High quality draft genome of Lactobacillus kunkeei EFB6, isolated from a German European foulbrood outbreak of honeybees

The lactic acid bacterium Lactobacillus kunkeei has been described as an inhabitant of fructose-rich niches. Here we report on the genome sequence of L. kunkeei EFB6, which has been isolated from a honeybee larva infected with European foulbrood. The draft genome comprises 1,566,851 bp and 1,417 predicted protein-encoding genes.


Introduction
Honeybees are the most economically valuable pollinators of agricultural crops [1]. A disappearance of honeybees would result in an approximately 90% decrease in production of some fruits [2]. European foulbrood (EFB) and American foulbrood (AFB) are the two most important honeybee diseases affecting the brood [3]. While the AFB is caused by the spore-forming, Gram positive bacterium Paenibacillus larvae [4], EFB is caused by the capsuleproducing Melissococcus plutonius [5]. It has been shown that members of the lactic acid bacteria (LABs) inhibit the growth of M. plutonius [6] and P. larvae [7]. LABs are found in a variety of habitats, including human and animal microbiomes, and are used as food additives.
The honeybee crop microbiome consists of 13 bacterial species belonging to the genera Lactobacillus and Bifidobacterium [8]. These bacteria play a key role in the production of honey and bee bread. The latter serves as long-term food storage for adult honeybees and larvae. L. kunkeei is a common symbiont for Apis and the dominating LAB member in bees [6]. The organism is a specialist for colonization of the honeybee crop and interacts with the epithelial layer of the crop. L. kunkeei has been described as a fructophilic LAB [9]. Initially, it was isolated from wine [10], but it has also been found on flowers and in honey.
L. kunkeei EFB6 is the first LAB isolated from a German EFB-diseased larva. Here, we describe genomic features of this organism, focusing on factors that improve competition with bacteria such as M. plutonius and P. larvae. In addition, potential cell surface proteins that might play a role in cellular adhesion and biofilm formation are analyzed.

Organism information
In October 2012, an EFB outbreak in Bavaria (Germany) was confirmed. EFB-diseased larvae from this outbreak were collected, immediately frozen in liquid nitrogen and stored at −80°C for further investigation. Several EFBinfected larvae were dissected under sterile conditions. To obtain LAB the guts of the larvae, which formed a yellow, glue-like slime, were suspended in MRS medium (Carl Roth GmbH & Co KG, Karlsruhe, Germany) and subsequently streaked on solidified MRS to isolate single colonies. Strain L. kunkeei EFB6 (Table 1, Additional file 1:  Table S1) was isolated from these agar plates after aerobic incubation at 35°C.
L. kunkeei EFB6 is a non-sporulating, low G + C Gram positive member of the Lactobacteriaceae and taxonomically related to the genus Pediococcus. The strain exhibited a 100% 16S rRNA gene nucleotide sequence identity to the type strain L. kunkeei YH-15 (Table 1, Figure 1). Cells harvested in exponential growth phase exhibited a length ranging from 0.7 to 1.3 μm and a diameter ranging from 0.3 to 0.5 μm as determined by transmission electron microscopy (TEM) of either negatively stained or ultrathin-sectioned samples ( Figure 2). Preparations for ultrathin sectioning and negative staining of cells were performed as described by [23]. The L. kunkeei EFB6 cell wall is approximately 12 nm thick. This value is rather thin compared to cell walls of other Gram positives [24]. Three distinct wall layers of L. kunkeei EFB6 (two darker stained outer and inner layers and a brighter layer in between) could be distinguished by TEM. Surface layers and cellular appendages (pili, fimbriae) were not detected.

Genome project history
The organism was selected for sequencing on the basis of its use as potential inhibitor for the primary agents of AFB and EFB [6,7]. The aim was to investigate potential factors to increase bacterial competition fitness and cell surface proteins, which might be important for cellular adhesion and biofilm formation.
A summary of the project information is shown in Table 2.

Growth conditions and DNA isolation
To isolate genomic DNA L. kunkeei EFB6 was grown aerobically in 50 ml MRS medium at 35°C with shaking at 150 rpm (Lab-Therm Lab-Shaker, Adolf Kühner AG, Birsfelden, Switzerland). Cells were harvested in

Genome sequencing and assembly
Whole-genome sequencing of L. kunkeei EFB6 was performed by employing the Genome Analyzer II (Illumina, San Diego, CA). The shotgun library was prepared according to the manufacturer's protocols. For de novo  Asterisks indicate that a consensus sequence was calculated from all 16S rRNA gene sequences present in the corresponding genome. L. kunkeei EFB6 is boxed. Sequences were aligned using ClustalW 1.6 [25]. The phylogenetic tree was obtained by using the UPGMA method within MEGA 6.06 software [26]. Numbers at nodes are bootstrap values calculated from 1,000 resamplings to generate a majority consensus tree. Bacillus subtilis DSM10 was used as outgroup. The scale bar indicates the nucleotide sequence divergence.
assembly, we used 2,000,000 paired-end Illumina reads (112 bp) and the SPAdes 2.5 software [27]. The final assembly contained 55 contigs larger than 500 bp and revealed an average coverage of 142.96.

Genome annotation
For automatic gene prediction the software tools YACOP [28] and Glimmer [29] were used. Identification of rRNA and tRNA genes was performed by employing RNAmmer [30] and tRNAscan [31], respectively. The annotation provided by the IMG-ER system [32] was corrected manually. For this purpose, data obtained from different databases (Swiss-Prot [33], TrEMBL [34] and InterPro [35]) were used to improve the quality of the annotation.

Genome properties
The genome statistics are provided in Table 3. The high quality draft genome sequence consists of 55 contigs that account for a total of 1,566,851 bp and a G + C content of 37 mol%. Of the 1,455 predicted genes, 1,417 were putatively protein-encoding, 35 represented putative tRNA genes and three putative rRNA genes. For the majority of the protein-encoding genes (75%) a function could be assigned. The distribution of these genes into COG functional categories [36] is shown in Table 4.

Insights into the genome
Five different Lactobacillus species were used for genome comparisons with L. kunkeei EFB6 based on blastp [37]. Results are shown in Figure 3. All five species are of interest as probiotics, part of the gastrointestinal tract of animals or humans, or used in the production of fermented food. The identification of orthologous proteins was performed with the program Proteinortho 5.04 [39] by using    [40]. With an identity cutoff of 50%, we identified 425 proteins in L. kunkeei EFB6 without orthologs in any other Lactobacillus species. Among these unique L. kunkeei EFB6 proteins, we selected 7 proteins for detailed analyses. Analysis of the 89-kb region shown in Figure 3 revealed five ORFs (LAKU_4c00030-LAKU_4c00070) without orthologs in any genomes derived from lactobacilli deposited in GenBank (as of 28.02.2014). Furthermore, no homologs could be identified in any other sequenced microbial genome (NCBI nr-database as of 05.03.2014) by using blastp (e-value cutoff of 1e-20). Except for LAKU_4c00060 (7,521 amino acids), we could identify an N-terminal signal peptide and a noncytoplasmic domain ( Figure 4A) using Phobius' domain prediction software [41]: LAKU_4c00040 (4,579 amino acids) and LAKU_4c00070 (3,129 amino acids) contain   [41] was used to determine putative domains. The yellow blocks represent signal peptides, the white color of the arrows show the non-cytoplasmic part. Red blocks represent transmembrane regions and blue blocks predicted coiled-coil structures. To test whether this region exists in other L. kunkeei strains, we designed specific primer-pairs for each ORF (Table 5, Figure 4A). Predicted PCR product sizes are depicted in white boxes. The presence of the genes were tested for L. kunkeei EFB6, L. kunkeei HI3 (isolated from honey), L. kunkeei DSM 12361 (isolated from wine), and L. johnsonii DSM 10533 (isolated from human blood) ( Figure 4B). The obtained PCR product sizes correlated with the predicted sizes (Table 5, Figure 4A). For L. johnsonii DSM 10533, no PCR product could be obtained.
kunkeei EFB6 is the first sequenced genome harboring these cluster, we designed specific primer pairs for detection of each ORF in other Lactobacillus strains by PCR ( Table 6). As shown in Figure 4B, all five ORFs were present in other L. kunkeei strains isolated from honey and wine. On the basis of domain prediction and IMG's bidirectional best hits [32], we assume that this gene cluster encodes cell surface or secreted proteins involved in cell adhesion or biofilm formation.
During genome comparison, we identified two additional proteins (LAKU_24c00010 and LAKU_24c00050) without a homolog in any of the publicly available genome sequences. These proteins show only weak sequence similarity to known proteins and might be involved in cellular adhesion. LAKU_24c00010 contains a signal peptide, transmembrane helices and 29 DUF1542 domains, which are typically found in cell surface proteins. In Staphylococcus aureus, it has been shown that some DUF1542containing proteins are involved in cellular adhesion and antibiotic resistance [42]. LAKU_24c00010 showed the highest sequence identities to the matrix-binding protein (WP_010490864) of "Lactobacillus zeae" KCTC 3804 (40%) [43] and the extracellular matrix binding protein (YP_005866289) of Lactobacillus rhamnosus ATCC 53103 (36%) ( Figure 5).
Additionally, LAKU_24c00050 contains N terminal transmembrane helices, two mucin-binding protein domains as well as a C terminal Gram positive-anchoring domain. Proteins with this domain combination are usually associated with bacterial surface proteins. LAKU_24c00050 showed similarity to the Mlp protein (WP_004239242) of Streptococcus mitis and other mucus-binding proteins ( Figure 6). Due to the mucosal surface-colonizing properties of lactobacilli, they have been investigated as potential recombinant mucosal vaccines [45].
In the genome of L. kunkeei EFB6, we identified genes encoding all proteins of the general secretory (Sec) pathway and putative polysaccharide biosynthesis proteins, which may participate in capsule or S layer formation. Recently, Butler et al. (2013) [47] detected a lysozyme produced by L. kunkeei Fhon2N and suggested a bacteriolysin or class III bacteriocin function. In L. kunkeei EFB6, we identified four genes belonging to the glycoside hydrolase family 25. Enzymes of this family are known  Figure 5 Tblastx comparison of L. kunkeei ORF LAKU_24c00010 to matrix binding proteins of L. rhamnosus ATCC 53103 and "L. zeae" KCTC 3804. The graphical presentation was done with Easyfig software (minimum blast hit length of 200 bp and a maximum e-value of 1e −100 ) [44]. LAKU_24c00010 shows similarities to WP_010490869, WP_010490864 and WP_010490862 of "L. zeae" KCTC 3804, but also to YP_005866289 (L. rhamnosus ATCC 53103). The ORFs used for comparison are labeled with NCBI accession numbers. The blast identity is shown in a colored scale ranging from 31 % (yellow) to 100 % (red).
to possess lysozyme activity. Two of the deduced proteins (LAKU_13c00160 and LAKU_32c00010) contain a signal peptide, indicating secretion of the proteins. LAKU_19c00290 harbors transmembrane helices and is probably anchored in the cell wall. LAKU_6c00080 did not contain a putative signal peptide or transmembrane helices.

Conclusion
In this study, we characterized the genome of L. kunkeei strain EFB6 isolated from an EFB-diseased larva. In a recent study was shown that L. kunkeei has the potential for biofilm formation and adhesion to the honey crop [6]. Our genome analysis supports these results. Using large surface proteins or extracellular matrix-binding proteins, L. kunkeei might be able to attach to eukaryotic epithelial cells. Furthermore, due to the presence of polysaccharide biosynthesis proteins and several enzymes with lysozyme activity, it is possible that L. kunkeei is actively protecting its niche against bacterial competitors. As LABs have been shown to have an inhibitory growth effect on M. plutonius, the use of LABs as probiotic additive against the EFB-causing agent is conceivable.

Additional file
Additional file 1: Associated MIGS Record.

Competing interests
The authors declare that they have no competing interests.
Authors' contributions MD, AP and RD designed research, MD, JS and FJT isolated and characterized strain EFB6, MD, AP and AL carried out genome analyses, MH performed electron microscopy, MD and RD wrote the manuscript with help of AP. All authors read and approved the final manuscript. Figure 6 Tblastx comparison of MucBP domain-containing proteins. Comparison of MucBP domain-containing proteins were performed using the program Easyfig (mininum blast hit length of 50 bp and maximum e-value of 1e −10 ) [44]. LAKU_24c00050 shows similarity to ORFs of Streptococcus mitis NCTC 12261 (NCBI accession numbers inside arrows, which represent ORFs used for comparison). Additionally, LAKU_24c00050 shows similarity to WP_003144513 of Gemella haemolysans ATCC 10379 and CCC15643 of Lactobacillus pentosus IG1 [46]. The blast identity is shown in a colored scale ranging from 20% (yellow) to 100% (red).