First report of the types TEM, CTX-M, SHV and OXA-48 of beta-lactamases in Escherichia coli, from Brazzaville, Congo

The objective of this study was to evaluate the emergence of beta-lactam resistance related to extended spectrum beta-lactamase (ESBL) encoding genes (TEM, SHV, CTX-M-1 and OXA-48) in Escherichia coli from Brazzaville. In the period between January 2016 and May 2017, 89 strains in E. coli were isolated from hospitalized patients, outpatients and domestic sewage. The E. coli strains were identified by the API 20E system. An antibiogram was performed on isolated trains by the disk diffusion method. The ESBL phenotype was detected using the synergistic technique according to CA-SFM (ESBL). Genes were detected using PCR and characterized by sequencing. An overall prevalence of 48.31 (43/89) and rates of 74.42, 23.26, 9.30, and 6.97% for blaTEM genes blaCTX-M-1, blaSHV and blaOXA-48 were detected. 25.58% were community strains and 74.42% hospital. The majority were isolated urine (81.40%) and the urology department was more represented. Except for imipenem, colistin strains of ESBL showed high resistance to other antibiotics than non-yielding ones (p <0.05). This high prevalence of detected ESBL, the high level of resistance to antibiotics raises fears of a high risk of multidrugresistant bacteria and call on the authorities for a policy of monitoring resistance.


INTRODUCTION
Infectious diseases remain a major public health problem, especially in Africa (Ouedraogo et al., 2017).Bacterial infections are the most dominant and are responsible for most of the nosocomial and community infections.Based on their wide spectrum, mechanism of action, low toxicity and low cost, antibiotics including the beta-lactam family such as penicillins, cephalosporins and carbapenems are the most commonly used for treating infection diseases (Livermore, 1995).Over the past year, a significant increase in antibiotic resistance to Enterobacteriaceae has been documented as well.According to the WHO (2014) report on global surveillance of antimicrobial resistance, antibiotic resistance is a serious public health problem.In competitive environments, microorganisms including Streptomyces, Nocardia, Actinomadura, and Penicillium species, can produce natural molecules allowing the inhibition of other competitors (Baquero and Coque, 2013).To avoid self-toxicity, these microorganisms have intrinsically developed antibiotic resistance.Enterobacteriaceae are classified in different groups according to antibiotic resistance (Philippon et al., 1989).Indeed, the abuse of antimicrobials led to the selection of multi-resistant strains, particularly those resistant to beta-lactams isolated from urinary infections, pulmonary infections, and septicemia, with an increasing frequency in hospitals (Lucet et al., 1996).In Enterobacteriaceae, different enzymes are produced depending on their membership in the different groups.The most commonly cited are chromosomal penicillinase, chromosomal cephalosporinase, chromosomal cefuroximase and chromosomal extended spectrum betalactamase (ESBL).Escherichia coli is a dominant Enterobacteriacea found in the human commensal flora, especially in the digestive tract (Ahoyo et al., 2007).
ESBLs are serine-type inactivation enzymes.With the exception of OXA SSBLs (class D), ESBLs are class A beta-lactamases according to the Ambler classification.More than 300 ESBLs have been described to date (Elhani, 2012).They are characterized by a great diversity.The majority of ESBLs are derived from TEM and VHS enzymes but new ESBLs have been described such as CTX-M (for cefotaximase), OXA (for oxacillinase), PER (for Pseudomonas aeruginosa), VEB (for Vietnam ESBL), GES (for Guyana extendedspectrum beta-lactamase), TLA (for TEM Like activity), BES (for Brazilian extended spectrum beta-lactamases), SFO (for Serratia fonticola) and FEC (fecal E. coli) (Elhani, 2012).
ESBLs constitute a large family of bacterial enzymes belonging to classes A, C and D based on the Ambler classification (Hall and Barlow, 2018), and capable of hydrolyzing penicillins, cephalosporins and aztreonam.They do not hydrolyze carbapenems or cephamides.This capacity to hydrolyse antibiotics has also been demonstrated in E. coli and Klebsiella species strains.Since the discovery of ESBL-producing bacteria in 1990, most of those detected are the conventional TEM and SHV types, which spread predominantly in hospital settings, including Klebsiella pneumoniae and Enterobacter species; they are associated with nosocomial outbreaks in intensive care units (Hall and Barlow, 2018).Recently, a new blaCTX-M gene encodes CTX-M enzyme has been detected.Community strains of E. coli, mainly responsible for urinary tract infections, could express blaCTX-M genes encoding CTX-M enzyme (Elhani, 2012) The blaTEM, blaSHV, blaCTX-M and blaOXA genes have been described in several epidemiological studies in Europe, Asia, the USA and South America (Bush and Jacoby, 2010;Winokur et al., 2001;Villegas et al., 2008).
Organisms expressing these genes are widespread throughout the world but some geographical regions have a significantly higher prevalence rate such as South America, Asia and Europe.These prevalence rates have been evaluated by large microbial resistance surveillance programs such as SENTRY and MYSTIC and there is a steady increase in there prevalence (Goossens, 2005;Yano et al., 2013).
Classes A and D enzymes are commonly found in Africa with a predominance of blaCTX-M-15 (Storberg, 2014).In Congo-Brazzaville, a prevalence of 73.8% of E. coli strains harboring beta-lactamase phenotype was reported (Moyen et al., 2014).To our knowledge, no study has been performed on the characterization of Enterobacteriaceae resistance genes.
Based on the aforementioned, this study was carried out to evaluate the emergence of beta-lactam resistance related to genes encoding ESBL (TEM, SHV, CTX-M-1 and OXA-48) in E. coli from Brazzaville.

Bacterial strains
From January 2016 to May 2017, E. coli isolates were isolated from various pathological samples (urine, vaginal swabs excretions, sperm, blood and pus), environmental samples (domestic sewage).
Collected samples were inoculated on Mac Conkey agar or methylene blue eosin (EMB) agar and incubated for 24 h at 37°C in the Bacteriology-Virology Laboratory of the CHU in Brazzaville and the Exau-Kenn Laboratory.
The suspected E. coli isolates were purified and biochemically identified using API 20E® galleries (bioMerieux, France) and the strains were stored in Luria-Bertani (LB) liquid medium (Sigma-Aldrich, France) supplemented with 10% glycerol at -80°C for the duration of the study.

DNA isolation
Bacteria isolates were separately grown overnight on a MacConkey agar plates at 37°C.Subsequently, colonies were picked up using a sterile 10 μL plastic loop and transferred into 200 μL sterile water (Nouria et al., 2016).Total DNA was extracted with the NucliSENS easyMag automat (Biomérieux, France) according to the manufacturer's protocol.DNA was eluted in a final volume of 200 µL.The extracted DNA was stored at -70°C for further analysis.
The PCR was carried out in a final volume of 50 μL containing: 1 μL of each primer at 20 μM, 5 μL of 10X PCR reaction buffer, 1 μL of the dNTPs, 0.25 μL of Taq DNA polymerase (Eurogentec, Belgium), 36.75 μL of distilled water sterile, and 5 μL of crude DNA extract.Standard PCR conditions at 95°C for 15 min; followed by 35 cycles of amplifications each comprising a hybridization step at 55°C for 50 s, an elongation step at 72°C for 1 min, and a final extension step at 72°C for 7 min.
The amplification products were detected by electrophoresis using 1.5% agarose gels containing BET ethidium bromide diluted to 125 μL per 50 ml agarose gel heated in an electrophoresis vat containing buffer TBE (TRIS, Borate, EDTA) at 0.5%.The migration is done for 25 min at a voltage of 135V (Invitrogen, Leek, The Netherlands), as well as a DNA molecular weight marker (BenchTop pGEM®DNAMarker, Promega, Madison, Wisconsin, USA).Visualization of the gels was performed using the leaning ADNPPEM marker (Promega, Madison, Wisconsin, USA) under ultraviolet illumination.

DNA sequencing
Classical PCR products positive for ESBL genes were purified using the NucleoFast 96 PCR plate (Machery-Nagel EURL, France) and sequenced using BigDye terminator chemistry on a automatic sequencer ABI3730 (Applied Biosystems, Foster City, Californie, États-Unis).
The sequences obtained were assembled and corrected with Codon Code Aligner software.Sequence alignment and analyses were performed using the basic local alignement search tool (BLAST) software available on the website of the National Information Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov).
Phylogenetic and molecular evolutionary analyses were conducted using MEGA version 7 (Kumar et al., 2017).

Statistical analysis
For the data analysis, intermediate category strains were counted among the resistant (I + R).The data were analyzed using GraphPad Prism 7 software.The Chi-square test (²) was used to compare the resistance frequencies between the different parameters studied.The difference between frequencies was considered significant when the p-value was less than 0.05.

Ethical clearance
The study was approved by the ethics committee of Marien Ngouabi University.The protocol was reviewed and accepted by the authorities of Brazzaville hospitals (this allowed us to conduct our study).During the study period, we ensured the confidentiality of the results and the anonimate of the patients.

Bacterial strains
Eighty-nine (89) isolates of E. coli non-redundant were collected from two types of samples: an environmental sample consisting of household wastewater collected in a few neighborhoods of Brazzaville and a clinical sample taken from samples for the diagnostic purpose of hospitalized patients and those seen externally at the Makelekele base hospital and at the University Hospital Center of Brazzaville.These samples consisted of five (5) blood cultures, sixty four (64) urine, one (1) vaginal sample, one (1) spermoculture and four (4) pyocultures, distributed as follows (Table 4): 14 E. coli originated from household sewage and 75 from hospital, twenty 20 strains were from outpatients and fifty five (55) straines from hospital services including 3 surgical, 9 urology care, 5 pediatric care (infant and grandchild), 4 infectious deseases, 10 cardiology care, 9 metabolic deseases, 9 general medicine, 3 neonatal and 3 intensive care unit (Table 3).
The resistance profiles of 89 isolates of E. coli studied on the 16 antibiotics used show a 100% sensitivity for imipenem and colistin, in E. coli ESBL+ and E. coli ESBLisolates.
Low resistance was observed for Ertapenem 9.30% (3/43).The frequencies of resistance to amoxicillin, amoxicillin/clavulanic acid and cephalothin, which are antibiotics used in the first-line treatment of urinary tract infections, are relatively high in the two groups, but moreso for the strains of ESBL+ than ESBL-(Table 2).This difference is not statistically significant (p > 0.05).Aside from amoxicillin, amoxicillin/clavulanic acid, cephalothin, ertapenem, imipenem, colistin, and trimethoprim + sulfaphametxazole, there is a statistically significant difference between E. coli ESBL+ and E. coli ESBL-(Table 2).
Regarding to aminoglycosides, there was fairly marked resistance for gentamicin in ESBL E. coli isolates only at 34.88%, but it was only 9.30% for amykacin.In addition, with respect to fluoroquinolones, a resistance level of 13.95% was detected for ciprofloxacin; Furans remain resistant to 69.77%; as for the sulfamides and tetracyclines, the latter remain resistant in the two groups of strains studied (Table 2).

Characterization of ESBL genes
Figure 1 shows the blaTEM, blaSHV, blaCTX-M-1 and blaOXA-48 genes with bands of respective sizes of 861, 1051, 944 and 744 base pairs.Figure 2 shows the evolutionary history of ESBL genes (blaTEM, blaSHV, blaCTX-M-1).The optimal tree with the sum of branch length = 3.01339548 is shown.Evolutionary distances were calculated using the composite maximum likelihood method and are expressed in units of the number of base substitutions per site.All positions with gaps and missing data have been eliminated.There was a total of 479 positions in the final dataset.
The distribution of strains according to the hospital service shows a predominance of ESBL producing strains for the surgical and pediatric departments at 100%.The urology department comes in second place with 77.78%; followed by infectious diseases 75%, metabolic diseases (55.56%) and cardiology (40%).General medicine, pediatric intensive care and neonatology departements come last (33.33%)(Table 3).
Sequencing of PCR amplification products and after comparison based on NCBI data revealed that: all 32 blaTEM-positive strains were all TEM-1 AAR25033.1 (Table 5); the 12 CTX-M strains of group 1 were all CTX-

DISCUSSION
ESBL-producing bacteria are a major concern in community and hospital settings because of their epidemic spread and multidrug resistance.In fact, ESBLs are found in a large proportion of Gram-negative bacilli, but enterobacteria are the most incriminated organisms (Gniadkowski, 2001).The present study, the first conducted in the Republic of Congo, detected 43 strains of E. coli producing ESBL on 89 strains tested, a prevalence of 48.31% between January 2016 and May 2017.
The high prevalence of ESBL-producing E. coli isolates observed in the present study is probably a consequence of selection pressure due to inappropriate prescribing and misuse of broad-spectrum antibiotics in both hospital and community setting (dispensing without prescription, selfmedication).
The rate of isolation of ESBL E. coli isolates was greater in the hospital (74.42%) than in the community (25.58%) with a statistically significant difference (p-value <0.05).These results corroborate with the literature on the epidemiology of ESBLs.The duration of hospitalization, the severity of the disease, the surgical intervention, the wearing of arterial or urinary catheters are the risk factors for the acquisition of ESBL in hospitalized patients (Bradford et al., 1997;Paterson, 2000).
Also, these species are distributed differently according to the services and sites of sampling.Urinary tract infection is a common pathology in daily practice.The main bacterial species involved in this infection is E. coli since it represents 50 to 80% of the agents involved (Matute et al., 2004).This corresponds to the results obtained in the present study with a rate of 81.40%.This is related to the physiopathology of urinary tract infection, which is usually ascending, and there is a strong colonization of the perineum by Enterobacteriacae of digestive origin, and in particular E. coli.In addition, there are specific factors of uropathogenicity.Thus, Escherichia coli has adhesins, capable of binding the bacterium to the urinary epithelium and preventing its removal (Matute et al., 2004).
In the present study, most ESBL producers were collected from patients in the surgical ward and the pediatry than other reported services.In these wards, isolates are exposed to great antibiotic pressure.Furthermore, many of these patients are particularly vulnerable to infection because they are immunocompromised or have an easy avenue of access for bacteria (Xiong et al., 2002).
In addition, the problem related to ESBL is mainly the frequent presence of co-resistances making multiresistant strains (Touati et al., 2012).
Indeed, ESBLs are usually carried by large plasmids which also carry resistance genes to non-β-lactam antibiotic classes, such as aminoglycosides, quinolones and trimethoprim/sulfamethoxazole.As well, the use of these antibiotics contributes to the selection of producing strains from ESBL (Paterson and Bonomo, 2005).
The incidence of occurrence of resistances aminoglycoside has increased in recent years and particularly in producer of ESBL (Spanu et al., 2002).The resistance levels of these strains in the present study is more important for gentamycin (34.88%).Amikacin remains the most effective molecule with 90.7% of susceptible strains as reported in several studies (Gangoue et al., 2005;Touati et al., 2012).
The fluoroquinolones show a fairly good activity, the overall sensitivity of the strains falling considerably in the case of strains producing ESBL.For the majority of strains, acquired resistance is the consequence of a mutation, which limits its diffusion (Larabi et al., 2003).A very low activity was detected for trimethoprim/sulfamethoxazole (93.02%) and tetracyclines (with resistant strains).This correlates with other studies, where many ESBL producers are multiresistant to non-β-lactam antibiotics, including fluoroquinolones and aminoglycosides (Livermore et al., 2007), Consequently, effective antibiotic therapy for treating these infections is limited to a small number of drugs, such as carbapenems and thus increasing the chance of resistance to carbapenems among the Enterobacteriaceae (Pitout, 2010).
As far as colistin is concerned, it is one of the molecules with the highest sensitivity levels on naturally occurring species sensitive (apart from Proteus).This is likely related to a lower use of this antibiotic in the current practice.This correlates with results of the susceptibility rate to this molecule were also reported by Nouria et al. (2016).
Carbapenems are the treatment of choice for severe infections caused by ESBL-producing enterobacteria (Nordmann et al., 2008).The resistance of enterobacteria to these molecules is still a marginal phenomenon as in the epidemiological data obtained for a large number of strains, with sensitivity percentages of 99 to 100% (Nordmann et al., 2008).In the present study, 9.3% strains had resistance to ertapenem and 100% were sensitive to imipenem.This decrease in sensitivity to ertapenem may be related to carbapenemase production since carbapenem activity may be compromised by the emergence of these enzymes (Munoz-price et al., 2013 ;Sekhri et al., 2010).
ESBLs are divided into five types: TEM, SHV, CTX-M, OXA-48, and others, based on the homogeneity of coding genes.Most ESBLs derived from plasmid-mediated penicillinases belonging to TEM or SHV famillies (Xue et al., 2012).Recently, the CTX-M group with a typical ESBL resistance phenotype but does not originate from TEM or SHV families have been described.The CTX-M group is a new family of plasmid-mediated ESBLs that preferentially hydrolyse cefotaxime (Xiong et al., 2002).
TEM and SHV-type ESBLs remain more common in North American and in Africa.CTX-M type ESBLs have been mainly in South America, Eastern Europe, Japan and more recently in Spain, Kenya (Xue et al., 2012) and Algeria (Nouria et al., 2016).
In this study, it was determined that most of 32 E. coli isolates were prevalent TEM-type ESBLs.TEM was the main type of beta-lactamase and CTX-M was the second.SHV was detected in five isolates.In addition to OXA-48 was detected in 3 isolates.Another interesting point is that the results modify the current epidemiology of ESBL in Enterobacteriaceae, the CTX-M that represented the majority of ESBLs in any region of the world, both in hospitals and community settings, to such an extent that their spread of pandemic (Elhani, 2012).

Conclusion
This can be the first study revealed the real existence of the genes encoding ESBL in E. coli isolates from the community and hospital patients in Brazzaville.A high prevalence (48, 31%) of ESBL E. coli isolates with predominance of the blaTEM gene.Fighting against this phenomenon is multidisciplinary and should integrate the rationalization of district compliance with the prescription of antibiotics, with hygiene measures.Justification of antibiotherapy policy and/or a restriction of the prescription of third-generation cephalosporin and even all beta-lactams would lead to a significant decrease in the frequency of ESBL.
Monitoring of antibiotic resistance of bacterial strains should be continuous and systematic to define therapeutic strategies adapted to local epidemiology data.

Figure 2 .
Figure 2. Evolutionary relationships of ESBL genes.Dendrogram generated using BioEdit and MEGA7.1 from the blaTEM gene set (blue square), blaCTX-M-1 (Red dot) and blaSHV (Pink Triangle) by comparing their relationship to each other and to other genes ESBL deposited at GenBank (black text).The lengths of branches are indicative of kinship.

Table 1 .
Primers used for conventional PCR and sequencing.

Table 4 .
Frequency of E. coli (N=89) and nature of samples.

Table 5 .
Characterization of the ESBL genes of the 43 E. coli isolates of hospital and community origin.