Characterization of the Emerging Enteropathogen Escherichia Albertii Isolated from Urine Samples of Patients Attending Sapporo Area Hospitals, Japan.

Recently, Escherichia albertii has been identified as a causative agent of diarrhea in humans and is often misidentified as diarrheagenic Escherichia coli (DEC), a lactose-nondegrading bacterium. In this study, we performed biochemical characterization, gene possession status, drug susceptibility testing, and sequencing analysis of the strains detected in urine samples. One urea-degrading strain was detected in terms of biochemical characteristics, but was found to be nonurea-degrading by another method, leading to conflicting results. All target strains possessed the E. albertii-specific gene, the DEC common gene eae, and the E. coli 16S rRNA gene. In the drug susceptibility test, all urine-derived strains were sensitive to tetracycline (TC), whereas the JCM 17328 strain was resistant to TC, suggesting that TC is effective against urine-derived E. albertii strains. In 16S rRNA sequencing analysis, the E. albertii strains were ranked at the top of homology, but not in the top one, making it difficult to differentiate them from other strains. In summary, if a suspected lactose-nondegrading E. coli strain was isolated from a urine sample, it could be differentiated from E. albertii by the presence of E. albertii-specific genes.

1. Introduction

Humans and numerous bird species are susceptible to Escherichia albertii, a developing enteropathogen responsible for epidemic mortality [1, 2]. Because of their phenotypic and genetic similarities, such as numerous biochemical traits and the possession of a type III secretion system encoded by the locus of enterocyte effacement, this bacterium has been commonly misdiagnosed as enteropathogenic or enterohemorrhagic E. coli [3]. Some E. coli strains have acquired virulence genes that confer pathogenicity and are referred to as diarrheagenic E. coli (DEC) [4]. However, in general, E. coli that cause urinary tract infections (UTI) are said to be different from those that cause diarrhea [5] and there have been reports of DEC detection in the urine [6].

E. albertii has biochemical properties similar to those of atypical E. coli, making it difficult to identify the bacterial species based on their biochemical properties alone [7, 8]. Since E. albertii may be misidentified by mass spectrometry, therefore, the detection of specific genes via PCR is the practical method for identifying E. albertii [9]. In recent years, we reported the isolation of E. albertii from fecal and urine samples [10]. However, in the present study, E. albertii strains isolated from urine samples were characterized in details through biochemical, gene possession status, drug susceptibility testing, and sequencing analysis. It was suggested that E. albertii could be identified by searching for the E. albertii-specific genes like mdh, lysP, and clpX.

2. Materials and Methods

A total of 48,516 bacterial strains were isolated from patients with fecal and urinary tract infections attending different private hospitals around the Sapporo area in Japan. Among the 35,646 urine-derived strains identified as E. coli by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry(MALDI-TOF MS, MALDI Biotyper; Bruker, Billerica, MA, USA) and MBT Compass Library (ver. 8.0.0, Bruker), we used six strains that were ranked in the top 10 as E. albertii [10]. Note that fully unlinkable anonymized strains were used in this study.

2.1. Biochemical Properties

Biochemical properties of E. albertii were examined using the API 20E system (bioMérieux, Marcy-l'Étoile, France). Glycolysis tests, which were not performed on the API 20E system, were performed by adding xylose to the BTB lactose agar plates (Nissui, Tokyo, Japan) at a rate of 1% E. albertii JCM 17328 was used as the reference strain. The strains those were positive for the urea degradation tests using the API 20E system, were confirmed to be degraded by rapid urea broth (BD, East Rutherford, NJ, USA).

2.2. PCR Amplification

Six strains were examined for their genetic characteristics. We amplified the malate dehydrogenase gene (mdh), lysine-specific transporter gene (lysP) [9], and heat-shock protein gene (clpX) [2] as specific genes for E. albertii by multiplex PCR. In addition, we used the common gene eae between E. albertii and DEC, and the E. coli 16S rRNA gene for PCR. Oligonucleotide primers used in this study are listed in Table 1.

Table 1

PCR primers used in this study.

Target genes		Sequences (5′ to 3′)	Sizes of PCR product (bp)	Reference
Escherichia albertii	mdh	CTGGAAGGCGCAGATGTGGTACTGATT	115	[9]
		CTTGCTGAACCAGATTCTTCACAATACCG
	lysP	GGGCGCTGCTTTCATATATTCTT	252	[9]
		TCCAGATCCAACCGGGAGTATCAGGA
	cplX	TGGCGTCGAGTTGGGCA	384	[2]
		TCCTGCTGCGGATGTTTACG
Diarrheagenic E. coli	eae	TGGCGTCGAGTTGGGCA	881	[11]
		TCCTGCTGCGGATGTTTACG
Escherichia coli	16S rRNA	GGAAGAAGCTTGCTTCTTTGCTGA	544	[12]
		AGGCCCGGGGATTTCACATCTGACTT

For DNA extraction, each strain was cultured in 3 mL of Luria–Bertani broth (Sigma–Aldrich Japan, Tokyo, Japan) and incubated overnight at 37°C with shaking at 115 rpm. A 100 µL aliquot of bacterial culture was suspended in 900 µL of Tris-EDTA buffer (Nippon gene, Tokyo, Japan), boiled for 5 min, centrifuged at 10,000 × g for 5 min, and the supernatant was collected.

The multiplex PCR was performed as follows: each 25 µL of the reaction mixture contained 20 mM Tris-HCl, 100 mM KCl, 1.5–2.0 mM MgCl2, 2.5 mM of each deoxynucleoside triphosphate(dNTP) mixture, 1.25 U of Taq DNA polymerase (TaKaRa Bio Inc., Shiga, Japan), 25 µM of each primer, and 2.5 µL of the DNA template. These samples, which had been preheated at 94°C for 1 min, were amplified for 25 cycles using a thermal cycler (TaKaRa PCR Thermal Cycler Dice touch, TaKaRa, Shiga, Japan). Each cycle consisted of denaturation at 94°C for 1 min, annealing at 55°C for 1 min, and extension at 72°C for 1 min. The final extension step was performed at 72°C for 10 min. PCR products were separated by electrophoresis on a 2.5% agarose gel.

2.3. Antimicrobial Susceptibility Tests

Antimicrobial susceptibility was determined using the Kirby–Bauer disk diffusion method [13] on Mueller–Hinton agar plates (Eiken, Tochigi, Japan). The antibiotics used in this study were 5 µg ofloxacin (OFLX), 5 µg ciprofloxacin (CPFX), 5 µg erythromycin (EM), 1 µg fosfomycin (FOM), and 30 µg tetracycline (TC) (Sensi disk, BD Fukushima, Japan). The inoculated plates were incubated for 24 h aerobically at 37°C. The diameters of the zones of inhibition were interpreted according to the Clinical Laboratory Standards Institute guidelines [14].

2.4. 16S rRNA Sequence Analysis

Direct DNA sequencing was performed to accurately identify the six E. albertii strains. The template DNA was prepared by suspending one colony of the target E. albertii, which had been purely cultured on heart infusion agar medium (Eikenn, Tochigi, Japan) in 1 mL of TE buffer, heating at 100°C for 5 min, followed by centrifugation at 10,000 rpm for 5 min, and collecting the supernatant. The following primers were used for PCR and sequencing; 8UA:5′-AGAGTTTGATC(A/C)TGGCTCAG-3′, 519A:5′-CAGC(A/C)GCCGCGGTAAT-3′, 519B:5′-ATTACCGCGGC(T/G)GCTG-3′, 907A:5′-AAACT(T/C)AAA(T/G)GAATTGACGG-3′, 907B:5′-CCGTCAATTC(A/C)TTT(A/G)AGTTT-3′, and 1485B:5′-TACGGTTACCTTGTTACGAC-3′ [15]. The PCR assay was performed as follows: each 50 µL reaction mixture contained 25 µL of PrimeSTAR polymerase (TaKaRa, Shiga, Japan), 10 µM of each primer, and 1 µL of the DNA template. Each cycle consisted of denaturation at 94°C for 15 sec, annealing at 55°C for 5 sec, and extension at 72°C for 10 sec were amplified for 30 cycles using a thermal cycler. PCR products were separated by electrophoresis on a 2% agarose gel. After confirming the bands by electrophoresis, DNA was extracted from agarose gels for samples with extra bands and from PCR reaction solution for samples without extra bands (gel PCR extraction kit, Nippon Gene, Tokyo, Japan). The purified DNA and the four 9.6 μM sequence primers [15] were subject to 16S rRNA sequence analysis by an outside supplier(Eurofins Genomics, Tokyo, Japan). The obtained sequences were done pairwise comparison by MEGA X software [16] and EZBioCloud [17].

3. Results

3.1. Biochemical Properties

The biochemical properties of the six E. albertii strains are listed in Table 2. All E. albertii strains were nonmotile and positive for lysine decarboxylase, ornithine decarboxylase, indole, glucose, mannitol, and L-arabinose, while being negative for arginine dihydrolase, citrate, H2S, tryptophan deaminase, Voges–Proskauer, gelatinase, inositol, L-rhamnose, D-melibiose, amygdalin, lactose, and D-xylose. Furthermore, three ONPG-positive E. albertii strains, one urea-positive strain, four D-sorbitol-positive strains, and one sucrose-positive strain were detected. In addition, the strain that was positive for urea degradation with API 20E was negative in the rapid urea broth. The six E. albertii strains and the JCM17328 strain were identified as E. coli using the API system. There were five different combinations of test results from the six isolates, so isolate no. 1 and no. 2 strains have the same results.

Table 2

Biochemical properties of E. albertii strains.

Agents or tests	E. albertii strains isolated from urine samples						Reactions for type strain E. albertii
	No. 1	No. 2	No. 3	No. 4	No. 5	No. 6	JCM17328
Motility	－	－	－	－	－	－	－
ONPG	－	－	+	+	+	－	+
Arginine dihydrolase	－	－	－	－	－	－	－
Lysine decarboxylase	+	+	+	+	+	+	+
Ornitine decarboxylase	+	+	+	+	+	+	+
Citrate	－	－	－	－	－	－	－
H2S	－	－	－	－	－	－	－
Urea (API 20E/urea broth)	－	－	－	－	－	+/-†	－
Tryptophan deaminase	－	－	－	－	－	－	－
Indole	+	+	+	+	+	+	－
Voges proskauer	－	－	－	－	－	－	－
Gelatinase	－	－	－	－	－	－	－
Acid from glucose	+	+	+	+	+	+	+
D-mannitol	+	+	+	+	+	+	+
Inositol	－	－	－	－	－	－	－
D-sorbitol	+	+	+	－	－	+	－
L-rhamnose	－	－	－	－	－	－	－
Sucrose	－	－	－	－	+	－	－
D-melibiose	－	－	－	－	－	－	－
D-amygdalin	－	－	－	－	－	－	－
L-arabinose	+	+	+	+	+	+	+
Lactose	－	－	－	－	－	－	－
D-xylose	－	－	－	－	－	－	－

+, positive; −, negative. †: API 20E, positive; rapid urea broth, negative.

3.2. PCR Amplification

The results showed that all six E. albertii strains and the E. albertii JCM 17348 strain harbored E. albertii-specific genes mdh, lysP, and clpX along with the eae and E. coli 16S rRNA genes.

3.3. Antimicrobial Susceptibility Tests

The antimicrobial susceptibility of the E. albertii strains to five antibiotics was investigated. Six strains were susceptible to OFLX, CPFX, FOM, and TC but were resistant to EM. The JCM 17328 strain was susceptible to OFLX, CPLX, and FOM but was resistant to TC and EM.

3.4. 16S rRNA Sequence Analysis

The sequence homologies with E. albertii JCM 17328 is shown in Table 3. Four out of the six strains had the highest homology with E. albertii, ranging from 99.44% to 99.52%. One strain had the highest homology with E. marmotae (98.92%) and the fifth-highest homology with E. albertii (98.41%). The other strain was Shigella dysenteriae, with the highest homology (98.97%), followed by E. albertii (98.70%). The four strains with the highest homology to E. albertii included E. fergusonii and Shigella among the top five homologs. One strain with the highest homology to E. marmotae and one strain with the highest homology to S. dysenteriae included not only E. albertii but also E. fergusonii and Shigella in the top five homologous strains. Five different sequences were analyzed from the six strains, and no. 4 and no. 5 strains have the same sequence analysis.

Table 3

16S rRNA sequence analysis.

Hit taxon name (hit strain name)	Strain no. (%)
	1	2	3	4	5
Escherichia albertii (TW07627)	99.22	98.41	98.70	99.45	99.45
Escherichia marmotae (HT073016)		98.92	98.63
Escherichia fergusonii (ATCC35469)	98.96	98.56	98.43	99.04	99.04
Shigella flexnari (ATCC29903)	98.88	98.48	98.36	98.98	98.98
Shigella boydii (GTC779)	98.80	98.41		98.91	98.91
Shigella sonnei (CECT4887)	98.72			98.84	98.84
Shigella dysenteriae (ATCC13313)		98.48	98.97

Noted: bold number is the highest homology in each E. albertii strain.

4. Discussion

E. albertii was first isolated from pediatric stools in Bangladesh in 1991 and was initially identified as Hafnia alvei [18]. Subsequently, Huys et al. [1] registered the strain as a new species of the genus Escherichia by 16S rRNA sequencing analysis, DNA-DNA hybridization and property analysis. There have been reports of infection cases worldwide, including in Japan [19, 20], but all cases have been isolated from diarrhea-derived stools; there have been no reports of isolation cases from urine samples. In the present study, detailed biochemical properties of the E. albertii strain isolated from specimens derived from patients with UTIs were analyzed. Nataro et al. [21] classified E. albertii into biotypes 1 and 2 using lysine decarboxylation and the indole reaction. The JCM 17328 strain used in this study was classified as biotype 1 based on lysine decarboxylation positivity and indole reaction negativity. However, six urine-derived strains were positive for lysine decarboxylation but not for the indole reaction and were not classified as either biotype 1 or 2. It was suggested that different types of strains from the JCM 17328 strain derived from diarrheal patients may be the cause of community-acquired UTI. In addition, one urea-degradation-positive strain was detected in API 20E. E. albertii is known to be negative for urea degradation [22] and no positive strains have been reported to date. Most E. coli strains, which have biochemical properties similar to those of E. albertii, are negative for urea decomposition. However, there have been studies using other urea decomposition tests that have found urea decomposition-positive E. coli strains [23]. Therefore, it was suggested that different results would be obtained if different urea decomposition test methods were used.

In the drug susceptibility test, the susceptibility of OFLX and CPFX to the new quinolones was good. The antibacterial spectrum of new quinolones is broad and they are indicative of systemic infections, such as pneumonia and UTIs [24]. As such, we believe that they are effective against UTIs caused by E. albertii. Stock et al. [25] reported that E. albertii is sensitive to FOM, which we have corroborated here, indicating that FOM is effective against E. albertii infection. Macrolide EM is a hydrophobic drug with a high molecular weight that does not pass through the hydrophilic porin pores formed in the outer membrane of Gram-negative rods and is considered to have little antibacterial activity against Gram-negative rods but is a known treatment for Campylobacter enteritis [26]. E. albertii and its close relatives, E. coli and Salmonella, are considered naturally resistant to EM [27], and E. albertii is also EM-resistant. In the present study, the six target strains and JCM17328 strains were EM-resistant, confirming that EM is ineffective against E. albertii. Chopra and Roberts [28] reported that Tetracyclines are broad-spectrum agents, exhibiting activity against a wide range of Gram-positive and Gram-negative bacteria, atypical organisms such as Chlamydia, Mycoplasma, and Rickettsia, and protozoan parasites, E. albertii is naturally resistant to TC as well as EM. In this study, JCM strains were TC-resistant, but all six strains were TC-sensitive, suggesting that TC is effective against urine-derived strains.

The intimin gene eae is known to cause attaching and effacing (A/E) lesions in intestinal tissues [29]. eae possessed by enterohaemorrhagic and enteropathogenic E. coli (EHEC and EPEC, respectively), which are classified as DECs, is known to be a common gene in E. albertii [7]. As such, E. albertii may be misidentified as EHEC or EPEC in DEC searches of diarrheal-derived strains. However, DEC-related genes have not been identified in strains suspected to be E. coli in urine specimen-derived strains, suggesting that E. albertii may have been misidentified as lactose-nondegrading atypical E. coli. Abberton et al. [30] reported that the E. coli 16S rRNA gene enabled differentiation between E. coli and Shigella. We tested whether it is possible to distinguish E. albertii from E. coli by targeting the E. coli 16S rRNA gene. All strains, including the JCM17328 strain, harbored the E. coli 16S rRNA gene, suggesting that a search for this gene also misidentified the strain as E. coli. Currently, E. albertii is identified by PCR targeting specific genes: mdh, lysP, and clpX [31-33]. The six E. albertii strains in this study possessed these specific genes and were identified as E. albertii. Therefore, eae-harboring strains need to be differentiated from E. albertii, and targeting the specific genes mdh, lysP, and clpX is effective.

Generally, 16S rRNA sequencing analysis is performed for bacteria that are difficult to culture or cannot be differentiated by conventional tests [34] and as such, we differentiated E. albertii from the 16S rRNA region. In the present study, we analyzed these strains using direct sequencing, which is used for rapid species identification in clinical laboratories [15]. One strain showed a high homology with E. marmotae, which was isolated from guinea pigs living on the Tibetan plateau and was registered as a new species in 2015 [35]. However, there have been no reports to date of infections caused by this organism. E. marmotae is characterized by ornithine decarboxylase negativity, rhamnose, and xylose degradation [35]. The strain in the current study did not possess any of these characteristics and was distinguishable from E. marmotae because it contained a specific gene for E. albertii. In addition, strains with high homology to S. dysenteriae could also be differentiated by possession of these specific genes and E. coli 16S rRNA genes, and all strains could be identified as E. albertii. The 16S rRNA sequence is similar in Escherichia species [36], which makes differentiation difficult. Of the strains included in this study, E. fergusonii was included within the top five of the six E. albertii strains. E. fergusonii has been associated with only a few case reports of disease in individuals of animal or human origin [37]. Therefore, it is difficult to identify all E. albertii by homology alone in 16S rRNA analysis, and differentiation from E. fergusonii is required; however, identification is possible by searching for E. albertii-specific genes.

In this study, the two strains showed the same pattern in biochemical properties, but another two strains showed the same pattern in sequence analysis. Even though the properties were the same, the sequence analysis yielded different results, suggesting that E. albertii does not show a relationship between biochemical properties and sequence analysis. We believe this makes identification more difficult.

5. Conclusion

Diarrhea-causing E. albertii was isolated from urine samples. E. albertii is difficult to identify using biochemical characterization or 16S rRNA sequencing. To identify E. albertii, it is necessary to differentiate it from EHEC and EPEC when the DEC-related gene eae is detected in stool samples. In addition, when strains suspected of being lactose-nondegrading atypical E. coli were detected in urine samples, it was suggested that E. albertii could be identified by searching for the E. albertii-specific genes mdh, lysP, and clpX.