Construction of Plasmids and Strains. The integrative plasmid (pHPV472) from which LacI-CFP and TetR-YFP were expressed in Caulobacter was constructed by cloning a 3.2-kb BamHI/HindIII fragment harboring the coding sequence for LacI-CFP and TetR-YFP from pLAU53 (1), along with a 2.3-kb NheI/BamHI fragment harboring the xylX promoter (P_xylX) from pXGFP4 (M. R. K. Alley, unpublished work) into SpeI/HindIII-restricted pHPV465. Plasmid pHPV465 is a pXGFP7C1-derived integration vector (M. R. K. Alley, unpublished work) containing the RP4 oriT, a spectinomycin/streptomycin resistance determinant and the pUK21 (2) polylinker. To create pHPV465, the polylinker was released from pUK21 by using SpeI/BglII and was cloned into SpeI/BamHI-restricted pXGFP7C1. To construct plasmids for labeling the origin of replication with (lacO)_n or (tetO)_n, ORF CC0006 was amplified by PCR, adding a NdeI restriction site at the 5' and a XbaI site at the 3' end of the fragment, and was cloned into the corresponding sites of vectors pLAU43 and pLAU44 (1), respectively, yielding plasmids pMT56 and pMT57. The resulting plasmids were transformed into CB15N by electroporation (3). Clones in which the plasmids had integrated by single homologous recombination onto the chromosome were selected and were transduced with ФCr30-dervived lysates (3) from strain PV2653, a CB15N derivative carrying the expression plasmid pHPV472 integrated at the xylX locus, yielding strains MT15 [ori::(lacO)_n; xylX::pHPV472] and MT16 [ori::(tetO)_n; xylX::pHPV472].

To create the two mariner plasmids pHPV499 and pHPV560, the RP4 oriT was first isolated from pPM927 (4) by restriction with PstI, and was blunted and subcloned into EcoRV/SmaI-restricted pOK12 (2). The resulting plasmid, pHPV412, was then cleaved with SpeI to liberate the oriT fragment, which was subsequently cloned into the XbaI site of pMiniHimar-LacZ (H. B. Kaplan, unpublished work), creating pHPV414. The arrays of operators, (lacO)_n or (tetO)_n, were then cloned into pHPV414. First, the Ωhyg cassette was isolated from pHP45Ωhyg (5) by restriction with BamHI, and was blunted and ligated into the (lacO)_n-bearing plasmid pLAU43 (1) that had been digested with SmaI/NsiI and was blunted with T4 DNA polymerase, creating pHPV534. The arrays of operators, (lacO)_n or (tetO)_n, were liberated from pHPV534 or pLAU44 (1), respectively, by restriction with NheI/XbaI and were ligated separately into the XbaI site of pHPV414, creating pHPV560 and pHPV499. To create the mariner strains, transposition was induced by mobilizing the mariner (tetO)_n and (lacO)_n plasmids, pHPV499 and pHPV560, from S17-1/ l pir into CB15N and selecting for kanamycin-resistant Caulobacter clones. A generalized transducing lysate was made from the pHPV499- and pHPV560-derived pools of kanamycin-resistant clones, and the mariner insertions were transduced into MT15 and MT16, selecting for gentamycin and kanamycin resistance, respectively. We also enriched for mariner insertions in the pilA-cpaABCDEF locus, encoding the structural components of the polar pili and their assembly machinery, by selecting for mariner strains that confer resistance to the pilus-specific bacteriophage ΦCbK. These insertions comprise mariner strains 20-40 and 60-70. No significant growth impediments were observed when the mariner strains were grown in the absence of xylose.

Additional strains harboring (lacO)_n insertions in the origin-proximal region of the chromosome were created by using a strategy based on homologous recombination of pLAU43-derived plasmids. They carry MPO prefixes to distinguish them from the transposon-derived strains. Suitable cosmids from an ordered C. crescentus chromosomal library [kindly provided by M. Laub (Harvard University, Cambridge, MA) and C. Stephens (Santa Clara University, Santa Clara, CA)] in pLAFR5 were digested with EcoRI, BamHI, HindIII, SacI, PstI, XhoI, and KpnI to fragment the inserts. Fragments of 1-3 kb were isolated, were blunted with T4 DNA polymerase, and ligated into the blunted NdeI site of pLAU43. Plasmids bearing an insert were identified by PCR and were subsequently transformed into CB15N by electroporation. Cointegrants were identified by selection for kanamycin resistance, were transduced with lysates from PV2653, and were used, in part, for the time-lapse FM experiments shown in Fig. 4B in the main text. All strains were subsequently transduced with lysates from MT16, selecting for gentamycin resistance to introduce the ori::(tetO)_n insertion into the strains already harboring (lacO)_n at an ori-proximal position. These strains along with those harboring the mariner insertions were analyzed in Fig. 2B in the main text.

Targeted insertions of (lacO)_n into the pilA, pleC, and podJ locus on the Caulobacter chromosome were made by transforming CB15N to kanamycin resistance by using pHPV459, pHPV484, or pHPV489, respectively. The resulting strains were transduced with lysates from PV2653, selecting for spectinomycin/streptomycin resistance, and subsequently with lysates from MT16, selecting for gentamycin resistance to transduce the ori::(tetO)_n insertion. To create pHPV459, pilA was amplified by PCR, was subcloned into pCR-Blunt II-TOPO (Invitrogen), and was ligated into the EcoRI site of pLAU43. To create pHPV484, a 710-bp fragment harboring the 3' end of CC2483, the gene downstream of pleC (CC2482), was amplified by PCR, was subcloned into pCR-Blunt II-TOPO, and was ligated into the EcoRI site of pLAU43. To create pHPV489, a 450-bp fragment harboring the 3' end of podJ was amplified by PCR, was subcloned, and was ligated into the EcoRI site of pLAU43.

The podJ probe used in FISH experiments was from pHPV58. To create pHPV58, a 5.2-kb NdeI/BamHI fragment was isolated from cosmid K28 (6), was blunted, and was cloned into the SmaI site of pJS71 (7).

Transposon Mapping. The mariner transposon derivatives used here deliver the E. coli R6K origin of replication along with the kanamycin resistance gene and either (tetO)_n or (lacO)_n into the Caulobacter chromosome. Thus, a fragment of the transposon containing the R6K origin, the kanamycin resistance determinant, and Caulobacter genomic DNA that lies adjacent to the site of the insertion can be retrieved by restricting genomic DNA with SacII, followed by self-ligation and transformation of an E. coli host expressing the l pir protein (E. coli EC100D^TM pir-116) to kanamycin resistance. The mariner insertion site was determined by sequencing across the junction of the transposon into Caulobacter chromosomal DNA with primer Him_up (5'-GAACTATGTTGAATAATAAAAACGA-3') of the retrieved plasmids.

Image processing. To process phase contrast and FM images, we used the MATLAB Image Processing Toolbox. For analysis of G₁ (swarmer) cells by FROS and FISH, three images were taken: (i) a phase contrast image to find and classify the cells and define the cell outline (the "mask", see below) for each cell, (ii) an image of the fluorescently labeled ori to identify the flagellated pole of the G₁ cells, and (iii) another fluorescence image using the appropriate filter to identify the intracellular position of the second fluorescently (mariner)-labeled chromosomal locus. The (lacO)_n- or (tetO)_n-marked origin, as well as the McpA chemoreceptor served as markers for the flagellated pole of the cells in FROS and FISH/IF experiments, respectively. In the time-lapse microscopy experiments, only a phase contrast picture and a fluorescence picture were acquired at each time point because the strains used only harbored (lacO)_n. In this case, the flagellated end of the cell was identified by inspection of the phase contrast images as the pole at which the stalk grew later in the cell cycle.

Cell identification. Phase contrast pictures were corrected for uneven illumination. The average value and SD for the intensity of the phase contrast picture were calculated. Thereafter, the mask for the cell was found by grouping all adjacent pixels that had an SD value of 1.5 lower than that of the average value. Only the resulting cell outlines that met the length, width, and area constraints were included in further analysis.

Cell length and fluorescent focus measurements. The cell length was determined as follows: (i) The centroid of the cell was found. (ii) The two pixels on opposite sides of the cell mask that were furthest from the centroid were identified and a line was drawn between these two points. Another line was drawn perpendicular to the previous line at the midpoint between the two cell extreme points. This line intersected the cell mask edge at two points; the midpoint between these two intersecting points was marked as the center of the cell. (iii) The points on opposite sides of the cell furthest from the center of the cell were found and marked as the two cell poles. An arc was then drawn through the two cell poles and the center of the cell. The length of this arc was designated as the cell length. (iv) The distance that each fluorescent spot was located from a pole was found by projecting the fluorescent focus on the cell arc and measuring the subarc created by the projected spot and the cell pole. The fluorescent focus was identified by aligning cell images with the masks. The local maximum in the interior of each cell mask that was greater than 5 SD above background was considered a fluorescent focus.

G₁ cell analysis. Each mask that contained one marker (ori-derived) focus and one mariner-derived focus was analyzed. Typically, for each strain several hundred masks met this criterion, taken from approximately six sets of images. The flagellated pole was identified as the cell pole closest to the ori focus. The distance to the other focus was measured from that cell pole. This number was divided by the cell length to find the relative (fractional) cell length position. The average and SD of all relative cell positions were calculated. Cells with relative positions greater than 3 SD away from the mean were excluded. Typically, less than 2% of cell masks were excluded. The mean and SD were then recalculated. All strains with greater than 50 cell masks and an SD of less than 0.11 were included in Fig. 2B in the main text. From a total of 127 mariner strains that were analyzed, 15 strains did not meet this requirement and were excluded from further analyses (see Table 1). We suspect that in many of these strains the mariner insertion has impaired or altered the function of one or several genes located near the site of insertion, resulting in pleiotropic effects on cell morphology and/or chromosome arrangement that, in turn, would give rise to a high SD. An alternative possibility is that the mariner insertions are unstable in some strains or that strains have acquired more than one mariner insertion by superinfection with two different generalized transducing phage particles or by transduction from a single particle that contained two mariner insertions located in the vicinity of each other. Finally, we cannot exclude the formal possibility that few chromosomal loci are not consistently located at a defined subcellular position.

Time-lapse analysis. Cells were aligned and analyzed for fluorescent foci as described above. Cells with discrete fluorescent foci in most of the time points were chosen by eye, and the stalked pole was identified visually. Focus positioning was also verified by inspection, and foci not recognized by the algorithm were chosen by hand. All cells with usable masks and loci markers over the period of observation, typically 15-25 cells per strain, were grouped together. The average time point at which the transition from one focus to two foci was observed (reflecting the time at which segregation has begun) was calculated for each strain. An offset time was added to each individual cell to obtain a population of cells for each strain that has a synchronized pattern of focus movement relative to the average time point at which focus separation occurred. The cell length, the position of the "old" and the "new" focus were then averaged between all of the cells at each time point to create the data shown in Figs. 3B and 4 A and B in the main text. Average rates of movement of foci in each cell were determined by identifying the first time point the "new" focus crossed the midplane of the cell. This result was compared with the last time point at which the cell had only one (the "old") focus. The distance covered between these two pictures divided by the difference in time between these two pictures was defined as the rate of movement for that cell.

1. Lau, I. F., Filipe, S. R., Soballe, B., Okstad, O. A., Barre, F. X. & Sherratt, D. J. (2003) Mol. Microbiol. 49, 731-743.

2. Vieira, J. & Messing, J. (1991) Gene 100, 189-194.

3. Ely, B. (1991) Methods Enzymol. 204, 372-384.

4. Smokvina, T., Mazodier, P., Boccard, F., Thompson, C. J. & Guerineau, M. (1990) Gene 94, 53-59.

5. Blondelet-Rouault, M. H., Weiser, J., Lebrihi, A., Branny, P. & Pernodet, J. L. (1997) Gene 190, 315-317.

6. Crymes, W. B., Jr., Zhang, D. & Ely, B. (1999) J. Bacteriol. 181, 3967-3973.

7. Skerker, J. M. & Shapiro, L. (2000) EMBO J. 19, 3223-3234.