Recruitment of patients and urine collection
All urine samples were collected from January 2008 to January 2010 at the Military Medical University, Hospital 103 in Ha Dong, Hanoi. Nearby pharmacies were instructed to refer patients to Hospital 103 that reported at least one of the following clinical symptoms of UTI: frequent urination, painful urination, hematuria, cloudy urine, pain in pelvic area or lower back. A midstream urine sample was collected at the hospital under supervision of a nurse and bacteriological culture initiated immediately after collection. Based on the established aetiology and results of antimicrobial susceptibility testing, patients were informed about which antimicrobial agent should be used for treatment. All patients were informed orally and in writing in their own mother tongue about the possibility to participate in the study and withdraw at any time without explanation. After the information was given, patients signed an approved consent form. The study protocols were approved by the ethical committee at the Hospital 103. No patients refused to participate.
Bacterial culture of urine
Flexicult agar plates (Statens Serum Institute, Copenhagen, Denmark) were used for culturing urine
. According to the manufacturer pure cultures with ≥ 103 CFU per ml urine should be regarded positive, however, as a conservative measure and due to the findings of unusual aetiology we used a definition of ≥ 104 CFU per ml urine as a positive urine sample. The selective and indicative Flexicult agar plate allows for the detection of the 10 most common pathogens associated with UTI, including E. coli, Klebsiella spp., Enterobacter spp., and Proteus spp.. Urine samples from a total of 276 patients were poured over Flexicult plates immediately after collection and incubated at 37°C for 18–24 hrs and if plates showed no visible growth for another 18–24 hrs. If plates showed growth of E. coli, E. faecium or E. faecalis in pure culture with ≥ 104 CFU per ml urine, three individual colonies were randomly picked from the control compartment in the plate and sub-cultured on non-selective LB agar, Lennox plates (DifcoTM, Becton, Dickinson and Company, Sparks, USA) which were incubated overnight at 37°C to obtain pure cultures. Colonies were subsequently grown overnight at 37°C in Brain Heart Infusion broth (Oxoid, Basingstoke, Hampshire, England) and stored for further characterization at – 80°C in Cryo tubes containing 30% glycerol. The identity of the three colonies was determined to confirm pure culture in the urine samples.
All patients were interviewed when urine samples were taken. Registered personal data included: age and sex as well as self-reported underlying diseases, including hematological disorders, respiratory infections, diarrhoea, diabetes, cancer, HIV/AIDS, liver cirrhosis, alcoholism, anatomical malformations of urinary tract, and history of nephro- or urolithiasis. Patients reporting underlying diseases as described above as well as hospital-acquired infections were excluded. Thus, study subjects included those with acute, recurrent and/or self-medicated UTI. The following clinical symptoms were recorded: frequent urination, painful urination, cloudy urine, blood in urine, pain in pelvic area, pain in lower back and fever. In addition, information was recorded about duration of symptoms, previous episodes of UTI, number of UTI episodes during the last year and how many of these episodes were diagnosed by a doctor. Finally, medical treatment of UTI before enrolment in the study was registered, e.g. type of antimicrobial used.
PCR and 16S rRNA identification of enterococci
E. faecalis and E. faecium were identified by species-specific PCR
. Isolates tentatively identified as E. faecalis or E. faecium based on appearance on the Flexicult agar plate, but testing negative in the E. faecalis- /E. faecium-specific PCR (n=9), were characterized by 16S rRNA sequence analysis (Macrogen®, Seoul, South Korea)
 using primers modified after
. Species identification was done by BLAST search in GenBank by use of the EzTaxon server (
Partial sequencing of groEL
The identity of Strepcococcus gallolyticus subsp. pasteurianus as shown by 16S rRNA sequence analysis was confirmed by partial sequencing (Macrogen®) of the groEL gene
 which encodes a heat-shock protein. Sequences were assembled using CLC Main Workbench 5.2 software (CLC bio, Aarhus, Denmark) and species identification was done by blasting the sequences in the NCBI database (
http://www.ncbi.nlm.nih.gov/). To determine a more precise similarity between the obtained sequences and the sequence of the type strain, local pair wise sequence alignment was done, using EMBOSS water alignment (
Antimicrobial susceptibility testing
MIC were determined for 27 E. faecalis to ampicillin (AMP, 2–32 μg/ml), penicillin (PEN, 2–32 μg/ml), vancomycin (VAN, 1–32 μg/ml) and high-level gentamicin (GEN, 16–1024 μg/ml) using the Sensititre® system (Trek Diagnostics Systems, East Grindstead, England). The following breakpoint values as proposed by EUCAST (European Committee on Antimicrobial Susceptibility Testing;
http://www.eucast.org/) were used: AMP R > 8 μg/ml, VAN R > 4 μg/ml where as R > 128 μg/ml was used for high-level GEN resistance.
Identification of aac (6’)-Ie aph (2”)-Ia resistance gene
All strains which had a MIC ≥ 512 μg/ml and three control strains which had a MIC ≤ 16 μg/ml against gentamicin were tested by PCR for the presence of the aac (6’)-Ie aph (2”)-Ia gene
 which encodes high-level aminoglycoside resistance associated with transposons in E. faecalis (Tn5281), Staphylococcus aureus (Tn4001) and Staphylococcus epidermidis (Tn4031)
MLST of E. faecalis
All urine isolates identified as E. faecalis (n=27) were characterized by MLST to investigate if certain STs were associated with UTI. One of the three E. faecalis isolates collected from each patient was randomly selected and characterized by full MLST sequence analyses of all seven genes (gdh, gyd, pstS, gki, aroE, xpt, and yqil). To confirm pure culture in the urine samples, a second isolate from each of 16 (28.1%) samples was typed by sequence analyses of the gki and yqil genes. Since the DNA sequence of the gki and yqil genes were identical for the two isolates characterized from individual urine samples further sequencing were not done on isolates from the remaining urine samples. The PCR primers and conditions used were those described on the E. faecalis-MLST website (
http://efaecalis.mlst.net/). Amplicons were sequenced in both directions (Macrogen®). DNA sequences were assembled and a sequence type was assigned to each strain using CLC Main Workbench 5.2 software (CLC bio, Aarhus, Denmark) and compared to published alleles (
All urine isolates of ST 16 (n=12) were characterized by Pulsed-Field Gel Electrophoresis (PFGE) to determine their clonality and degree of similarity to four E. faecalis ST 16 strains isolated from Danish endocarditis patients
. DNA was prepared directly in a solid agarose plug (SeaKem Gold Agarose; Lonza, Basel, Switzerland) for restriction endonuclease digestion with the enzyme SmaI (BioLabs, New England, USA). The separations of fragments were done on CHEF-DR III (Bio-Rad, Richmond, CA, USA) at the following conditions: 6V/cm at 14°C for 20 h at a filed angel of 120°. The electrophoresis was carried out at switch times of 2.2 to 54.4 s. Following electrophoresis the gels were stained for 15 min in ethidium bromide (2 mg/ml water; Sigma-Aldrich Denmark A/S, Brøndby, DK), then destained in water for 15 min and visualized under UV light (Gel Doc, Bio-Rad)