Efficient formation of single-copy human artificial chromosomes

Abstract

Large DNA assembly methodologies underlie milestone achievements in synthetic prokaryotic and budding yeast chromosomes. While budding yeast control chromosome inheritance through ~125-base pair DNA sequence-defined centromeres, mammals and many other eukaryotes use large, epigenetic centromeres. Harnessing centromere epigenetics permits human artificial chromosome (HAC) formation but is not sufficient to avoid rampant multimerization of the initial DNA molecule upon introduction to cells. We describe an approach that efficiently forms single-copy HACs. It employs a ~750-kilobase construct that is sufficiently large to house the distinct chromatin types present at the inner and outer centromere, obviating the need to multimerize. Delivery to mammalian cells is streamlined by employing yeast spheroplast fusion. These developments permit faithful chromosome engineering in the context of metazoan cells.

Fig. 1.. 760 kb HAC constructs efficiently acquire centromeres and exist as autonomous chromosomes.

(A) Schematic of approach to generate a HAC. (B) Representative images of a single-copy HAC generated in HT1080 and U2OS cells. Insets: 5x magnification. Bar, 5 μm. See also Table S1. (C) Quantification of proportion of “small HACs” (FISH signal spans less than 1 μm), “large HACs” (FISH signal spans greater than 1 μm), “integrations” and “no signal” spreads generated from HAC formation assays. The mean (+/− SD) is shown. (D) Comparison of size of a HAC made from YAC-Mm-4q21^LacO and a multimerized HAC made from 4q21 BAC^LacO. Both HACs are shown at the same scale. Coloring is identical to what is shown in the merged images in panel B. Bar, 1 μm.

Fig. 2.. YAC-Mm-4q21^LacO-based HACs are inherited as autonomous chromosomes with functional kinetochores and robust CPC recruitment.

(A) Representative image of a single copy HAC that has been isolated in a monoclonal cell line. Inset: 5x magnification. Bar, 5 μm. (B) Quantification of fraction of spreads with a HAC in monoclonal cell lines. The mean (+/− SD) is shown. Colors indicate individual HAC lines that were assessed in HAC retention assays. Red, orange, green and blue correspond to HAC clones 1, 2, 3, and 5. (C) Quantification of HAC loss rate after culturing without selection for 30 days. The mean (+/−SD) is shown. Experiments are color coded to correspond to the clones shown in panel B. Grey shading indicates the range of loss rates for prior generations of HACs (12, 16, 29). (D) Representative image of HACs synchronized in mitosis showing Aurora B and ACA. The image shows 8 0.2 μm z-projected stacks (see also Fig. S4 for centromere delineation in the z-dimension). Inset: 5x magnification. Bar, 5 μm. (E) The radial position of HACs was measured relative to endogenous centromeres. The position of 20 HACs, each endogenous centromere and the center of DNA mass was measured. The distance between HAC or endogenous centromere and the center of DNA mass was calculated. The distance of each HAC from the center was normalized based on the total length across (i.e. the diameter) of mitotic chromosomes. The inner black circle represents the mean radial position of endogenous centromeres, while the dotted line represents one standard deviation from the mean. An illustration is shown below the graph.

Fig. 3.. YAC-Mm-4q21^LacO HACs are functional as single copy DNA.

(A) Schematic of approach used to enrich single copy HACs. (B) Representative image of HACs isolated by sucrose gradient with either a single or two foci of LacO. The proportion of HACs with a single or two foci is noted. HACs with two LacO foci also had two CENP-A foci suggesting that they are mitotic. Bar, 1 μm. (C) Representative image of a multimerized HAC (Clone 27 from (16)) from mitotic chromosome spreads. Bar, 1 μm.

Fig. 4.. YAC-Mm-4q21^LacO-based HACs are intact 760 kb circles with similar chromatin stretching properties as natural chromosomes.

(A) Southern bot analysis of the indicated HAC lines using a LacO probe. The multimerized HAC line was diluted 20-fold in HT1080 cells to reduce signal relative to the YAC-Mm-4q21^LacO-based HAC so that they could be in a similar range of detection. (B) Schematic showing extent of stretching HACs in our experiments (panels D-H), with indicated regions detected by FISH. The number of foci shown is in the range predicted by prior stretching experiments with natural chromosomes, with actual outcomes measured in panels C-G and fig. S6. (C) Representative images of a stretched YAC with both the 4q21 and M. mycoides sequence labeled via FISH compared to endogenous 4q21 in asynchronous cells. Bar, 1 μm. (D) Representative images of an unstretched and stretched HAC with both the 4q21 and M. mycoides sequence labeled via FISH compared to endogenous 4q21 in asynchronous cells. Bar, 1 μm. (E) Quantification of the length of 4q21 FISH in the YAC, HAC and the endogenous chromosome after stretching chromatin in asynchronous cells. The mean (+/− SD) is shown. p< 0.001 and p < 0.05 based on an unpaired, two-tailed t-test. (F) Quantification of the number of foci from 4q21 FISH in the YAC, HAC and in the endogenous chromosome after stretching chromatin in asynchronous cells. The mean (+/− SD) is shown. p values > 0.05 based on an unpaired, two-tailed t-test and is marked as not significant (n.s.). (G) Representative images of a stretched HAC with both the 4q21 and M. mycoides sequence labeled via FISH after enriching for cells in metaphase. Bar, 1 μm. (H) Quantification of the number of foci from 4q21 FISH in the HAC and the endogenous chromosome after stretching chromatin and enriching for cells in metaphase. The mean (+/− SD) is shown. p < 0.0001 based on an unpaired, two-tailed t-test. (I) Model illustrating how construct size influences HAC formation outcomes. Single-copy BAC-based HACs have never been isolated and are shown as too small to harbor a functional inner centromere.