Acinetobacter sp. HM746599 isolated from leatherback turtle blood


  • Editor: Craig Shoemaker

Correspondence: Gerald Soslau, Office of Professional Studies in the Health Sciences, Drexel University College of Medicine, 245 N 15th Street, Philadelphia, PA 19102, USA. Tel.: +1 215 762 7831; fax: +1 215 762 7434; e-mail:


A newly described bacterial isolate, Acinetobacter sp. HM746599, has been obtained from leatherback sea turtle hatchling blood. The implication is that the hatchling was infected during development in the egg, which is substantiated by other studies to be reported by us in the future. The 16S rRNA gene sequence of the bacterium (GenBank accession number: HM746599) showed the greatest similarity to the identified species, Acinetobacter beijerinckii (97.6–99.78%) and Acinetobacter venetianus (99.78%). Acinetobacter sp. HM746599 are gram-negative, rod-shaped coccobacilli and are hemolytic/cytotoxic to human and sea turtle red blood cells (RBCs). Hemolysis is not the result of any detectable soluble toxin. Acinetobacter beijerinckii and A. venetianus hemolyze sheep RBCs while Acinetobacter sp. HM746599 does not, and unlike A. venetianus, the growth of Acinetobacter sp. HM746599 and A. beijerinckii is not supported by l-arginine. Many Acinetobacter species, especially hemolytic ones, are pathogenic to immunologically compromised humans and it is possible that, in addition to sea turtles, this bacterium might also be a danger to susceptible humans who handle infected hatchlings. The bacteria are available from CCUG (Culture Collection, University Gothenburg, Göteborg, Sweden) and from NRRL (Agricultural Research Service Culture Collection, Peoria, IL).


By 1995, at least 18 species had been identified within the genus Acinetobacter (Vaneechoutte et al., 1995). Acinetobacter species are most commonly found in soil and water; however, they may also be found on surfaces in hospitals. They are generally nonpathogenic to healthy humans, but may result in life-threatening infections in debilitated patients (Dijkshoorn et al., 1993; Juni, 2001; Kanafani et al., 2003; Starakis et al., 2006). At least one species, Acinetobacter baumannii, has been identified as a superbug in some infected humans (Liang et al., 2011). Other Acinetobacter species can be found in terrestrial, fresh water and marine habitats and as pathogens or symbionts of other animals. In this study, we utilize a polyphasic approach to characterize a species of Acinetobacter isolated from the blood of a leatherback sea turtle hatchling.

The leatherback turtle (Dermochelys coriacea) is an endangered species (Spotila et al., 1996) with a major nesting site at Parque Marino Nacional Las Baulas, Costa Rica. Turtles from this population nest primarily from October through February and are the only sea turtle species that cannot be maintained in captivity. Unfortunately, eggs laid on these beaches have a very low (50%) hatching success rate (Bell et al., 2002), which, along with human activities, contributes to their declining numbers. As part of a broader research effort aimed at the physiology, ecology and conservation of leatherback turtles, we extracted samples of blood in an aseptic, nonharmful way from leatherback adults and from hatchlings in order to study platelet aggregation and coagulation (Soslau et al., 2004, 2005). One pooled sample of hatchling whole blood contained numerous bacteria, and yet no red blood cells (RBCs) after storage at room temperature for 24 h. Hemolytic/cytotoxic bacteria were isolated from this sample for the studies described here. Future studies on the prevalence, pathogenicity and modes of transmission of this and other microorganisms from leatherback turtle samples may ultimately assist workers in the conservation of this critically endangered species.

Materials and methods

Isolation of blood samples

We extracted 0.1-mL samples of blood in an aseptic, nonharmful fashion into heparinized syringes from alcohol-swabbed hatchlings for platelet aggregation and coagulation studies (Soslau et al., 2004, 2005) with approval from the University IACUC Committee. Light and electron microscopy revealed that one pooled sample of whole blood from 10 hatchlings contained numerous bacteria, but no RBCs after 24 h of storage at room temperature (data not shown). The likelihood of contamination was deemed to be small because only one bacterial species was isolated from the blood sample and because all hatchlings were handled with gloves and carefully swabbed with sterile alcohol pads before blood extraction with a sterile heparinized syringe. All hatchlings appeared healthy at the time of blood collection. Because of our IACUC protocol and the limitations imposed by the US Fish and Wildlife, Department of Interior, with animals on the endangered species list, the hatchlings must be released into the wild shortly after being used briefly for nonharmful experimentations. As such, we could not document the long-term health outcome for the infected hatchling(s).

Growth and culture of bacteria

A 50-μL sample of the hydrolyzed blood with bacteria was added to 10 mL of Luria–Bertani (LB) broth, incubated at 37 °C overnight on a rocker plate. Bacteria were then subcultured on LB agar plates at 37 °C. Ten colonies were isolated from these plates for subsequent assays of hemolytic activity on human and sheep blood agar plates. Human blood agar plates (5% blood) were prepared by dissolving 19 g trypticase soy agar in 475 mL of ddH2O in a microwave oven, cooled to 50 °C and then mixing in 25 mL of freshly drawn human blood from a student volunteer (in accord with our IRB Committee) before pouring into sterile Petri dishes. The sheep blood agar plates were purchased from All 10 colonies of the sea turtle bacteria were found to be hemolytic. The 16S RNA genes of three of these were amplified and partially sequenced (methods described below), all yielding essentially identical sequences. It would appear that the hatchling was infected with a single bacterial species. One clone, 2-04LB-Cl-5, was then selected for complete sequencing as described below. The chemical and growth characteristics of the bacteria were kindly assessed by the US Centers for Disease Control, Washington, DC.

Test for hemolysin toxin(s)

To detect any soluble toxins with hemolytic activity, bacteria were grown overnight in an LB broth and 1.5 mL was centrifuged at 18 500 g in a microcentrifuge for 4 min. The bacterial supernatant was then filtered through a 0.45-μm filter twice to ensure the removal of all bacteria. Removal was confirmed through the absence of bacterial growth after incubation of a filtered sample overnight in LB media. Freshly drawn human blood was then diluted 1 : 1 with a sterile isotonic saline and 200 μL was incubated with 10, 50, 100 and 200 μL of the bacterial supernatant or equivalent volumes of LB broth. Samples were observed microscopically for lysis after 1, 4, 24 and 48 h.

rRNA gene sequencing

DNA was isolated from bacterial pellets obtained from 10 mL cultures. Bacteria were lysed in 1 mL of DNAzol and DNA isolated according to the manufacturer's protocol (Invitrogen). The virtually complete rRNA gene sequence was established by sequencing multiple PCR samples run in the forward and reverse directions (four to six runs in each direction) with two sets of previously described universal primer pairs [P0mod (forward) and PC3 (reverse) gene location 18–32 and 787–806, respectively; P3 (forward) and PC5 (reverse) gene location 787–806 and 1487–1507, respectively] (Wilson et al., 1990). Sequences from the middle of the 16S RNA gene were established with an additional primer pair (forward: 5′-ATC TGG AGG AAT ACC GAT-3′, and reverse: 5′-AGG CGG TCT ACT TAT CGC-3′) because the region between the 5′ and the 3′ end fragments could not be sequenced reliably. DNA sequencing was conducted by the Nucleic Acid Protein Research Core Facility at Children's Hospital of Pennsylvania, Philadelphia, PA. The complete DNA sequence was submitted to GenBank (accession number: HM746599).

Phylogenetic analysis

To place this bacterium into an evolutionary context, we performed blastn searches using the 16S rRNA gene sequence of the identified Acinetobacter species as the query (2-04LB-Cl-5, GenBank accession number: HM746599). Sequences of all hits with >99% identity to the query were downloaded, as were 16S rRNA gene sequences from other representatives within the Acinetobacter genus. Sequences were aligned using the Ribosomal Database Project II Sequence aligner (Cole et al., 2009), and small manual adjustments to these alignments were subsequently made in macclade (Maddison & Maddison, 2003). A maximum likelihood search was implemented with garli (Zwickl, 2006) via the CIPRES Portal (Miller et al., 2009), using a GTR+G+I model of nucleotide substitution, and model parameters were estimated during the run. A separate likelihood analysis, with 100 bootstrap replicates, was performed using this same approach. Using paup* v4.0b10, we also performed a bootstrap analysis using parsimony (Swofford, 2002). Parsimony trees were constructed using stepwise addition to generate starting trees, followed by the tree bisection reconnection approach for branch swapping and 1000 bootstrap replicates. Optimal trees for our analyses were visualized and labeled using the Interactive Tree of Life website (Letunic & Bork, 2007).

Results and discussion

We set out to identify hemolytic bacteria isolated from the blood of leatherback sea turtle hatchlings due to their likely effects on hatchling survival and susceptible humans who handle them. The bacteria grew as round, smooth, translucent/semi-opaque colonies on LB agar plates. Electron microscopic analysis of the hemolytic bacterial sample showed the presence of pairs and rods of coccobacilli consistent with Acinetobacter among lysed cellular debris (data not shown). While hemolysis was not seen on sheep blood agar plates, hemolytic activity was observed with bacteria grown on human blood agar plates that produced clear halos around colonies. Hemolytic activity was also observed with human and turtle RBCs in whole blood. However, the human RBCs in whole blood were totally unaffected by a bacterial supernatant prepared from a 24-h sample of Acinetobacter sp. HM746599 grown in LB broth. It would thus appear that RBC lysis by these bacteria does not depend on a soluble toxin. Hemolytic Acinetobacter strains have been reported to excrete a phospholipase (Lehmann, 1973; Juni, 2001) and it is possible that Acinetobacter sp. HM746599 retains a lytic phospholipase with the cell membrane. This bacterium may also be unique in its ability to hydrolyze human and turtle RBCs, but not sheep RBCs as all previously tested Acinetobacter species were either not hemolytic or lysed both human and/or sheep RBCs (Lehmann, 1973; Juni, 2001). We propose that the species' selectivity for RBCs may be related to the nature of the hemolysin associated with this bacterium.

In Table 1, we compare the characteristics of this bacterium with those of previously identified Acinetobacter species. While these data are not meant to be an exhaustive comparison with all known Acinetobacter, they do reveal the characteristics of Acinetobacter sp. HM746599 that are either similar to or different from those reported in at least one other Acinetobacter species. Not listed in the table are the following: dextran; lactulose; d-maltose; d-sucrose; l-sorbose; d-tagatose; d-trehalose; glycerol; and d-mannitol, which did not support the growth of Acinetobacter sp. HM746599 as found previously for all other tested strains of Acinetobacter (Kampfer et al., 1993). While Kampfer et al. (1993) reported the variable growth of different strains of Acinetobacter with d-arabinose and d-ribose, we did not detect the growth of Acinetobacter sp. HM746599 with either of these sugars.

Table 1.   Characteristics of Acinetobacter sp. HM746599 compared with other selected Acinetobacter species
CharacteristicsAcinetobacter HM746599A. beijerinckii*A. venetianusA. calcoaceticus‡,§,¶A. haemolyticus∥,‡,**,††A. johnsonii∥,**Unknown Acinetobacter‡‡A. brisouii sp. nov.§§A. anitratus¶¶A. baumannii∥∥8–9***A. lwoffii‡,*,¶
From blood+  +        
Hemolytic human+  +++     +
Hemolytic sheep++ ++      
Hemolytic turtle+NDNDNDNDND     ND
Extracellular hemolysin  + +     +
Growth at 10°C     +     
Growth at 37°C+++ ++      
Growth at 44–45°C        
NaCl+      +    
l-Arginine+ +      +
Catalase+++ +       
Hippurate       +   
Sensitive to optochin           
Sensitive to bacitracin          
Sensitive to vancomycin           
Sensitive to pen-streptomycin+        + 

The rRNA gene sequence (GenBank accession number: HM746599) has been established by sequencing four to six different PCR samples of DNA isolated from one clone, run in the forward and reverse directions. Among the species described from the Acinetobacter genus, the complete 16S rRNA gene sequence had a maximum sequence identity of 99.8% to Acinetobacter venetianus and Acinetobacter beijerinckii which both exhibit hemolytic activities (Nemec et al., 2009; Vaneechoutte et al., 2009). Acinetobacter sp. HM746599, like A. beijerinckii isolated from humans, does not grow on l-arginine; however, A. venetianus does (Nemec et al., 2009). The 16S rRNA gene maximum likelihood phylogeny revealed close relatedness between Acinetobacter sp. HM746599 with uncultured bacteria and several members of the Acinetobacter genus, including A. beijerinckii and A. venetianus (Fig. 1). Among the 16 closest relatives of the turtle-associated sequence that had described isolation habitats, all except two were free-living, symbiotic (i.e. hemolytic bacteria from coral), or pathogenic (i.e. bacteria from sole) bacteria from marine environments. The remaining two were obtained from terrestrial habitats (i.e. black sand and an insect from the order Hemiptera). Support for the monophyly of this group was modest based on the maximum likelihood analysis (bootstrap support=68), but considerably higher according to parsimony (bootstrap support=97). Bacteria from other lineages on this tree came predominantly from human clinical specimens, other vertebrates and activated sludge.

Figure 1.

 Maximum likelihood phylogeny of Acinetobacter species. This rooted maximum likelihood phylogeny depicts the relatedness between Acinetobacter sp. HM746599 isolated from the blood of a leatherback turtle hatchling with other strains and species of Acinetobacter. Moraxella lacunata (AF005160) and Alkanindiges illinoisensis (AF513979) were used as outgroups to root the tree (not shown). Support from bootstrap analyses is indicated above (maximum likelihood) and below (parsimony) the branches leading to the corresponding nodes.

We postulate that bacterial infections of leatherback sea turtle eggs in the wild may contribute to embryonic death and may also present as bacterial infections in hatchlings that may harm the young turtle as well as susceptible humans who handle them. The characterization of a putatively pathogenic bacterium that was likely present in the embryos lays the groundwork for future studies on the relationship between bacterial infections of leatherback turtle eggs with low hatching rates on the beaches in Costa Rica. The establishment of the etiology of low egg viability may ultimately lead to a treatment modality to increase the hatching rate of this critically endangered species. Indeed, recent reports demonstrated that bacteria (Awong-Jaylor et al., 2008) and the fungus, Fusarium solani (Sarmiento-Ramirez et al., 2010), were responsible for/associated with failed loggerhead sea turtle eggs, making it clear that egg-associated pathogens are an area of concern for leatherback turtles.

The Acinetobacter sp. HM746599 bacteria are available from the Culture Collection, University Gothenburg, Göteborg, Sweden (CCUG-600049), and from the Agricultural Research Service Culture Collection, Peoria, IL (NRRL-B-59471).


We would like to thank Dr Richard Facalam at the CDC, Washington, DC, for the analysis of several characteristics of the bacteria and Dr David Collins of the University of Reading, UK, for the initial partial sequencing of the rRNA gene in the bacteria.