Population Genetics of Vibrio vulnificus: Identification of Two Divisions and a Distinct Eel-Pathogenic Clone

Genetic relationships among 62 Vibrio vulnificus strains of different geographical and host origins were analyzed by multilocus enzyme electrophoresis (MLEE), random amplification of polymorphic DNA (RAPD), and sequence analyses of the recA and glnA genes. Out of 15 genetic loci analyzed by MLEE, 11 were polymorphic. Cluster analysis identified 43 distinct electrophoretic types (ETs) separating the V. vulnificus population into two divisions (divisions Ι and ΙΙ). One ET (ET 35) included all indole-negative isolates from diseased eels worldwide (biotype 2). A second ET (ET 2) marked all of the strains from Israel isolated from patients who handled St. Peter's fish (biotype 3). RAPD analysis of the 62 V. vulnificus isolates identified 26 different profiles separated into two divisions as well. In general, this subdivision was comparable (but not identical) to that observed by MLEE. Phylogenetic analysis of 543 bp of the recA gene and of 402 bp of the glnA gene also separated the V. vulnificus population into two major divisions in a manner similar to that by MLEE and RAPD. Sequence data again indicated the overall subdivision of the V. vulnificus population into different biotypes. In particular, indole-negative eel-pathogenic isolates (biotype 2) on one hand and the Israeli isolates (biotype 3) on the other tended to cluster together in both gene trees. None of the methods showed an association between distinct clones and human clinical manifestations. Furthermore, except for the Israeli strains, only minor clusters comprising geographically related isolates were observed. In conclusion, all three approaches (MLEE, RAPD, and DNA sequencing) generated comparable but not always equivalent results. The significance of the two divisions (divisions Ι and ΙΙ) still remains to be clarified, and a reevaluation of the definition of the biotypes is also needed.     

Occurrence of Vibrio vulnificus Biotypes in Danish Marine Environments

During the unusually warm summer in Denmark in 1994, 11 clinical cases of Vibrio vulnificus infection were reported. These reports initiated an investigation of the occurrence of V. vulnificus biotypes in Danish marine environments. Samples of coastal water, sediment, shellfish, and wild fish were analyzed by preenrichment in alkaline peptone water amended with polymyxin B (2.0 × 10⁴ U/liter) followed by streaking onto modified cellobiose-polymyxin B-colistin agar. V. vulnificus-like colonies were tested with a V. vulnificus-specific DNA probe. Low densities of V. vulnificus were detected in water (0.8 to 19 CFU/liter) from June until mid-September and in sediment (0.04 to >11 CFU/g) from July until mid-November. The presence of V. vulnificus was strongly correlated with water temperature. However, we isolated V. vulnificus from water from a mussel farm at a lower temperature than previously reported (7°C). In 1 of the 13 locations studied, V. vulnificus was found in mussels in 7 of 17 samples analyzed; this is the first report of V. vulnificus in European shellfish. V. vulnificus was also isolated from gills, intestinal contents, and mucus from wild fish. Although biotyping of 706 V. vulnificus strains isolated during our investigations revealed that the majority of the strains (99.6%) belonged to biotype 1, biotype 2 was detected in seawater at a low frequency (0.4%). Our findings provide further evidence that seawater can serve as a reservoir and might facilitate spread of V. vulnificus biotype 2 to eels, with subsequent spread to persons handling eels. In conclusion, our data demonstrate that V. vulnificus is ubiquitous in a temperate marine environment and that V. vulnificus biotype 2 is not strictly confined to eels.     

Genetic relatedness of clinical and environmental isolates of lactose-positiveVibrio vulnificus

Environmental isolates of lactose-positiveVibrio vulnificus from different geographic areas were compared with clinical strains ofV. vulnificus on the basis of their phenotypic traits, virulence, DNA base composition, and DNA-DNA reasociation. EnvironmentalV. vulnificus strains were phenotypically indistinguishable from clinical isolates. These strains had a DNA base composition of 47–48 mol% guanine + cytosine and 85% reassociation at stringent temperature with DNA from clinicalV. vulnificus strains. These result indicateV. vulnificus strains from widely separated regions of the marine environment are indistinguishable from strains that have been agents of septicemia associated with shellfish consumption and of wound infections associated with seawater exposure.     

Genotypic Diversity and Virulence Characteristics of Clinical and Environmental Vibrio vulnificus Isolates from the Baltic Sea Region

The genetic diversity of Vibrio vulnificus isolates from clinical and environmental sources originating from the Baltic Sea region was evaluated by multilocus sequence typing (MLST), and possible relationships between MLST clusters, potential genotypic and phenotypic traits associated with pathogenicity, and source of isolation were investigated. The studied traits included genotyping of polymorphic loci (16S rRNA, vcg, and pilF), presence/absence of potential virulence genes, including nanA, nab, and genes of pathogenicity regions, metabolic features, hemolytic activity, resistance to human serum, and cytotoxicity to human intestinal cells. MLST generated 35 (27 new) sequence types and divided the 53 isolates (including four reference strains) into two main clusters, with cluster I containing biotype 1 and 2 isolates of mainly environmental origin and cluster II containing biotype 1 isolates of mainly clinical origin. Cluster II isolates were further subdivided into two branches. Branch IIB included isolates from recent cases of wound infections that were acquired at the German Baltic Sea coastline between 2010 and 2011 and isolates from seawater samples of the same regions isolated between 1994 and 2010. Comparing the MLST data with the results of genotyping and phenotyping showed that strains of MLST cluster II possess a number of additional pathogenicity-associated traits compared to cluster I strains. Rapid microbiological methods such as matrix-assisted laser desorption ionization–time of flight (MALDI-TOF) mass spectrometry combined with typing of selected virulence-associated traits (e.g., serum resistance, mannitol fermentation, nanA, and pathogenicity region XII) could be used for risk assessment purposes regarding V. vulnificus strains isolated from the Baltic Sea region.     

Distribution of Vibrio vulnificus and Other Lactose-Fermenting Vibrios in the Marine Environment

During the summer of 1981, 3,887 sucrose-negative vibrios were isolated from seawater, sediment, plankton, and animal samples taken from 80 sites from Miami, Fla., to Portland, Maine. Of these, 4.2% were able to ferment lactose. The lactose-positive strains isolated from the various samples correlated positively with pH and turbidity of the water, vibrios in the sediment and oysters, and total bacterial counts in oysters. Negative correlations were obtained for water salinity. Numerical taxonomy was performed on 95 of the lactose-fermenting environmental isolates and 23 reference strains. Five clusters resulted, with the major cluster containing 33 of the environmental isolates and all of the Vibrio vulnificus reference strains. The 33 isolates, which produced an acid reaction in lactose broth within hours of initial inoculation, represented 20% of all lactose-fermenting vibrios studied. These isolates were nearly identical phenotypically to clinical strains of V. vulnificus studied by the Centers for Disease Control, Atlanta, Ga., and by our laboratory, and their identification was confirmed by DNA-DNA hybridization studies. V. vulnificus was isolated from all sample types and from Miami to Cape Cod, Mass., and comparison of the environmental parameters of the eight subsites yielding this species with those of all 80 subsites revealed no significant differences. The majority of the isolates were obtained from animals, with clams providing most (84%) of these. On injection into mice, 82% of the V. vulnificus isolates resulted in death. Members of the remaining four clusters contained strains which differed from V. vulnificus in such phenotypic traits as luminescence and in urease or H₂S production. None of the other reference cultures, including nine other Vibrio species, were contained in the remaining clusters, and these isolates could not be identified. Most of these were also lethal for mice. Phenotypic differences, potential pathogenicity, and geographic distribution of the five clusters were examined. It is concluded that V. vulnificus is a ubiquitous organism, both geographically and in a variety of environmental sources, although it occurs in relatively low numbers. The public health significance of this organism and of the other unidentified lactose-fermenting Vibrio species is discussed.     

Genome-Wide SNP-Genotyping Array to Study the Evolution of the Human Pathogen Vibrio vulnificus Biotype 3

Vibrio vulnificus is an aquatic bacterium and an important human pathogen. Strains of V. vulnificus are classified into three different biotypes. The newly emerged biotype 3 has been found to be clonal and restricted to Israel. In the family Vibrionaceae, horizontal gene transfer is the main mechanism responsible for the emergence of new pathogen groups. To better understand the evolution of the bacterium, and in particular to trace the evolution of biotype 3, we performed genome-wide SNP genotyping of 254 clinical and environmental V. vulnificus isolates with worldwide distribution recovered over a 30-year period, representing all phylogeny groups. A custom single-nucleotide polymorphism (SNP) array implemented on the Illumina GoldenGate platform was developed based on 570 SNPs randomly distributed throughout the genome. In general, the genotyping results divided the V. vulnificus species into three main phylogenetic lineages and an additional subgroup, clade B, consisting of environmental and clinical isolates from Israel. Data analysis suggested that 69% of biotype 3 SNPs are similar to SNPs from clade B, indicating that biotype 3 and clade B have a common ancestor. The rest of the biotype 3 SNPs were scattered along the biotype 3 genome, probably representing multiple chromosomal segments that may have been horizontally inserted into the clade B recipient core genome from other phylogroups or bacterial species sharing the same ecological niche. Results emphasize the continuous evolution of V. vulnificus and support the emergence of new pathogenic groups within this species as a recurrent phenomenon. Our findings contribute to a broader understanding of the evolution of this human pathogen.     

Protocol for Specific Isolation of Virulent Strains of Vibrio vulnificus Serovar E (Biotype 2) from Environmental Samples

The eel pathogen Vibrio vulnificus biotype 2 comprises at least three serovars, with serovar E being the only one involved in both epizootics of eel vibriosis and sporadic cases of human infections. The virulent strains of this serovar (VSE) have only been recovered from clinical (mainly eel tissue) sources. The main objective of this work was to design and validate a new protocol for VSE-specific isolation from environmental samples. The key element of the new protocol is the broth used for the first step (saline eel serum broth [SEB]), which contains eel serum as a nutritive and selective component. This approach takes advantage of the ability of VSE cells to grow in eel serum and thus to separate themselves from the pool of competitors. The growth yield in SEB after 8 h of incubation was 1,000 times higher for VSE strains than for their putative competitors (including biotype 1 strains of the species). The selective and differential agar Vibrio vulnificus medium (VVM) was selected from five selective media for the second step because it gave the highest plating efficiency not only for the VSE group but also for other V. vulnificus groups, including biotype 3. The entire protocol was validated by field studies, with alkaline peptone water plus VVM as a control. V. vulnificus was isolated by both protocols, but serovar E was only recovered by the new method described here. All selected serovar E isolates were identified as VSE since they were virulent for both eels and iron-overloaded mice and resisted the bactericidal action of eel and iron-overloaded human sera. In conclusion, this new protocol is a suitable method for the isolation of VSE strains from environmental samples and is recommended for epidemiological studies of the pathogenic serovar E.     

Pulsed-Field Gel Electrophoresis Analysis of Vibrio vulnificus Strains Isolated from Taiwan and the United States

Vibrio vulnificus is a marine bacterium that causes human wound infections and septicemia with a high mortality rate. V. vulnificus strains from different clinical and environmental sources or geographic regions have been successfully characterized by ribotyping and several other methods. Pulsed-field gel electrophoresis (PFGE) is a highly discriminative method, but previous studies suggested that it was not suitable for examining the correlation of V. vulnificus strains from different origins. We employed PFGE to determine its efficacy for characterizing V. vulnificus strains from different geographic regions, characterizing a total of 153 strains from clinical and environmental origins from the United States and Taiwan after SfiI or NotI digestion. V. vulnificus strains showed a high intraspecific diversity by PFGE after SfiI or NotI digestion, and about 12% of the strains could not be typed by the use of either of these enzymes. For PFGE with SfiI digestion, most of the clinical and environmental strains from the United States were grouped into cluster A, while the strains from Taiwan were grouped into other clusters. Clinical strains from the United States showed a higher level of genetic homogeneity than clinical strains from Taiwan, and environmental strains from both regions showed a similarly high level of heterogeneity. PFGE with NotI digestion was useful for studying the correlation of clinical strains from the United States and Taiwan, but it was not suitable for analyzing environmental strains. The results showed that PFGE with SfiI digestion may be used to characterize V. vulnificus strains from distant geographic regions, with NotI being a recommended alternative enzyme.     

Isolation of bacteriophage infectious for Vibrio vulnificus

Nine phage isolates infectious for Vibrio vulnificus and falling into four morphological groups were isolated from estuarine waters collected in Louisiana. Of the 60 V. vulnificus strains tested, 87% were susceptible to one or more of the isolates. With the exception of V. fluvialis, Vibrio species other than vulnificus were resistant to infection. A spectrum of enteric bacterial strains were similarly resistant. Susceptibility differences were seen between opaque (virulent) V. vulnificus strains and those with translucent (nonvirulent) colony types, with the former being more susceptible. Susceptibility patterns to infection by the nine phage isolates among the V. vulnificus test strains suggest that the latter may fall into several groups. Other aspects relating to the phage isolates are presented.     

Molecular Typing of Environmental and Clinical Strains of Vibrio vulnificus Isolated in the Northeastern USA

Vibrio vulnificus is a ubiquitous marine bacterium that is responsible for infections and some seafood-related illnesses and deaths in the United States, mainly in individuals with compromised health status in the Gulf of Mexico region. Most phylogenetic studies focus on V. vulnificus strains isolated in the southern United States, but almost no genetic data are available on northeastern bacterial isolates of clinical or environmental origin. Our goal in this study was to examine the genetic diversity of environmental strains isolated from commercially-produced oysters and in clinical strains of known pathogenicity in northeastern United States. We conducted analyses of a total of eighty-three strains of V. vulnificus, including 18 clinical strains known to be pathogenic. A polyphasic, molecular-typing approach was carried out, based upon established biotypes, vcg, CPS, 16S rRNA types and three other genes possibly associated with virulence (arylsulfatase A, mtlABC, and nanA). An established Multi Locus Sequence Typing (MLST) method was also performed. Phylogenetic analyses of these markers and MLST results produced similar patterns of clustering of strains into two main lineages (we categorized as ‘LI’ and ‘LII’), with clinical and environmental strains clustering together in both lineages. Lineage LII was comprised primarily but not entirely of clinical bacterial isolates. Putative virulence markers were present in both clinical and environmental strains. These results suggest that some northeastern environmental strains of V. vulnificus are phylogenetically close to clinical strains and probably are capable of virulence. Further studies are necessary to assess the risk of human illness from consuming raw oysters harvested in the northeastern US.     

Ecology ofVibrio vulnificus in Galveston Bay oysters,suspended particulate matter,sediment and seawater: Detection by monoclonal antibody — immunoassay — most probable number procedures

Summary Oysters, suspended particulate matter (SPM), sediment and seawater samples were collected from West Galveston Bay, Texas over a 16-month period and analyzed for the presence ofVibrio vulnificus, a naturally-occurring human marine pathogen. Detection and enumeration ofV. vulnificus was performed using a species-specific monoclonal antibody (mAb FRBT37) in an enzyme immunoassay (EIA)-most probable number (MPN) procedure capable of detecting as few as 2000 target organisms.V. vulnificus was not detected in seawater, oyster or SPM samples during the cold weather months, but was detected at low levels in several sediment samples during this time period. Increased levels of the organism were first observed in early spring in the sediment, and then in SPM and oysters. The major increase inV. vulnificus occurred only after the seawater temperature had increased above 20°C and the winter-spring rainfall had lowered the salinity below 16. The highestV. vulnificus levels at each site were associated with suspended particulate matter. These results are consistent with the hypothesis that (1)V. vulnificus over-winters in a floc zone present at the sediment-water interface, (2) is resuspended into the water column in early spring following changes in climatic conditions, (3) colonizes the surfaces of zooplankton which are also blooming during early spring and (4) are ingested by oysters during their normal feeding process.     

Vibrio mimicus are the reservoirs of the heat-stable enterotoxin gene (nag-st) among species of the genus Vibrio

Using a 0.27 kb DNA probe specific for the heat-stable enterotoxin gene (nag-st) of Vibrio cholerae non-O1, 1109 strains representing 17 species of the genus Vibrio, isolated from clinical and environmental sources were examined. The nag-st gene was preponderantly associated with strains classified as V. mimicus; 16.8% of these strains hybridized. It was more frequent in the clinical isolates (22.6%) than in the environmental isolates (13.7%). The incidence of nag-st gene-positive strains of V. mimicus isolated from different countries was uniformly high and ranged between 8.7% (Bangladesh) and 57.1% (environmental strains from USA). The incidence of the nag-st gene was much lower among strains of V. cholerae non-O1 (3.6%). Probe-positive and-negative strains of V. mimicus and V. cholerae non-O1 were used to evaluate the performance of the conventional suckling mouse assay for detection of the NAG-ST enterotoxin. Of the 31 probe-positive strains, only five (16.1%) yielded a positive fluid accumulation ratio (FA ratio) when neat heated culture supernatant was used to perform the suckling mouse assay. All the 31 probe-positive strains gave a positive FA ratio when 20-fold concentrated and heated culture supernatants of the strains were used to perform the suckling mouse assay. The need to concentrate (by at least 20-fold) the culture supernatant of strains of V. mimicus and V. Cholerae non-O1 was identified as an important step to obtain consistent results when using the suckling mouse assay for detection of NAG-ST.P. Yuan, A. Ogawa and T. Takeda are with the Department of Infectious Disease Research, National Children's Medical Research Center, 3-35-31 Taishido, Setagaya-ku, Tokyo 154, Japan; P. Yuan is also with the National Institute for the Control of Pharmaceutical and Biological Products, Beijing, China. T. Ramamurthy and G.B. Nair are with the National Institute of Cholera and Enteric Diseases, Calcutta, India. T. Shimada is with the National Institute of Health, Tokyo 141, Japan. S. Shinoda is with the Faculty of Pharmaceutical Sciences, Okayama University, Japan.     

Intraspecific Diversity of Vibrio vulnificus in Galveston Bay Water and Oysters as Determined by Randomly Amplified Polymorphic DNA PCR

Randomly amplified polymorphic DNA (RAPD) PCR was used to analyze the temporal and spatial intraspecific diversity of 208 Vibrio vulnificus strains isolated from Galveston Bay water and oysters at five different sites between June 2000 and June 2001. V. vulnificus was not detected during the winter months (December through February). The densities of V. vulnificus in water and oysters were positively correlated with water temperature. Cluster analysis of RAPD PCR profiles of the 208 V. vulnificus isolates revealed a high level of intraspecific diversity among the strains. No correlation was found between the intraspecific diversity among the isolates and sampling site or source of isolation. After not being detected during the winter months, the genetic diversity of V. vulnificus strains first isolated in March was 0.9167. Beginning in April, a higher level of intraspecific diversity (0.9933) and a major shift in population structure were observed among V. vulnificus isolates. These results suggest that a great genetic diversity of V. vulnificus strains exists in Galveston Bay water and oysters and that the population structure of this species is linked to changes in environmental conditions, especially temperature.     

Inhibition of <Emphasis Type="Italic">Vibrio</Emphasis> biofilm formation by a marine actinomycete strain A66

China remains by far the largest aquaculture producer in the world. However, biofilms formed by pathogenic Vibrio strains pose serious problems to marine aquaculture. To provide a strategy for biofilm prevention, control, and eradication, extracts from 88 marine actinomycetes were screened. Thirty-five inhibited the biofilm formation of Vibrio harveyi, Vibrio vulnificus, and Vibrio anguillarum at a concentration of 2.5% (v/v). Thirty-three of the actinomycete extracts dispersed the mature biofilm. Six extracts inhibited the quorum-sensing system of V. harveyi by attenuating the signal molecules N-acylated homoserine lactones’ activity. Strain A66, which was identified as Steptomyces albus, both attenuated the biofilms and inhibited their quorum-sensing system. It is suggested that strain A66 is a promising candidate to be used in future marine aquaculture.     

A potential bacterial biocontrol agent,strain S2V2 against pathogenic marine <Emphasis Type="Italic">Vibrio</Emphasis> in aquaculture

We isolated a marine bacterium strain S2V2 which inhibited the growth of pathogenic marine Vibrio spp. The aims of this research were to identify a new antibiotic-producing marine bacterium strain S2V2, and evaluate its spectrum activity and pathogenic property. Analysis of 16S rDNA sequence placed strain S2V2 in the genus Pseudoalteromonas, but the sequence similarity was low (95.46%) implying the strain might be a new species in this genus. Strain S2V2 inhibited the growth of 67.9% of 28 Vibrio strains tested. This strain inhibited V. alginolyticus, V. anguillarum, V. fluvialis, V. harveyi, V. metschnikovii, V. splendidus, V. ordalii, V. parahaemolyticus, and V. vulnificus, but inactive against V. campbellii, Aeromonas hydrophyla and Staphylococcus aureus. Strain S2V2 produced extracellular non proteinaceous antibacterial substances. The highest antibacterial activity was found when strain S2V2 was cultured for 96 h in ZoBell broth medium. An artificial infection to post larvae of Lithopenaeus vanname indicated that strain S2V2 was a non pathogenic bacterium. Non pathogenic property and specific antibacterial activity against a broad range of fish pathogenic marine Vibrio of strain S2V2 suggest that this strain is a prospective source of unique antibiotic and a potential biocontrol agent in marine aquaculture.     

Analysis of Vibrio vulnificus from Market Oysters and Septicemia Cases for Virulence Markers

Representative encapsulated strains of Vibrio vulnificus from market oysters and oyster-associated primary septicemia cases (25 isolates each) were tested in a blinded fashion for potential virulence markers that may distinguish strains from these two sources. These isolates were analyzed for plasmid content, for the presence of a 460-bp amplicon by randomly amplified polymorphic DNA PCR, and for virulence in subcutaneously (s.c.) inoculated, iron-dextran-treated mice. Similar percentages of market oyster and clinical isolates possessed detectable plasmids (24 and 36%, respectively), produced the 460-bp amplicon (45 and 50%, respectively), and were judged to be virulent in the mouse s.c. inoculation-iron-dextran model (88% for each). Therefore, it appears that nearly all V. vulnificus strains in oysters are virulent and that genetic tests for plasmids and specific PCR size amplicons cannot distinguish between fully virulent and less virulent strains or between clinical and environmental isolates. The inability of these methods to distinguish food and clinical V. vulnificus isolates demonstrates the need for alternative subtyping approaches and virulence assays.     

Occurrence of Vibrio vulnificus in mussel farms from the Varano lagoon environment

Aims: Monitoring the occurrence of the human pathogen Vibrio vulnificus in a mussel farm located in the lagoon of Varano (Italy). Methods and Results: A total of 72 samples of mussel, water and sediment, collected from two locations of Varano lagoon in the Gargano peninsula, during a 7‐ month survey, were analysed. Isolation and PCR characterization of six V. vulnificus environmental genotype strains revealed that this pathogen was isolated when with T was above 22°C and salinity ranged between 22·7 and 26·4‰. No significant correlation of the occurrence of V. vulnificus with water pH or salinity was observed. Moreover, 8% of mussel samples were found to be contaminated by V. vulnificus. All of that positive mussel samples originated from the same sampling station. Conclusion: It is suggested that warmer season are risky to eat raw or undercooked bivalve molluscs in the local area. Significance and Impact of the Study: To increase knowledge about environmental conditions that may affect the occurrence of waterborne pathogen Vibrio vulnificus in seafood.     

Multi-site Analysis Reveals Widespread Antibiotic Resistance in the Marine Pathogen <Emphasis Type="Italic">Vibrio vulnificus</Emphasis>

Vibrio vulnificus is a serious opportunistic human pathogen commonly found in subtropical coastal waters, and is the leading cause of seafood-borne mortality in the USA. This taxon does not sustain prolonged presence in clinical or agricultural settings, where it would undergo human-induced selection for antibiotic resistance. Therefore, few studies have verified the effectiveness of commonly prescribed antibiotics in V. vulnificus treatment. Here we screened 151 coastal isolates and 10 primary septicaemia isolates against 26 antimicrobial agents representing diverse modes of action. The frequency of multiple resistances to antibiotics from all sources was unexpectedly high, particularly during summer months, and a substantial proportion of isolates (17.3%) were resistant to eight or more antimicrobial agents. Numerous isolates demonstrated resistance to antibiotics routinely prescribed for V. vulnificus infections, such as doxycycline, tetracycline, aminoglycosides and cephalosporins. These resistances were detected at similar frequencies in virulent and non-virulent strains (PCR-based virulence typing) and were present in septicaemia isolates, underlying the public health implications of our findings. Among environmental isolates, there were no consistent differences in the frequency of resistance between pristine and anthropogenically impacted estuaries, suggesting natural rather than human-derived sources of resistance traits. This report is the first to demonstrate prevalent antibiotic resistance in a human pathogen with no clinical reservoirs, implying the importance of environmental studies in understanding the spread, evolution and public health relevance of antibiotic resistance factors. Electronic supplementary material  The online version of this article (doi:) contains supplementary material, which is available to authorized users.