Clinical characteristics of Elizabethkingia species colonization or infection
From January 2019 to December 2021, a total of different 71 strains of the non-repetitive Elizabethkingia strains were isolated and identified in our hospital. As shown in Table 1, the specimens were mainly isolated from the respiratory tract, including sputum (49/71, 69.0%), tracheal secretory fluid (10/71, 14.1%), and bronchoalveolar lavage fluid (2/71, 2.8%). There were only few samples of sterile body fluids, including cerebrospinal fluid (CSF) (3/71, 4.3%), drainage fluid (2/71, 2.8%), blood (1/71, 1.4%), etc. The specimens were mainly isolated from the department of ICU (26/71, 36.6%) and emergency (23/71, 32.4%), and less from the neurosurgery (7/71, 9.9%). The 71 patients included 46 males and 25 females. The average age was 56.9 years (range, 0–92 years), 69.0% of the patients were over 50 years. Most of the patients were also found to be suffering from the various underlying diseases, such as pulmonary infection, hypertension, diabetes, etc. 85.9% (61/71) patients were hospitalized for more than two weeks. In addition, most patients had been subjected to invasive treatment, such as mechanical ventilation (51/71, 171.8%), catheter insertion (42/71, 59.2%) and nasogastric tube (45/71, 63.4%). All patients had a history of antibiotic exposure, whereas sixty-seven patients (94.4%) had used broad-spectrum antibiotics for more than one week.
Table 1
Characteristics of 71 patients with EME colonization or infection
Characteristics | Value |
Age(years) | |
Rang | 0ཞ92 |
Mean ± SD | 56.9 ± 20.7 |
Gender, n (%) | |
Male | 46(64.8) |
Female | 25(35.2) |
Hospitalization duration(days), mean ± SD | 49.1 ± 40.7 |
Comorbidity, n(%) | |
Hypertension | 21(29.6) |
Diabetes mellitus | 13(18.3) |
Chronic obstructive pulmonary disease | 9(12.7) |
Cardiovascular disease | 23(32.4) |
End-stage renel disease | 14(19.7) |
Mechanical ventilation, n(%) | 51(71.8) |
Indwelling device, n(%) | 50(70.4) |
Nasogastric tube | 45(63.4) |
Urinary catheter | 42(59.2) |
Surgical puncture or drain | 4(5.6) |
Surgery, n(%) | |
Transplantation | 1(1.4) |
Chemoradiotherapy, n(%) | 1(1.4) |
Ward, n (%) | |
Geriatrics | 2(2.8) |
Nerosurgery | 7(9.9) |
Orthopaedics | 1(1.4) |
Intensive care unit | 26(36.6) |
Neurology | 2(2.8) |
Emergency | 23(32.4) |
Urology | 1(1.4) |
Thoracic surgery | 1(1.4) |
Cardiac Surgery | 2(2.8) |
Cardiology | 3(4.3) |
Hematology | 1(1.4) |
Nephrology | 2(2.8) |
Side of isolation, n(%) | |
Respiratory tract | 61(85.9) |
Blood | 1(1.4) |
Urine | 1(1.4) |
Ascites | 1(1.4) |
Drainage fluid | 2(2.8) |
Pus juice | 2(2.8) |
Cerebrospinal fluid | 3(4.3) |
Antimicrobial Susceptibility Test
As shown in Table 2, antimicrobial susceptibility of the Elizabethkingia isolates with the corresponding MICs towards tested antibiotics revealed very similar antibiotic susceptibility spectra with only small variations between the various strains. All 71 strains were extremely resistant to 13 ~ 16 antimicrobial agents, thus showing that they were multidrug resistance strains, including aminoglycosides, macrolides, and carbapenems (imipenem and meropenem), demonstrated that they were multidrug-resistant strains. Additionally, they also exhibited higher resistance rates to trimethoprim-sulfamethoxazole. The resistance rate of piperacillin was 100%, but increased sensitivity rates were observed when combined with β-lactamase inhibitors (cefoperazone/sulbactam and piperacillin/tazobactam) (90.1% and 66.2%, respectively). Finally, all isolates were found to be susceptible to minocycline and colistin.
Table 2
Antimicrobial susceptibilities of 71 Elizabethkingia isolates
Agents | R(%) | I(%) | S(%) |
Piperacillin | 71(100%) | 0 | 0 |
Piperacillin-tazobactam | 64(90.1) | 1(1.4) | 6(8.5) |
Cefoperazone-sulbactam | 47(66.2) | 4(5.6) | 20(28.2) |
Ceftazidime | 69(97.2) | 0 | 2(2.8) |
Ceftriaxone | 62(87.3) | 2(2.8) | 7(9.9) |
Cefotaxime | 69(97.2) | 2(2.8) | 0 |
Aztreonam | 69(97.2) | 2(2.8) | 0 |
Imipenem | 70(98.6) | 1(1.4) | 0 |
Meropenem | 68(95.8) | 2(2.8) | 1(1.4) |
Gentamicin | 68(95.8) | 2(2.8) | 1(1.4) |
Amikacin | 68(95.8) | 1(1.4) | 2(2.8) |
Minocycline | 0 | 0 | 71(100) |
Ciprofloxacin | 34(47.9) | 5(7.0) | 32(45.1) |
Levofloxacin | 25(35.2) | 0 | 46(64.9) |
Trimethoprim-sulfamethoxazole | 26(36.6) | 0 | 45(63.4) |
Tetracycline | 57(80.3) | 10(14.1) | 4(5.6) |
Tigecycline | 24(33.8) | 21(29.6) | 26(36.6) |
General features and clinical descriptions of seven representative Elizabethkingia strains
The next-generation sequencing and assembly of 7 genome has been reported in Table 3. The genome size ranged from 4.04Mb to 4.31Mb, with an average size of 4.10Mb, which is consistent with the genome size among the selected 83. Elizabethkingia strains (Supplementary materials Table S1). The number of contigs per genome ranged from 43 to 75, with a mean of 63.6. The read depth ranged from 23.26 to 80.63, with a mean of 38.79. The average GC content in Elizabethkingia strains was 35.84%. The seven distinct isolates were collected from 7 (5 male and 2 female) patients, with a mean age of 61.42 years. The sources of isolation included BALF (n = 1), sputum (n = 1), blood (n = 3), CSF (n = 2). All infections were healthcare associated, two patients had septic shock, and all patients had > 3 complex disease, such as hypertension, diabetes mellitus, or a malignancy.
Table 3
General features of selected Elizabethkingia strains.
Isolate | Specimen | Patient age | Patientgender | Whole-genome sequencing | Hospital ward | Length of hospital stay(days) | Size (Mb) | GC % | CDS | Contig | Protein |
HX WHF | CSF | 50 | F | E. miricola | Neurosurgery | 17 | 4.31 | 35.83% | 4034 | 63 | 4079 |
HX YK | Blood | 51 | M | E. anophelis | Liver surgery | 20 | 4.04 | 35.61% | 3720 | 70 | 3764 |
HX ZCH | Sputum | 47 | M | E. anophelis | Orthopedics | 22 | 4.04 | 35.61% | 3720 | 76 | 3764 |
HX XZB | CSF | 24 | M | E. miricola | Emergency | 8 | 4.08 | 35.81% | 3727 | 43 | 3769 |
HX WYD | Blood | 92 | M | E. meningoseptica | Emergency | 5 | 4.08 | 36.52% | 3706 | 59 | 3750 |
HX QKY | BALF | 77 | M | E. miricola | Respiratory medicine | 5 | 4.13 | 35.83% | 3766 | 59 | 3813 |
HX CGY | Blood | 89 | F | E. anophelis | Intensive care unit | 10 | 4.02 | 35.70% | 3706 | 75 | 3750 |
Resistance and virulence associated genes of the various Elizabethkingia strains
Antimicrobial resistance genes in the 7 Elizabethkingia strains has been shown in Table S2, which carried all the three previously described β-lactamase genes in Elizabethkingia, including the extended-spectrum β-lactamase blaCME and metallo-β-lactams blaB and blaGOB. Seven isolates also carried aminoglycoside resistance gene rnaA/rnaB, chloramphenicol resistance gene catB11, aminoglycoside resistance gene such as aadS, aac (3)-IVb and aac (3)-IIIc were only present in few strains. Comprehensive analysis with other 83 isolates from GenBank, all strains contain distinct AMR genes and mainly comprise three general resistance mechanism group: antibiotic efflux, antibiotic inactivation, and antibiotic target alteration. Individual strains carried from 7 to 10 AMR determinants (Supplementary materials Table S2). The various putative virulence factors were successfully predicted by VFDB, and all carried elongation factor tufA gene involved in Elizabethkingia adherence (Supplementary materials Table S3). HX WYD without the stress adaption gene katA, other 6 strains all carried the katA gene. A number of virulence genes were predicted to be involved in capsular polysaccharide biosynthesis, such as protein of Cap8G and Cps4J involved in immune modulation, chaperonin protein GroEL, catalases, peroxidase, heat shock protein and many others. However, eight-two of them in the bacterial virulence database also displayed high identity with this seven Elizabethkingia, although there were some minor variations (Table S3). We found that some virulence genes were widely distributed in 83 strains from GenBank, which may not be present in our 7 strains, such as tviB, fcl, wbtF, cps4J, and they all were responsible for immune modulation and regulation of the different metabolic processes. In addition, the htpB gene was detected in 24 Elizabethkingia strains, but it was not found in 7 isolates from our study. Elizabethkingia are non-motile bacteria without the flagellin structure proteins, which was consistent with the lack of the virulence gene flp in all the genome sequences.
Core and Pan genome analysis: core and accessory genes.
To analyze the core and pan genome of Elizabethkingia, the whole-genome sequences of the 90 strains were examined, which led to identification of the 83 Elizabethkingia species isolates strains from GenBank (Supplemental material Table S1). The core and pan-genomes were classified and 90 Elizabethkingia genomes were selected for the comprehensive analysis (Fig. 1). The core genome analysis showed that the number of shared genes decreased with addition of the additional input genomes. Overall, Elizabethkingia displayed an open pan-genome feature because new genes appeared when more sequenced genomes were added to the analysis. The distribution of the various gene families and the number of new genes were shown in Fig. 2A and 2B, respectively. In the 90 Elizabethkingia strains, 2,077 core (conserved) genes were recognized. In each strain, the number of accessory genes ranged from 1,098 to 1,925, and the number of unique genes ranged from 0 to 368.
The whole-genome comparisons possess the power to discriminate between the different strains and species with high resolution. Genome sequences were compared by using a pairwise method, calculating and comparing the ANI. Figure 2C depicts the results of ANI for the 90 Elizabethkingia strains. The pair-wise comparisons showed a minimum ANI of ∼80.72% for the most distant strains, but the strains of the E. anophelis subspecies showed an ANI of > 98.0%. Additionally, we noted that E. meningoseptica and other four species ANI values were obviously lower than other species. In the dendrogram, the two species, E. ursingii and E. occulta, were relatively close to E. miricola. The demarcation of the five species in the Elizabethkingia genus was clearly visible in the heat map produced from the similarity matrix.
Functional Cogs And Kegg
Figure 3A and 3C represents the COGs of the five Elizabethkingia type species. The total number of the genes in the five species ranged from 3,066 to 3,629. A total of 2,609 shared COGs were identified among the five Elizabethkingia type species. Among all the seven strains, the E. anophelis strain ZCH exhibited no unique gene families, and the E. miricola strain QKY showed the highest number of unique gene families (n = 60). The difference in the number of unique gene families might reflect the phenotypical traits that were observed to be specific to the group of bacteria. Functional analysis of the COGs in the 90 Elizabethkingia genomes was assigned into core, accessory and unique genes related to the regulation of the metabolism, cellular processes and signaling, as well as the different poorly characterized functions. Specifically, the result revealed that the majority of unique genomes were associated with metabolism, and the unique gene families were mostly related to “metabolism” (Fig. 3C), including the general functions, amino acid transport, cell wall/membrane/envelope and ribosomal structure biogenesis, as well as translation, transcription, replication, processing and modification. In addition, compared with unique genome, the core and accessory genes could function in extra cellular processes and signaling. The function of COGs with “information storage and processing” might be associated with intracellular survival in different environment. However, the exact reason for the large presence of genes with function of “poorly characterized” in unique gene families is not clear. In the KEGG analysis, the various genes associated with metabolism accounted for the largest part (Fig. 3B and Fig. 3D). The majority of these distinctive and core genes were linked to carbohydrate metabolism, but among the accessory genes, the regulation of amino-acid metabolism was the focus of the greatest number of genes, followed by energy metabolism. In addition to the metabolic functions, these were also linked to membrane transport, drug resistance, replication
and repair, translation, as well as modulation of cellular growth and death. All of these functions can confer the bacteria the ability to survive against the external environment.
Figure 3. The clusters of the orthologous groups (COGs) in core, accessory, and unique genomes and their associated. Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis of the 90 Elizabethkingia strains. (A) Distribution of functional COGs in each core, accessory, and unique genome. (B) The detailed distribution of KEGG with their functions. (C) The majority of core, accessory, and unique genes were found to be associated with metabolism. (D) Functional annotations showed that the gene families associated with carbohydrate metabolism, amino acid metabolism, cofactor and vitamin metabolism, as well as energy metabolism accounted for the largest part in these 90 Elizabethkingia strains.