Bacterial Hybrids¶
This work is licensed under a Creative Commons Attribution 4.0 International License.
Section 5, session 30, the last session in the schedule. The reading for the day is Judson, Chapter 9: Theory.
The problem
This statement is the editors' reconstruction. Winfree's own problem sheet does not survive (see the note at the foot of the page). The numbers were made up for this page; they are not anyone's lab data.
Two strains of a microbe, A and B, cannot grow on a bare nutrient plate. Strain A cannot make two substances it needs; strain B cannot make two different ones. Each strain also carries two traits you can read from a colony: Lac, whether it ferments lactose (Lac+) or not (Lac−), and Vir, whether it resists (Virʳ) or is killed by (Virˢ) a certain virus.
| needs | Lac | Vir | |
|---|---|---|---|
| strain A | substances 1 and 2 | Lac+ | Virˢ |
| strain B | substances 3 and 4 | Lac− | Virʳ |
Experiment 1. Two billion cells of A alone on a bare plate: nothing grows. B alone: nothing. Two billion of each, mixed, washed and spread: 312 colonies.
Experiment 2. The 312 colonies, scored for Lac and Vir:
| colony type | count |
|---|---|
| Lac+ Virˢ (like A) | 128 |
| Lac− Virʳ (like B) | 141 |
| Lac+ Virʳ | 21 |
| Lac− Virˢ | 22 |
Experiment 3. A third trait, Gal (galactose fermentation), is added. The cross is repeated with strains that differ in each pair of traits, and each time the share of colonies with a mixed combination (one trait from each parent) is counted: Lac and Vir, about 14%; Lac and Gal, about 5%; Gal and Vir, about 9%.
Your task.
- List every explanation for the 312 colonies, with an observation that would rule each out.
- What would Experiment 2 show if the two traits were inherited independently? What does it actually show?
- What relation do the three percentages of Experiment 3 obey, and what picture does that relation force on you?
- Do the cells pass material by touching, or through the broth? Design one apparatus that decides.
- Predict one thing your explanation says must happen that nobody has looked at yet.
Experiment 1 as three plates. Schematic: the dots stand for colonies and are not a count. Drawn for this site (CC BY 4.0).
Why it is in the course¶
Section 5, "Inferences, Hypotheses, Explanations", opens with Chamberlin's The Method of Multiple Working Hypotheses (session 25) and Platt's Strong Inference (session 27). It ends in session 30 with Judson's chapter on theory and this problem, the last one in the schedule. The version reconstructed here asks for both readings on one set of numbers.
Colonies where there should be none have several honest explanations. "Facts before explanations of facts" (the syllabus's phrase for session 04) means listing them all before choosing one. Then traits nobody selected narrow the choice, and three measurements obey a relation that none of them shows alone. No biology is needed. The syllabus says the exercises "depend as little as possible on knowledge of any particular subject area".
Where it comes from¶
The phrase "bacterial hybrids" was already in print in 1924, in the title of a paper by E. Almquist. Until the mid-1940s, though, bacteria were widely assumed to lie outside genetics: no visible chromosomes, no sex, nothing to cross.
In 1946 Joshua Lederberg, then a graduate student, and Edward Tatum mixed two strains of Escherichia coli K-12, each unable to make several substances it needed. They recovered rare colonies that grew where neither parent could. What those colonies meant was disputed for years. Lederberg shared the 1958 Nobel Prize in Physiology or Medicine for this work, and in 1996 he looked back on how sharply it had been contested. How the dispute was settled is in the Resolution below.
Escherichia coli at about 10,000×. Photo by Eric Erbe, digital colourization by Christopher Pooley, USDA Agricultural Research Service. Public domain, via Wikimedia Commons.
Hints
- First explain the two plates that grew nothing. What do they rule out?
- Independent traits behave like two coin flips. What counts would that give?
- Add two of the percentages and compare with the third.
- For question 4, you need a vessel where broth is shared but cells are not.
Resolution
Linkage. Independence predicts about 78 in each class. Instead the parental types make up 269 of 312 (86%): Lac and Vir usually travel together. They are linked. In 1947 Lederberg reported this kind of linkage among unselected traits, including lactose fermentation and phage resistance.
Order. 5 + 9 = 14, and no other pairing works. Frequencies that add like distances lie on a line, with Gal between Lac and Vir. What passes between the cells is an ordered arrangement, not a bag of separate factors.
The rivals. The control plates all but exclude back-mutation. Re-streaking one colony alone on a bare plate tests cross-feeding: if it grows by itself, it is a new kind of cell. The diffusible substance fell to Bernard Davis's 1950 U-tube: a sintered glass filter let broth through but no cell, and no recombinants appeared. The cells must touch.
What it turned out to be. Transfer by contact, now called bacterial conjugation. William Hayes showed in 1953 that it runs one way, from a donor to a recipient. Élie Wollman, François Jacob and Hayes then interrupted matings at timed intervals (1956) and saw markers arrive in sequence. The map eventually closed on itself: the E. coli chromosome is a circle.
A caveat. The tidy sum was built into the made-up data; real recombination frequencies add only approximately.
Sources¶
- Arthur T. Winfree, The Art of Scientific Discovery (ECOL 479/579), course handout; the session-30 schedule line and its link markup — Wayback Machine capture, 20 April 2002 🔓
- Joshua Lederberg and Edward L. Tatum, "Gene Recombination in Escherichia coli", Nature 158, 558 (1946) — doi:10.1038/158558a0 🔒
- Edward L. Tatum and Joshua Lederberg, "Gene Recombination in the Bacterium Escherichia coli", Journal of Bacteriology 53(6), 673–684 (1947) — PubMed Central 🔓
- Joshua Lederberg, "Gene Recombination and Linked Segregations in Escherichia Coli", Genetics 32(5), 505–525 (1947) — PubMed Central 🔓
- Bernard D. Davis, "Nonfiltrability of the Agents of Genetic Recombination in Escherichia coli", Journal of Bacteriology 60(4), 507–508 (1950) — PubMed Central 🔓
- William Hayes, "Observations on a Transmissible Agent Determining Sexual Differentiation in Bacterium coli", Journal of General Microbiology 8(1), 72–88 (1953) — doi:10.1099/00221287-8-1-72 🔒
- Élie L. Wollman, François Jacob and William Hayes, "Conjugation and Genetic Recombination in Escherichia coli K-12", Cold Spring Harbor Symposia on Quantitative Biology 21, 141–162 (1956) — doi:10.1101/sqb.1956.021.01.012 🔒
- Joshua Lederberg, "Genetic Recombination in Escherichia coli: Disputation at Cold Spring Harbor, 1946–1996", Genetics 144(2), 439–443 (1996) — PubMed Central 🔓
- E. Almquist, "Investigations on Bacterial Hybrids", Journal of Infectious Diseases 35(4), 341–346 (1924) — doi:10.1093/infdis/35.4.341 🔒 (cited only for the phrase in its title)
- Arthur T. Winfree, The Art of Scientific Discovery, original course syllabus — PDF 🔓
How sure are we that this is Winfree's problem?
The syllabus gives only the name, the session and the reading. In
Winfree's handout, the link on this item points to a bookmark named
Lac_Vir; its target was never archived. Such bookmarks elsewhere name
the subject behind a playful title ("Paired Observations" points to
Keplers_Laws). Lac (lactose fermentation) and virus resistance were
the traits scored in the early E. coli crosses, so the subject is
fairly secure. That reading is the editors' inference, not Winfree's
text. What is lost is the exercise itself: his wording, his data and
its form. Identification: probable.
- Reconstructing the Lederberg–Tatum reasoning (high): the reading used on this page.
- The same episode as a multiple-working-hypotheses exercise (medium): compatible with the first.
- Later E. coli genetics (not weighed): Lac as the lac operon and Vir as a virulent phage such as lambda vir.
- An invented "toy genetics" table (low): does not explain why Lac and Vir are named.
Back to Section 5 · All problems · The schedule
Machine-readable: this page as Markdown · raw source on GitHub · llms.txt · llms-full.txt (whole site). See For AI agents.