Beyond Phylogeny: Evolutionary analysis of a mosaic pathogen

Dr Rosalind Harding

Departments of Zoology and Statistics, Oxford University,UK

Research Collaborators

Naiel Bisharat Dept of Epidemiology and Preventative

Medicine, Tel Aviv University, Israel Derrick Crook

Nuffield Dept of Clinical Laboratory Sciences, John Radcliffe Hospital, University of Oxford, UK

Martin Maiden Dept of Zoology, University of Oxford

Bisharat et al. (2005) Hybrid Vibrio vulnificus Emerg Infect Dis 11:30-35

Population Genetics

Interplay of micro-evolutionary processes Mutation and recombination Population structure and demography Natural selection

Questions and strategy concern: Understanding steady-state patterns of diversity Learning about ancestral history (genealogy) Understanding dynamics: emergence of new strains

Major technical problem Trees don’t show recombination events

Vibrio vulnificus

Globally wide-spread inhabitant of marine and estuarine environments

Dangerous waterborne pathogen: case fatality rate for V. vulnificus septicemia may reach 50%

Typically, cases of V. vulnificus infection are sporadic Human infection acquired through eating

contaminated raw or undercooked sea food, or via contamination of wounds by seawater or marine animals

Disease Outbreak in Israel

Major outbreak of systemic V. vulnificus infection among fish market workers and fish consumers

Epidemiology 1995: first case 1996: 32 patients 1997: 30 patients all handled fresh Tilapia fish cultivated in inland fish

farms 1998: marketing policy changed to prevent sale &

handling of live Tilapia fish New biotype identified

Distinctive biochemistry, eg salicin-negative, lactose-negative (5 atypical characteristics for the species).

Severe soft tissue infections/ Necrotizing fasciitis

V. vulnificus diversity

Biotype 1: sampled from environment, healthy fish, shellfish etc; associated with sporadic human infection

Biotype 2: associated with disease in eels Biotype 3: new cause of human disease

outbreak in Israel. Where did Biotype 3 come from?

Biotypes have been defined based on biochemical tests of phenotype.

Initial genetic analysis

MLST: multi-locus sequence typing Sequences of fragments of ‘housekeeping’ genes

(dN/dS ratios < 1.0) 10 genes, 5 from each of the two chromosomes,

each fragment ~400 bp Concatenated sequence of 4,326 bp defines

sequence types (STs) Isolates:

Biotype 1: n=82 isolates (39 from human disease, 43 from environment

Biotype 2: n=15 isolates (13 from eels) Biotype 3: n=61 isolates (60 from human disease, 1

from fish-pond water)

11- Environment (Denmark) 65- Environment (Germany) 13- Environment (Denmark) 19- Human (USA) 59- Healthy fish (Israel) 17- Oyster (USA) 49- Environment (Germany) 44- Environment (Germany) 6- Diseased eels (Spain, Japan, Sweden, Taiwan),a

12- Environment (USA) 66- Environment (Germany) 9- Diseased eels (Denmark),b

47- Sea water (Japan) 41- Oyster (USA) 62- Oyster (USA) 35- Oyster (USA) 15- 1(Environment), 1(human) (USA) 51- Oyster (USA) 24- Oyster (USA) 29- Human (USA) 43- Human (Germany) 26- Oyster (USA) 28- Oyster (USA) 39- Oyster (USA) 48- Diseased eel (Denmark)b

53- Oyster (USA) 63- Oyster (USA) 10- Diseased eel (Denmark)b, healthy fish (Israel) 30- Oyster (USA) 38- Oyster (USA) 31- Oyster (USA) 52- Human (USA) 27- Oyster (USA) 34- Oyster (USA) 23- Oyster (USA) 54- Oyster (USA) 25- Oyster (USA) 4- Environment (USA) 3- Human (USA) 16- Human (USA) 22- Oyster (USA) 8- Human (61), healthy fish (1) (Israel) 45-Environment (Germany) 57- Human (Spain) 14- Environment (Denmark) 61- Human (Sweden) 69- Shrimp (Indonesia) 70- Human (Sweden) 1- Human (USA) 2- Human (USA) 56- Human (South Korea) 68- Human (Sweden) 55- Human (Singapore) 46- Human (Japan) YJ016- Human (China) 18- Human (USA) 67- Human (Japan) 40- Human (USA) 32- Human (4), oyster (1) (USA) 42- Human (USA) 5- Environment (Spain) 58- Healthy fish (Israel) 20- Human (USA) 50- Human (Singapore) 7- Shrimp (Thailand) CMCP6- Human (South Korea) 60- Oyster (USA) 21- Human (USA) 64- Human (USA) 36- Human (USA) 33- Human (USA) 37- Human (USA)

82

99

97

70

100

76

78

94

72

84

98

80

84

99

88

77

0.002

II

Vibrio parahemolyticus

UPGMA tree of concatenated sequences of 10 genes: two major groups: I & II, plus ST8

I

ST8=Biotype 3

All Biotype 3 isolates were identical at level of MLST resolution.

Output from STRUCTURE analysis, assuming K= 3 populations

Genetic differentiation into two ‘populations’ is not explained by geographic location of isolates

Genetic differentiation into two ‘populations’ is not explained by biotype distribution.

Output from STRUCTURE analysis, assuming K= 3 populations

Biotype 3

Biotype 1 occurs in both populations

However, Biotype 3 does have a distinctive intermediate genetic identity between the populations.

Output from STRUCTURE analysis, assuming K= 3 populations UPGMA

Group I

UPGMA Group II

Two populations: different disease associations

Population B is associated with disease in humans

Population A is associated with eel disease

Biotype 3 is a hybrid between parents from Population A and Population B

Infe

rre

d a

nce

stry

I

II

A

B

Biotype 3 is a mosaic genome

Clonal expansion of Biotype 3

Maynard Smith, J et al (2000) BioEssays 22:1115-1122

Disease outbreak clones emerge from a background of low frequency variation connected by mutation and recombination.

Progress summary

The disease outbreak in Israel (Biotype 3) was caused by a clonal expansion of Sequence Type 8

ST 8 is a mosaic sequence created by recombination between parents from Populations A and B

Next questions How much recombination? How did the genetic differentiation between

Populations A and B arise? Population A = UPGMA Group I = Eel

disease associated Population B = UPGMA Group II =

Human disease associated

Splits graph of concatenated sequences from 10 genes

Cluster I = Population A

Association with eel disease (biotype 2)

ST8 = Biotype 3

Cluster II = Population B

Association with human disease

Recombination exchange between groups I & II is rare

ST8 (Biotype 3) has a glp allele from Population B/group II

Alleles 12 and 38 from Cluster II STs are more closely related to Cluster I

Splits graph of allelic sequences from glp gene

II

I

Recombination rates within genes within groups are high

Minimum of 9 recombination events

Ancestral history is not as simple as a tree.

II

I

Evidence of recombination from Beagle: www.stats.ox.ac.uk/~lyngsoe/beagle

Splits graph of alleles from dtdS gene

Polymorphism for a complex trait?

Is the genetic differentiation related to pathogenicity phenotype? higher odds for causing either human or eel

disease

Next Question.

Isolation in a metapopulation?

Is the genetic differentiation caused by isolation between populations?

Any clues from diversity in individual genes?

If polymorphism, perhaps expect differentiation to localise to one or a subset of genes?

If differentiation is due to isolation between populations, expect all genes to show the same patterns.

USA-Env USA-ENV Denmark-EEL Israel-Env Denmark-Env Baltic Sea USA-Env USA-clinical

USA-Env USA-Env USA-Clinical USA-Env Baltic sea Baltic sea Japan-EEL Denmark-eel Denmark-Env USA-Env Japan-Env USA-Env Germany-Clinical USA-Env USA-Env USA-Env USA-Env Denmark-EEL USA-Env USA-Env USA-Clinical USA-Env USA-Env USA-Env USA-Env USA-Env Baltic sea USA-Env USA-Env USA-eNV USA-Clinical USA-Clinical USA-Env Israel-Clinical Denmark-Env Spain-Clinical Baltic sea Sweden-Clinical USA-Env -S.Korea-Clinical Japan-Clinical Indonesia-Env USA-Clinical Israel-Env Spain-eel farm Singapore-Clinical USA-Clinical Sweden-Clinical USA-Clinical USA-Clinical USA-Clinical USA-Clinical Thailand-Env USA-Clinical USA-Clinical USA-Clinical Japan-Clinical Taiwan-Clinical

USA-Clinical Singapore-Clinical Sweden-Clinical S. Korea-Clinical USA-Clinical USA-Clinical

0.005

Biotype 3

UPGMA group I (Population A)

UPGMA group II (Population B)

In Biotype 3, genes 1, 2, 4, & 10 are from group II, i.e. human disease associated.

The same split is preserved across genes 1, 2, 4 & 10

1. Large chromosome: glp

2. Large chromosome: gyrB

4. Large chromosome: metG

10. Small chromosome: tnaA

But the same split is also preserved across the other 6 genes, e.g.

6. Small chromosome: dtdS

9. Small chromosome: pyrC

5. Large chromosome: purM8. Small chromosome: pntA

Conclusions

Differentiation between populations is evident across all 10 genes. Recombination exchange between populations is rare across all genes.

Within populations: Large numbers of alleles related through recombination as well as mutation history

Isolation by distance? Polymorphism? Recombination is key to generating diversity

in Vibrio vulnificus

Clonal Expansion

In expansions of clonal complexes, new mutations are evident before recombination. (Linkage disequilibrium due to selective sweep.)

Differentiation is shaped by selection: clonal complexes emerge as new adaptations

Meta-population structure

Old population diversity generated by mutation and recombination is sustained.

Differentiation is shaped by isolation: outbreaks emerge as new recombinants