Agrobacterium tumefaciens – the tumour causer

Agrobacterium tumefaciens is a bacterium which, with the help of the Ti megaplasmid, manipulates plants into providing it with a safe and nutritious environment. The discovery of this interaction has been important not only for understanding the natural history of the bacterium, but for its use as a tool to transform plants. This has both allowed better understanding of plant biology and allowed genetic modification for agriculture.

The bacterium was first isolated by Erwin Smith. He was investigating a daisy with Crown Gall disease and found he could isolate a bacterium which produced a slimy exudate. He inoculated healthy plants with the bacterium and found that they developed tumours. This property gave the bacterium its species name – the tumour causer.

In the 1930s, Braun characterised the bacterium further. He took a tumorous plant and heat killed the bacteria by incubating it at 65° C for three days. The plant cells survived and so cells from the tumour could be grafted onto healthy plants. They continued to divide and formed new tumours. This indicated that the bacteria had passed some determinant to the plants, the Tumour Inducing Principle, rather than constantly producing some chemical to cause the tumours.

Research somewhat stalled here until the growth of molecular biology in the 1970s. Gel electrophoresis allowed the detection of plasmids in bacteria at this time, and scientists attempted to isolate them from Agrobacterium. They found that strains causing Crown Gall disease had a 250 kb megaplasmid. This property gave the plasmid its name the Ti plasmid (tumour inducing).

What is transferred?

Mary Dell Chilton prepared RNA from tumour cells, which she labelled and used to probe restriction fragments from the Ti plasmid. They hybridised, demonstrating that DNA is present in plant cells and transcribed into RNA.

Forward genetic mutant screens helped to shed light on this transferred (T-) DNA. Normally infected plants form tumours, this is due to high expression of two hormones. Cytokinin and auxin. In tms mutants less auxin was produced triggering shoot production. Conversely in tmr mutants less cutokinin was produced causing roots to form. When these loci were cloned it was shown that they coded for proteins involved in the biosynthesis of their respective hormones.

Another fascinating feature of the T-DNA is that is codes for the synthesis of a protein synthesing a particular opine. The exact opine depends on which strain of bacterium is involved. The non-transferred portion of the Ti plasmid codes in part for the metabolism of the respective opine. This means that any bacterium which loses the plasmid will no longer be able to benefit from resources produced by the plant.

What triggers DNA transfer?

In a mutant screen a series of virulence genes were identified. They were named vir A-G. These were probed by transforming the bacterium with a merodiploid reporter vector.

Analysis showed that virA was expressed constitutively, but that the other vir genes required plant washate for their activation. The inducers involved were wound phenolic compounds – acetylsyringone was found to be especially powerful. This allowed the discovery of an elegant regulatory mechanism involving a positive feedback loop. virA and virG are needed to activate the operon but do so only upon activation by a wound phenolic compound. This increases the rate of virG expression from a very small level and so in turn allows faster activation of the operon. This allows the virulence genes to be expressed in an all-or-nothing fashion.

What is the mechanism of transfer

Mary Dell Chilton sequenced the regions on either side of the T-DNA and found a 25bp repeat in direct orientation. Patty Zambryski showed that these borders were crucial for virulence. They attempted a Southern blot with a probe for T-DNA on induced Agrobacterium cells and found that they could identify this DNA even if they did not denature their sample before blotting, indicating the T-DAN was single stranded. They proposed a conjugation like mechanism. virD1D2 forms a nuclease which cuts at the borders. It is coated in the virE2 product and targeted to the plant through a type IV secretion system (the same type used by Heliobacter pylori). Once inside the plant importins target the protein DNA complex to the nucleus where it inserts.

DNA repair

Each of us has around 10^13 cells in our body, a mutation in just one of these can lead to a cancer. Thus it is crucially important that each cell has an advanced system for the detection of DNA damage and the control of appropriate response measures. These may be repair, via a number of pathways dependent on the type of damage, or senescence or even apoptosis when damage is severe.

Extent of damage

As many as one million lesions may occur per day in a single human cell. Damage can occur from endogenous sources, such as reactive oxygen species produced by metabolism. But it also created exogenously, by ultraviolet light and other ionising radiation, mutagens and viruses.

These can cause

  1. Oxidation of bases
  2. Alkylation of bases
  3. Hydrolysis – e.g. deamination, depurination, depyrimidation
  4. Formation of ‘bulky adducts’
  5. Mismatched bases

Every time a cell undergoes division its telomeres shorten. When these reach the Hayflick limit, the cell goes into senescence, an irreversible dormant state.

DNA repair

Direct reversal

  • Pyrimidine dimers, which are formed by UV irradiation, are repaired by photolyase on its activation by UV (photoreactivation).
  • Guanine methylation is reversed by methyl guanine methyl transferase
  • Some C and A methylation can also be directly reversed
  • The adaptive response in E.coli, upon alkylation of DNA backbone. Ada suicidally transfers alkyl groups onto its cysteine residues (not catalysis, stoichiometric). Methylation of ada -> activation of itself, + other alkylation responsive genes.

Single strand damage

  • Base-excision repair (BER) after oxidation, alkylation, hydrolysis or deamination
    Damaged base is removed by DNA glycosylase
    AP endonuclease recognises “missing tooth” and cuts phosphodiester bond, allowing resynthesis by DNA polymerase. Then ligase seals backbone.
  • Nucleotide excision repair (NER) recognises bulky lesion s which distort the helix (e.g. thymine dimers). Requires Uvr system in E. coli, = UvrABC + DNA Helicase II. UvrA-UvrB scans DNA, UvrA recognises distortions, leaves complex to be replaced by UvrC. UvrB cleaves bond downstream of DNA damage while UvrC cleaves upstream. Helicase II breaks hydrogen bonds to remove excised segment. The gap is filled by DNA Pol I.
  • DNA mismatch repair (MMR) is strand-specific, the daughter strand is recognised as the one to be corrected. Three proteins are involved mutS, mutH and mutL. mutS forms a dimer that recognises the daughter strand’s mismatched base and binds, mutS recruits mutL dimers. MutH binds to hemimethylated sites, and is activated by mutL. It nicks the daughter strand and recruits DNA helicase II. MutSHL then travels behind the helix along the single strand it creates towards the mismatch. An exonuclease digests the tail left behind. The process ends past the mismathed site and PolIII can then repair the daughter strand.

Double strand breaks

  • NHEJ – DNA Ligase IV uses microhomology to rejoin two ends (also used in VDJ recombination. Ku?
  • Micro-homology mediated end joining (MMEJ) uses the Ku protein too
  • Homologous recombination-uses sister chromatid/ homolgous chromosome


I’ve recently become a fan of the BBC gameshow OpenConnect. I’ve raced through the walls that can be played on the BBC website but now that I have run out I thought I would make some more and launch a version to which other people could submit.
Visit PuzzGrid!

Bacteriophage Lambda


Infection begins when a phage adsorbs to the lamB gene product, a receptor in the outer membrane of E. coli. The phage stores its DNA under pressure and so it can be injected into the cell, with the help of a cellular protein – ptsM. The DNA in the phage’s head is linear and double stranded but it has 12 base complementary single strang overhangs on each 5’ prime end, rich in G and C. These “sticky ends” mean that the DNA rapidly circularises. In most circumstances cII levels in the cell will not be high and so the phage will enter the lytic cycle typical of bacteriophages.

Lytic cycle

Immediate early (IE) stage

The phage’s lytic cycle uses three promoters PL, PR and PR’. These need to be strong promoters so that they can recruit the host polymerase. The host polymerase binds preferentially to these promoters to such an extent that if many copies of the promoters are artificially inserted the cell is unable to transcribe its own DNA and dies. PL promotes a transcript which initially terminates after transcription for N, an anti-termination protein. Transcription from PR produces cro, which is important in continuing the lytic cycle rather than lysogeny (see cI section). The PR’ transcript at this stage only produces a small inactive protein.

Figure 1: Immediate early stage
(click on any diagram to enlarge)

Delayed early (DE) stage

The immediate early stage produces the anti-termination factor N. There are areas around the PL and PR promoters called N utlilisation sites or nut sites. They cause N to bind to the transcriptional apparatus at the promoter. This prevents some of the the early termination that occurs in the immediate early stage. Most importantly for lysis, it allows the transcription of O, P and Q. O and P are needed for DNA synthesis to occur – they recruit the host’s machinery for DNA replication to a point near their locus. P prevents the host from synthesising bacterial DNA, working with cII. The machinery produces a number of copies of the phage’s circular DNA, each of which is used in rolling circle replication to produce around 10 phage genomes. Q is an anti-termination factor needed for progression to the late stage. Although int is part of the PL transcript, integrase is not synthesised because the transcript also contains the sib region which folds in a hair-pin loop and results in this section of the transcript being cleaved by RNaseIII.

Figure 2: Delayed early stage

Luzzati first showed that N was an anti-termination factor. He used a mutant lysogen which had a defect in N and a temperature sensitive transcriptional repressor which allowed transcription from PL and PR at 30ºC but not 40ºC. He superinfected the lysogen with a phage providing N at 30ºC but no transcription of the DE region of the original prophage took place. It was not until the temperature was raised to 40ºC that transcription began from PL and PR and was allowed to continue without termination by N, producing both IE and DE products. This proved that N did not create new promoters around the DE region but acted as an anti-termination factor.

Late stage

Once replication of the phage DNA has been initiated the other components of the new phages, heads and tails, are needed, as well as proteins for lysis. Q extends the PR’ transcript by binding to the transcriptional apparatus at a Q utilisation (qut) site close to the promoter. Once Q is bound termination does not occur as easily and so S, R and a number of genes coding for the heads and tails of the phage particles are transcribed. The head and tail proteins self assemble. R is an endolysin which is needed to cleave the cell wall for lysis.

Figure 3: Late stage

S produces mRNA with dual start sites for translation: one of these start sites produces S105, a holin; the other produces S107 which has an additional basic residue and acts as an antiholin because it cannot interact with the intact membrane correctly. Holin is produced twice as frequently as antiholin but binds preferentially to the anti-holin. This means that half of the holin molecules are disabled by antiholin. It is believed that this mechanism allows the sudden and rapid creation of many holes after a delay. Initially many S products are produced but there are few functional holin-holin dimers, not enough to create a single hole. When enough holin-holin dimers aggregate to form a hole in the membrane, its properties change and allow S107 to function as a holin. This immediately triples the number of S products available for hole creation, allowing rapid lysis.


λ has an alternative lifestyle to the violence of lysis. It can integrate into the host’s genome and form a stable lysogen which continues to grow and divide. This is a stable state, even superinfection by the same phage willnot lead to lysis. There are some related phages which are heteroimmmune (e.g. 21 and 434); these can infect lambda lysogens and trigger lysis. Various experiments have been done to hone in on the ‘immunity substance’that provides protection from self superinfection.

Hfr strains of E. coli lysogenic for λ can be mated with wild type bacteria. They do not form recombinants, the Hfr strain transfers it’s entire genome and the prophage becomes lytic (zygotic induction). If the wild type cells are first made lysogenic then recombinants are formed. This demonstrates that the immunity substance is a cytoplasmic factor. In another experiment, λ is added to a population of E. coli cells on a petri dish. A turbid plaque is formed for wildtype λ which is made up of lysogens and periodically some lysis. Mutations in some genes (cI, cII and cIII) give clear plaques, meaning that all cells have gone into lysis. Every one of these factors is needed to establish lysogeny. Using temperature sensitive mutations to each gene it has been shown that only cI is needed to maintain lysogeny once established. It is the repressor, or immunity substance. A very few mutants are unable to establish lysogeny even in the presence of cI and always go into lytic cycle, these are dubbed virulent.

cI was isolated using a double labelling experiment. Wild type phage protein was labelled with one radioactive amino acid marker and protein from a cI strain was labelled with another. Proteins were pooled and separated by chromatography. The protein fraction which only had the wildtype label was cI. This was confirmed by the fact that it did not bind to the DNA of virulent mutants.


The ultimate factor which determines whether λ will go into a lytic or lysogenic cycle after infection is the level of cII in the cell. If cII levels are high enough a lysogen will be formed, otherwise a lytic cycle will ensue. But cII is broken down by ftsH, a protease. cIII inhibits ftsH by binding to it competitively and so allows cII levels to remain high. This is why both cII and cIII are essential for the establishment of lysogeny.

Figure 4: Initial establishment of lysogen – cII active but low cI

cII is a transcriptional activator which acts on three promoters, PRE, Pint and PQ’. PRE promotes a transcript which produces cI and also the antisense to the normal mRNA to cro (which is being transcribed as usual in the delayed early stage.) This is important as it binds to the sense mRNA and prevents translation. PQ’ performs the same task for Q. This is necessary to stop the extension of the PR’ transcript which would send the cell into lysis. Pint has a short transcript which is translated to an integrase. Since it is not affected by N it terminates before the sib region and so no hairpin loop forms, so the RNA is not a target for RNase III.


The areas around PL and PR are the operators OL and OR. These allow control of the promoters: each has three sites to which transcription factors can bind. cI forms a dimer and binds best to site 1, then site 2 and worst to site 3. When it binds to OL1 it is well positioned to interact with other cI dimers to allow them to bind cooperatively to OL2. Once these two sites are bound, transcription from PL is repressed. This is important to prevent xis being produced which would excise the integrated prophage, and generally to avoid a waste of resources on unnessessary products. There is a similar story for OR, with initial binding to OR1 and then cooperative binding to OR2. But here the binding to OR2 activates another promoter PRM in addition to repressing PR. Repression of PR stops the production of Cro and Q, removing the dependence of the phage on PRE and PQ’ (and hence cII) and prevents O and P from initiating replication on the integrated genome.


Figure 5: Lysogenic state – cII levels now low, cI levels at optimum

PRM is the promoter for the maintenance for lysogeny. Its transcript produces cI, which we have established is the only factor needed to maintain the lysogenic state. In lytic state thispromoter is repressed by the action of cro which cannot bind cooperatively and binds best to OR3.This is why cro levels must be kept down for lysogeny. At very high concentrations in lytic cells, cro will bind to OR2 and OR1 to regulate its own production by negative feedback. cI has a similar regulatory mechanism, at high concentrations it will bind to OR3 and so repress PRM.

Integration and induction

The integrase produced by the Pint transcript results in the insertion of the prophage at attB, which lies between the galactose catabolic operon and the biotin biosynthetic operon. The inserted prophage is a circular permutation of the phage DNA in the particle, with int on one side and the sib region on the other.Once inserted the prophage will be replicated with the rest of the bacterial genome and so lie dormant, being passed into all the bacterial progeny.

In general the lysogenic state is observed to be very stable, 1 in 105 cells would normally go into lysis in a culture. This process is called induction. Lysis can be reliably induced by exposure to UV light. This damages DNA, which is one of the triggers for the cell’s SOS response – a postreplication mechanism for DNA repair.The genes for the SOS response are repressed in normal cells by a dimer called lexA. In the SOS response, recA is made active by binding to single stranded DNA and cleaves lexA, removing the repression.But cI has the same motif that makes lexA a target for cleavage and so in lysogens it is cleaved too.

cI levels fall and so cI no longer binds to the operators; PR and PL are no longer repressed. PRM is no longer activated so no further cI can be made. The situation now looks very like Figure 1 except that the DNA is now part of the bacterial genome and not a small loop of phage DNA. The same process happens – N istranscribed which allows the elongation of the PR and PL transcripts. The PL transcript no longer has the sib region because of the circular permutation, and so no hairpin loop is formed and RNase III attack does not occur. This means that xis (and int) are expressed. Both these factors and a bacterial gene product called fis are needed to allow excision. The excised circular DNA proceeds as in the lytic cycle.

Figure 6: Induction of a lysogen – cI has been cleaved

Lysis or lysogeny?

Not all the factors that influence the levels of cII – and so the fate of an infected cell – are understood. The nutritional status of cells is certainly important and there are proteases other than ftsH which can breakdown cII and are regulated by the cell cycle. Although a model organism for the study of bacterial infection, λ still holds its mysteries.