ErmF and ereD Are Responsible for Erythromycin Resistance in Riemerella anatipestifer

To investigate the genetic basis of erythromycin resistance in Riemerella anatipestifer, the MIC to erythromycin of 79 R. anatipestifer isolates from China and one typed strain, ATCC11845, were evaluated. The results showed that 43 of 80 (53.8%) of the tested R. anatipestifer strains showed resistance to erythromycin, and 30 of 43 erythromycin-resistant R. anatipestifer strains carried ermF or ermFU with an MIC in the range of 32–2048 μg/ml, while the other 13 strains carrying the ereD gene exhibited an MIC of 4–16 μg/ml. Of 30 ermF + R. anatipestifer strains, 27 (90.0%) carried the ermFU gene which may have been derived from the CTnDOT-like element, while three other strains carried ermF from transposon Tn4351. Moreover, sequence analysis revealed that ermF, ermFU, and ereD were located within the multiresistance region of the R. anatipestifer genome.


Introduction
Riemerella anatipestifer is one of two species in the genus Riemerella within the family Flavobacteriaceae, which is one of the largest branches in the phylum Bacteroidetes [1,2]. R. anatipestifer infection is a contagious disease of domestic ducks, geese, turkeys, and various other domestic and wild birds that poses a substantial threat to the duck industry worldwide and accounts for significant economic losses [3].
Erythromycin inhibits bacterial protein synthesis by binding at the exit tunnel of the 50S ribosomal subunit, resulting in the subsequent abortion of the growth of nascent peptide chains. Years ago, many R. anatipestifer clinical isolates were found to be sensitive to erythromycin [4], thus this antibiotic has been used to successfully treat R. anatipestifer infection in some duck flocks. However, erythromycin treatment failures have been noted on several occasions in the past few years. The growing increase in the rates of erythromycin-resistant R. anatipestifer isolates in recent years is alarming and the mechanism of erythromycin resistance in R. anatipestifer has not been described.
Three major mechanisms of erythromycin resistance have been identified in Gram-negative and-positive bacteria [5,6]. The most well-known mechanism is the target-site modification of the 50S ribosomal subunit, which is mainly mediated by methylases encoded by the erythromycin ribosomal methylase (erm) gene and this methylation also causes resistance to lincosamides and streptogramin B antibiotics (MLS) [7]. The second described mechanism is the synthesis and activity of erythromycin inactivating enzymes, such as erythromycin esterase [8,9]. The third known cause of erythromycin resistance is the active removal of the antibiotic by efflux systems, which maintains the intracellular antibiotic concentration at a subtoxic level that does not affect bacterial cell growth [5,10].
So far, there have been relatively few studies on erythromycin resistance in bacteria in the family Flavobacteriaceae. However, in Bacteroides, which belong to the family Bacteroidaceae in the phylum Bacteroidetes, several erm genes (ermB, ermF and ermG) have been found [11]. In addition, our previous results demonstrated that the ermF gene may be expressed in R. anatipestifer and the conjugative transposon Tn4351 can be transferred to and randomly inserted into the genome of R. anatipestifer [12]. In this study, for the first time, our results showed that erythromycin resistance in R. anatipestifer was due to the presence of the ermF, ermFU, and ereD genes in the bacterial genome.

Bacterial strains and growth conditions
From 1996 to 2014, a total of 79 R. anatipestifer isolates were isolated from sick ducklings in China (S1 Table). R. anatipestifer type strain ATCC11845 and Escherichia coli strain ATCC25922 were obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA). R. anatipestifer strains were cultured at 37°C in tryptic soybean broth (TSB; TSB, Difco Laboratories, Detroit, MI, USA) or agar in an atmosphere of 5% CO 2 . E. coli strains were grown in Luria-Bertani broth or agar at 37°C.

Antibiotics susceptibility testing
Erythromycin susceptibility tests for each strain were performed in 96-well microtitre plates (Corning Incorporated, Corning, NY, USA) by determination of the minimum inhibitory concentration (MIC) value of erythromycin, as described previously [5]. E. coli ATCC 25922 isolates were used for quality control. Erythromycin was serially diluted two-fold in TSB broth to obtain antibiotic solutions with concentrations ranging between 4096 and 0.0625 μg/ml. The turbidity of the inoculum was adjusted to 10 7 CFU/ml (100 μl/well). An inoculated broth containing no antibiotic was included as a growth control and a tube of uninoculated broth was used as a sterility control. The microplates were incubated at 37°C for 24 h. The lowest concentration of erythromycin that inhibited bacterial growth was considered as the MIC. Due to the lack of Clinical and Laboratory Standards Institute (CLSI)-approved erythromycin breakpoints applicable to R. anatipestifer and, moreover, in our previous study, 1 μg/ml of erythromycin was successfully used to select random Tn4351 transposon mutants of R. anatipestifer strain CH3 [12], strains with an MIC of erythromycin of 0.25 μg/ml were considered susceptible, 0.5 μg/ml as intermediate, and 1 μg/ml as resistant as per the CLSI-approved criteria for Streptococcus spp. [13]. These criteria were further confirmed using the disc diffusion test with 15-μg erythromycin disks (Hangzhou Microbiological Co., Hangzhou, China). The final MIC value was estimated based on the average of at least three measurements.
To determine whether ermF or ereD was harboured in R. anatipestifer, DNA extracted from boiled R. anatipestifer bacteria was used as a DNA template for detection of the ermF and ereD genes by PCR. In addition, to further determine that the identified erythromycin resistance gene cassettes in R. anatipestifer were located within the genome or plasmid, the genomic DNA of erythromycin-resistant strains was isolated using the TIANamp Bacteria DNA Kit (Tiangen Biotech (Beijing) Co., Ltd., Beijing China) and the plasmids were extracted using the TIANprep Midi Plasmid Kit (Tiangen Biotech). The ermF and ereD genes were PCR-amplified using genomic DNA or plasmid as templates, respectively. In addition, to determine the type of ermF carried by different ermF + R. anatipestifer strains, the upstream sequence of ermF gene was PCR-amplified using primers ermFU P1 plus ermFU P2 for ermFU, and ermF P1' plus ermFU for ermF, respectively. DNA sequencing was used to identify the PCR products.
In addition, to determine whether R. anatipestifer carried other genes that convey erythromycin resistance in other bacterial species, PCR was used to test for the presence of the ermA, ermB, ermC, ermD, ermE, ermG, ermT, ermX, mphA, mphR, and msrA genes in all R. anatipestifer isolates. The primers used in this study are listed in S2 Table. Genomic walking of the full ermF cassette of strain HXb2 and its flanking sequences The open reading frames (ORFs) of the ermF gene were PCR-amplified from different R. anatipestifer strains and sequenced. Genome walking was performed to clone the upstream flanking sequence from strains with different erythromycin resistance levels according to the manufacturer's instructions (Takara Biotech Co., Ltd, Dalian, China). In addition, differences in the upstream sequence of ermF types (ermF, ermFU, or ermFS) carried by ermF + R. anatipestifer strains were determined by PCR. Sequence analysis was performed using Lasergene 7.0 software (DNASTAR Inc., Madison, WI, USA). To determine how the ermF gene was transferred into the genome of R. anatipestifer, the genetic environment of the ermFU cassette within the genome of HXb2, which is the most virulent isolate identified in ducklings so far, was sequenced and analysed using the BLAST program (http://blast.ncbi.nlm.nih.gov/Blast.cgi).

Real-time PCR
To evaluate the expression of resistance genes at the transcriptional level in the wild-type and transferred strains, the bacteria were grown to log phase (OD 600nm 0.8-1.0), and total RNA was extracted using the RNeasy Mini kit (Qiagen, Hilden, Germany), followed by first-strand cDNA synthesis using the Sensiscript RT kit (Qiagen), according to the manufacturer's instructions. Quantitative real-time PCR was performed to measure the mRNA levels of ermF, ermFU, and ereD using SYBR green PCR master mix (Applied Biosystems, Foster City, CA, USA) and the primers listed in S2 Table. Relative quantification of gene expression was calculated using the ΔCT method based on the signal intensity of the PCR products according to the following formula: 2 -ΔCT = 2 -(sample Ct-normalizer Ct) (Ct = threshold cycle of real-time PCR). TbdR1 (Riean_1024) was used as an endogenous control for sample normalization. Results are presented as fold-change relative to mRNA expression levels of the wild type strains.

Nucleotide sequence accession numbers
The sequences of the ermF region of different R. anatipestifer strains, the ereD region of R. anatipestifer strain SX, and the multiresistance region (MRR) of strain HXb2, were deposited in the GenBank database under the accession numbers KP265714-KP265722 and KR857248-KR857269, respectively.

Erythromycin susceptibility testing
The erythromycin MICs of 80 R. anatipestifer strains ranged from 0.125 to 2048 μg/ml. As shown in Fig 1, 43 (53.8%) of the 80 tested strains exhibited resistance to erythromycin, while 37 (46.2%) were susceptible. Furthermore, 2 (8.3%) of 24 R. anatipestifer strains, which were isolated from sick ducklings in China between 1996 and 2004, were resistant to erythromycin, while 41 (77.4%) of 56 R. anatipestifer isolates obtained from 2005 to 2014 exhibited erythromycin resistance. These findings suggested that the percentage of isolates showing resistance to erythromycin greatly increased over the last 10 years. The increased use of erythromycin and other macrolides may have increased the selective pressure on bacterial populations [16], although the percentage of erythromycin-resistant R. anatipestifer isolates may vary among different duck farms in different regions in China.

Prevalence of the ermF and ereD genes in R. anatipestifer strains
To evaluate the prevalence of the ermF and ereD genes in R. anatipestifer strains, the presence of these two genes in 80 R. anatipestifer strains were determined by PCR. The results showed that 30 (70.0%) of 43 erythromycin-resistant strains encoded the ermF gene. As shown in Fig 1, the range of MICs of 30 ermF + erythromycin-resistant strains were widely distributed from 32 to 2048 μg/ml, and 43.3% (13/30) of ermF + strains exhibited very high resistance (MIC 1024 μg/ml). On the other hand, all 13 ermFerythromycin-resistant strains encoded the ereD gene, which exhibited relatively lower resistance with an MIC in the range of 4-16 μg/ml. Our results showed that both ermF and ereD were encoded by the R. anatipestifer genome, while none of the erythromycinsusceptible strains encoded the ermF and ereD genes. Moreover, none of the 80 tested R. anatipestifer strains harboured mphA, mphR, msrA, or other erm genes. These results suggested that ermF and ereD may be involved in erythromycin resistance in R. anatipestifer strains and the ermF gene, which codes for ribosomal methylase, was the most frequently encoded gene that determines erythromycin resistance in R. anatipestifer.

Sequence analysis of the ermF gene in R. anatipestifer
Sequence analysis of the ermF ORFs revealed that the ermF gene from ermF + R. anatipestifer strains, which exhibit different erythromycin resistance levels, shared an amino acid homology of 98.5%-100%. There have been three ermF genes sequenced from the Bacteroide frangilis group: ermF (Gen-Bank accession number: M14730), which is encoded by transposons Tn4351 and Tn4400 [17], ermFS (M17808), which is encoded by transposon Tn4551 [18], and ermFU (M62487, AJ311171), which is encoded by transposon Tn5030 and the conjugative transposon CTnDOT [19,20]. In this study, according to differences within the upstream sequences, the ermF genes from different R. anatipestifer strains could be classified into two types: ermF and ermFU. As shown in Fig 2, 27 out of 30 (90.0%) ermF + R. anatipestifer strains tested in this study carried the ermFU gene, while only three strains carried ermF, which indicated that the ermFU gene was more widespread in R. anatipestifer than ermF, and the CTnDOT-like element (including Tn5030) significantly contributed to the dissemination of erythromycin resistance determinants in R. anatipestifer. Conjugal transfer is mostly responsible for the spread of resistance genes within the Bacteroides group [21]. Two types of conjugative elements have been identified in Bacteroides: plasmids and chromosomal elements. Transposons Tn4351, Tn4400, and Tn4551, which carry the ermF or ermFS gene, were part of Bacteroides plasmids pBF4, pBFTM10, and pBI136, respectively [21], while Tn5030 or CTnDOT, which carry ermFU, were normally integrated into the bacterial chromosome [19,20]. However, it was odd that all the erythromycin resistance genes identified in R. anatipestifer strains (ermF, ermFU, and ereD genes) were only found within the genomes. This may be the result of the evolution of R. anatipestifer because Bacteroides plasmids, such as pBF4, pBFTM10, and pBI136, may not be stably maintained in R. anatipestifer strains. In addition, CTnDOT contains essential mobilization genes (mobA and mobB) and other transfer (tra) genes [19], while Tn4351 does not. Therefore, CTnDOT-like elements may spread more easily in bacteria than Tn4351-like elements.
Sequence homology analysis showed that the upstream sequence of the ermFU gene from R. anatipestifer strains (from -1 to -220*-262) shared 85.8%-88.7% homology with that of CTnDOT (from -1 to -256), and we found that the region at -29 to -70 of the upstream sequence of ermFU contained a 37-nt deletion (strains HXb2 and HGb1), a 2-nt deletion (strains GuiZ-1, YXb15, CH-1), or a 5-nt insertion (strain NJ-4), as compared with that in CTnDOT (Fig 2). In addition, there was a 25-nt replacement upstream of ermFU in different R. anatipestifer strains, as compared with that in CTnDOT. The -1 to -220*-262 region upstream from the ermFU gene in R. anatipestifer strains may play roles in transcriptional control as a promoter. The above-mentioned sequence differences suggested that expression of the invading ermFU gene in R. anatipestifer may be dependent on certain sequence rearrangements within the transcription and/or translation start signals to accommodate ermFU gene expression in this host [20].
Conversely, as compared with the upstream sequence of ermF from transposon Tn4351, differences of only 1, 5, and 12 nt were found in strains YZ-1 (-1 to -590), ZJb2 (-1 to -677), and JY-6 (-1 to -618), respectively. Moreover, 564 bp (-27 to -590), 651 bp (-27 to -677), and 592 bp (-27 to -618) of the upstream sequences of the ermF gene from strains YZ-1, ZJb2, and JY-6 were the truncated forms of the IS4351R sequence of transposon Tn4351. Although the existing part of an IS4351 element may also be because the transcriptional start site for the ermF gene was contained within this insertion sequence (IS) element [17], our findings suggested that the ermF regions of these R. anatipestifer strains were derived from Tn4351. Both the Bacteroides and Riemerella genera belong to the phylum Bacteroidetes, which compose the most substantial portion of animal and human gastrointestinal flora. Therefore, for R. anatipestifer, Bacteroides could serve as reservoirs of antibiotic resistance genes and other genes, while possibly passing them on to R. anatipestifer.

The genetic environment of the ermFU gene in HXb2
To gain insight into the genetic environment in HXb2, the ermFU gene and the flanking sequences were sequenced and compared with the corresponding regions in R. anatipestifer strains CH-1 (accession number: CP003787), CH-2 (CP004020), CH3 (CP006649), RA-GD (CP002562), and DSM15868 (CP002346). As shown in Fig 3, except for the type strain DSM15868, there was an 8-29-kb MRR in R. anatipestifer strains, and the ermFU gene of strains HXb2 and CH-1 were located in the MRR. Analysis of the MRR in R. anatipestifer strains HXb2, CH-1, CH-2, CH3, and RA-GD, and the corresponding region in strain DSM15868, indicated that the MRR in these strains was inserted into the non-coding region between the gene (psbs, Riean_1797) coding for polysaccharide biosynthesis protein and a gene (hypA, Riean_1798) coding for hypothetical protein. The insertion mechanism of the resistance genes at this specific region on the R. anatipestifer chromosome remains unknown, and only a few short DNA fragments (<45 bp) in the MRR of various other strains were found to share a high homology with certain ISs by online IS Finder software (https://www-is.biotoul.fr//), but we found a 601-bp  non-coding direct repeat (designated DR1) located at the intergenic region between psbs and hypA in DSM15868, and a total five copies of the DR1 were found scattered within a 30-kb region within this genome. DR1 was also found within corresponding regions of strains CH-1(2 copies), CH-2 (5 copies), CH3 (2 copies), RA-DA (5 copies), and HXb2 (5 copies in the MRR). Therefore, we speculated that the non-coding intergenic sequence between these two genes may be a recombinant hot spot for the insertion of antibiotic resistance genes. In addition, we found a non-coding tandem direct repeat (a 517-bp repeat unit designated DR2) located upstream of ermFU in the MRR of strain HXb2, and one copy of the DR2 repeat unit was also found between bla and orf2 in this MRR (Fig 3). Moreover, tandem DR2 repeats were also found in the MRR of strains CH-1 (two copies) and CH3 (one copy), and 1-4 copies of the DR2 repeat unit were found in the MRR or other regions within the genomes of other R. anatipestifer strains.
In addition, the aads-ermFU region in the MRR in strain HXb2 shared an identity of 99.4% to that of strain CH-1, and the 199-bp junction sequences between the two genes were almost identical, with only a 1-nt replacement found, while nt sequences at the right and left junctions of the aads-ermFU region were obviously different. These findings suggest that the aads and ermFU genes may be co-transferred into R. anatipestifer. However, the aads-ermFU region was not found in other plasmids, CTns or bacterial genomes, while both the aads and ermFU genes were present in CTnDOT and shared 97%-99% identity to those in HXb2 and CH-1, while in CTnDOT, they were separated by genes tetX1 and tetX2, and their orientations were also different from those in R. anatipestifer strains HXb2 and CH-1. These findings suggest that perhaps the aads-ermFU region in R. anatipestifer was derived from a new CTnDOT-like element which has not yet been discovered.
Sequence analysis showed that the ereD determinant in the CH-2 genome was flanked by the 217-bp non-coding direct repeat sequence DR3. In R. anatipestifer plasmids pRA0511 and pRA0846, we also identified the DR3 sequence, which was located at the left and right flanking regions of the text-cat genes and floR-truncated sat genes, respectively [22,23]. The appearance of DR1-DR3 in the MRR indicated that these direct repeats may play an important role in transferring resistant genes or others into the genome of R. anatipestifer. In addition, the antibiotic resistance gene order and copy number in the MRR from different R. anatipestifer strains also differed, which suggested that these antibiotic resistance genes may be introduced into the MRR though several recombination events.
Supporting Information