The Reverse-­‐Evolution Experiment (REE) 20

The   experiment   presented   here   took   advantage   of   the   ability   of  P.  fluorescens  to   quickly  diversify  in  a  static  microcosm.  New  variants  (WS)  arise  by  mutations  in   genes   that   are   involved   in   the   production   of   cellulose   (see   section   1.4.3).   The   polymer   enables   the   cells   to   colonise   the   air-­‐liquid   interface.   Within   this   niche   oxygen  and  nutrients  are  equally  available  and  provide  a  benefit  enabling  WS  to   increase  in  frequency  (see  section  1.4.2).  Under  shaking  conditions,  however,  this   novel  trait  is  lost  due  to  the  costs  of  cellulose  production  without  it  providing  any   benefits.  One  of  the  fundamental  questions  that  were  addressed  by  the  REE  was   how  repeatable  and  predictable  is  evolution.  Can  a  novel  trait  (WS),  once  it  is  lost   under  shaking  conditions,  repeatedly  evolve  when  the  bacteria  are  grown  again  in   a  static  environment?  What  is  the  genetic  basis  of  repeated  WS  evolution?  

Dr.  Hubertus  Beaumont  and  Professor  Paul  Rainey  initiated  the  REE  in  2006  with   12   replicate   lineages   of   the   bacterium  P.   fluorescens   SBW25.   The   lineages   were   kept  in  glass  microcosms  containing  6  ml  liquid  KB  media.  All  lines  were  treated   equally   and   were   exposed   to   two   alternating   environmental   conditions   –   static   and   shaken.   Under   static   conditions   the   microcosms   were   left   undisturbed,   whereas  under  shaken  conditions  the  microcosms  grew  under  constant  shaking  at   160  rpm.    

The  experiment  started  by  growing  12  replicate  cultures  of  P.  fluorescens  SBW25   in   the   static   environment.   After   the   first   selection   round   (three   days),   the   static   cultures  were  diluted  and  plated  on  KB  agar  plates,  which  were  incubated  for  two   days.  The  plates  were  then  examined  for  colonies  with  morphologies  different  to   that  of  the  ancestral  colony  phenotype.  One  colony  of  the  most  common  new  type   was   picked   from   the   agar   plate,   transferred   into   a   fresh   microcosm   and   subsequently   exposed   to   shaken   conditions.   After   three   days   (second   selection   round)  the  procedure  described  above  was  repeated,  and  the  most  common  new   colony  type  now  growing  under  shaken  conditions  was  selected  and  transferred   into   a   new   microcosm.   In   essence,   every   time   a   new   type   was   found,   it   was   transferred  into  new  media,  and  kept  under  the  opposite  environmental  condition  

    Figure   1.8:   Experimental   design   of   the   REE   of   one   of   the   12   replicated   lines   in  P.   fluorescens  SBW25.  The  initial  inoculation  was  carried  out  with  12  replicates  of  SBW25.   Cultures   were   kept   in   microcosms   under   static   conditions.   After   one   selection   round   (three  days),  the  cultures  were  plated  on  agar  plates  and  screened  for  new  colony  types.   The   most   common   new   type   was   used   to   inoculate   new   KB   microcosms   that   were   kept   under  shaken  conditions.  After  three  days  the  cultures  were  plated  and  screened  for  the   most  common  new  colony  type.  The  new  type  was  then  used  to  inoculate  the  new  static   microcosms  starting  a  new  cycle  of  evolution.  All  12  replicate  lines  went  through  three  to   eight  cycles  of  evolution  during  the  REE  (Beaumont  et  al.,  2009).  

Where  no  new  type  could  be  detected  after  three  days,  6  µl  of  the  bacterial  culture  

were   transferred   into   fresh   microcosms.   These   were   kept   under   the   original   conditions   parallel   to   the   plating   step   (see   above),   either   static   or   shaken,   for   additional   three   days   (same   selection   round).   Cultures   were   diluted,   plated   and   screened  for  new  types.  Where  no  new  colony  morphology  type  was  found  on  the   plates,  the  procedure  described  above  was  repeated  until  a  new  colony  type  was   observed.   If   no   new   type   was   observed   over   a   long   period   of   time   (10   transfers   within  the  same  selection  round)  the  lineage  was  terminated.  Every  newly  evolved   type  was  stored  in  small  tubes  containing  a  saline-­‐glycerine  solution  at  -­‐80°C.  The  

‘frozen  fossil  collection’  was  later  revived  and  used  for  further  analysis  (Beaumont  

1.4.4.1 Whole-­‐genome  sequencing  of  evolutionary  endpoints  

Sequencing  the  entire  genome  of  an  organism  has  become  an  indispensable  and   powerful   tool   for   evolutionary   biologists.   Primarily   it   has   been   used   to   discover   changes  in  the  DNA  sequence  of  an  evolutionarily  derived  genome  by  comparing  it   with   the   known   reference   ancestral   genome   sequence.   Solexa   sequencing   is   an   efficient   technology   that   yields   high   quality   sequences   and   was   used   in   this   experiment.   Here   the   genotypes   of   the   evolutionary   endpoint   in   the   12   lineages   that   did   not   produce   any   more   new   colony   types   was   revived   from   the   freezer   stock  and  used  for  Solexa  sequencing.  This  way  it  was  possible  to  reconstruct  the   evolutionary   pathway   that   was   taken   by   each   lineage   depending   of   the   environmental   conditions   (static   or   shaken).   The   phenotype   was   mapped   to   the   genotype   by   identifying   all   causative   mutations   and   the   order   in   which   they   appeared.    

1.5

The  rise  of  a  stochastically  switching  phenotype  

The  experimental  design  of  the  REE  made  it  possible  to  follow  real-­‐time  evolution   of   12   parallel   lineages   of  P.   fluorescens   that   evolved   alternately   in   a   static   and   shaken  environment,  starting  from  the  same  clonal  founder  population  (Fig.  1.8;   see  section  1.4.4).  Increased  selection  pressure  under  static  conditions,  due  to  the   lack   of   oxygen   in   the   media,   favours   diversification   into   multiple   new   types,   usually  different  WS  phenotypes.  Propagation,  typically  of  a  WS  type,  in  a  shaken   environment  often  reverses  the  phenotype  to  one  that  resembles  the  ancestral  SM   type  (Fig.  1.1;  Rainey  &  Travisano,  1998).  Each  phenotypic  change  was  caused  by   a   mutational   modification.   In   two   of   the   12   lineages,   Line   1   and   Line   6,   a   novel   phenotype   evolved   independently   after   nine   rounds   of   selection.   This   new   type   showed  distinctive  colony  morphology  on  agar  plates  and  was  not  detected  in  any   of  the  12  lineages  during  earlier  selection  rounds.  Phenotypic  and  genetic  analysis   revealed  that  the  new  type  had  properties  of  a  stochastically  switching  genotype   (Beaumont  et  al.,  2009;  Gallie,  2009).