The Lim­i­ta­tions of LB Medium

by Hi­roshi Nikaidō

LB medium, also known in­cor­rectly as Luria-Bertani medium, is widely used to grow bac­te­r­ial cul­tures, mainly be­cause it is easy to pre­pare and pro­vides a broad base of nu­tri­ents. LB broth con­tains, per ml, 10 mg tryp­tone (a mix­ture of pep­tides formed by the di­ges­tion of ca­sein with the pan­cre­atic en­zyme, trypsin), 5 mg yeast ex­tract (an au­tolysate of yeast cells), and 5 or 10 mg NaCl. It was for­mu­lated by Giuseppe Bertani in 1951 for study­ing lysogeny in Es­cherichia coli. He called it "Lysogeny Broth," or LB. No­tably, the orig­i­nal for­mu­la­tion in­cluded 1 mg/ml glu­cose, which has been dropped in more re­cent times. This medium was de­signed for work at low bac­te­r­ial densi­ties, a key point of this ar­ti­cle.

While this may ap­pear a tasty dish for many bac­te­ria of re­search in­ter­est, it is an in­ap­pro­pri­ate choice for phys­i­o­log­i­cal stud­ies wherein re­pro­ducibil­ity is re­quired. Since only bac­te­r­ial cul­tures in bal­anced growth (achieved by suf­fi­cient time in ex­po­nen­tial growth) have a re­pro­ducible av­er­age cell size and chem­i­cal com­po­si­tion, none of the com­po­nents of liq­uid me­dia should be­come ex­hausted dur­ing growth of the cul­ture. Is this the case with LB broth? To an­swer this ques­tion, we must know what lim­its bac­te­r­ial growth in LB broth.

Growth of E. coli usu­ally stops, even in the pres­ence of the large to­tal con­cen­tra­tion of or­ganic nu­trients in LB broth, when the OD₆₀₀ reaches around 2, cor­re­spond­ing to about 0.6 mg of E. coli (dry weight) per ml. The rea­son is not dif­fi­cult to find: LB medium pro­vides only a scant amount of car­bo­hy­drates, and sur­pris­ingly small amounts of other uti­liz­able car­bon sources. Tryp­tone and yeast ex­tract are mostly com­posed of pep­tides of vary­ing length. In their de­fin­i­tive 1968 study of Bacto Neopep­tone us­ing gel fil­tra­tion, Payne and Gilvarg found that there was a clear size limit for the us­able pep­tides at about 650 dal­tons – which cor­re­sponds to the ex­clu­sion limit of porin chan­nels de­ter­mined sev­eral years later. The smaller, us­able pep­tides were a mi­nor­ity, per­haps a quar­ter of the en­tire mix­ture. Free amino acids were an even smaller mi­nor­ity, ap­prox­i­mately 1% or less of the en­tire prepa­ra­tion. If we as­sume a sim­i­lar size dis­tri­b­u­tion for the pep­tides in tryp­tone and yeast ex­tract, we can pos­tu­late that the yield of E. coli is lim­ited pri­mar­ily by the avail­able car­bon sources.

Growth of E. coli MG1655 in Luria-Bertani broth. An overnight cul­ture was di­luted 5,000-fold in Luria-Bertani broth and cul­ti­vated at 37°C with vig­or­ous aera­tion. A The OD₆₀₀ (tri­an­gles) and cell con­centration (crosses) were mea­sured per­iodically. B The ra­tio of the OD₆₀₀ to the cell con­cen­tra­tion was cal­cu­lated at each point and mul­ti­plied by 10⁹. Source

In 2007, D'Ari and his col­leagues did a care­ful study of growth of E. coli in LB broth. They showed that very early in the pe­riod usu­ally con­sid­ered to be well within the ex­ponential phase, when the OD₆₀₀ reached around 0.3, there is an abrupt change in phys­i­ol­ogy and the cell size be­gins to de­crease. This ob­ser­va­tion is not new. As ac­knowledged by the D'Ari group, Wang and Koch in 1978 had sim­i­larly shown that, very early in the "ex­po­nen­tial" growth phase, E. coli in LB broth goes through a di­auxie-like change in growth be­hav­ior. D'Ari's group went fur­ther and sug­gested that what mat­ters is not the gen­eral avail­abil­ity of amino acids gen­er­ated by the in­tra­cel­lu­lar hy­drol­y­sis of im­ported pep­tides, but the avail­abil­ity of amino acids eas­ily uti­lized as car­bon sources. Al­though these work­ers mon­i­tored only the free amino acids in the medium (whereas the cells were un­doubt­edly uti­liz­ing pep­tides for the most part), they con­cluded that dur­ing growth in LB broth E. coli cells un­dergo al­ter­ations in car­bon nu­tri­tion, us­ing the eas­ier-to-uti­lize amino acids un­til they were de­pleted, then switch­ing to the harder-to-use amino acids.

As an aside, I'd like to men­tion that the con­clu­sion of D'Ari et al. is at­trac­tive to me for sen­ti­men­tal rea­sons. Back in the 1950s, I was in a lab­o­ra­tory in Japan that stud­ied the se­quen­tial uti­liza­tion of amino acids in com­plex me­dia. My men­tor, Kiyokatsu Kinjo, was im­mensely im­pressed by Jacques Monod's study of di­auxic growth done un­der an­other re­source-strained en­vi­ron­ment – Paris dur­ing the WWII – and tried to mimic Monod's work by study­ing the uti­liza­tion of amino acids as the sole ni­tro­gen source. In­deed, us­ing a Vib­rio strain he showed di­auxic growth with var­i­ous mix­tures of amino acids. He es­tab­lished that as­par­tate and argi­nine (and later ser­ine and cys­teine were also added to this group) are used first, while other amino acids (such as leucine and va­line) are not used un­til the cells en­ter the sec­ond phase.

Al­though Kinjo was study­ing the pref­er­en­tial uti­liza­tion of amino acids as sources of ni­tro­gen, prob­a­bly their uti­liza­tion as car­bon sources fol­lows a sim­i­lar pat­tern since as­par­tate and ser­ine would gen­er­ate C4 and C3 com­pounds that read­ily en­ter the cen­tral me­tab­o­lism. We should note, how­ever, that al­ter­ations in car­bon nu­tri­tion oc­cur even in some syn­thetic (min­i­mal) me­dia. As is well known, E. coli grow­ing in glu­cose-min­i­mal medium con­verts glu­cose into ac­etate and other or­ganic acids dur­ing the first phase; then later in the "ex­po­nen­tial growth" phase it grows prima­rily by ox­i­diz­ing these or­ganic acids. Thus, changes in car­bon nu­tri­tion are per­haps an in­evitable com­pli­ca­tion of batch cul­ture tech­niques.

Are the shift­ing car­bon sources the only cause of al­ter­ations in a young E. coli cul­ture grow­ing in LB broth? I would like to pro­pose that an­other fac­tor may be an even more pro­found in­flu­ence: the avail­abil­ity of di­va­lent cations. In the 1980s, Brian Wilkin­son noted that the prop­er­ties of the outer mem­brane of Para­coc­cus den­i­tri­f­i­cans were very dif­fer­ent de­pend­ing on whether the bac­teria were grown in a di­va­lent-cation-suf­­fi­cient syn­thetic medium or a com­plex medium contain­ing Bacto-Pep­tone and yeast ex­tract. He and his stu­dent, Sechan Wee, mea­sured the Mg²⁺ con­tent of these com­po­nents as 5.4 and 29 µmol/g. (The Ca²⁺ con­tent was much lower.) This very import­ant pa­per is not in­cluded in the Med­line data­base, and has been cited only twice (once by my­self). If we as­sume that Difco Tryp­tone has a sim­i­lar Mg²⁺ con­tent, the to­tal Mg²⁺ con­tent of LB will be only about 200 µM.

How­ever, E. coli needs a lot of Mg²⁺ and Ca²⁺ to bridge the highly neg­a­tively charged LPS mole­cules in its outer mem­brane. The to­tal amount of Mg²⁺ as­so­ci­ated with E. coli cells is es­ti­mated to be around 100 mM. (This in­for­ma­tion is ac­ces­si­ble by sub­scrip­tion here.) If a liter of LB broth even­tu­ally pro­duces 2 g (dry weight) or about 6 ml (cel­lu­lar vol­ume) of E. coli, we would need 600 μmol of Mg²⁺, three times more than what is avail­able in LB. The LPS struc­ture can be mod­i­fied through the PhoPQ sys­tem and E. coli un­doubt­edly sur­vives even un­der these Mg²⁺-star­va­tion con­di­tions, but the Mg²⁺ con­tent of the cells, es­pe­cially of the outer mem­brane, changes through­out the "ex­po­nen­tial phase" and the cells are not ex­pected to be too happy. The sit­u­a­tion would be much worse if we ac­cept the ex­per­i­men­tally de­ter­mined value for the Mg²⁺ con­tent of LB broth cited by Papp-Wal­lace and Maguire in the EcoSal chap­ter (ac­ces­si­ble by sub­scrip­tion here): 30–40 µmol per liter. My feel­ing is that this se­ri­ous prob­lem with LB broth has been ne­glected for many years. It may even be that E. coli grown in LB broth of­ten comes to the sta­tion­ary phase due to the lack of Mg²⁺! Cer­tainly one can­not blindly use LB medium if one is ob­serv­ing phe­nom­ena in­volv­ing the in­tegrity of the outer mem­brane or ionic in­ter­ac­tions of LPS, such as the sus­cep­ti­bil­ity of bac­teria to poly­ca­tionic drugs (e.g., amino­gly­co­sides or polymyxin). In this sense, clin­i­cal microbiolo­gists are ahead of phys­i­ol­o­gists; for such pur­poses they use ex­clu­sively "cation-ad­justed" Mueller-Hin­ton broth that con­tains about 500 µmol per liter of both Mg²⁺ and Ca²⁺. I also note that the MOPS medium of Nei­d­hardt, per­haps the only ra­tio­nally-de­signed min­i­mal medium, con­tains more than 500 µmol of Mg²⁺ per liter. I hope that my col­leagues will pay more at­ten­tion to the avail­abil­ity of di­va­lent cations in the cul­ture medium.

Fi­nally, one must point out that LB me­dia sold as pre­mixed cap­sules by cer­tain com­pa­nies con­tain bile salts as a con­t­a­m­i­nate of the "beef" prod­uct present in the medium. The con­t­a­m­i­na­tion is suf­ficient to pre­vent the growth of mu­tants that lack the mul­tidrug ef­flux pump (AcrAB-TolC) of E. coli, one func­tion of which is to pro­tect the cells from bile salts in their en­vi­ron­ment. This unpleas­ant fact speaks of the dif­fer­ences be­tween batches of pre­mixed medium. Un­less made from scratch, LB broth can vary from batch to batch, thus can­not be con­sid­ered re­pro­ducible.

The bot­tom line is that, from a mul­ti­tude of per­spec­tives, the use of LB medium is to be discour­aged, es­pe­cially for use in any stud­ies in which the phys­i­o­log­i­cal state and meta­bolic func­tions of the cell mat­ter. How­ever, I must con­fess that I have been guilty of com­mit­ting this sin many times in my ca­reer.

Hi­roshi Nakaido is PBD Fac­ulty Sci­en­tist, Struc­tural Bi­ol­ogy De­part­ment, and Pro­fessor of Bio­chem­istry and Mol­e­c­u­lar Bi­ol­ogy, UC Berke­ley.