Diversity of Cronobacter genus isolated between 1970 and 2019 on the American continent and genotyped using multi-locus sequence typing

Abstract

This study aimed to evaluate the Cronobacter spp. strains isolated on the American continent and characterized using multi-locus sequence typing (MLST) available in the PubMLST database and current literature. From 465 Cronobacter spp. strains, the majority (n = 267, 57.4%) was from North America, mainly from USA (n = 234) and 198 (42.6%) were from South America, mainly from Brazil (n = 196). A total of 232 (49.9%) were isolated from foods, 102 (21.9%) from environmental, 87 (18.7%) from clinical, 27 (5.8%) from PIF, one from water (0.2%) and 16 (3.5%) from unknown sources. A total of five species were represented: Cronobacter sakazakii (374, 80.4%), Cronobacter malonaticus (41, 8.8%), Cronobacter dublinensis (29, 6.2%), Cronobacter turicensis (16, 3.5%) and Cronobacter muytjensii (5, 1.1%). The strains with complete MLST profile (n = 345) were assigned to 98 STs, a ratio of 3.5 strain by ST found and the calculated Simpson`s index was 0.93. The strains showed a high diversity and after eBURST analysis, 30 STs (n = 189) formed 12 single and/or double-locus variant clonal complexes (CC). A total of 38 STs (38.7%) were associated with clinical cases of infection, including well established C. sakazakii CC 1, 4, 8 and 83; C. malonaticus ST60, 307, 394 and 440; and C. sakazakii ST 12 and 494.

Cronobacter, MLST, genetic diversity, epidemiology, foodborne pathogens, bacterial infections

INTRODUCTION

The Cronobacter genus belongs to the family Enterobacteriaceae, for which seven species have been described (Joseph et al. 2012a; Iversen et al. 2008). The genus has come to the attention of the food industry, especially infant formula manufacturers and regulators due to its association with life-threatening infections of neonates. According to Forsythe (2018), the Cronobacter species can be grouped according to their clinical relevance, with Cronobacter sakazakii and Cronobacter malonaticus forming Group 1 and is composed of the majority of clinical isolates; Group 2 comprises Cronobacter turicensis and Cronobacter universalis, which have been rarely reported clinically; Group 3 comprises Cronobacter dublinensis, Cronobacter muytjensii and Cronobacter condimenti, which are primarily environmental commensals and are probably of little or no clinical significance. The majority of cases are in the adult population (Patrick et al. 2014; Alsonosi et al. 2015; Kadlicekova et al. 2018) however only one outbreak has been attributed to contaminated food. This was due to C. sakazakii (Yong et al. 2018).

There has been considerable concern related to the presence of Cronobacter spp. in powdered infant formula (PIF) due to their highlighted association with neonatal infections (Flores et al. 2011; Pan et al. 2014; Fei et al. 2017). The Food and Agriculture Organization of the United Nations and the World Health Organization (FAO/WHO) has undertaken three risk assessments of Cronobacter spp. in PIF, powdered follow-up infant formula (FUF) and other infant food (FAO/WHO 2004, 2006 and 2008). Their summary recommendations included the use of internationally validated detection and molecular typing methods for Cronobacter spp. to gain a better understanding of the ecology, taxonomy, virulence and other characteristics of Cronobacter spp. and on ways to reduce its levels in reconstituted PIF. Consequently, there has been considerable research into improved detection methods, more reliable identification procedures, genotyping schemes and analysis (Forsythe 2018).

The application of conventional (7-loci) multilocus sequence typing (MLST) has revealed considerable information on the emergent bacterial pathogen. The Cronobacter spp. MLST scheme develop by Baldwin et al. (2009) requires the partial sequence analysis of seven housekeeping genes: atpD, fusA, glnS, gltB, gyrB, infB and ppsA. This DNA-sequence approach, supported by an open access curated database, has established new Cronobacter species and recognition of the clonal lineages (pathovars) attributed to the majority of neonatal meningitis and necrotizing enterocolitis cases (Joseph et al. 2012b; Joseph and Forsythe 2012). As an example of the benefits of rapid DNA sequencing, even without considering Next Generation Sequencing of genomes, the study of just 7-loci is a model for modern microbial epidemiology and food microbiology ( McMullan et al. 2018; Lepuschitz et al. 2019; Strysko et al. 2020). The majority of strains deposited in the PubMLST Cronobacter spp. database come from Asia, in particular China (n = 1315; 44.4%) and Europe (n = 969; 32.7%; PubMLST, 2020). However, despite the high number of entries from China (n = 1100; 39%), there have been few reported clinical cases in China with which to compare the distribution of Cronobacter spp. between different sources (Ling et al. 2021). In comparison, there are 492 (16.6%) strains in the database from the American continent from a wide range of sources.

Because reporting is not mandatory in most countries, the true incidence of invasive infant Cronobacter spp. infections is unknown (Strysko et al. 2020). Strysko et al. (2020) reviewed all cases of blood-stream infection or meningitis among infants that had been reported to the Center for Disease Control and Prevention (CDC) and in the literature (1961–2018, n = 183). They reported that global annual reporting was significantly higher during the final quarter of the study period, increasing from a mean of 1.2 cases/year before 2004 to 8.7 cases/year from 2004–2018.

Many cases of Cronobacter spp. infection have been reported in America, including severe meningitis cases of neonates (CDC 2009; Asato et al. 2013; Hariri, Joseph and Forsythe 2013; Umeda et al. 2017; Chaves et al. 2018; Sundararajan et al. 2018; Strysko et al. 2020) and adult infections (Patrick et al. 2014; Alsonosi et al. 2015; Kadlicekova et al. 2018). Consequently, accurate microbial source tracking is needed to support epidemiological, regulatory actions surveillance and preventive measures. For example in identifying routes of contamination in foods and in particular PIF. The use of a centralized genotyping database enables further analysis of Cronobacter spp. strains isolates from foods, environmental and clinical samples on the American continent. This will lead to a better understanding of the epidemiology of these pathogens in this specific geographic area.

In this study, the genetic diversity and epidemiological profile of Cronobacter spp. strains isolated on the American continent were characterized using MLST, and compared with other Cronobacter spp. strains deposited in the PubMLST database.

MATERIALS AND METHODS

Bacterial strains

This study analyzed 480 Cronobacter spp. strains isolated on the American continent and deposited in the Cronobacter PubMLST open access curated database (Jolley, Bray and Maiden 2018; PubMLST 2020). General metadata available included ST, CC, species, country, year of isolation, source and association with human infections (last access 10/15/2020). The list and details of all the isolates included in this study, along with their MLST profile details, are in supplementary file 1.

MLST of Cronobacter spp. isolates from the American continent

The Cronobacter species were confirmed using the fusA allele sequence (Forsythe 2018). A neighbor-joining tree based on multiple alignments of the concatenated sequences of the seven loci (atpD, fusA, glnS, gltB, gyrB, infB and ppsA; 3036 nucleotides concatenated length) was constructed with the strains with complete MLST allelic profiles in the database (n = 345 of 465 strains) using the Interactive tree of life (iTOL) v3 (Letunic and Bork 2016). The MLST profiles were also clustered with GrapeTree (Zhou et al. 2018) using a categorical coefficient and visualized using the minimum spanning tree (MST) tool and also analyzed using the eBURST algorithm (Feil et al. 2004).

The Simpson`s index (SI) was applied to measure the genetic diversity among isolates using the fusA allele and also the concatenated 7-loci MLST sequences Hunter and Gaston (1988).

Investigation of Cronobacterspp. infections and occurrence in PIF in America

Cases of infection due to Cronobacter spp. and isolation of the organism from PIF in each American’ country were compiled using the search terms ‘Cronobacter’ or ‘Enterobacter sakazakii’ in combination with ‘infection’, ‘newborn’, ‘infant’, ‘neonate’ or ‘the name of each America country’ to conduct a literature search of the Medline, Scientific Electronic Library Online (SciELO), Scopus and ResearchGate databases and review associated bibliographies.

RESULTS AND DISCUSSION

A total of 465 Cronobacter spp. strains isolated on the American continent in the period of 1970–2019 had been deposited in the PubMLST database. The majority (57.4%) were from North America especially from USA (50.3%), followed by Canada (5.6%) and few strains (1.5%) from Mexico. Regarding South America, Brazil was the country of origin for almost all strains (n = 196), and Chile and Uruguay had only one strain each (Table 1). A total of 232 (49.9%) were isolated from foods, 102 (21.9%) from environmental, 87 (18.7%) from clinical samples, 27 (5.8%) from infant formula, one from water (0.2%) and 16 from unknown sources (3.4%). The distribution of Cronobacter spp. STs according to the source is shown in Fig. 1.

Figure 1.

Minimum spanning tree showing the genetic relationship of 345 Cronobacter strains isolated from the American continent and deposited in the PubMLST database (http://www.pubmlst.org/cronobacter/, last access 10/15/2020). Strains are distributed according to ST classification. The size of the circle is proportional to the number of the strains; strains sharing five or more alleles are surrounding by a gray halo and represents the 12 groups found in the eBURST algorithm analyses and clonal complex (CC) described in the database. The groups that not belongs to any CC described in the database are represented as ‘not determined’ (ND). Lines connecting ST groups, indicate that they differ in one allele (thick solid line), or three to five locus (thin and dotted lines).

Table 1.

Cronobacter strains (n = 465) isolated from the American continent and deposited in the PubMLST database (http://www.pubmlst.org/cronobacter/, last access 10/15/2020).

Continent (Number of strains)	Country (Number of strains)	Species (Number of strains)	Source (Number of strains)
North America (267)	United States of America (234)	C. sakazakii (193)	Clinical (55), environment1 (70), food (48), infant formula (11), unknown (8), water (1)
		C. dublinensis (15)	Environment (14), clinical (1)
		C. malonaticus (13)	Clinical (9), environment1 (3), food (1)
		C. turicensis (10)	Environment1 (6), clinical (2), food (2)
		C. muytjensii (3)	Environment1 (1), unknown (2)
	Canada (26)	C. sakazakii (24)	Clinical (11), unknown (5), environment (3), infant formula (3), food (2)
		C. malonaticus (2)	Clinical (1), unknown (1),
	Mexico (7)	C. sakazakii (7)	Infant formula (4), clinical (2), environment (1)
South America (198)	Brazil (196)	C. sakazakii (148)	Food (136), infant formula (6), clinical (3), environment2 (3)
		C. malonaticus (26)	Food (21), clinical (3), infant formula (1), environment (1)
		C. dublinensis (14)	Food (14)
		C. turicensis (6)	Food (6)
		C. muytjensii (2)	Food (2)
	Chile (1)	C. sakazakii (1)	Infant formula (1)
	Uruguay (1)	C. sakazakii (1)	Infant formula (1)

1

Including strains isolated from insects;

2

Including strains isolated from utensils.

The isolation of Cronobacter spp. from PIF purchased in American countries and reported in the literature is shown in Table 2. A total of seven countries reported the isolation of the pathogen in the period of 1996–2018. Unfortunately, the majority of isolates were not genotyped by MLST partially due to the technique not being developed until 2009. In many studies, the isolation occurred after 2008, the year when Codex Alimentarius approved the microbiological criteria for PIF that include the absence of Cronobacter species (CAC 2008). Similarly, FAO/WHO (2006) recommended the establishment of genotyping methods for Cronobacter spp. to facilitate tracing the organism.

Table 2.

Presence of Cronobacter in powdered infant formula commercialized in American countries 1996–2018.

Country	Year of isolation	Microorganism (Number of strains)	Source	Comments	Reference
Canada	1996	Cronobacter spp. (8)	PIF1	A total of 24 samples from five different companies were analysed (n = 120). Cronobacter was isolated from four companies.	Nazarowec-White and Farber (1997)
USA2	2001	Cronobacter spp. (not determined)	PIF and reconstituted PIF	Cronobacter spp. was isolated from both opened and unopened cans during the investigation of an outbreak.	CDC (2002)
Argentina	2005–2008	C. sakazakii (22) C. malonaticus (1)	PIF	The PIF samples were from three separate companies. Samples had been imported from two different Latin American countries and one from Europe.	Terragno et al. (2009)
Brazil	2006	C. sakazakii (6) C. malonaticus (1)	PIF and reconstituted PIF	A total of 99 samples were analyzed, of which 42 were unopened cans of powdered infant formula (PIF), 25 reconstituted infant formulas in feeding bottles, 27 utensils used in the preparation of infant formula and five samples of fortified cows milk	Siqueira et al. (2013)
Chile	2008	Cronobacter spp. (4)	PIF	A total of 80 samples were analyzed	Sáez, Llanos and Tamayo (2012)
	2013–2014	C. sakazakii (2)	PIF	A total of seven sample for premature neonates and 65 samples for neonates 0–6 months. 21 produced in Chile, 44 in México and seven in Holland. The two positive samples for C. sakazakii were produced in Chile.	Parra et al. (2015)
	2016–2017	C. sakazakii (8)	PIF	A total of 90 samples from four countries (United States, Singapore, Chile and Holland), from three manufacturers (1, 2 and 3), seven commercial dairy brands, of which three were powdered and four were liquid products. C. sakazakii was isolated from seven samples produced in Chile and one in Singapore, and all PIF.	Parra-Flores et al. (2018)
Mexico	2011	C. sakazakii (2)	PIF and reconstitute PIF	A total of 21 samples of PIF, 10 for premature infants and 11 for children aged 0–6 months; and 29 samples of reconstitute PIF, 11 for premature infants and 18 for infants aged from 0–6 months. C. sakazakii was isolated from one sample of PIF and one of reconstitute PIF.	Flores et al. (2011)
Honduras	2018	C. sakazakii (2)	PIF and reconstitute PIF	A total of 100 samples of imported PIF from five commercial brands, from six different countries. C. sakazakii was isolated from two samples of reconstitute PIF.	Márquez et al. (2019)

1

Powdered infant formula;

2

United States of America.

Some studies reported that contaminated PIFs had been imported from different countries within the Americas and also from Europe (Table 2). In South America, the increased commercialization of food was stimulated by the Southern Common Market (MERCOSUL 2020). Such international food trade present a challenge in order to avoid the spread of pathogens, as in the case of PIF, where the low water activity associated with the characteristics of the Cronobacter genus facilitates the survival of the pathogen in these products even though the microbiological criteria are very strict (Yan et al. 2015; Umeda et al. 2017).

In Argentina, Asato et al. (2013) reported two cases of neonatal infections associated with C. malonaticus and one case associated with C. sakazakii. A total of two of these were fatal, however the route of transmission was not identified. Similarly in Brazil, where cases of Cronobacter spp. infections in neonates were reported but neither PIF nor other types of foods were identified as the vehicle of contamination (Brandao et al. 2015; Chaves et al. 2018). Further studies in Brazil have reported an occurrence of Cronobacter spp. in 44.4% of corn-based farinaceous foods samples (Costa et al. 2020a) and 56.7% of oat and linseed samples (Silva et al. 2019). In Cuba, Leyva et al. (2008) isolated Cronobacter spp. from one imported skimmed powdered milk sample. In Colombia, Vanegas, Rugeles and Martinez (2009) analyzed 222 samples from milk feeders including sterile and non-sterile surfaces, utensils used for the formula preparation and food handlers and Cronobacter spp. was identified in eight samples. Morato-Rodríguez et al. (2018) analyzed 102 samples of breast milk substitutes made from corn and plantain starch and identified 27 samples contaminated by Cronobacter spp. The presence of this organism in these foods and their potential risk to neonates could be a significant public health problem in America, especially in developing countries, where a large proportion of neonates are fed breast milk substitutes (Morato-Rodríguez et al. 2018; Silva et al. 2019; Costa et al. 2020a).

A total of five Cronobacter species were reported: C. sakazakii (374, 80.4%), C. malonaticus (41, 8.8%), C. dublinensis (29, 6.2%), C. turicensis (16, 3.5%) and C. muytjensii (5, 1.1%). The diversity of Cronobacter species observed was similar to the data from the PubMLST database for other countries; highlighting the domination of Group 1 species (C. sakazakii and C. malonaticus) (Forsythe 2018).

A total of 452 strains (97.2%) had been identified based on fusA allele sequencing and the calculated SI was 0.84. Cronobacter spp. genetic diversity using neighbor-joining tree analysis indicating the species, source and country of origin is presented in Fig. 2. The 345 strains were assigned to 98 STs, a ratio of 3.5 strain per ST and the calculated SI was 0.93. C. sakazakii strains were found in almost all countries and represented 64 out of a total of 98 STs (65.3%). The SI is used to measure the genetic diversity among isolates, since it calculates the probability of unrelated strains being classified into different groups (Hunter and Gaston 1988). The acceptable level of discrimination will depend on a number of factors, but an index of greater than 0.90 would seem to be desirable for a typing scheme (Hunter and Gaston 1988). Consequently, the MLST technique was considered efficient for typing strains isolated within America. This diversity was also observed using the eBURST algorithm. A total of 30 STs formed 12 single-locus variant (SLV) and/or double-locus variant (DLV) groups which shared five or more allelic profiles, and nine CCs had already been defined in the database (Table 3; Fig. 1). A total of 68 STs were identified as singletons and belonged to C. sakazakii (n = 100), C. malonaticus (n = 24), C. dublinensis (n = 18), C. turicensis (n = 9) and C. muytjensii (n = 5).

Figure 2.

Neighbor-joining phylogenetic tree based on the seven MLST loci (3036 base pair concatenated length) of the 345 Cronobacter strains isolated from the American continent and deposited in the PubMLST database (http://www.pubmlst.org/cronobacter/, last access 10/15/2020). This tree was generated using iTOL v3. The first inner circle shows Cronobacter species, the second circle shows the country from which strain was isolated; the outer circle refers to the source of the isolation from each strain.

Table 3.

Groups resulting among the 75 STs of the Cronobacter strains (n = 345) isolated from the American continent and deposited in the PubMLST database (http://www.pubmlst.org/cronobacter/, last access 10/15/2020) following eBURST analysis.

Group (CC1)	Species	ST2	Number of isolates	Source (Number of isolates)	Geographic distribution	Period of isolation
1 (CC4)	C. sakazakii	4	64	Clinical (30), environmental3 (13), infant formula (9), food (9), unknow (3)	Brazil, Canada, USA	1973–2014
		15	1	Clinical (1)	Canada	1990
		107	1	Clinical (1)	USA	2011
		108	1	Infant formula (1)	USA	2011
		121	1	Food (1)	USA	2003
		218	1	Infant formula (1)	USA	Unknown
2 (CC1)	C. sakazakii	1	59	Environmental (25), food (19), clinical (10), unknown (4), infant formula (1)	Brazil, USA, Canada, Mexico	1988–2019
		391	1	Unknown (1)	Canada	Unknown
		693	1	Food (1)	Brazil	2018
3 (CC8)	C. sakazakii	8	18	Clinical (12), infant formula (2), environmental³ (4)	USA, Canada, Uruguay	1977–2012
		226	2	Clinical (1), food (1)	Brazil, Mexico	2014–2016
4 (CC40)	C. sakazakii	40	10	Food (7), environmental (3)	Brazil, USA	2013–2018
		128	1	Environmental (1)	USA	Unknown
		395	1	Food (1)	Brazil	2014
5 (CC3)	C. sakazakii	3	5	Food (5)	Brazil, USA	2014–2018
		396	1	Food (1)	Brazil	2009
6 (CC83)	C. sakazakii	65	1	Infant formula (1)	USA	1988
		83	5	Environmental (3), food (2)	Brazil, USA	2006–2014
7 (CC23)	C. sakazakii	23	3	Food (3)	Brazil, USA	2005–2013
		687	1	Food (1)	Brazil	2018
8 (CC24)	C. turicensis	24	2	Environmental (2)	USA	2014
		387	1	Environmental (1)	USA	2014
9 (CC113)	C. sakazakii	56	1	Infant formula (1)	Brazil	2007
		113	1	Infant formula (1)	Brazil	2007
10 (ND4)	C. sakazakii	222	1	Food (1)	Brazil	2018
		402	1	Food (1)	Brazil	2014
11 (ND)	C. malonaticus	60	1	Environmental3 (1)	USA	2012
		394	1	Clinical (1)	Brazil	2013
12 (ND)	C. turicensis	738	1	Food (1)	Brazil	2016
		740	1	Food (1)	Brazil	2016

1

Clonal complex;

2

Sequence type;

3

Including strains isolated from insects and/or utensils;

4

Not determined.

Group 1 (n = 69) comprised C. sakazakii of six STs from CC4; 4, 15, 107, 108, 121 and 218. ST4 is a dominant ST in the PubMLST database, consisting of 64 strains isolated from three countries over a period of more than 40 years and the majority of strains was isolated from clinical specimens and infant formula (Joseph et al. 2012b; Siqueira et al. 2013). STs 15, 107, 108, 121 are SLVs of ST4, differ in the fusA allele, whereas ST 218 is another SLV but differs in the gltB allele. A high number of strains in the PubMLST database have been reported as CC4, and it has been recognized as a stable clonal lineage which is particularly associated with neonatal meningitis (Joseph and Forsythe 2011; Joseph et al. 2012b; Joseph and Forsythe 2012; Hariri, Joseph and Forsythe 2013). ST15 was isolated in 1990 from a hospitalized child in Canada (Baldwin et al. 2009; Joseph and Forsythe 2012). ST107 was isolated from cerebrospinal fluid in a brain abscess of a term infant in USA (Joseph et al. 2012b; Joseph and Forsythe 2012). ST108 was isolated from an infant formula in USA and had previously been isolated from PIF in Germany. ST121 and ST218 had been isolated in USA from soy protein and from an infant formula, respectively (Ivy et al. 2013).

Group 2 (n = 61) comprised C. sakazakii STs 1, 391 and 693. These belong to CC1. ST1 was a dominant ST (59/61) isolated from four countries over a period of more than 30 years, including strains from clinical cases and infant formula (Siqueira et al. 2013; Pan et al. 2014). ST1 is the central ST and has been associated with infections in infants and adults in many countries (Joseph et al. 2012b; Shi et al. 2018; Li et al. 2020). ST391 and ST693 are DLV and SLV of ST1, respectively. They had been isolated from an unknown source in Canada and from a flake corn flour in Brazil, respectively (Paula et al. 2020).

Group 3 (n = 20) comprises C. sakazakii STs 8 and 226 that belong to CC8, and were isolated from clinical (n = 13), environmental (n = 4) infant formula (n = 2) and food (n = 1) samples. ST8 is associated with neonatal infections in neonates over a long time period (1977–2019) across the world, including four continents (North America, South America, Europe and Asia). The most recent cases were reported in 2019 in China (Kadlicekova et al. 2018). ST8 has been isolated from PIF in Uruguay and USA (Forsythe 2015), and from insects (Musca domestica) in USA, indicating that these flies can act as reservoir of this pathogen (Pava-Ripoll et al. 2012). ST226 is a SLV from ST8 and was associated with clinical infection in Mexico and also isolated from food samples in China, Turkey, South Korea, Czechoslovakia and Brazil but not in PIF (Xu et al. 2015; Vojkovska et al. 2016; Brandao et al. 2017; Ling et al. 2018; Li et al. 2019).

Group 4 (n = 12) comprises C. sakazakii STs 40, 128 and 395 that belong to CC40 and were isolated from food and environmental samples. ST40 have been isolate from samples of human clinical samples (feces, urine and cervix) in Europe and Asia (Lepuschitz et al. 2019) and foods (Gičová et al. 2014; Fei et al. 2017; Li et al. 2019). STs 128 and 395, both SLV of ST40, were isolated from environmental in USA (Ivy et al. 2013) and from flaxseed flour in Brazil (Brandao et al. 2017), respectively.

Group 5 (n = 6) comprised of C. sakazakii STs 3 and 396. These belong to CC3 and were food isolates. ST396 is a DLV of ST3 and had been isolated from a milk powder in Brazil (Brandao et al. 2017). ST3 has been isolated from enteral feeding tube of neonates in a neonatal intensive care unit survey in United Kingdom (Hurrell et al. 2009) and also from foods and environmental samples in China (Xu et al. 2015; Ling et al. 2018; Hu et al. 2019; Li et al. 2019).

Group 6 (n = 6) comprises C. sakazakii STs 65 and 83 that belong to CC83. ST83 strains were isolated from environmental and food samples in Brazil and USA (Brandao et al. 2017). However, this ST have been associated bacteremia cases in Israel and China (Block et al. 2002). ST83 is recognized as a persistent contaminant in PIF manufacturing facilities (Chase et al. 2017) and have been isolated from infant formula samples in many countries (Gičová et al. 2014; Fei et al. 2017). ST 65 is a SLV from ST83 (differ in infB allele) and was isolated from an infant formula in 1988 in USA.

The Groups 7–12 have only few strains each one (n ≤ 4) but Group 11 comprise relevant STs associated with severe clinical infections in neonates. Groups 7–10 and 12 comprises STs that were isolated from foods and environmental samples but not associated with human infections (Joseph and Forsythe 2012; Siqueira et al. 2013; Gičová et al. 2014; Pan et al. 2014; Ling et al. 2018; Silva et al. 2019; Paula et al. 2020).

Group 11 (n = 2) comprises the SLV C. malonaticus STs 60 and 394. ST60 was first reported in clinical cases in Czechoslovakia in 1983. In 2014, a ST60 strain (Chcon-9, ID 1494) was isolated from cerebral spinal fluid of an infant born in China, who had been fed fortified breast milk before developing clinical symptoms of meningitis (Ogrodzki and Forsythe 2017). Ogrodzki and Forsythe (2015) reported that ST60 had the same capsular profile (K2:CA2:Cell+) as C. sakazakii CC4 strains which are associated with meningitis cases. ST60 have been isolated from foods in Europe and Asia (Fei et al. 2018; Ling et al. 2018; Li et al. 2019), including follow up formula and PIF (Joseph and Forsythe 2012). In America, ST60 was isolated from an insect (Musca domestica) in USA highlighting the potential risk of these flies in the contamination of food chain with this pathovar (Pava-Ripoll et al. 2012). The ST394 is a SLV of ST60 and was isolated from hemoculture of a neonate with bacteremia in Brazil (Brandao et al. 2015; Umeda et al. 2017). A further phenotypical characterization of the strain showed the ability to producing of biofilm in glass tubes, degradation of casein in milk agar and β-hemolysis against erythrocytes from guinea, pig, horse and rabbit (Umeda et al. 2017). However, no study evaluating if ST394 strains also encode for the K2:CA2:Cell + capsular profile has been undertaken, which will be important to predict if ST394 could also presented the same virulence potential as ST60.

Regarding the singletons, from the 68 STs, 26 (38.2%) were associated with clinical cases, and 14 (20.6%) of them occurred in America and are discussed below. The majority was found in C. sakazakii (n = 7) and C. malonaticus (n = 5) strains.

C. sakazakii ST12 was isolated from a clinical sample in 1975 in USA, from clinical cases in former Czechoslovakia (Baldwin et al. 2009), from a fatal outbreak in a neonatal intensive care unit in France and the most recent was from a hemoculture of 65-years-old patient in Ireland (Masood et al. 2015). C. sakazakii ST41 (CC281) was isolated from a foot wound sample in 1975 in USA (Joseph et al. 2012b) and no other clinical reported was found. C. sakazakii ST50 were isolated from a sputum or spinal fluid sample in 1977 (Joseph et al. 2012b) and septum in 2009, both in USA. No other clinical case was reported, and the ST were found in foods from China and Czechia, including infant formula (Killer et al. 2015; Fei et al. 2017; Li et al. 2017, 2019; Ling et al. 2018). C. sakazakii ST64 (CC64) was isolated from a bronchial sample in 1973 in USA (Joseph et al. 2012b). Later it was isolated from clinical samples in Czechia and Slovakia, in 2013 and 2016, respectively (PubMLST, 2020). Regarding food samples, it has been isolated from in many countries (including Brazil) in different type of foods that include PIF and FUF (Pan et al. 2014; Xu et al. 2015; Brandao et al. 2017; Li et al. 2019; Costa et al. 2020a). C. sakazakii ST110 (CC4), unique in the database, was isolated from a CSF of a neonate in 2011 in USA and is a triple locus variant (TLV) of ST4 (Joseph et al. 2012b). C. sakazakii ST233 was isolated from a faecal human sample in USA. After, this ST was found in environmental and food samples in Slovenia, Czechoslovakia, China (Ling et al. 2018; PubMLST, 2020), USA and Brazil (Brandao et al. 2017). C. sakazakii ST494 has previously been isolated from clinical samples related to a fatal case of meningitis in a newborn in Brazil in 2017 (Chaves et al. 2018). Costa et al. (2020b) reported that this ST can produce cytotoxic compounds that induced several cell death characteristics, including loss of cell-cell contact, microvilli reduction and cellular lysis. This ST was also isolated from spices and edible mushroom in China (Li et al. 2019) and in 2018 from a corn meal sample in Brazil (Costa et al. 202a).

C. malonaticus ST7 (CC7) strains were isolated in clinical cases in 1973 and 1977 in USA (Baldwin et al. 2009), from PIF samples in Brazil (Siqueira et al.2013), as well as insects and weaning foods in USA (Baldwin et al. 2009; Pava-Ripoll et al. 2012). In other countries, ST7 strains were isolated from clinical, environmental and foods (Pan et al. 2014; Xu et al. 2015; Ling et al. 2018; Hu et al. 2019; Li et al. 2019), and from clinical samples from adults in Europe (Alsonosi et al. 2015; Kadlicekova et al. 2018). C. malonaticus ST10 (CC63) has only one strain isolated from a clinical sample in Canada and the others four strains were isolated from foods in China and Czechoslovakia (Baldwin et al. 2009; Li et al. 2019). C. malonaticus ST129 (CC129) strains were isolated from a blood culture in USA in 1977 (Joseph et al. 2012b; Ivy et al. 2013). Afterwards, this ST was found in environmental samples from PIF production facilities in Ireland (Yan et al. 2015) and from foods in China and Brazil (Li et al. 2017; Costa et al. 2020a). C. malonaticus ST307 (CC112) was isolated from blood culture samples of a <1 month infant preterm that died due to fatal meningitis in USA in 2011 (Hariri, Joseph and Forsythe 2013). C. malonaticus ST440 (CC611) strains have already been isolated from two cases of infections in neonates in Brazil in 2013 (Brandao et al. 2015; Umeda et al. 2017). ST440 was found in milk and edible mushrooms in China (Li et al. 2019) and natural mineral drinking water in Brazil (Vasconcellos et al. 2019). In 2020, one strain isolated from CSF of a neonate not premature was reported in Japan, but no more information is available in the PubMLST database (PubMLST 2020).

Regarding the remaining species, C. dublinensis ST80 (CC80) was isolated from an abscess in 1979 in USA (Joseph et al. 2012b). It was also found in water, foods and infant formula samples in Europe (Grim et al. 2013; Gičová et al. 2014; Vojkovska et al. 2016). C. turicensis ST5 is the main ST of the species isolated from clinical infections. The strains were isolated in the USA from blood in 1970 and from bone marrow in 1975 in the USA. Furthermore, it was isolated from UK milk powder in 2004 (Joseph et al. 2012b; Czerwicka et al. 2013).

Cronobacter spp. strains isolated on the American continent showed a high diversity when characterized by MLST and therefore a powerful tool for epidemiological investigations. In the majority of reported Cronobacter spp. clinical cases, the source of contamination could not be determined. This could be due to many reasons such as inadequate investigation, lack of appropriate analytical methodologies, or insufficient sample collection. This reinforces the need for a more effective risk communication strategy in relation to the risk to neonates from Cronobacter spp. The notification of Cronobacter spp. infections in American countries should be encouraged in order to understand the real incidence in the continent. The sending of strains to reference laboratories to perform molecular characterization is also of great importance so that investigations and routes of contamination of pathogenic strains can be defined and control measures can be better planned and implemented.

ACKNOWLEDGEMENTS

We thank CNPQ for financial support to P. Vasconcellos (PIBIC/Fiocruz). This publication made use of the Cronobacter multilocus sequence typing database site (http://pubmlst.org/cronobacter/), hosted by the University of Oxford. The authors thank the many contributors to the Cronobacter PubMLST database (http://www.pubMLST.org/cronobacter).

Conflicts of Interest

None declared.

Author notes

Authors contribution: Paula Vasconcelos Costa– conception and design or the acquisition and analysis of data, drafting the manuscript and approval of the final submitted version.

Luiza Vasconcellos– conception and design or the acquisition and analysis of data, drafting the manuscript and approval of the final submitted version.

Stephen James Forsythe– conception and design or the acquisition and analysis of data, drafting the manuscript and approval of the final submitted version.

Marcelo Luiz Lima Brandão– conception and design or the acquisition and analysis of data, drafting the manuscript and approval of the final submitted version.