Transcriptional effect of HSA on gene expression
Transcriptome analysis of A. baumannii AB5075 and AB5075 Δhns cultured in LB broth revealed a differential expression profile of 183 genes (22). When A. baumannii AB5075 was cultured in LB broth with or without HSA there were 30 genes differentially expressed (Table S1). In both cases, an FDR-adjusted P-value of < 0.05 and log2-fold change >1 was considered as the differentially expressed-genes (DEGs). Inspection of the DEGs in both analyses showed that there were 11 genes in common (hhDEG) (Fig. 1A and Table S2). These genes are suitable candidates to be regulated by the presence of HSA through modification of the intracellular H-NS concentration. Six and five of these DEGs were up- or down-regulated, respectively (Fig. 1B). Three of these genes are involved in quorum sensing (acdA, kar and fadD), and one in quorum quenching (aidA) (24). Other hhDEGs are associated with biofilm formation (csuAB) (25), β-oxidation of fatty acid (fadA) (26), stress response (katE, ywrO) (27), and metabolism (phaC, yedL1, and yedL2) (Fig. 1B).
Cultures of the multidrug resistant, hypervirulent A. baumannii AB5075 model strain in the presence or absence of a physiological concentration of HSA were next subjected to quantitative RT-PCR (qRT-PCR) using total RNA to confirm that the presence of HSA in the growth medium results in a reduction of the H-NS synthesis (Fig. 1C).
We next determined the levels of H-NS in A. baumannii AB5075 pMBLe-hns, a strain that carries a plasmid that overexpresses H-NS. Fig. S1 shows that expression was elevated in cells grown in LB medium, but the H-NS levels were drastically reduced when the medium was supplemented with 3.5% HSA. This strain and the wild type A. baumannii AB5075 were used to assess the expression of nine selected hhDEGs by qRT-PCR using RNA obtained from cultures carried out in LB broth supplemented with and without 3.5% HSA. Expression of acdA, kar, fadD, and aidA were upregulated in A. baumannii AB5075 grown in the presence of HSA. However, it was of interest that increasing the concentration of H-NS did not modify the expression of these genes (compare bar AB5075LB with AB5075 pMBLe-hns in Fig. 2). Expression of csuAB was downregulated in A. baumannii AB5075 growing in the presence of HAS.
In contrast, the fadA gene behaved unexpectedly; significant changes were not observed in the presence or absence of HSA in A. baumannii AB5075 and a reduction in expression was observed in the overproducing strain that was driven by the presence of HSA, i.e., when overexpression of H-NS was reversed (Fig. 2). The molecular basis of these modifications and changes are unknown at this time. The stress-response related genes katE and ywrO were downregulated in the presence of HSA and differences were not noted when H-NS was overexpressed (Fig. 2). The metabolic gene phaC a key enzyme in the polymerization of polyhydroxyalkanoates (PHAs), showed a significant up-regulation in both strains, but significant modifications to levels of expression when H-NS was overexpressed was not seen (Fig. 2).
Since biofilm and quorum sensing are intimately related to A. baumannii’s virulence (and we identified genes involved in these two processes that are regulated by HSA and H-NS), we expanded the qRT-PCR analysis to two other genes associated with these processes. The expression of csuE, associated to biofilm formation, was downregulated when A. baumannii AB5075 was cultured in the presence of HSA. In A. baumannii AB5075 pMBLe-hns was overexpressed; but a significant downregulation was observed when HSA was present in the culture medium (Fig. S2). In the case of abaI, which codes for the acyl homoserine lactone (AHL) synthase, the presence of HSA in the milieu produced a significant reduction in expression. A. baumannii AB5075 pMBLe of H-NS growing in the presence or absence of HSA expressed the gene at low levels (Fig. S2).
HSA plays a role in modulating the biofilm formation, the quorum sensing network, and the oxidative stress through H-NS
A. baumannii AB5075 grown in the presence of HSA produced a reduced mass of biofilm, which was similar to that produced by the hns deficient mutant (Fig. 3A and B). The A. baumannii AB5075 pMBLe-hns strain as well as the mutant complemented by the plasmid pMBLe-hns cultured in the absence of HSA showed elevated levels of biofilm production. In A. baumannii AB5075 pMBLe-hns a regulatory effect was noted in HSA-containing medium. Taken together, these results support a regulatory role of HSA through modifying H-NS expression.
Phenotypic modifications in quorum sensing were assessed determining levels of acyl-homoserine lactone (AHL) using the Agrobacterium tumefaciens-based solid plate assays (28, 29). Rodgers et al. demonstrated that the supernatants from A. baumannii AB5075 or AB5075 Δhns produced similar intensity of color, likely caused by an increase in both AHL synthesis (quorum sensing) and lactonase activity (quorum quenching) in AB5075 Δhns strain (22). To study the effect of HSA on AHL secretion into the surrounding medium, A. baumannii AB5075, AB5075 Δhns, and the strains complemented with pMBLe-hns were cultured in different conditions before assessing the AHL concentration in the growth medium. A. baumannii AB5075 produced much lower amount of AHL when it was cultured in the presence of HSA (Fig. S3). A. baumannii AB5075 pMBLe-hns growing in the presence or absence of HSA and A. baumannii AB5075 Δhns pMBLe-hns cultured in LB broth, all produced low levels of AHL. The low AHL concentration in the growth medium could be explained by cancelling modifications in expression of genes associated with AHL synthesis (acdA, kar, fadD, and abaI) and lactonase, which is responsible for quorum quenching activity (aidA) (Fig. 1, 2, and S2).
Hydrogen peroxide is a disinfectant with potent bactericidal activity that is used in vaporized form to control outbreaks of multi-resistant A. baumannii infections (30) (31, 32). Since A. baumannii is catalase-positive and the katE gene was a hhDEG, down-regulated in the presence of HSA, we assessed the production of catalase activity, as the decrease in absorbance at 240 nm resulting from the consumption of H2O2 (Fig. 4). Addition of HSA to the growth medium resulted in a reduction of catalase activity in the wild type strain. Equally low activity was produced by A. baumannii AB5075 Δhns growing in the absence of added HSA, which suggest that the action of HSA occurs through reduction of H-NS synthesis (Fig. 4). These results agreed with both RNA-seq analyses and RT-qPCR results. As expected, the complemented A. baumannii AB5075 Δhns pMBLe-hns produced higher catalase activity when compared to A. baumannii AB5075 (Fig. 4). Surprisingly, overexpressing hns showed reduced catalase activity when compared to the wild type strain; we still do not know the molecular bases for this unexpected results (Fig. 4).
The results described in this section support the hypothesis that some A. baumannii genes are regulated by H-NS, whose expression is modified by the presence of HSA in the milieu. H-NS is a nucleoid-associated protein that binds DNA in a relatively nonspecific manner (it shows preference for AT-rich and curved regions) and alters its topology, which modifies levels of transcription. The effects of H-NS in A. baumannii have not been thoroughly studied (17, 20, 22). An early analysis found that an A. baumannii mutant containing a disrupted hns gene was modified in the expression of several virulence genes such as those associated with the autotransporter Ata, the type VI secretion system, a type I pilus cluster, the acetoin metabolism, and phenylacetic acid degradation (17). This strain also showed altered adherence to biotic surfaces and increase virulence as determined using the Caenorhabditis elegans infection model system (17). A previous study showed that exposure of the model A. baumannii A118 to HSA was accompanied by a reduction of expression of H-NS and a wide variety of genes related to antibiotic resistance and stress response as well as genes that encode functions associated with natural competence (14, 22, 23, 33).
The global nature of H-NS as transcriptional regulator makes it challenging to definitely associate genes that are regulated by H-NS through variations in levels of expression of the regulator when the cells are exposed to HSA. However, despite these limitations, the results described in this work strongly suggest that there is a group of genes whose expression is indirectly regulated at the transcriptional level by HSA through reducing H-NS expression. We conclude that HSA-mediated reduction of H-NS levels may be one regulatory circuit utilized by A. baumannii to adapt to selected environments such as those where HSA-containing human fluids are abundant.