Widespread duplications in the genomes of laboratory stocks of Dictyostelium discoideum

Abstract

Background: Duplications of stretches of the genome are an important source of individual genetic variation, but their unrecognized presence in laboratory organisms would be a confounding variable for genetic analysis.

Results: We report here that duplications of 15 kb or more are common in the genome of the social amoeba Dictyostelium discoideum. Most stocks of the axenic 'workhorse' strains Ax2 and Ax3/4 obtained from different laboratories can be expected to carry different duplications. The auxotrophic strains DH1 and JH10 also bear previously unreported duplications. Strain Ax3/4 is known to carry a large duplication on chromosome 2 and this structure shows evidence of continuing instability; we find a further variable duplication on chromosome 5. These duplications are lacking in Ax2, which has instead a small duplication on chromosome 1. Stocks of the type isolate NC4 are similarly variable, though we have identified some approximating the assumed ancestral genotype. More recent wild-type isolates are almost without large duplications, but we can identify small deletions or regions of high divergence, possibly reflecting responses to local selective pressures. Duplications are scattered through most of the genome, and can be stable enough to reconstruct genealogies spanning decades of the history of the NC4 lineage. The expression level of many duplicated genes is increased with dosage, but for others it appears that some form of dosage compensation occurs.

Conclusion: The genetic variation described here must underlie some of the phenotypic variation observed between strains from different laboratories. We suggest courses of action to alleviate the problem.

Figure 1

Relationships between the most commonly used Dictyostelium strains. (a) Simplified genealogical tree showing the relationships between common laboratory strains derived from NC4. The branch marked 'Ax3' is more complex than shown here: sub-lineages have been given the names KAx3 and Ax4. The auxotrophic strain DH1 was engineered in an 'Ax3' background, and JH10 from 'Ax4.' (b) Plaque morphologies. Cells were plated clonally in association with Klebsiella aerogenes on SM agar. Plaques were photographed after 4 days. Small DH1 plaques are indicated with arrowheads. Variation in diameter is a function of the rate of feeding and of the motility of the amoebae. Where the bacteria are cleared the amoebae aggregate in streams; this process had not yet begun in the slow-growing DH1 plaques. (c) Fruiting bodies. Wild type cells - in this instance NC4(Dee) - form larger, more robust fruiting bodies than axenic mutants.

Figure 2

Duplications are frequent in 'wild type' axenic strains. (a-e) Log₂ ratios (each strain compared to the Ax2(Ka) reference) are indicated by vertical lines; array probes are ordered according to their chromosomal location given by dictyBase assembly version 2.5. The previously known Ax3 duplication involves the region of chromosome 2 between approximately 2.25 and 3 Mb, which is wholly contained within the region duplicated in Ax2(Wee).

Figure 3

The distribution of amplifications across the genome. For each chromosome (depicted as arrows, with scale indicating Mb of sequence), different colored bars represent the segments duplicated, approximately to scale. Each feature is named according to the first column of Table 2, in which more precise data concerning size and location are given, along with the strains involved.

Figure 4

A duplication common to Ax2 lines. (a) All Ax2 strains in our study plus selected other strains of NC4 and non-NC4 backgrounds are displayed. Each block is colored according to the log₂ ratio for the comparisons of each strain with reference Ax2(Ka). Since log₂ ratios are consistently greater than zero for the duplicated genes in examples of Ax2 other than the reference, we suggest that this region is amplified further in these strains. The genes plotted are: a, DDB0190411; b, DDB0190412; c, DDB0190413; d, DDB0201787 (probe 1); e, DDB0201787 (probe 2); f, DDB0190415; g, DDB0190416; h, DDB0201789; i, DDB0190418; j, DDB0216669; k, DDB0190421; l, DDB0190422; m, DDB0190424; n, DDB0190426; and o, DDB0190427. (b) The breakpoints of the duplication in Ax2(Ka) were confirmed by real-time quantitative PCR, in comparison with Ax4(Ku). Mean log₂ ratios ± standard error are shown, summarizing, per gene, four pairwise comparisons of threshold cycles.

Figure 5

A novel duplication present in a subset of the Ax3 lineage. Nine strains in our study are lineal descendants of Ax3, and one other carries one or more chromosomes from it. NC4A2, based on our evidence, also descends from Ax3. Of these 12 lines, 7 carry a near identical duplication of chromosome 5 sequence. The breakpoints are not entirely clear because of noise in the data, and it is possible that there is some difference between strains. The genes plotted here are: a, DDB0188657; b, DDB0219507; c, DDB0188659; d, DDB0188660; e, DDB0188661; f, DDB0188665; g, DDB0188667; h, DDB0216146; i, DDB0188671; j, DDB0188673; k, DDB0188674; l, DDB0188677; m, DDB0188678; n, DDB0188686; o, DDB0188687; and p, DDB0188688.

Figure 6

NC4A2 lines contain a duplication of the same segment of chromosome 2 that is duplicated in Ax3. The duplication appears for the most part identical in all strains derived from Ax3. We show here (a) Kax3(U), (b) NC4A2(Kn), and (c) NC4A2(SC) because they display points of similarity not observed in the other examples of this lineage in our study. The point of inversion of this tandem inverse duplication is to the right of the plot, where some genes (log₂ ratios negative) appear to have been deleted in both copies in NC4A2 and KAx3(U). At least one of these genes appears to have been lost in both copies in several other of the Ax3-lineage strains in our study, but unfortunately some of the probes for these genes were not printed well and so our data do not permit us to assess exactly how frequent these deletions are. A segment within the duplication towards the left-hand side appears to be present as a single copy in both NC4A2 lines and in KAx3(U); this runs from DDB0233427 to DDB0191242, and appears to be present in the expected two copies in all other Ax3 derived strains we have studied.

Figure 7

Duplications are also frequent in non-axenic wild types in the NC4 lineage. (a) NC4(L) chromosome (chr) 2, (b) NC4(B) chromosome 4, (c) NC4(S) chromosome 6. As well as the three strains shown, NC4(Wi) has a duplication overlapping that observed in NC4(S). It is not clear why duplications near the chromosome ends are especially frequent in these non-axenic lines, or indeed whether this is merely a sampling artifact. The parasexual segregant XP99 is not axenic and yet carries two duplications away from termini; however this strain inherited some of its chromosomes ultimately from Ax3 and we can be confident that one of these duplications occurred in that axenic ancestor.

Figure 8

Two adjacent genes are duplicated in wild isolates but not in a subset of laboratory stocks, including all axenic strains. Shown here are all wild isolates in our study, plus selected laboratory strains descended from NC4. Note that some NC4 derived strains group with the wild strains, having two copies of these genes, but a subset (including all axenic strains) possess them in single copy.

Figure 9

Potentially deleted sequences. Probes with extreme log₂ ratios (below -3 in any strain) are considered to reflect deletions in that strain. The data have been clustered both by gene and by strain, to display relationships better (but information about groupings of genes along chromosomes is scrambled somewhat). The dendrogram gives an indication of relationships between strains, but note that this is based on the extremely small set of 36 genes shown here, and so should not be over-interpreted. The candidate genes are: 1, DDB0205403; 2, DDB0188003; 3, DDB0188004; 4, DDB0168894; 5, DDB0218478; 6, DDB0188007; 7, DDB0188002; 8, DDB0215073; 9, DDB0191949 (probe 1); 10, DDB0191949 (probe 2); 11, DDB0206115; 12, DDB0206106; 13, DDB0206110; 14, DDB0206111; 15, DDB0191930; 16, DDB0206108; 17, DDB0206112; 18, DDB0206109; 19, DDB0167672; 20, DDB0219404; 21, DDB0187848; 22, DDB0185937; 23, DDB0188514; 24, DDB0203140; 25, DDB0186442; 26, DDB0206404; 27, DDB0218143; 28, DDB0219338; 29, DDB0202734; 30, DDB0188991; 31, DDB0217456; 32, DDB0184376; 33, DDB0184375; 34, DDB219746; 35, DDB0206525; 36, DDB0217158.

Figure 10

Duplications are apparent also at the mRNA level. (a) Data from an experiment comparing mRNA from growing Ax2(Ka) and Ax2(Wee) ordered by chromosomal location, for chromosome 2 only, where Ax2(Wee) has a duplication. This is the only region displaying a striking shift away from log₂ ratio = 0. (b) Histogram of log₂ ratios within the region duplicated in Ax2(Wee). The distribution tends towards bimodality, with one clear peak near log₂ ratio = 0 and a less distinct one towards log₂ ratio = 1. Thus, a portion of genes is dosage-insensitive. (c) Correlation between log₂ ratios from independent mRNA comparisons. The duplication carried by Ax2(Wee) entirely encompasses the segment duplicated in Ax3 and its derivatives, allowing us to compare the dosage-sensitivity of genes in this overlap in different strains. The data are clearly correlated, with genes sensitive to dosage in Ax2(Wee) tending also to be sensitive in DH1 (unpublished work with T Soldati; the Soldati laboratory strains are indicated by So). A similar set of genes appears to undergo dosage compensation in both strains.