each block to zero ( one block being defined as the group of spots printed by the same pin ) . 
Data visualization and data mining were performed using GeneSpring 7.1 ( Silicon Genetics ) . 
The complete dataset is available at GEO ( Accession Number GSE18452 ) . 
Real-time PCR
To determine the levels of rpoS mRNA in wild type and DstpA strains through growth in batch-culture , overnight cultures were diluted 1:1000 in LB and cells were grown as for the transcriptomic experiment at 37 °C under aeration ( 250 r.p.m. ) . 
Aliquots were removed at regular intervals and treated with 0.2 vol . 
of stop solution ( 90 % EtOH ; 10 % water-saturated phenol ) . 
Total RNA was isolated as described above , using the Promega SV total RNA purification kit and RNA concentrations were determined on a Nanodrop machine ( NanoDrop Technologies ) . 
The RNA ( 5 mg ) was reverse-transcribed in 25 ml of Stratascript first-strand buffer in the presence of 0.5 mM dNTPs , 1 mg of random hexamers and 50 U Stratascript ( Stratagene ) . 
The relative amounts of target mRNA were then determined by real-time PCR using SYBR Green JumpStart Taq ReadyMix following the manufacturer 's instructions ( Sigma ) . 
The real-time PCR was performed using genespecific primer pairs ( Table S2 ) designed in silico ( http : / / frodo.wi.mit.edu/primer3/ ) to generate amplicons in the 100 -- 120 bp range . 
ampD was used as an internal standard as it generally displays little variation in the transcriptional studies performed in our lab . 
In all the transcriptomic data presented in this article , ampD did not change by > 1.3-fold . 
Environmental stress survival
The ability to survive various environmental challenges was measured on S. Typhimurium bacterial cultures grown in LB until MEP ( A600 = 0.12 -- 0.15 ) or LEP ( A600 = 1.0 -- 1.2 ) . 
At the appropriate time , hydrogen peroxide ( 20 mM final concentration ) or polymyxin B ( 4 mg ml-1 ) was added . 
Survival at low pH was tested by diluting MEP cultures 1:100 into LB poised to pH 3.0 with HCl . 
Resistance to salt was examined by diluting MEP or LEP cultures 1:10 into LB containing 3.3 M NaCl . 
Samples were taken at various time points and spread on LB agar , except for the samples subjected to acid shock that were spread on Tryptone Soya Agar ( Oxoid CM0131 ) to promote optimal recovery . 
Colony forming units were enumerated and the relative survival of each strain was determined at each time point . 
For each test , at least three biological replicates were quantified . 
To determine the amounts of StpA3 ¥ FLAG or s38 protein levels , cells were lysed by sonication ( 10 mm amplitude ; MSE Soni-prep 150 ) . 
Total protein amounts were then determined using the bicinchoninic acid assay ( Sigma , Cat . 
BCA-1 ) . 
For each sample , 15 mg of total protein were re-suspended in SDS sample buffer ( Sigma , Cat . 
S3401 ) and run on NuPAGE 12 % Bis-Tris gels ( Invitrogen ) . 
Transfer to PVDF membrane was performed as described by the manufacturer ( Invitrogen ) and the membranes were then blocked for 45 min in 10 % Marvel Milk prepared in TBS-Tween 20 ( 25 mM Tris ; 0.8 % NaCl ; 0.02 % KCl ; 0.05 % Tween 20 ; pH 7.5 ) . 
The membranes were subsequently probed for 2 h with 1:1000 anti-s38 ( Neoclone , Cat . 
W0009 ) or 10 mg ml-1 anti-FLAG M2 antibody ( Sigma , Cat . 
F3165 ) serum in 10 % Marvel Milk prepared in TBS-Tween 20 . 
Washes and chemiluminescent immunodetection were performed as described previously ( Hautefort et al. , 2008 ) . 
Spectinomycin assay of protein stability
The S. Typhimurium SL1344 cultures were grown to LEP , at which point spectinomycin was added to a 100 mg ml-1 concentration to stop protein synthesis ( Zhou and Gottesman , 2006 ) . 
Bacterial samples were taken at regular intervals and immediately precipitated by adding 1/6 volume 30 % TCA . 
The samples were then left on ice for 20 min and centrifuged at 20 000 g for 20 min at 4 °C . 
Pellets were washed with cold acetone and centrifuged again at 20 000 r.p.m. for 20 min . 
After air-drying , the pellets were re-suspended in SDS sample buffer for Western blot analysis . 
Chromatin immunoprecipitation
Overnight cultures of S. Typhimurium stpA3 ¥ FLAG or S. Typh-imurium SL1344 were diluted 1000-fold into 250 ml flasks containing 25 ml of LB and grown at 37 °C under aeration . 
After reaching mid-exponential growth ( OD600 0.12 ) , formaldehyde ( Sigma ) was added to reach a 1 % final concentration and the cultures incubated for 15 min at 37 °C . 
The cross-linking was then stopped by adding 1/4 vol . 
1 M glycine . 
Cells were washed three times in ice-cold PBS ( pH 7.4 ) , re-suspended in 500 ml lysis buffer ( 10 mM Tris pH 8.0 ; 50 mM NaCl ; 10 mM EDTA ; 20 % sucrose ; 20 mg ml-1 lysozyme ) and incubated 45 min at 37 °C at which point 500 ml 2 ¥ RIPA ( 100 mM Tris pH 8.0 ; 300 mM NaCl ; 2 % Nonidet P40 ; 1 % sodium deoxycholate ; 0.2 % SDS ) was added . 
Chromatin was solubilized by sonication ( 10 mm amplitude ; MSE Soniprep 150 ) until DNA fragments were between 300 and 750 bp , and the lysate centrifuged at 12000 g for 10 min to remove debris . 
To determine the binding of StpA3 ¥ FLAG to the S. Typhimurium genome , the cell extract from S. Typhimurium stpA3 ¥ FLAG was incubated with anti-FLAG M2 antibody for 4 h at 4 °C . 
Chromatin immuno-precipitation with the anti-FLAG M2 antibody on the SL1344 wild-type strain provided the negative control . 
Co-immunoprecipitation of H-NS and DNA was performed by incubating S. Typhimurium SL1344 lysate with H113 monoclonal anti-H-NS antibody ( Sonnenfield et al. , 2001 ) . 
As a negative control , immunoprecipitation was obtained by incubating the cell extract without the addition of any antibody.After the incubation of the cell extracts with or without antibody , 50 ml protein G beads ( Sigma , E3405 ) were added and left for 16 h at 4 °C . 
The protein G beads were then washed twice in 1 ¥ RIPA , twice in wash solution ( 10 mM Tris pH 8.0 ; 250 mM LiCl ; 1 mM EDTA ; 0.5 % Nonidet P40 ; 0.5 % sodium deoxycho-late ) and twice in TE ( 10 mM Tris pH 8.0 ; 1 mM EDTA ) . 
The immunoprecipitate was eluted in 150 ml of elution buffer ( 50 mM Tris pH 8.0 ; 10 mM EDTA ; 1 % SDS ) prewarmed at 65 °C . 
Cross-linking was reversed by incubating the eluate in 0.5 ¥ elution buffer containing 0.8 mg ml-1 pronase ( Sigma ) and DNA purified using the Qiagen PCR purification kit . 
The StpA and H-NS ChIP data are available at GEO ( Accession Number GSE18452 ) . 
Acknowledgements
We thank Ida Porcelli and Gary Rowley for useful discussions , and Martin Goldberg for assistance with the construction of plasmid pMDH20 . 
We are indebted to Lionello Bossi for providing plasmid pSUB11 for the construction of the epitope-tagged StpA . 
We are grateful to Roy Bongaerts and Isabelle Hautefort for their help with the analysis of the stpA-gfp + transcriptional fusion . 
We also thank Fran Mulholland and Nigel Belshaw for technical assistance . 
We acknowledge funding from the BBSRC Core Strategic Grant to J.H. . 
Supporting information
Additional supporting information may be found in the online version of this article . 
Please note : Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors . 
Any queries ( other than missing material ) should be directed to the corresponding author for the article .