Diversity of Amoeba-Associated Giant Viruses Isolated in Algeria

Abstract

The discovery of several giant amoeba viruses has opened up a novel area in the field of virology. Despite this, knowledge about ecology of these viruses remains patchy. In this study, we aimed to characterize the diversity of giant viruses in Algeria by inoculating 64 environmental samples on various amoeba strains. After isolation by co-culture with nine amoeba supports, flow cytometry and electron microscopy were used to putatively identify viruses. Definitive identification was performed by PCR and sequencing. Mimiviruses, marseilleviruses, faustoviruses and cedratviruses were the main viruses isolated in this study. Moreover, a new virus, which we named fadolivirus, was also isolated and was found to belong to the recent metagenomic descriptions of Klosneuvirinae. Despite the use of 9 amoeba supports for co-culture, most of the isolates were obtained from two amoebas: Acanthamoeba castellanii Neff and Vermamoeba vermiformis CDC 19. Finally, the viruses most frequently isolated were marseilleviruses (55.5%) and Mimiviruses (22.2%). This work shows that the isolation of viruses previously detected by metagenomic analyses can be tedious, but possible.

1. Introduction

Megavirales order has recently been proposed to replace the current nucleocytoplasmic large DNA viruses (NCLDV) based on capsid and genome sizes, ancestral genes encoding autonomy transcription and even translation components [1,2,3,4,5,6]. This order contains viruses considered giant viruses and the story began at the beginning of the 21st century with the discovery of Acanthamoeba polyphaga mimivirus (APMV) [1,7]. These viruses modified the definition of what is a virus regarding their genomic contents, size and various phenotypic characteristics [1,7]. Six years later, a second isolation of giant viruses infecting amoeba was reported with the description of marseillevirus [1,7,8]. These last two families are currently recognized by the International Committee of Taxonomy of Viruses (ICTV).

Later, many other giant viruses infecting amoeba have been isolated and described in several geographical areas worldwide, such as Pandoravirus, Pithovirus, Mollivirus, faustovirus, cedratvirus, kaumoebavirus and pacmanvirus which have also recently been included into the proposed order of Megavirales [2,9,10,11,12,13,14,15]. Recently, several high throughput procedures were implemented to speed up the process of isolating giant virus in amoeba co-cultures and flow cytometry sorting was used to separate mixtures of these viruses and putatively identified the viral population [9,16,17,18].

Despite this, their natural hosts remain unknown, they seem to be ubiquitous in the environment and capable of surviving in highly diverse ecosystems [2,4]. They were isolated in Europe, Asia, Africa and America [5,9,14,17]. The viruses isolated to date appear to be the tip of the iceberg, as recent metagenomic studies have led to the discovery of a large number of new families [6,19]. The introduction of novel amoeba is a strategy consisting to potentially increase number of novel isolates. By the introduction of novel amoeba as V. vermiformis we isolated faustovirus follow by others as kaumoebavirus, orpheovirus. In a previous work [17], we compared different Acanthamoeba and we observed variations in the lineages and virus isolated depending the amoeba hosts. Based on those both experiences, we decided to test more novel hosts.

In the present study, we aimed to characterize the diversity of giant viruses in the Algerian environment using co-culture strategy on different amoeba cell supports from different types of environmental samples.

2. Materials and Methods

2.1. Samples Collection

A total of sixty-four environmental samples were collected in the environment of 4 different cities, including Mostaganem (35°55’59.999’’ N, 0°4’59.999’’ E), Chlef (36°8’26.664’’ N, 1°19’50.124’’ E), Sidi bel Abbes (35°12’0’’ N, 0°38’29’’ W) and Tlemcen (34°53’18.262’’ N, 1°19’4.815’’ W) located in north-west Algeria. Samples were collected between May and September 2017. The nature and origin of the samples are summarized in

Table 1. Distribution of giant virus isolates with their relative amoeba and samples.

No.	Sample	Site	Amoebas
			A. polyphaga	A. castellanii Douglas	A castellanii Neff	V. vermiformis	A. mauritaniensis	A. quina
1	sewage	Mostaganem		marseillevirus	marseillevirus
2	sewage	Mostaganem	marseillevirus		marseillevirus		marseillevirus	marseillevirus
3	sewage	Mostaganem			marseillevirus
4	sewage	Mostaganem
5	well water	Mostaganem	mimivirus		marseillevirus
6	well water	Mostaganem	mimivirus		marseillevirus
7	sewage	Mostaganem			marseillevirus
8	sewage	Mostaganem	mimivirus		marseillevirus
9	sewage	Mostaganem			marseillevirus
10	water dam	Mostaganem		marseillevirus		faustovirus
11	sewage	Mostaganem	marseillevirus		mimivirus
12	sewage	Mostaganem	marseillevirus
13	sewage	Mostaganem	mimivirus	mimivirus
14	sewage	Mostaganem	mimivirus
15	well water	Mostaganem
16	well water	Mostaganem			marseillevirus
17	spring water	Mostaganem
18	spring water	Mostaganem
19	soil	Mostaganem			mimivirus
20	soil	Mostaganem
21	swamp	Chlef			cedratvirus
22	swamp	Chlef			cedratvirus
23	swamp	Chlef
24	spring water	Chlef					marseillevirus
25	spring water	Chlef					marseillevirus
26	sewage	Chlef			mimivirus
27	soil	Chlef			cedratvirus
28	lake	Chlef
29	lake	Chlef	mimivirus
30	water well	Chlef
31	wastewater	Chlef
32	wastewater	Chlef
33	wastewater	Chlef
34	wastewater	Chlef
35	wastewater	Chlef
36	wastewater	Chlef			marseillevirus
37	soil	Chlef	marseillevirus
38	soil	Chlef			cedratvirus
39	wastewater	Chlef
40	wastewater	Chlef
41	wastewater	Sidi bel Abbes			marseillevirus
42	wastewater	Sidi bel Abbes
43	wastewater	Sidi bel Abbes
44	wastewater	Sidi bel Abbes
45	wastewater	Sidi bel Abbes
46	sewage	Sidi bel Abbes
47	sewage	Sidi bel Abbes
48	sewage	Sidi bel Abbes
49	sewage	Sidi bel Abbes
50	sewage	Sidi bel Abbes
51	sewage	Sidi bel Abbes			marseillevirus
52	sewage	Sidi bel Abbes
53	sewage	Sidi bel Abbes
54	sewage	Sidi bel Abbes				fadolivirus
55	sewage	Sidi bel Abbes
56	sewage	Telmcen				faustovirus
57	Water dam	Telmcen			marseillevirus	faustovirus
58	sewage	Telmcen			marseillevirus	faustovirus
59	water well	Telmcen
60	sewage	Telmcen
61	sewage	Telmcen			marseillevirus
62	sewage	Chlef
63	sewage	Chlef				faustovirus
64	sewage	Chlef

. All samples were stored in sterile tubes in a dark room at 4 °C until their inoculations. The soil samples were prepared by adding 5 mL of sterile distilled water and by vortex before inoculations.

2.2. Preparation of Amoeba Support

A collection of nine amoeba was used to perform co-cultures: Acanthamoeba castellanii strain Neff (ATCC^® 30010), A. castellanii strain Douglas (ATCC^® 50374), Acanthamoeba polyphaga (strain Linc AP1), Acanthamoeba mauritaniensis (ATCC^® 50253), Acanthamoeba quina (ATCC^® 50241), Acanthamoeba culbertsoni (ATCC^® 30171), Acanthamoeba divionensis (ATCC^® 50238), Vermamoeba vermiformis (strain CDC 19) and Willaertia magna (ATCC^® 50035). Fresh amoeba strains were grown in 75 cm² cell culture flask (Corning^®, Corning, NY, United States) containing 25 mL of peptone-yeast extract-glucose medium (PYG, Eurobio^®, France) and incubated at 28 °C during 48 h for Acanthamoeba spp. and at 25 °C during 72 h for V. vermiformis train CDC 19) and W. magna (ATCC^® 50035).

2.3. Isolation of Giant Viruses Using Co-Cultures

We used the protocol previously reported by Reteno et al. [12]. Briefly, amoeba cells suspensions were prepared at 5 × 10⁵ cells/mL for Acanthamoeba strains and 1 × 10⁶ cells/mL for V. vermiformis and W. magna. After centrifugation of amoebas (700× g during 10 min), we replaced the medium by starvation medium also named TS (homemade) [12] for all amoebas species except W. magna for which we added 10% of fetal bovine serum (FBS) in the TS medium. Doxycycline (20 µg/mL), vancomycin (10 µg/mL), ciprofloxacin (10 µg/mL), imipenem (10 µg/mL) and voriconazole (20 µg/mL) were added to cell suspension in order to eliminate contaminant overgrowth including intracellular bacteria and fungi. Then, 500 µL of each amoeba suspension was transferred to 24-wells plates and inoculated with 50 µL of each sample then incubated at 30 °C for 4 days for primo-culture. First and second blind subcultures were performed in new plates of 24-well, by inoculating 50 µL of the primo-culture in 500 µL of fresh cell suspension and incubated at 30 °C for 3 days. If fungal contamination was observed, supernatant was passed through 0.80-μm-pore sized filter before subculture. Third, the sub-culture was prepared as described above on fresh amoebae with doxycycline only (20 µg/mL). In each plate, four wells were used as negative controls with sterile water inoculation instead of sample.

2.4. Preliminary Characterization

We used Hemacolor^® staining and optical microscopy for screening wells presenting a cytopathic effect. In this case, the contents of the wells where a cytopathic effect was detected were transferred to a 25 cm² cell culture flask containing amoeba species used for co-culture with 5 mL of appropriate medium (as described above). For optical microscopy observation, 50 μL of amoeba cells before complete lysis were centrifuged and were then stained using Hemacolor^® Rapid staining Kit according to the manufacturer’s recommendations. In parallel, 50 μL of the supernatants were passed to flow cytometry, as previously described, and population analyses were superimposed on those of the previously known viruses using FlowJow software [18]. In the event that the viral population did not overlap perfectly, or rare populations were detected, we performed transmission electron microscopy and molecular identification for confirmation.

2.5. Electron Microscopy Observations

We confirmed the nature of amoeba lytic agents (virus/bacteria) by electron microscopy. We performed a negative staining using a fixed supernatant from co-culture with glutaraldehyde following the same procedure previously described [18]. Embedding procedure was carried out as previously reported [15]. Electron micrographs were obtained on a Tecnai G20 F20 TEM (FEI, Bonn, Germany) operated at 200 keV. Image J software was used to determine particle size.

2.6. Molecular Analysis and Sequencing

Regarding flow cytometry profiles obtained (see section above), we identified potential viral sub-populations as described in previous studies [13,15,18]. Then, we confirmed the identification of viral isolates by both real-time using probes and by standard PCR methods and using specific primer to confirm marseilleviruses, mimiviruses, faustoviruses and cedratviruses (

Table 2. Primer and probe sequences used for qPCR.

Name of Virus	Name of Probes	Target Genes	Primer Sequences (5’ to 3’)	Estimate Size of the Amplicons
Faustovirus	Fstv-photol-F1	Photolyase	GTCGCGGACGAGATGAGATT	712 bp
	Fstv-photoL-R1		TCACGCATACCGGCATCTAC
Marseillevirus	MV-F	D5 helicase	TCTGGGAGTGGGCTTTATCT	180 bp
	MV-R		AGGGTAATGACCTCGGGTA
	probe FAM		AGGATTGAACCTTCGCTGTTAC
Mimivirus	mimi_polB_DNApol_R322	DNA polymerase b	AAACAGGTGCACCAACATCA	230 bp
	mimi_polB_DNApol_F322		GGTTTCCATTTTGACCCAAG
Cedratvirus	CedV_Rpb10_F1	RNA polymerase 10 Rpb10	GGAAAGAATAGGTGCAGTGCG	240 bp
	CedV_Rpb10_R1		AAGAGATGGAAGTGGGGTTGC
	CedV_DNApolb_F1	DNA polymerase b	CACAGTCTCACCTCTTGCGT	700 bp
	CedV_DNApolb_R1		GCACAGCTCTTCTTCCGAGT

). The procedure of real-time PCRs of DNA extraction and PCR amplification were performed following the work of Ngounga et al. [20]. None sequencing of amplicons was done. Regarding results obtained by flow cytometry and by PCR we decided to do genomes by Next-genome sequencing.

2.7. Genomic Sequencing and Phylogenetic Analyses

For viruses not identified by molecular analysis, 3 faustoviruses and cedratvirus N38, we carried out the sequencing of the genome. Cloned viruses were produced in 15 flasks (75 cm²) in starvation medium containing fresh A. castellanii or V. vermiformis monolayer which were cultivated for 48–72 h at 30 °C. After observation of lysis, supernatants were collected and centrifuged at 60,000× g for 45 min. Then, a 25% gradient sucrose was made following the same procedure previously reported [13,18]. Stocks were kept at − 80 °C. Genome sequencing was performed using MiSeq Technology (Illumina, Inc., San Diego, CA, USA) and using the paired-end application in parallel in a 2 × 251 bp run for each bar-coded library. Genome assembly was performed with hybrid spades with default parameter [21]. Phylogenetic tree were built using MEGA 6.0 package (https://www.megasoftware.net), using MUSCLE for alignments and maximum likelihood construction with JTT model in 1000 bootstrap replicates using all sites in the alignments. Sequences of predicted proteins used for phylogenetic analyses were available on NCBI (MT394892-MT394893-MT394894).

3. Results

3.1. Amoeba Lysis and Identification of Viruses

Among the 64 samples inoculated on 9 amoeba hosts, representing 576 wells, a cytopathic effect was observed in 63. Analysis by flow cytometry identified 45 viruses, while other amoeba lyses were due to intracellular bacteria and were not studied. Among these, 25 were identified as marseilleviruses, 10 as mimiviruses, 5 as faustoviruses, 4 as cedratviruses and one was a new virus that we named fadolivirus (Figure 1). The 4 cedratviruses showed a typical morphology of this family of viruses (Figure 2). Rates of isolation as the virus species isolated varied dramatically according to amoeba used as support. Twenty-two viruses (48.9% of the total number of viruses) were isolated in A. castellanii Neff, 10 (22.2%) viruses were isolated in A. polyphaga, 6 (13.3%) in V. vermiformis, 3 (6.7%) in A. castellanii Douglas, 3 (6.7%) in A. mauritaniensis and 1 (2.2%) in A. quina. No viruses were isolated in A. culbertsonii, A. divionensis and W. magna.

Cedratviruses were isolated with A. castellanii Neff only. Faustoviruses and fadolivirus were isolated in V. vermiformis only. Marseilleviruses were also isolated with A. mauritaniensis and A. quina, as well as mimiviruses that were isolated with three different amoeba species: A. polyphaga, A. castellanii Douglas and A. castellanii Neff (Table 2). Among the 31/64 (48.43%) samples positive for virus isolation, 17 were from sewage samples, 10 from water samples and 4 from soil samples. Comparing three types of samples, we found 25 (55.6%) isolates from sewage samples. Of these 25 isolates: 6 mimiviruses, 15 marseilleviruses, 3 faustoviruses and 1 with a new virus, followed by 18 (40%) isolates from water samples: 4 mimiviruses, 9 marseilleviruses, 2 faustoviruses and 3 cedratviruses.

Regarding the distribution of viruses, most isolates came from Mostaganem city 25 (55.6%) followed by Chlef 11 (24.4%), Tlemcen 6 (13.33%) and Sidi bel Abbes 3 (6.66%). For the sample cedratvirus N38, we performed MiSeq sequencing and brief genome analyses. The genome is too partial to lead to a complete analysis. Indeed, we obtained more than 40 scaffolds with an average coverage from 130 to 440 depending scaffold considered and with a range of length from 1600 base pairs to 199,764 base pairs. However, and according to the phylogenetic trees based on the predicted proteins RNA polymerase subunits 1 and 2 (Figure 3 and Figure 4)—we observed the separation between the Brazilian cedratvirus on one side (Lineage B) and the members of lineages A on the other (cedratvirus A11, cedratvirus lausannensis and cedratvirus Zaza) [13,22,23] and another probable lineage with cedratvirus kamchatka (466,767 bp). This suggests that cedratvirus N38 could be a prototype member of a novel lineage and expand the known diversity between cedratvirus strains.

3.2. Isolation of a New Virus Belonging to the Klosneuvirinae Family

Fadolivirus was isolated from a sewage collected in Sidi bel Abbes on V. vermiformis. Ultrathin section permitted to observe icosahedral viral particles of about 300 nm (Figure 1) and seems to possess short fibrils on its capsid. Inside the capsid, the DNA packaging shows a progressive acquisition of density (dark electron dense core) with finally multi-layers, as observed for Bodo saltans virus [24]. The first genomic draft confirms the isolation of Klosneuvirinae-like viruses with a current draft genome higher than 1.5 Megabases and with an average coverage of 121. Indeed, the phylogenetic analysis based on the RNA polymerase subunit 1 shows a relative proximity to viruses recently reconstructed from metagenomics [6,19]. In particular, maximum-likelihood tree of fadolivirus with this protein show a relatively close distance with Indivirus and Klosneuvirus (Figure 5). Currently, this isolation needs further description in order to characterize a clear and complete genomic comparison.

Red color indicates nodes containing mimivirus lineages A, B and C and Tupanviruses, orange was used for Faunusvirus, purple for Klosneuvirinae, Blue for Cafeteria roenbergensis virus, green for Mimiviridae infecting algae. We deleted branches supported by bootstrap value of less than 0.6. No homolog of Rpb1 was found in the draft genome of dasosvirus, barrevirus, hyperionvirus. harfovirus, satyrvirus, homavirus, geavirus genomes present a truncated homolog of Rpb1. Blue labels concern viruses detected by metagenomic analyses.

4. Discussion

The results of this study are in agreement with those previously reported regarding giant viruses’ distribution in ecosystems and confirmed their ubiquitous presence [2,18]. We identified 45 giant viruses from 64 Algerian environmental samples. The higher rates of isolation of giant virus were found in sewage samples (54.8%) and are in accordance with those reported in previous studies investigating viruses in the Old world [12,25] and in contrast to studies in the New World [17,26] which showed an increased rate of isolation in water samples. Numerous Mimiviruses and Marseilleviruses were isolated as well as 5 new strains of faustovirus still undergoing genome analysis. In addition, we isolated four strains of cedratvirus with elongated shape, all these cedratvirus-like being identified correctly with real time PCR. Among these, we identified a probable novel lineage of cedratvirus, named cedratvirus N38. These new isolates need a wider complete genome sequencing to determine their exact position. Furthers studies are needed to characterize this novel giant virus named fadolivirus presenting a 300-nm icosahedral capsid and having a close relationship with the recent described putative Klosneuvirinae [19,27]. Recent detection of new families of viruses using a smart approach of mini-metagenomics rather than bulk metagenomics suggests that some of these new families could be in extremely low abundance [6]. As a result, their isolation is probably due to a longer incubation period, very careful well observations and randomness. However, the development of very high throughput procedures for the isolation of giant viruses [28] and the combination with sorting before inoculation, as is done for mini-metagenomics, could be a good strategy in the future to isolate these hidden giant viruses.

Meanwhile, we observed trends similar to those found in a previous study regarding the rate of isolation according to the cell support [17]. Indeed, Dornas et al. reported variation in the mimivirus isolations using 3 different strains of Acanthamoeba sp., A. polyphaga, A. castellanii Neff and A. griffini. Despite this, we observed herein that the rate of isolation could be extremely variable, from about 48% of well lysis with A. castellanii Neff than 0% in W. magna, A. culbertsonii or A. divionensis. This evidence brings complexity in current hypothesis of host prediction. Indeed, bioinformatics and metagenomics studies enable the detection and recovery of many giant viruses [29]. However, the detection of their hosts is currently probabilistic, mainly based on the presence of 18S rRNA genes. We currently do not know their real ability to infect hosts, but our co-culture experience has shown us that this range could be restrictive. To date, only Tupanvirus is capable of infecting two amoebas of different genera, V. vermiformis and Acanthamoeba spp [5]. It is also possible that some amoebae have a high permissiveness to viruses that are not their natural host, as was observed with VERO cells and human viral pathogens.