The LysR‐type transcriptional regulator LelA co‐regulates various effectors in different Legionella species

The intracellular pathogen Legionella pneumophila translocates more than 300 effector proteins into its host cells. The expression levels of the genes encoding these effectors are orchestrated by an intricate regulatory network. Here, we introduce LelA, the first L. pneumophila LysR‐type transcriptional regulator of effectors. Through bioinformatic and experimental analyses, we identified the LelA target regulatory element and demonstrated that it directly activates the expression of three L. pneumophila effectors (legL7, legL6, and legU1). We further found that the gene encoding LelA is positively regulated by the RpoS sigma factor, thus linking it to the known effector regulatory network. Examination of other species throughout the Legionella genus revealed that this regulatory element is found upstream of 34 genes encoding validated effectors, putative effectors, and hypothetical proteins. Moreover, ten of these genes were examined and found to be activated by the L. pneumophila LelA as well as by their orthologs in the corresponding species. LelA represents a novel type of Legionella effector regulator, which coordinates the expression of both adjacently and distantly located effector‐encoding genes, thus forming small groups of co‐regulated effectors.

The Fis repressors also silence the expression of EEGs and their local regulators after they are acquired via horizontal gene transfer (HGT) in the Legionella genus (Linsky et al., 2020;Linsky & Segal, 2021).(v) The Fur regulator was shown to control the expression of a single EEG and several other proteins involved in iron acquisition (Chatfield et al., 2012;Hickey & Cianciotto, 1997;Isaac et al., 2015;Robey & Cianciotto, 2002).More recently, two local regulatory systems of EEGs, found in only a few Legionella species, were identified in L. pneumophila.(i) The LciRS TCS that is activated by copper and controls the expression of a single adjacent EEG located in the same genomic island with it (Linsky et al., 2020).(ii) The RegK3 LuxR family transcriptional regulator, which controls the expression of two adjacent EEGs located in the same genomic island as RegK3 (Linsky & Segal, 2021).These direct global and local regulators of EEGs form a highly interconnected regulatory network, with regulators regulating the expression of EEGs as well as other EEG regulators (Linsky et al., 2020;Linsky & Segal, 2021;Rasis & Segal, 2009), and accessory components such as LerC, LetE, and PTS Ntr that connect and coordinate the EEGs regulatory systems, and balance the expression of different groups of effectors (Bachman & Swanson, 2004b;Feldheim et al., 2016Feldheim et al., , 2018;;Speiser et al., 2017).Other regulators, such as LqsR, LvbR, ArgR, and OxyR, were also shown to affect the expression levels of EEGs (Hochstrasser et al., 2019;Sahr et al., 2017;Tanner et al., 2017;Tiaden et al., 2007), but not to directly regulate their expression.
LysR-type transcriptional regulators (LTTRs) are abundant prokaryotic DNA-binding proteins (Baugh et al., 2023;Maddocks & Oyston, 2008;Mayo-Perez et al., 2023).The basic structure of the LTTR family members includes an N-terminal helix-turn-helix (HTH) domain, which binds to a specific DNA regulatory element, and a C-terminal co-inducer binding domain.To mediate their function, members of this family typically form homotetramers (Henikoff et al., 1988;Maddocks & Oyston, 2008;Oliver et al., 2016;Schell, 1993).LTTRs can function as either activators, repressors or dual regulators of gene expression, and either as local regulators that control the expression of a few genes located adjacent to them or as global regulators (Baugh et al., 2023, Maddocks & Oyston, 2008, Mayo-Perez et al., 2023).The LTTRs family contains numerous transcription factors involved in regulating various physiological processes in bacteria, including biosynthesis, metabolism, quorum sensing, oxidative stress, and virulence gene expression.
We hypothesized that one or more of the L. pneumophila LTTRs might function as a local regulator of EEGs.We found that orthologs of the L. pneumophila LTTR lpg2402 are present in several Legionella species, and in most of them, they are located adjacent to an EEG.Bioinformatic and experimental analyses revealed that lpg2402 is the first L. pneumophila LTTR that directly regulates the expression of EEGs.Moreover, the lpg2402 LTTR regulator represents a novel type of Legionella EEGs transcriptional regulator that controls the expression of EEGs located adjacent to it as well as EEGs located elsewhere in the genome, thus generating a small group of transcriptionally co-regulated effectors.

| The distribution of the lpg2402 LTTR orthologs in the Legionella genus
The notion that some of the L. pneumophila EEGs transcriptional regulators are located adjacent to their target genes (Linsky et al., 2020;Linsky & Segal, 2021), led us to seek additional such regulatory systems.This search resulted in the identification of the LTTR lpg2402 and the EEG legL7 (lpg2400) (de Felipe et al., 2008) located adjacently in L. pneumophila (Figure 1a).To gain insights into the lpg2402 orthologs and their putative target genes in the Legionella genus, we examined all the fully sequenced Legionella species for close orthologs of lpg2402.Close orthologs of lpg2402 (with the lowest homology of 59.2% identical, 78.6% similar amino acids, and BLAST e value of 10 −122 ) were found in 11 characterized Legionella species and two uncharacterized Legionella species (Figure 1a).In three of these species (L.pneumophila, L. bozemanii, and L. gratiana) the same EEG (legL7) was located adjacent to the regulator encoding gene.Interestingly, in five other Legionella species (L.spiritensis, L. wadsworthii, L. anisa, L. parisiensis, and the uncharacterized species PC1000) a different EEG (legA1) was located next to the lpg2402 LTTR ortholog and in L. clemsonensis an ortholog of the EEG lpg1986 is located adjacent to it (Figure 1a).
Careful examination of the upstream regulatory regions of these nine EEGs, led us to identify a putative regulatory element located at a similar distance from their putative −10 promoter element (Figure 1b).This conserved regulatory element constitutes a typical LTTR regulatory element (Baugh et al., 2023;Goethals et al., 1992;Maddocks & Oyston, 2008;Oliver et al., 2016) harboring the sequence motif AT-N 11 -AT-N 7 -AT-N 11 -AT and containing additional conserved positions which form an inverted-repeat sequence between the conserved ATs (Figure 1b).
The identification of this conserved putative regulatory element made it possible to search for it in the genomic sequences of all the Legionella species harboring a lpg2402 ortholog.This analysis revealed that three L. pneumophila EEGs (legL7, legL6, and legU1) harbor a very similar regulatory element (Figure 2a) and, altogether, 34 genes containing the putative regulatory element were identified in the 13 Legionella species harboring the regulator (Figure S1).These genes encode effectors, putative effectors, and hypothetical proteins (see below).Following these and the experimental results presented hereafter, lpg2402 was named LelA for Legionella EEGs (effector-encoding genes) LTTR (LysR-type transcription regulator).

Legionella genus
LelA orthologs are present in only 11 out of the 60 characterized Legionella species sequenced (LelA has no close orthologs in any other bacteria), and their distribution in the Legionella genus does not coincide with the Legionella species phylogenetic tree.This led us to examine whether LelA undergoes HGT in the Legionella genus.To this end, we reconstructed the phylogenetic tree of all the LelA orthologs (Figure S2A).We compared this tree to the F I G U R E 1 Distribution of lpg2402 in the Legionella genus.(a) Schematic representation of the genes located in the genomic region adjacent to the lpg2402 (lelA) LTTR orthologs in 11 characterized Legionella species, and two uncharacterized Legionella species (PC997 and PC1000), harboring the regulator.Homologous genes are marked by the same color and pattern, and non-homologous genes are marked in white.The genes are indicated by their locus tag number.The position of the conserved regulatory element predicted to be recognized by the lpg2402 LTTR is marked by red arrows pointing downward.The maximum-likelihood phylogeny tree on the left was reconstructed using amino acid alignment of the lpg2402 orthologous ORFs from the 13 Legionella species harboring the regulator.(b) The regulatory regions of genes located adjacent to lpg2402 and harboring a conserved regulatory element.The putative −10 and −35 promoter elements are in blue, the nucleotides representing the common LTTR motif are in red, other conserved nucleotides are in bold and the inverted-repeat sequence is marked by arrows.
phylogeny of the PmrA regulator present in all the Legionella species (Figure S2B).While the analysis using PmrA resulted in a tree structure similar to that of the Legionella species phylogenetic tree (Burstein et al., 2016;Gomez-Valero et al., 2019), the tree based on the LelA orthologs resulted in a very different topology (Figure S2).In addition, in L. bozemanii, two paralogs of lpg2402 were identified, and they are located in different clades on the LelA phylogenetic tree, even though they are found in the same Legionella species (Figure S2A).This situation can arise either from two independent HGT events or a single HGT event and duplication of the regulator in L. bozemanii.However, since the two L. bozemanii paralogs show a large difference in their GC content (Table S1) and are diverged (as evident from the tree), it is more likely that they arise from two independent HGT events.
Nevertheless, both L. bozemanii paralogs probably recognize the same regulatory element and, therefore, share the same regulons (see below).Moreover, the GC content of the genes encoding the LelA orthologs was considerably lower in some of the Legionella species in comparison to the genomic GC content (Table S1), suggesting a recent HGT event (see Section 3; Ochman et al., 2000;Zhang et al., 2014).Collectively, the gene distribution in the genus, its phylogeny, and the GC content differences, suggest that the LelA regulator underwent numerous HGT events in the Legionella genus.The levels of expression of the lacZ fusions were found to be significantly different (*p < 10 −5 , unpaired Student's t-test) between expression levels of the same lacZ fusion in the deletion mutants and the one in the wild-type strain.(d-f) LelA activates the expression of legL7, legL6, and legU1.The expression of wild-type lacZ fusions (white bars) of legL7 (d), legL6 (e), and legU1 (f), and the same fusions containing a mutation in the suspected LelA regulatory element (gray bars), were examined in L. pneumophila containing a deletion in lelA or fis3 (for legL7).The bacteria examined contained a plasmid with the L. pneumophila lelA gene cloned under the control of the Ptac promoter (activated by IPTG), and they were grown in media containing different concentrations of IPTG (indicated below the bars).Bacteria containing the same lacZ fusions without the lelA gene were used as controls (left columns in all panels).The levels of expression of the lacZ fusions were found to be significantly different (*p < 10 −5 , unpaired Student's t-test) between the fusions containing the wild-type regulatory region and the mutated regulatory region in each IPTG concentration.β-Galactosidase activity was measured as described in the Experimental Procedures section.Data (expressed in Miller units [M.U.]) are the average ± standard deviations (error bars) of the results from at least three different biological replicates.

| LelA regulates the expression of three L. pneumophila EEGs
To determine whether LelA regulates the expression levels of the three L. pneumophila EEGs (legL7, legL6, and legU1, Figure 2a) harboring its putative regulatory element, we constructed a deletion mutant of the lelA gene.Examining this deletion mutant in Acanthamoeba castellanii, an amoeba host of Legionella indicated that the lelA gene is dispensable for intracellular growth in this host cell (Figure S3), and a similar results was also reported in other host cells (Park et al., 2020).Previously other local regulators of EEGs were also found to be dispensable for intracellular growth (Linsky et al., 2020;Linsky & Segal, 2021), and these results are explained by the functional redundancy among effectors (O'Connor et al., 2011(O'Connor et al., , 2012)).Next, we examined whether LelA affects the expression levels of the three EEGs indicated above by determining the expression levels of EEGs-lacZ fusions in the L. pneumophila wildtype strain and comparing them to the expression in the lelA deletion mutant (Figure 2b).In the lelA deletion mutant, we observed a minor (but statistically significant) reduction in the expression levels of two EEGs (legL6 and legU1) and no effect on legL7, indicating that LelA might function as a positive regulator of legL6 and legU1.
The above analysis also indicated that the expression levels of the three EEGs examined are very low, which might suggest that they are subjected to Fis repression, as was previously shown for numerous EEGs (Linsky et al., 2020;Linsky & Segal, 2021;Zusman et al., 2014).Therefore, we examined the expression levels of the three EEGs in the fis1 (lpg0542) and fis3 (lpg1743) deletion mutants (Figure 2c) (a double deletion of fis1 and fis3 is inviable).All three EEGs were found to have higher expression levels in the fis3 deletion mutant, indicating that Fis3 represses their expression, and might hinder our ability to observe the effect of LelA on their expression.
These results led us to use a controlled expression system to examine the LelA regulation of the three EEGs.In this system, the lelA gene was cloned under the regulation of the Ptac promoter (induced by IPTG), which was introduced into the L. pneumophila lelA deletion mutant together with the three EEGs-lacZ fusions indicated above (Figure 2d-f and Figure S4).Using this analysis, all three EEGs showed dose-dependent expression levels as the concentration of IPTG (controlling the expression level of the lelA gene) was increased.Since the expression levels obtained with LegL7 were relatively low (Figure S4), its activation by LelA was also examined in the fis3 deletion mutant (which represses its expression, Figure 2c).Indeed, stronger activation by LelA was observed (compare Figure 2d and Figure S4A) in the fis3 deletion mutant.To completely avoid the effect of Fis repression on these EEGs, we also examined the controlled expression system in the surrogate host E. coli, and the concentration of IPTG used determined the expression levels of the three EEGs-lacZ fusions in E. coli (Figure S5A-C, white bars), proving that LelA regulates their expression.Furthermore, we examined the effect of LelA overexpression on two EEGs (sidH and lubX) that do not harbor the regulatory element we identified and are known to be regulated by PmrA and Fis, respectively.Both EEGs in both L. pneumophila and E. coli showed no activation by LelA (Figure S6A-D), supporting the specificity of LelA to the activation of its target genes using the overexpression system.Therefore, we concluded that the LelA regulator activates the expression level of three L. pneumophila EEGs, indicating that it might function as their direct positive regulator.

| LelA requires the conserved regulatory element located upstream of the three EEGs to activate their expression
To examine the connection between LelA and the regulatory element identified, we constructed an AT to TA mutation in the third AT of the consensus sequence in each of the three EEGs (legL7, legL6, and legU1; Figure 2a).The lacZ fusions harboring these mutations were cloned into the controlled expression system described above and examined in L. pneumophila lelA deletion mutant (legL7 was also examined in the fis3 deletion mutant, as described above; Figure 2d-f and Figure S4A).The results obtained indicated that the AT to TA mutations in the regulatory element identified completely abrogated the ability of LelA to activate the expression of the three EEGs.The lacZ fusions harboring the mutations in the regulatory region of these three EEGs were also examined in E. coli, and similar results were obtained (Figure S5A-C).Collectively, our results further suggest that the conserved regulatory element we identified plays a critical role in the activation of legL7, legL6, and legU1 by LelA.

| LelA directly binds to the regulatory region of the three effectors it regulates
To further support the results presented, the L. pneumophila LelA protein was His-tagged, overexpressed, purified, and used for gel mobility shift assays with the legL7, legL6, and legU1 regulatory regions.The L. pneumophila LelA-His 6 protein bound to the regulatory region of these three EEGs, as evident by a migration shift of the DNA probe (Figure 3a-c).The amounts of shifted probes positively correlated with the amounts of LelA-His 6 (Figure 3a-c, lanes 2-6).
In addition, competition with an unlabeled probe reduced the band shift (Figure 3a-c, compare lanes 5 and 7).To further validate the binding specificity, we performed additional competition assays with the same amount of unrelated DNA, which showed no reduction in the band shift (Figure 3a-c, compare lanes 7 and 8).The mobility shift assays (Figure 3), together with the LelA overexpression analysis (Figure 2), and the analysis of the mutations in the LelA consensus sequence (Figure 2), establish LelA as a direct activator of legL6, legL7, and legU1 EEGs in L. pneumophila.

| lelA expression is regulated by the RpoS sigma factor
We were interested to explore the way by which the lelA gene itself is regulated.The lelA gene was found to be expressed at very low levels (Figure 4), which might suggest that it is repressed by Fis, as was previously described for the two other L. pneumophila EEGs local regulators (Linsky et al., 2020;Linsky & Segal, 2021).Analyzing the regulatory region of different Legionella lelA orthologs (Figure 4a), revealed in a few of them putative Fis regulatory elements (G-N 13 -C, Zusman et al., 2014), but not in a conserved position (Figure 4a).
When the lelA expression levels were examined in the fis1 and fis3 deletion mutants no significant effect was observed (Figure 4b).
Since it was previously shown that Fis1 and Fis3 can compensate for the absence of one another (Zusman et al., 2014), we mutagenized the two putative Fis regulatory elements identified, but in this case too, no change in the lelA expression level was observed, indicating that lelA is probably not regulated by Fis (Figure 4c).The expression level of the L. pneumophila lelA was also unaffected by its own deletion or overexpression (Figure 4b and Figure S4B), indicating that it is not an autoregulator.This is in line with the absence of the LelA target regulatory element in the lelA upstream regulatory region.
It was previously shown, using a rpoS deletion array, that several L. pneumophila LTTRs, including LelA, are downregulated in the absence of RpoS (Hovel-Miner et al., 2009).Examination of the −10 promoter region of the lelA orthologous genes in the Legionella species indicated that the extended −10 promoter sequence closely matches that of the consensus sequence of the E. coli RpoS (Landini et al., 2014;Schoebel et al., 2010).To examine whether RpoS indeed regulates the expression level of lelA, the expression level of lelA was examined in a rpoS deletion mutant and compared to the wild-type strain.A clear reduction in lelA expression level was obtained in the rpoS deletion mutant (Figure 4b).Furthermore, overexpression of RpoS increased the expression level of lelA (Figure 4d) indicating that RpoS positively regulates its expression.
The regulation of lelA by RpoS (Figure 4b,d), and the Fis repression of the EEGs regulated by LelA (Figure 2c), complement each other, since RpoS activates gene expression at stationary phase (Bachman & Swanson, 2001, 2004a;Hales & Shuman, 1999;Hovel-Miner et al., 2009) and Fis represses gene expression stronger at exponential phase (Zusman et al., 2014).Therefore, the combined regulation of lelA by RpoS and of the LelA target EEGs by Fis, predicts that the EEGs under LelA regulation will show increased expression at the stationary phase compared to the exponential phase.Examination of the three EEGs regulated by LelA and the lelA gene itself at the exponential and stationary phases revealed that two of the EEGs (legL6 and legU1) and the lelA gene have indeed increased expression levels at the stationary phase (Figure 4e).However, legL7 exhibited very low expression levels due to Fis repression, which is probably the reason why there was no observable effect of the growth phase on its expression (Figure 4e).This is similar to the lack of effect of the lelA deletion mutant on its expression levels, Figure 2b.Our results reveal that lelA itself and the EEGs it regulates are connected to the EEGs regulatory network by RpoS and Fis, respectively.This is the first reported connection between direct regulators of EEGs and the RpoS sigma factor in L. pneumophila.

| The activation by LTTRs depends on a specific regulatory element
As presented above, legU1 (lpg0171) was strongly activated by LelA.
Examination of the legU1 genomic region revealed another LTTR (lpg0173) located adjacent to it (Figure 5a).Bioinformatic analysis indicated that lpg0173 orthologs are found in 15 Legionella species (Figure S7), and they are always positioned adjacent to the lpg0174-lpg0175-lpg0176 operon, like in L. pneumophila (Figure 5a).This operon encodes a Pyoverdine/dityrosine biosynthesis protein, TauD/ TfdA Taurine family dioxygenase and a FAD-binding protein, and the function of these proteins in L. pneumophila is unknown.Examination of the intragenic region between lpg0173 and lpg0174 revealed a putative LTTR regulatory element with a sequence motif (AT-N 11 -AT-N 7 -AT-N 11 -AT) similar to the one recognized by LelA (Figure S7).However, in comparison to the LelA consensus sequence, the putative lpg0173 target regulatory element harbors different conserved nucleotides and a different inverted-repeat sequence between the conserved ATs (Figure 5b).To determine whether legU1 activation by The Legionella pneumophila LelA-His 6 protein binds to the legL7, legL6, and legU1 regulatory regions.Gel mobility shift assay was performed with purified LelA-His 6 protein and the DIG-labeled legL7 (a), legL6 (b), and legU1 (c) regulatory regions.The first lane did not contain any protein.The rest of the lanes contained increasing amounts of the LelA-His 6 protein, starting from 4 nM.The competition assay was performed using an unlabeled probe as a specific competitor (unlabeled probe) or the same amount of an unrelated DNA (Unrelated comp).
LelA is specific and if the lpg0173 LTTR located adjacent to legU1 can affect its expression, we generated a deletion mutant and overexpression system for the lpg0173 LTTR.We first examined whether lpg0173 affects the expression of lpg0174.The expression levels of lpg0174 were lower in the lpg0173 deletion mutant compared to its expression level in the wild-type strain.Overexpression of lpg0173 under Ptac control resulted in activation of the lpg0174 expression level (Figure 5c).These results indicate that the lpg0173 LTTR activates the expression of lpg0174.To determine whether lpg0173 affects the expression of legU1, it was examined using lpg0173 overexpression, and no activation was observed (Figure 5d).The opposite result was obtained with LelA, which activated the expression of legU1 (Figure 2f), but did not affect the expression level of lpg0174 (Figure 5e).Taken together, these results indicate that even though legU1 is located adjacent to the lpg0173 LTTR, lpg0173 does not affect the expression of legU1 and each of the two LTTRs (LelA and lpg0173) activates the expression of the gene harboring its specific regulatory element (legU1 and lpg0174, respectively).

| Properties and distribution of the LelA putative target genes in the Legionella genus
Previously we described the RegK3 and LciR local regulators of EEGs (Linsky et al., 2020;Linsky & Segal, 2021), which are present in six Legionella species (not the same species), each of them controls the expression of the same EEGs in all the Legionella species harboring them, and all these EEGs are located in the same "regulator-effector island" together with their regulator (see introduction).On the contrary, LelA is the first Legionella EEGs regulator predicted to control the expression of different EEGs located adjacent to it in different Legionella species (Figure 1), and it is also predicted to control the expression of additional genes located elsewhere in the genomes of these species (Figure 6).In addition, the LelA regulator is found in a larger number of Legionella species (11 characterized species; Figure 7, red boxes), in comparison to LciR and RegK3.Therefore, we were interested to examine the genes predicted to be regulated by the Legionella LelA orthologs in other Legionella species.To this end, we analyzed the 34 genes which harbor the putative LelA regulatory element (Figure S1), which is present in all the Legionella species encoding LelA (Figure 6).Altogether, genes from 11 orthologous groups were found to contain the regulatory element.Most of these genes are shared between the Legionella species harboring a LelA ortholog (Figure 7, green and blue boxes).The number of putative target genes identified in each of the Legionella species ranged from one to four (Figures 6 and 7).In most of these species, one of these genes is located adjacent to the regulator (Figure 7, green boxes), and the rest are scattered in the genome (Figure 7, blue boxes).In addition, in some of the species harboring a lelA ortholog, no gene with its regulatory element was located adjacent to it (Figure 6); however, in these species (namely L. hackelia, L. jamestowniensis, and L. lansingensis), a putative LelA regulatory element was found in several genes located elsewhere in the genome (Figure 7, blue boxes).
Besides the three validated EEGs (legL6, legL7, and legU1) found in L. pneumophila, and the two validated EEGs (legA1 and lpg1986) found adjacent to the lelA orthologous genes in the other Legionella species (Figure 7, green boxes), the putative regulatory element was also found upstream to genes encoding proteins with common effector domains (Burstein et al., 2016;Cazalet et al., 2004), such as the Ankyrin, F-box, and SidC domains (Figure 6).Altogether, out of the 34 genes which were found to contain the LelA regulatory element, are consistently located distantly from the regulator (for example the gene encoding the putative effector containing a SidC domain is present in seven Legionella species, and in none of them it is located adjacent to the regulator (Figure 7, blue boxes; orthologous group #5).This phenomenon occurs with six out of the 11 orthologous groups (Figure 7, blue boxes; orthologous groups #4-8, and 10).Few genes harboring the LelA regulatory element were found to be located adjacent to the regulator in one species but not in other species (for example, legL7 is located adjacent to the regulator in L. pneumophila and distantly in the genome in L. hackelia, Figure 7).
In the Legionella species that contain only a single gene harboring the regulatory element, it is always located adjacent to the regulator (Figure 6), and orthologous genes are also found adjacent to the regulator in other Legionella species (for example the Lspi_1396 gene is the only gene containing the regulatory element in L. spiritensis and its ortholog is located next to the regulator also in L. anisa and other Legionella species).In addition, most (all besides two) of the genes harboring the LelA regulatory element are found in more than a single Legionella species, even though these species are not closely related.Some of the genes harboring a LelA regulatory element are also found in Legionella species which do not contain a LelA ortholog, and in these species, the LelA target regulatory element is also missing (Figure 7, gray boxes).Collectively, the genus distribution of the genes regulated by LelA is dynamic and these genes seem to rapidly undergo changes in their regulation and variation in their presence/ absence pattern in different Legionella species.

| LelA orthologs from different Legionella species control the expression level of genes harboring the LelA target regulatory element
To experimentally examine how LelA orthologs regulate the genes described above, we chose LelA orthologs from three Legionella species (L.hackelia, L. anisa, and L. lansingensis).They were selected due to the evolutionary distance between the orthologs, providing a good sample of the LelA phylogeny (Figure 1a).Two of these lelA orthologs (from L. lansingensis and L. hackelia) have no putatively regulated gene located adjacent to them, and the putative effectors predicted to be regulated by these LelA orthologs (Lhac_0817, Lhac_0975, Lani_1686, Llan_0301, and Llan_2142), harbor common effector domains (Figure 6).We constructed lacZ fusions for ten genes containing the regulatory element (four from L. hackelia, three from L. anisa, and three from L. lansingensis) and examined their expression level in L. pneumophila.All the genes F I G U R E 7 Distribution of the LelA orthologs and their target genes in the Legionella genus.A maximum-likelihood tree of 60 characterized Legionella species was reconstructed based on the concatenated amino acid alignment of eight orthologous ORFs, present in all Legionella species.For each species, the presence of a LelA orthologous regulator (red), the presence of an EEG located adjacent to the regulator harboring the LelA target regulatory element (green), the presence of a gene located elsewhere in the genome harboring the LelA regulatory element (blue), and the presence of an orthologous gene which does not contain the LelA regulatory element (gray), are presented.Each column of boxes represents a single orthologous group, the orthologs are listed on the left.except one showed very low expression levels in L. pneumophila and all of them were not affected by the lelA deletion (Figure S8), similar to the results obtained with the L. pneumophila LegL7 EEG (Figure 2b).Therefore, to avoid Fis repression, we examined the effect of LelA on these genes using the controlled expression system in E. coli, which was found to be a good surrogate system for LelA analysis (Figures S5A-C and S6A-D).Each of the genes was examined using the L. pneumophila LelA ortholog and the corresponding LelA ortholog (Lhac_2674, Lani_1844, and Llan_1155).The expression levels of the genes were examined without the regulator, as well as with the regulator cloned under Ptac control in the absence and presence of IPTG (which controls the expression of the LelA orthologs).All ten genes examined were activated by the L. pneumophila LelA ortholog as well as by the LelA orthologs from their corresponding species (Figure 8a-c).The L. lansingensis LelA ortholog (Llan_1155) exhibited a unique expression pattern in this experiment: The expression levels of its target genes were very high even without the addition of IPTG, indicating that this ortholog might be more active (or stable) under the experiment's conditions and the leaky expression from Ptac was probably sufficient for it to strongly activate gene expression (Figure 8b).Collectively, our results suggest that all the LelA orthologs recognize the same regulatory element and activate the expression levels of the genes which harbor them.The genes activated by LelA encode effectors, putative effectors, and other hypothetical proteins, and most of them are shared between the Legionella species harboring the LelA regulator.
(ii) RegK3, a LuxR family regulator which controls the expression of two adjacent EEGs, that are located in the same genomic island with it (Linsky & Segal, 2021).We presented here a line of evidence showing that LelA regulates the expression of EEGs in L. pneumophila and other Legionella species (Figure 9).LelA is the first LTTR family member which is found to regulate the expression of EEGs in Legionella.
Similar to the other local regulators identified thus far, LciRS and RegK3, the LelA regulator was found to be dispensable for intracellular growth (Linsky et al., 2020;Linsky & Segal, 2021;Park et al., 2020).However, the LelA regulator differs from these two local regulators in several properties: The LelA LTTR represents a novel type of L. pneumophila EEGs local regulators, which regulate the expression of both adjacently and distantly located EEGs, which results in a small group of co-regulated effectors.As LciRS and RegK3 regulators, LelA underwent several HGT events in the Legionella genus (Figure S2 and Table S1).However, interestingly, comparing the GC content of lelA to the genomic GC content resulted in a large difference only in species in which LelA is predicted to regulate a single gene located adjacent to the lelA encoding gene (L.gratiana and L. spiritensis; Table S1), suggesting a more recent HGT event.In most of the Legionella species, the LelA regulator is also predicted to regulate the expression of genes located far from the lelA gene.In these species, the GC content of the lelA gene was more similar to the genomic GC content (Table S1).This suggests that the regulation of distant genes by LelA, and the lelA gene GC content amelioration occur in the same species and both are indications of earlier HGT events compared to the HGT events described above.
LelA was shown here to activate the expression of three L. pneumophila effectors (LegL6, LegL7, and LegU1) and this might suggest that these three effectors function together during infection.However, to date, only the function of LegU1 (harboring a eukaryotic F-box domain) was uncovered (Ensminger & Isberg, 2010).LegU1 was shown to interact with components of the host ubiquitination machinery and to form a functional Skp-Cullin-F-box complex that can specifically direct the ubiquitination of the host chaperone protein BAT3.Currently, there is no information regarding the function of the two other effectors regulated by LelA (LegL6 and LegL7), both harbor a eukaryotic leucine-rich repeat (LRR).Therefore, we cannot speculate whether these three effectors function together in the same pathway.Interestingly, in other Legionella species harboring LelA, two putative effectors encoded by genes regulated by LelA also harbor eukaryotic domains (ankyrin repeats and F-box domains, Figure 7), the function of these effectors is also unknown.
L. pneumophila harbors 13 regulators from the LTTR family and only one of them (OxyR) was studied thus far.OxyR (lpg1815), was shown to directly repress the expression of ahpC, an alkyl hydroperoxide reductase (LeBlanc et al., 2008;Tanner et al., 2017).In this study, we analyzed two additional L. pneumophila LTTRs.Our main focus was LelA (lpg2402) which was found to directly and positively regulate the expression of three EEGs.Due to the genomic proximity of one of the LelA target EEGs (legU1 -lpg0171) to another LTTR (lpg0173), the latter was studied as well.Lpg0173 was found to function as an activator of an operon located back-to-back to it (lpg0174-lpg0176), in all the Legionella species harboring it.Our results demonstrated that each of the two LTTRs (LelA and lpg0173) specifically regulate their target gene(s) (Figure 5c-e).Similarly to LelA and lpg0173 (as well as other LTTRs) OxyR also recognizes a conserved regulatory element which constitutes a typical LTTR regulatory element harboring the sequence motif AT-N 11 -AT-N 7 -AT-N 11 -AT (LeBlanc et al., 2008) and containing additional conserved positions which form an invertedrepeat sequence between the conserved ATs (Figure S9).Comparison of the regulatory elements of these three LTTRs (Figure S10), reveals the similar overall structure of their regulatory elements as well as differences in the position and sequence of the inverted-repeat located between the conserved ATs (Figure S10).These differences in the regulatory elements probably direct each of these LTTRs to recognize and control the expression of its specific target genes.
An additional interesting finding of our study is the regulation of the lelA regulator itself by the RpoS sigma factor.Previously, two microarray analyses of L. pneumophila found conditions affecting the expression level of several LTTRs, including lelA.One was a microarray comparing wild-type L. pneumophila to a rpoS deletion mutant (Hovel-Miner et al., 2009), and the second was a microarray comparing growth with and without nicotinic acid (Edwards et al., 2013).
The similarity between the genes affected in the two analyses suggests a connection between RpoS and nicotinic acid, but currently, there is no support for this hypothesis.The finding that the lelA gene is activated by RpoS and that the three EEGs regulated by LelA are repressed by Fis connects the LelA regulator and its target EEGs to the L. pneumophila effectors regulatory network in two points that were not known to be connected thus far.Fis was shown to repress F I G U R E 8 The LelA orthologs activate the expression of genes harboring the identified regulatory element.The bacteria examined contain a plasmid with a lacZ fusion and the same lacZ fusion with the lelA gene or the ortholog from the corresponding Legionella species (Lhac_2674, Lani_1844, and Llan_1155) cloned under the control of the Ptac promoter (activated by IPTG).The strains harboring the plasmids containing the regulator were examined in media containing (0.1 mM) or lacking IPTG.The expression levels of L. anisa lacZ fusions (Lani_0816, Lani_1686, and Lani_1843) with lelA or Lani_1844 cloned under Ptac control (a), the expression of L. lansingensis lacZ fusions (Llan_0301, Llan_1420, and Llan_2142) with lelA or Llan_1155 cloned under Ptac control (b), and the expression of L. hackelia lacZ fusions (Lhac_0817, Lhac_0975, Lhac_1802, and Lhac_2448) with lelA or Lhac_2674 cloned under Ptac control (c), were examined in E. coli.The levels of expression of the lacZ fusions were found to be significantly different (**p < 10 −4 ; *p < 10 −5 ; both by unpaired Student's t-test) between the expression of the same strain grown without IPTG and the one grown with IPTG, or between the lacZ fusion without the regulator and with the regulator grown without IPTG.β-Galactosidase activity was measured as described in the Experimental Procedures section.Data (expressed in Miller units [M.U.]) are the average ± standard deviations (error bars) of the results from at least three different biological replicates.many EEGs and regulators that underwent HGT with their EEGs as part of a "regulator-effector island" (Linsky et al., 2020, Linsky & Segal, 2021).In these cases, also the regulators (RegK3 and LciR) were found to be repressed by Fis.However, LelA is not repressed by Fis and instead, it is positively regulated by the RpoS sigma factor, which was previously shown to control the two sRNAs (RsmY and RsmZ) that are part of the LetAS-RsmYZ-CsrA cascade (Figure 9).These results further demonstrate the degree of complexity and connectivity of the L. pneumophila effectors regulatory network.
LelA is the first LTTR found to directly control the expression of EEGs in Legionella.LelA represents a novel type of effectors regulator, which controls the expression of adjacent and distant EEGs in L. pneumophila as well as in other Legionella species.This novel type of EEG regulation generates small groups of Legionella EEGs which are transcriptionally co-regulated during infection.This tight regulation suggests that the orchestration of small groups of EEGs might be crucial for the effective infection by Legionella.

| Bacterial strains, plasmids, and primers
The L. pneumophila wild-type strain used in this study was JR32, a streptomycin-resistant, restriction-negative mutant of L. pneumophila Philadelphia-1, which is a wild-type strain in terms of intracellular growth (Sadosky et al., 1993).In addition, mutant strains derived from JR32 as well as other Legionella species that were used in this study are listed in Dataset S1 in the supplemental material.The E. coli strains used in this work are also listed in Dataset S1.Plasmids and primers used in this work are listed in Datasets S2 and S3, respectively.
Site-directed mutagenesis was performed by regular PCR or the PCR overlap extension approach (Ho et al., 1989), as previously described (Zusman et al., 2007), to construct the following substitution mutations: a substitution mutation in the putative LelA binding site in the regulatory region of the legU1, legL6, and legL7 genes; a substitution mutation in the putative Fis binding site in the regulatory region of the lelA gene; and double substitution mutations of the two putative Fis sites in the regulatory region of To construct deletion substitution mutants in the L. pneumophila lelA, rpoS, and lpg0173 genes, a 1-kb DNA fragment located on each side of the planed deletions was amplified by PCR using the primers listed in Dataset S3.The resulting plasmids were digested with suitable enzymes, and the inserts were used for a four-way ligation containing the Kanamycin resistance cassette.The plasmids generated, pMG-pUC18-lpg2402-UP-Km-DW, pYF-RpoS-4way, and pJV-lpg0173-Km (Dataset S2), were digested with PvuII, and the resulting fragments were cloned into the pLAW344 allelic exchange vector digested with EcoRV to generate the plasmids pMG-pLAW344-UP-Km-DW-lpg2402, pYF-RpoS-pLAW, and pJV-lpg0173-pLAW (Dataset S2).The allelic exchange deletion substitution mutants were constructed as previously described (Segal & Shuman, 1997).
For the construction of the plasmid expressing the His-tagged LelA, the lelA gene was amplified by PCR using the primers listed in Dataset S3, cloned into pET-21a, and sequenced to generate the plasmid pZT-lpg2402-His6-4 (Dataset S2).

| β -Galactosidase assay
β-Galactosidase assays were performed as previously described (Zusman et al., 2007).L. pneumophila strains were grown for 48 hr on charcoal-yeast extract (CYE) plates containing chloramphenicol (Cm).The bacteria were scraped off the plate and suspended in ACES-yeast extract (AYE) broth, and the bacterial OD600 was calibrated to 0.1 in fresh AYE, containing different concentrations of IPTG (when indicated) and Cm.The resulting cultures were grown on a roller drum for about 18 h, until reaching an OD600 of about 3.2 (early stationary phase) and used for the β-galactosidase assay.
β-Galactosidase assays in E. coli were performed similarly, but the E. coli strains were grown for about 6 h in LB containing Cm and different concentrations of IPTG (when indicated) until reaching an OD600 of about 2.5 (early stationary phase) and used for the β-galactosidase assay.The assays were done for 20, 50, or 100 μL of culture, and the substrate for β-galactosidase hydrolysis was o-nitrophenylβ-D-galactopyranoside.

| Protein purification and gel mobility shift assay
LelA-His 6 was purified from E. coli BL21(DE3) using nickel bead columns (Qiagen), according to the manufacturer's instructions.
After purification, the fractions containing the protein were dialyzed overnight against a buffer containing 20 mM Tris (pH-7.5),50 mM KCl, 5 mM MgCl 2 , 0.1 mM EDTA, 0.1 mM dithiothreitol (DTT) and 30% Glycerol, then with the same buffer containing 50% Glycerol for few hours, and the purified protein was then stored at −20°C.A gel mobility shift assay was performed as previously described (Zusman et al., 2014), with a few modifications.
The putative regulatory region of legL6, legL7, and legU1 (160 bp, 161 bp, and 165 bp, respectively) was amplified by PCR using the primers listed in Dataset S3 and 3′ end-labeled with digoxigenin (DIG) by using DIG-11-ddUTP (Roche).Increasing amounts of the purified LelA-His 6 protein (between 4 and 360 nM) were mixed with 1.5 nM of the DIG-labeled probe in buffer containing 20 mM Tris (pH -7.5), 50 mM KCl, 30 mM NaCl, 5 mM MgCl 2 , 0.1 mM EDTA, 0.1 mM DTT, 60 μg/mL poly[d(I-C)], 60 μg/mL bovine serum albumin and 30 μg/mL herring sperm DNA.For the competition experiments, a 100-fold excess of the unlabeled probe or unrelated DNA was allowed to bind the LelA-His 6 protein for 15 min before the addition of the DIG-labeled probe.The binding reaction was carried out for 30 min at room temperature, and the samples were then loaded onto a 5% polyacrylamide-0.25×Tris-acetic acid-EDTA (TAE) gel in 0.5× TAE running buffer.Following electrophoresis, the gel was transferred to a nylon membrane and fixed by UV cross-linking.The DIG-labeled DNA fragments were detected

F
LelA regulates the expression of three Legionella pneumophila EEGs.(a) The regulatory regions of three L. pneumophila EEGs harbor the conserved regulatory element.The putative −10 and −35 promoter elements are in blue, the nucleotides representing the common LTTR motif are in red, other conserved nucleotides are in bold and the inverted-repeat sequence is marked by arrows.The nucleotides shaded in yellow were mutated (AT to TA). (b, c) Expression of legL7, legL6, and legU1 lacZ fusions was examined in wild-type L. pneumophila and the lelA deletion mutant (b) and in wild-type L. pneumophila and the fis1 and fis3 deletion mutants (c).

F
The expression of the lelA gene is activated by RpoS.(a) The regulatory regions of the lelA orthologous genes.The −10 promoter elements are in blue, the RpoS extended −10 promoter is in purple, the putative Fis regulatory elements are highlighted in yellow, relatively conserved nucleotides are marked in bold, the start codons are in red, the experimentally validated transcription start site of the L. pneumophila lelA gene is in bold and underlined, the position of the mutations constructed in the L. pneumophila lelA Fis putative regulatory elements (Mut-1 G to A and Mut-2 C to A) are marked by blue arrows, and the RpoS consensus sequence is indicated.(b) Expression of the lelA lacZ fusion in wild-type L. pneumophila and the fis1, fis3, lelA, and rpoS deletion mutants.The levels of expression of the lelA lacZ fusion were found to be significantly different (*p < 10 −5 , unpaired Student's t-test) between expression levels of the lelA lacZ fusion in the rpoS deletion mutant and the one in the wild-type strain.(c) Expression of lelA lacZ fusions containing mutations in the putative Fis regulatory elements (Mut-1 and Mut-2) and a combined mutation in both Fis regulatory elements (Mut 1 + 2).(d) The expression of the lelA lacZ fusion with RpoS cloned under Ptac control was examined in the L. pneumophila rpoS deletion mutant.The levels of expression of the lelA lacZ fusions were found to be significantly different (*p < 10 −5 , unpaired Student's t-test) between the expression of the lelA lacZ fusion without RpoS and the one grown under different IPTG concentrations.(e) Expression of the legL7, legL6, legU1, and lelA lacZ fusions in wild-type L. pneumophila at the exponential and the stationary phase.The levels of expression of the lacZ fusions were found to be significantly different (*p < 10 −5 , unpaired Student's t-test) between expression levels at the exponential and stationary growth phases.β-Galactosidase activity was measured as described in the Experimental Procedures section.Data (expressed in Miller units [M.U.]) are the average ± standard deviations (error bars) of the results from at least three different biological replicates.
68% (23 genes) encode validated effectors or proteins containing effectors domains.The other genes containing the LelA regulatory element encode hypothetical proteins that are present only in the Legionella genus and show no homology to any known proteins or known domains (Figure 6), and they thus could encode yet unknown effectors.Additionally, we identified several genes in different Legionella species which contain the LelA regulatory element, but F I G U R E 5 lpg0173 LTTR activates the expression level of lpg0174 and not the one of legU1.(a) Schematic representation of the genomic region of EEGs regulated by LelA.EEGs regulated by LelA are marked in light blue, genes predicted to be regulated by lpg0173 are marked in orange, the LTTR encoding genes (lelA and lpg0173) are marked in purple, and other genes are marked in white.The genes are indicated by their locus tag number.The positions of the conserved regulatory element recognized by the LelA LTTR are marked by red arrows pointing downward, and the regulatory element predicted to be recognized by the lpg0173 LTTR is marked by a green arrow pointing downward.(b) Comparison of the consensus sequences predicted to be recognized by LelA and lpg0173 (based on Figures S1 and S7, respectively).The conserved ATs are marked in red, additional conserved nucleotides are marked in bold and the inverted-repeat sequences are marked by arrows.The pink-shaded nucleotides mark the major differences between the two consensus sequences which overlap the inverted-repeat sequences.(c-e) lpg0173 LTTR activates the expression of lpg0174 and not the one of legU1.The bacteria examined contained a plasmid with the L. pneumophila lpg0173 or the lelA genes cloned under the control of the Ptac promoter (activated by IPTG), and they were grown in media containing different concentrations of IPTG (indicated below the bars).The expression of the lpg0174 lacZ fusion with lpg0173 cloned under Ptac control (c), the legU1 lacZ fusion with lpg0173 cloned under Ptac control (d), and the lpg0174 lacZ fusion with lelA cloned under Ptac control (e), were examined in L. pneumophila.The levels of expression of the lacZ fusions were found to be significantly different (*p < 10 −5 , unpaired Student's t-test) between the expression of the same strain grown without IPTG and the one grown under different IPTG concentrations.β-Galactosidase activity was measured as described in the Experimental Procedures section.Data (expressed in Miller units [M.U.]) are the average ± standard deviations (error bars) of the results from at least three different biological replicates.

F
I G U R E 9 Model of the regulatory network of direct regulators of Legionella pneumophila EEGs.The global EEGs regulatory systems-the CpxRA and PmrAB TCSs, the LetAS-RsmYZ-CsrA regulatory cascade, including RpoS, and the Fis1 and Fis3 NAPs, which were found to regulate numerous EEGs, are shown.In addition, the three local regulators of EEGs-the LciRS TCS and the LuxR type regulator RegK3, the first LTTR EEGs regulator LelA described in this study are shown, as well as the Fur regulator.DNA-binding proteins are in purple; connector proteins are in red; phosphorylation is indicated by a small yellow circle; RNA binding protein is in green; small RNAs are in blue; accessory components are in white and SHKs are illustrated as integral membrane proteins.Acetyl-P, acetyl phosphate; PEP, phosphoenolpyruvate; PTS, phosphotransferase system.The number of EEGs that were shown to be regulated by each of the regulatory systems is indicated in parentheses.Arrows and T-shaped symbols indicate activation and repression, respectively.Solid lines indicate direct regulation and broken lines indirect regulation.the lelA gene.The primers used for the site-directed mutagenesis are listed in Dataset S3, and the resulting plasmids are listed in Dataset S2.To construct IPTG-inducible lelA, the L. pneumophila lelA gene was amplified by PCR using the primers listed in Dataset S3.The PCR product was then digested with EcoRI and BamHI and cloned into pMMB207C to generate pML-pMMB207c-lpg2402.The insert of this plasmid was sequenced and then digested with BamHI and a Kanamycin resistance cassette (Pharmacia) was cloned into it to generate pNS-ptac-lpg2402-Km.The resulting plasmid was then digested with XbaI and NsbI, and the resulting fragment, containing Ptac-lelA together with the lacI gene, was cloned into the plasmids containing the lacZ fusions of the legU1, legL6, legL7, lelA, lubX, sidH, lpg0174, Lani_0816, Lani_1686, Lani_1843, Lhac_0817, Lhac_0975, Lhac_1802, Lhac_2448, Llan_0301, Llan_1420, and Llan_2142 genes digested with XmnI and XbaI, as well as plasmids containing the mutations in the LelA binding site in the regulatory region of legU1, legL6, and legL7 genes, resulting in plasmids listed in Dataset S2.In addition, the L. pneumophila LTTR lpg0173, the L. pneumophila RpoS sigma factor (lpg1284), and the three LelA orthologs from L. anisa, L. hackelia, and L. lansingensis (Lani_1844, Lhac_2674, and Llan_1155, respectively) were also cloned under Ptac control, to generate pNS-Ptac-lpg0173, pDT-pMMB-Ptac-rpoS, pNS-Ptac-Lani1844, pNS-Ptac-Lhac2674, and pNS-Ptac-Llan_1155.These plasmids were also digested with XbaI and NsbI, and the resulting fragment, containing Ptac-lpg0173 together with the lacI gene, was cloned into plasmids containing the lacZ fusions of lpg0174 and legU1 genes.The fragment containing Ptac-RpoS together with the lacI gene was cloned into the plasmids containing the lelA lacZ fusion.The fragment containing Ptac-Lani_1844 together with the lacI gene was cloned into plasmids containing the lacZ fusions of Lani_0816, Lani_1686, and Lani_1843 genes.The fragment containing Ptac-Lhac_2674 together with the lacI gene was cloned into plasmids containing the lacZ fusions of Lhac_0817, Lhac_0975, Lhac_1802, and Lhac_2448 genes.The fragment containing Ptac-Llan_1155 together with the lacI gene was cloned into plasmids containing the lacZ fusions of Llan_0301, Llan_1420, and Llan_2142 genes digested with XmnI and XbaI, resulting in plasmids listed in Dataset S2.