Acanthamoeba spp. can be a host for a wide range of pathogenic microorganisms such as Legionella pneumophila, Chlamydophila pneumoniae, Cryptococcus neoformans, Mycobacterium avium, Listeria monocytogenes, and Pseudomonas aeruginosa, etc [3,5–7]. Among these microorganisms, the interaction between Acanthamoeba spp. and Legionella spp. is one of the most investigated. After uptake of Legionella spp. by Acanthamoeba spp., Legionella spp. forms a specialized compartment called Legionella-containing vacuole (LCV). LCV avoids fusion with lysosomes to deter lysosomal digestion and also inhibits phagosomal maturation, thereby enabling L. pneumophila to actively replicate inside the LCV [8].

To date, LCV and a large number of effectors transferred by the intracellular multiplication/defective organelle transport (Icm/Dot) type IV system of Legionella spp. have been identified [9–12]. Although the roles of these genes from Legionella spp. have been evaluated, little research has been done on genes of Acanthamoeba spp. during endosymbiosis with Legionella spp. To understand the intracellular survival strategy of Legionella spp., inhibition of phagosome lysis in Acanthamoeba spp. needs to be studied. In this study, total transcriptional changes of A. castellanii in response to survival and replication of L. pneumophila during 12 hr and 24 hr were investigated by RNA sequencing analysis.

The LCV in the Legionella-infected A. polyphaga has been reported to remain intact for up to 8 hr post-infection (hr pi), disrupted by 12 hr pi, and eventually lysed to release the intracellular pathogens into the cytoplasm of the amoeba by 18 to 24 hr pi [13]. L. pneumophila infection incurred the lysis of more than 80% of A. polyphaga at 24 hr pi, and it has also been suggested that the intracellular condition may significantly differ between 12–18 hr pi and 24 hr pi [13]. Contrary to the previous findings, L. pneumophila-infected A. castellanii in the present study remained intact even at 24 hr pi. Therefore, gene expression patterns at 12 hr pi and 24 hr pi were compared to confirm whether drastic differences were present in L. pneumophila-infected A. castellanii at these 2 time points.

A. castellanii was infected with L. pneumophila [14], and the Legionella-infected Acanthamoeba (L+A) was incubated for 12 hr and 24 hr at 25°C incubator. mRNA-Seq reads were mapped using TopHat software [15], and differentially expressed gene were determined based on BEDtools and EdgeR [16–18]. And we used the FPKM (fragments per kilobase of exon per million fragments) as the method of determining the expression level of the gene regions. Gene classification was based on searches done by DAVID (http://david.abcc.ncifcrf.gov/).

RNA samples from different experimental conditions were sequenced to investigate the endosymbiosis-induced gene expression changes in A. castellanii (Fig. 1). A total of 7,108 genes whose expressions changed 12 hr pi and 24 hr pi were displayed using a heat map (Fig. 1A). Genes from each group were colorized based on their expression level. Strongly upregulated/downregulated genes, as indicated by intense red/blue colors, were more prevalent in the 12 hr pi group than the 24 hr pi group. Genes whose expression levels changed more than 2 fold were selected for further analysis (Fig. 1B). Among the 7,018 genes, 1,211 and 1,131 genes in the 12 hr pi group were upregulated and downregulated more than 2 fold, respectively. Similarly, 1,321 and 1,379 genes from the 24 hr pi group were upregulated and downregulated more than 2 fold, each respectively. Venn diagram revealed that a fraction of the DEGs found in the 12 hr pi overlapped with the DEGs from 24 hr pi group (Fig. 1C). Our results revealed that 2,342 and 2,700 DEGs in L. pneumophila-infected A. castellanii at 12 hr pi and 24 hr pi were changed more than 2 fold, respectively. More DEGs were observed at 24 hr pi than at 12 hr pi, which may indicate that more gene involvement is required for survival of L. pneumophila in the later stages of infection.

DEGs of L. pneumophila-infected A. castellanii were assigned an Entrez Gene ID and subsequently classified into 3 domains: biological process (BP), cellular component (CC), and molecular function (MF). Classified DEGs were subdivided further into various gene ontology (GO) terms under each of the domains. In the 12 hr pi group, DEGs were assigned to 10 subcategories in BP, 8 subcategories in CC, and 23 subcategories in MF (Fig. 2). During the 12 hr endosymbiosis, 17 genes in ‘regulation of transcription’ (domain: BP), 30 genes in the ‘nucleus’ (domain: CC), and 37 genes in ‘DNA binding’ (domain: MF) were determined to be the most upregulated genes. Within the CC domain, 200 downregulated genes were involved in the GO term ‘integral component of the membrane’. In the 24 hr pi group, DEGs were subdivided into 29, 13, and 41 GO terms under BP, CC, and MF domains, each respectively (Fig. 3). In the BP domain, 26, 39, and 32 genes from the GO terms ‘regulation of transcription’, ‘intracellular signal transduction’, and ‘cyclic nucleotide biosynthetic process’ were drastically downregulated, each respectively. In the CC domain, similar to the 12 hr pi A. castellanii, 270 downregulated genes were involved in the GO term ‘integral component of membrane’. In the MF domains, downregulated genes associated with the GO terms ‘zinc ion binding’ and ‘protein kinase activity’ were 57 and 42, respectively. From each of the domains, 26 genes from ‘carbohydrate metabolic process’ (domain: BP), 30 genes from the ‘nucleus’ (domain: CC), and 40 genes from ‘DNA binding’ (domain: MF) were mainly upregulated.

DEGs from A. castellanii, which were upregulated or downregulated more than 10 fold post-infection with L. pneumophila, were listed in Tables 1 and 2. In the 12 hr pi A. castellanii, 47 out of 1,211 DEGs were upregulated (Table 1) and 90 out of 1,131 DEGs were downregulated more than 10 fold (Table 2). Identities for several most upregulated proteins in this group were 2 hypothetical proteins (1,100 fold and 345 fold), S-adenosylmethionine-dependent methyltransferases (173 fold), and GDPD-mannose-3′, 5′-epimerase (87 fold) (Table 1). GO analysis of the assigned Entrez Gene IDs revealed that the DEG which underwent 1,100 fold increase was a hypothetical protein that belonged to the DNA binding (GO: 0003677) category. Similarly, the GDPD-mannose-3′,5′-epimerase which was increased 87 fold, was associated with catalytic activity (GO: 0003824). Based on these findings, it can be speculated that these DEGs may be of importance during the initial phase of infection. While xylosyltransferase 1 was downregulated more than 700 fold, sulfiredoxin 1 was downregulated more than 500 fold, and vacuolar sorting-associated protein 13 were downregulated more than 12 fold. (Table 2). GO analysis results revealed that the DEGs downregulated 10 fold or more were predominantly associated with the integral component of membrane (GO: 0016021). Findings are consistent with the changes in DEGs categorized under CC as illustrated in Fig. 3.

Although 132 out of 1,321 DEGs were upregulated and 54 out of 1,379 DEGs were downregulated more than 10 fold in A. castellanii 24 hr pi, approximately 60% of these DEGs (78 DEGs and 30 DEGs) were identified as hypothetical proteins. Strong inhibition of DEGs were observed in both 12 hr pi A. castellanii (90 DEGs) and 24 hr pi A. castellanii (53 DEGs). Among the DEGs demonstrating 10 fold or greater changes, 47 genes were upregulated while 90 genes were downregulated within the initial 12 hr pi (Tables 1, 2). Conversely, by 24 hr pi, 132 upregulated and 54 downregulated DEGs were observed. From these results, we supposed that L. pneumophila infection facilitated reduced A. castellanii gene expression during the early stage of infection to inhibit phagocytic digestion, while enhancing the expression of specialized A. castellanii genes during the late infection stage for LCV lysis and access to host cell machinery for intracellular replication.

Interestingly, L. pneumophila-infected A. castellanii showed differential expressions of methyltransferase-associated proteins. In addition to the S-adenosylmethionine-dependent methyltransferases and lysine methyltransferase enzyme domain-containing protein (Table 1), 11 DEGs associated with methyltransferase were upregulated, and 19 DEGs were downregulated upon infection with L. pneumophila for 12 hr. Furthermore, L. pneumophila-infected A. castellanii also demonstrated differential expressions of acetyltransferase-associated proteins. Histone acetyltransferase-associated protein was upregulated and 8 other acetyltransferases were downregulated 12 hr pi with L. pneumophila. Icm/Dot type IV secretion system and its effectors of L. pneumophila modulate host gene expression by altering the chromatin structure or by affecting the activities of transcription factors [19]. Post-translational modifications such as DNA methylation, histone acetylation, and histone methylation have been shown to play a critical role in the epigenetic regulation of eukaryotic gene expression [19,20]. Our results revealed that L. pneumophila-infected A. castellanii showed differential expressions of 30 kinds of methyltransferase-associated proteins and 9 kinds of acetyltransferase-associated proteins at 12 hr pi. Based on the changes to epigenetic regulatory gene expressions, it can be speculated that L. pneumophila can alter the gene expression of A. castellanii through epigenetic mechanisms.

A plethora of DEGs induced in A. castellanii by the endosymbiont L. pneumophila were revealed in this study. However, 38.3% (1,930 of the 5,042) of A. castellanii genes were identified to be hypothetical proteins. Proportions of theses hypothetical proteins in the 12 and 24 hr pi groups can be ascribed to the lack of Acanthamoeba spp. database. Our investigation of the DEGs in A. castellanii by an endosymbiont provides important information to understanding the survival strategy utilized by notable intracellular pathogen L. pneumophila in A. castellanii. Future studies investigating the presence of an endosymbiosis-specific gene may help elucidate the underlying mechanism involved in L. pneumophila pathogenesis, which would contribute to understanding the inhibition of phagocytosis within A. castellanii or even immune evasion mechanism in human macrophages.