Whole genome sequence analysis of Salmonella Typhi provides evidence of phylogenetic linkage between cases of typhoid fever in Santiago, Chile in the 1980s and 2010–2016

Typhoid fever epidemiology was investigated rigorously in Santiago, Chile during the 1980s, when Salmonella enterica serovar Typhi (S. Typhi) caused seasonal, hyperendemic disease. Targeted interventions reduced the annual typhoid incidence rates from 128–220 cases/105 population occurring between 1977–1984 to <8 cases/105 from 1992 onwards. As such, Santiago represents a contemporary example of the epidemiologic transition of an industrialized city from amplified hyperendemic typhoid fever to a period when typhoid is no longer endemic. We used whole genome sequencing (WGS) and phylogenetic analysis to compare the genotypes of S. Typhi cultured from acute cases of typhoid fever occurring in Santiago during the hyperendemic period of the 1980s (n = 74) versus the nonendemic 2010s (n = 80) when typhoid fever was rare. The genotype distribution between “historical” (1980s) isolates and “modern” (2011–2016) isolates was similar, with genotypes 3.5 and 2 comprising the majority of isolations, and 73/80 (91.3%) of modern isolates matching a genotype detected in the 1980s. Additionally, phylogenomically ‘ancient’ genotypes 1.1 and 1.2.1, uncommon in the global collections, were also detected in both eras, with a notable rise amongst the modern isolates. Thus, genotypes of S. Typhi causing acute illness in the modern nonendemic era match the genotypes circulating during the hyperendemic 1980s. The persistence of historical genotypes may be explained by chronic typhoid carriers originally infected during or before the 1980s.

The persistence of historical genotypes may be explained by chronic typhoid carriers originally infected during or before the 1980s.

Author summary
Studies of Salmonella Typhi (the cause of typhoid fever) rarely include isolates collected both before and after the interruption of hyperendemic transmission because this typically occurred decades before modern bacteria preservation methods. After substantial reduction in disease, it was assumed that sporadic cases and infrequent outbreaks were due to either chronic biliary carriers or importations, but this was difficult to characterize with low resolution bacterial typing methods. In Santiago, Chile, typhoid fever persisted at hyperendemic levels through the 1980s until organized control efforts in the 1980s and changes to wastewater policy in 1991 caused annual typhoid incidence to plummet.

Introduction
Typhoid fever is a systemic infection, caused by the bacterium Salmonella enterica subspecies enterica serovar Typhi (S. Typhi), which is spread through ingestion of food and water contaminated with human faeces containing S. Typhi. Typhoid fever in Santiago, Chile rose to hyperendemic levels in the mid-1970s with a striking annual incidence of 177-220 cases per 10 5 population per year from 1977-1983, partially declined during the years 1984-1991 (51-128 cases/10 5 per year), and finally plummeted below 8 cases/10 5 per year after 1991, reaching nonendemic levels of~0.5/10 5 in 2012 [1]. During the hyperendemic years, the highest incidence rate was observed in school age children, and case numbers peaked in the rainless summer months [2]. To address this public health issue, in 1979 the Ministry of Health of Chile implemented a Typhoid Fever Control Program involving strategic preventive interventions including large-scale field trials of live oral typhoid vaccine Ty21a in >500,000 schoolchildren [3][4][5][6], and identification and treatment of chronic carriers among food handlers [7]. Collectively, these interventions led to a steady decline in typhoid cases between 1983 and 1990 [3][4][5][6]8].
Epidemiologic and environmental bacteriology studies in the 1980s undertaken by the Chilean Typhoid Fever Control Program also implicated the use of raw untreated sewage water for irrigation of vegetable crops during the rainless summer months as a key risk factor responsible for amplified transmission of S. Typhi in Santiago [8,9]. In April 1991, the government of Chile abruptly prohibited the practice of irrigating crops with untreated sewage wastewater in Santiago [1,8,10,11]. This action was taken following the explosive epidemics of El Tor cholera in 1991 that occurred in neighbouring countries to the North, including in Peru, beginning in late January [12,13], followed by Ecuador in February [13], and Colombia in March [13]. The pivotal stimulus for the intervention was a small outbreak of cholera in Santiago, Chile, in April, 1991 [1,8,10,11]. Following imposition and strict enforcement of the prohibition of use of untreated sewage water for irrigation, the Santiago cholera epidemic quickly came to an end [10,11], and the incidence of typhoid dropped precipitously in subsequent years [1,8,10]. From 2000-2014, the annual incidence of confirmed typhoid fever in Santiago has ranged from 0.20-1.18 cases/10 5 population. The current rare sporadic cases of typhoid in Santiago include illnesses among: travellers who visit countries with high incidence rates of typhoid; immigrants from typhoid-endemic countries; and transmission from Chilean chronic carriers who were originally infected decades earlier (pre-1991) during the era of hyperendemic transmission.
Phylogenetic analysis of whole genome sequencing (WGS) data has become the gold standard for inferring relatedness between S. Typhi isolates [14]. WGS can also be used to assign genotypes, identify and characterize plasmids, compare gene content and homology, and identify molecular determinants of antibiotic resistance and virulence [15,16]. Whereas WGS data from many global isolates are available, few bacterial collections include isolates from both before and after the interruption of amplified hyperendemic typhoid fever. Also, WGS-derived information concerning the genotypes isolated from typhoid fever cases in South American countries is scarce. The majority of publicly available sequenced isolates come from hyperendemic regions in South Asia and East Africa where H58 (4.3.1) genotypes dominate [17].
Herein we describe a genomic analysis of S. Typhi isolates from Santiago, Chile from two distinct epidemiologic periods: the early 1980s (N = 74), when typhoid was hyperendemic, and from 2010-2016 (N = 80), a quarter century after the interruption of amplified transmission of S. Typhi in Santiago. We show that the genotypic composition of the population structure is similar between these two periods, even though the local epidemiology has changed drastically. In addition, we describe a potential carrier-related cluster of the "ancient" 1.1 clade, and the identification of a new variant of plasmid pHCM2 carried by 2 isolates of another ancient genotype (1.2.1).

Study design
In total, 74 "historical" isolates of S. Typhi from cases of acute typhoid fever in Santiago, Chile in the period 1981 to 1986, during the era of hyperendemic typhoid, were obtained from the culture collection of the Center for Vaccine Development and Global Health of the University of Maryland, Baltimore. "Modern" S. Typhi isolates (n = 80) were strains collected from acute cases of typhoid fever and referred to the Instituto de Salud Pública de Chile (ISP, Institute of Public Health) by clinical microbiology laboratories from 2011-2016; these referred strains were generally accompanied by only limited demographic and clinical information.

Phylogenetic analysis
Previously defined [21-23] recombinant regions such as prophages and plasmids were excluded manually, and any remaining recombinant regions were filtered using Gubbins (v2. 5

S. Typhi genotypes in Santiago, Chile
Genotypic analysis indicated that most of the isolates from Santiago in this study belonged to genotypes 2 (n = 43, 27.9%), 3.5 (n = 51, 33.1%) and 1.1 (n = 22, 14.3%) ( Table 1). While these  The genomic population structure of S. Typhi in Santiago appeared highly consistent between the hyperendemic and post-hyperendemic eras (Fig 1). Both rare "ancient" genotypes 1.1 and 1.2.1, as well as more commonly found genotypes 2 and 3.5, were frequently present across both study periods. Despite several decades having passed since the cessation of the era of hyperendemic transmission, most of the post-2010 sporadic S. Typhi isolates shared bootstrap-supported phylogenetic clusters with the 1980 hyperendemic isolates. Only within genotype 1.1 did recent and historical sequences form independent clusters. The intermingling of historical and recent sequences suggests that the 2010s isolates originated from progenitors within the same pool of S. Typhi circulating in the 1980s.
Amongst the 151 non-H58 isolates, few antimicrobial resistance mechanisms were detected. One genotype 2.0.0 isolate carried a tetB gene, known to encode a mechanism of resistance to tetracyclines (albeit not an antibiotic of clinical relevance for treatment of typhoid fever). Two genotype 2.2 isolates and one genotype 4.1 isolate had single non-synonymous point mutations in the Quinolone Resistance Determining Region (QRDR) of genes gyrA and gyrB, associated with reduced susceptibility to fluoroquinolones [36]; in the late 1980s, ciprofloxacin was the antibiotic of choice for treating patients with typhoid fever. Only 17 candidate plasmids were detected across all the isolates, including a Colicin plasmid, three related to plasmid pHCM2 and 13 of incompatibility type IncX1 (Fig 1).

Global transmission dynamics
To better understand how this local population structure fits into the global context of S. Typhi, a maximum likelihood phylogenetic tree was generated using a total of 2,013 non-H58 S. Typhi genomes from approximately 70 countries (Fig 2). The S. Typhi isolates from Santiago of genotypes 3.5, 2, 1.1 and 1.2.1 clustered together and were genetically distinct from those sequenced previously from other countries. The other genotypes present in Santiago were mostly found on long branches, indicating significant differentiation from their closest relatives in the global context phylogeny.

"Ancient" S. Typhi clades endemic to Chile
Two of the clades common in our dataset, 1.1 and 1.2, are found closer to the root of the global S. Typhi tree and thus can be considered "ancient" clades, which are detected with very low global frequency. Of the 1.1 isolates (n = 22), two were from the 1980s collection (Fig 3). Only six other isolates, 0.3% of the global collection described by Wong et al [14], belonged to clade 1.1. Interestingly, a cluster of closely related genotype 1.1 S. Typhi isolates was detected in the years 2011 (1 isolate) and 2012 (19 isolates). These isolates also exhibit lower genetic variation (median SNP distance of 8) than those observed among the two 1980s 1.1 isolates (119 SNPs). The smallest SNP distance between a 1980s 1.1 isolate and a 2010s 1.1 isolate was 79 SNPs. This suggests the 2011-2012 isolates have a recent common ancestor.
We sought to determine if there was a specific virulence gene associated with the 2011-2012 genotype 1.1 cluster, using SRST2 with the Salmonella VFDB. Although there were no additional or missing virulence factor-associated genes detected, this screening did reveal that all of the genotype 1.1 isolates from 2011-2012 cluster harboured a SNP in the ssrA gene. The alignment of this gene revealed the same missense mutation (G to A at position 1648968 of CT18, causing a glycine to glutamic acid amino acid change) in the ssrA gene (ssrA-G639E), (Fig 3). The other ancient clade present amongst the Santiago isolates was genotype 1.2.1. This was the only clade in our datasets (except from one recently imported H58 isolate) in which isolates harboured pHCM2 plasmids. pHCM2 plasmids are found in~15% of S. Typhi and are

PLOS NEGLECTED TROPICAL DISEASES
Lingering Salmonella Typhi genotypes in Santiago, Chile generally highly conserved (Fig 4A). To our knowledge, based on the data available in the S. Typhi public collection on Pathogenwatch: (https://pathogen.watch/collection/1y3cdsjq55hfpublic-genomes-accessed Dec 2021), pHCM2 plasmids have not been detected in ancient clades 0.0 to 1.2 previously. Two pHCM2 plasmids were found among our isolates from genotype 1.2.1; one in a 1984 isolate (ERR2710544) and the other from 2013 (ERR4289185). These plasmids were highly similar to each other, but significantly different from the pHCM2 found on S. Typhi reference CT18 (accession no: AL513384), and from another novel pHCM2 found in a 1973 isolate from Thailand [37]. The pHCM2 sequences described here differed by >2000 SNPs (98.4% nucleotide identity, 89% coverage) from the reference (Fig 4).

Discussion
This study utilizes WGS analyses to compare S. Typhi isolates from patients who developed acute typhoid fever in Santiago during the modern era (i.e., 2011-2016, decades post-interruption of amplified typhoid transmission in 1991) with isolates from acute cases during the hyper-endemic era of the 1980s. Using phylogenetics and genome-wide screening, we compared the genotypes, AMR determinants and plasmid content of S. Typhi isolates between these two time periods. Whilst comparisons of isolates from the 1980's with isolates from 1990 (post vaccinating interventions-but pre-irrigation ban, when incidences were still high 47 cases /10 5 [1]) have been undertaken previously with pulse field gel electrophoresis (PFGE) [38], to our knowledge, this is the first WGS-based analysis comparing S. Typhi collected during a period of hyper-endemicity and decades after typhoid control.
The observation that the 1980s and 2010s collections of S. Typhi isolates do not cluster independently within our commonly found genotypes (2, 3.5 and 1.2.1) suggests that many of the sporadic cases detected in the 2010s share common ancestry with S. Typhi cases from the 1980s. These may, in turn, be linked to shedding of S. Typhi from chronic typhoid carriers. This is expected decades after hyper-endemic typhoid is controlled, as chronic carriers will continue to shed sporadically until they are treated or die out of the population [39].
Although there have been several detected introductions of the globally dominant H58 S. Typhi clade (4.3.1) into Santiago, as occurs with most industrialized countries in North America, Europe, and Australia, this was not a frequent genotype found in the sporadic modern Chilean cases. This is distinct from countries such as the United Kingdom where imported

PLOS NEGLECTED TROPICAL DISEASES
Lingering Salmonella Typhi genotypes in Santiago, Chile cases of H58 S. Typhi have become common among travel-associated typhoid cases in recent years [40,41].
S. Typhi WGS data from a recent study in Colombia found genotypes 2, 2.5 and 3.5 to be the local endemic genotypes [42]. Genotype 2.5 was not detected among the Santiago isolates; however, genotype 3.5 and 2 were detected in Santiago, and were the nearest genomic neighbours of the Colombian isolates on the global tree. Even though the sequenced isolates from Colombia and Santiago, Chile both contained isolates from the 2010s and historical isolates, they clustered separately on the global tree, suggesting that there has not been obvious transmission between the two countries. More representative WGS data from other Central and South American countries are needed to determine if this is the case with other unsampled potentially endemic countries in the area.
One of the more interesting findings from this dataset was the cluster of 20 highly related strains of genotype 1.1 in 2011-2012. These isolates differed by a median of 8 SNPs in the core genome. For context, the average core genome mutation rate in S. Typhi is approximately 1 substitution/genome/year every year [21, [43][44][45]. As such, it is unlikely that these isolates originated from a clonal source and sequentially acquired this genetic variation in <2 years. Conversely, these strains are clearly highly related and were all collected during a short time span. A potential explanation to account for this cluster is that it was an outbreak caused by bacteria shed from a long-term carrier. S. Typhi colonizing the gallbladder or other long-term physiologic niche of a chronic typhoid carrier may evolve and mutate over time into a heterogeneous, yet closely related, community of bacteria [46]. Thus, the closely related nature of these isolates occurring mainly over a single year and during a period when typhoid was no longer hyperendemic in Santiago could be explained by an outbreak instigated by a chronic carrier who was shedding multiple highly related strains. If this hypothetical carrier was working as a food handler, intermittently contaminating food or water consumed by subjects during 2011-12, this could result in the cluster of closely related isolates we observe. The cessation of cases could then be explained by a change of employment or residence, a cholecystectomy procedure, a course of antibiotics that concentrate in the gallbladder (e.g., ciprofloxacin) prescribed for another infection or surgery, or death of the hypothetical carrier at the end of 2012.
Notably, the only detectable difference in virulence factor-associated genes from this potential carrier-related 2011-2012 cluster is the ssrA missense mutation. The ssrA/B two component response regulator is a principle regulator of Salmonella pathogenicity island-2 (SPI-2) [47,48]. SPI-2 has been shown to be insignificant in S. Typhi macrophage survival capability and S. Typhi's infection efficiency of humanized mice [49,50]. The ssrA gene has been linked to biofilm formation in Salmonella and persistence on gallstones [51,52]. This observation may be worth additional investigation to investigate the role of SPI-2 and ssrA in Typhoid carriage.
Our analysis identified a novel variant of cryptic plasmid pHCM2 in two sequences from genotype 1.2.1. A pHCM2 type plasmid was present in the S. Typhi reference isolate CT18 from Vietnam [19] and the plasmid family was described in some detail [53]. Interestingly, any role for pHCM2 in S. Typhi biology remains elusive. We know that almost all of the pHCM2 plasmids in the global S. Typhi collection are highly conserved in terms of overall nucleotide diversity. However, the Chilean pHCM2 from clade 1.2.1 is substantially different from most other family members. The only other reported pHCM2 plasmid with a similar level of variation is harboured in a S. Typhi Thailand isolate from 1973 [37].
The main limitation of this study is the limited sample size of S. Typhi genomes from both the historical hyperendemic and modern nonendemic periods in Santiago, Chile. Although not all isolates were stored during the hyperendemic period more than 40 years ago, we incorporated all available stored isolates from acute cases. For the modern time period we used a similar number of available isolates exclusively from acute cases to represent circulating infection. Among the approximately 7 million inhabitants in Santiago, Chile, few blood culture isolates from acute cases of typhoid fever were recorded in the 2010s, (<0.5/10 5 cases per year) and not all isolates are sent to the ISP; thus, only a limited number of S. Typhi isolates were available for sequencing. Nevertheless, the study design, based on S. Typhi causing acute disease between epidemiologically-distinct eras, lends support to our primary finding: the same genotypes circulating in the hyperendemic 1980s are found circulating in the modern 2010s period.
There are especially limited genomic data available for S. Typhi circulating in neighbouring countries, presenting another limitation of this study. Nevertheless, efforts to foster sequencing infrastructure and data generated in South America through the PAHO regional networks have accelerated during the Covid-19 pandemic. It is expected that whole genome sequencing of other pathogens, like S. Typhi, will follow in this region.
In conclusion, this genomics analysis suggests that asymptomatic typhoid carriers (or other unknown reservoirs) are likely responsible for continuing sporadic typhoid cases in Chile. The ancient clades present in the country have persisted. We recommend that all sporadic typhoid cases and clusters should be evaluated using epidemiological outbreak investigations and WGS as routine surveillance. This will help ensure the detection of any potential introduction or development of drug-resistant S. Typhi and may lead to the identification and treatment of carriers, which would further limit ongoing disease transmission.
Supporting information S1