Bridging the Gap Between Developmental Genetics and Paleontology

Contemporary Evolution of Threespine Stickleback in Loberg Lake, Alaska

Introduction: Studies of contemporary microevolution (evolution within species during < 100 years) can provide novel insight into some of the most difficult problems in evolutionary biology, including the pattern by which the traits of an ancestor are transformed into those of a descendant and the time required for such changes. Until recently, there have been few studies of contemporary microevolution, but in recent years it has become clear that populations can evolve substantially on contemporary time scales and that the magnitude of evolutionary divergence between ancestral and descendant populations can be comparable to differences among related species.

We have observed contemporary evolution in a population of threespine stickleback fish (Gasterosteus aculeatus) in Loberg Lake, Alaska. Loberg Lake is a small (~4.5 ha), relatively deep lake in the Matanuska-Susitna Borough, north of Anchorage, Alaska. The native Loberg Lake stickleback population was exterminated in 1982 by the Alaska Department of Fish and Game to improve the lake for recreational fishing. The native population was typical of resident lake populations in the area, exhibiting significant armor reduction relative to its anadromous ancestor. In 1990, as part of a broad survey of morphological variation of stickleback populations in Cook Inlet, Alaska, we sampled the lake and found stickleback in it. However, these stickleback contrasted strikingly with the native population. They were heavily armored, like anadromous (sea-run) stickleback, and generally resembled anadromous stickleback in several respects.

The most likely explanation is that anadromous stickleback colonized Loberg Lake sometime between 1983 and 1989. We have sampled this lake every year since 1990 and have documented relatively rapid evolution of numerous traits, including armor, trophic, and body form traits, in the direction of the extinct population that originally inhabited the lake. The newly established Loberg Lake population provides a rare glimpse into how highly derived lacustrine populations of threespine stickleback, often constituting the most common species in lakes around recently glaciated regions of the northern hemisphere, originate in nature.

Below, we present some of our findings related to the evolution of armor traits.

Fig. 1. Above: A typical preserved anadromous stickleback that has been stained red to highlight the bone. Notice the unbroken series of large lateral plates along its flank and large spines. Anadromous stickleback in this region always possess strong armor. Below: Two specimens collected in Loberg Lake in 2006, highlighting the lateral plate variation in the lake population. The scale bar in the photogrpahs is 10 mm.

Evolution of Lateral Plate Morph Frequencies: Lateral plates are enlarged bony scales that form a single row along the flanks of threespine stickleback. They are an important defense mechanism against predators, especially piscivorous fish. Ancestral anadromous stickleback have about 33 lateral plates covering the entire flank, and exhibit the "complete" morph. Most derived resident lake populations have only a few lateral plates (<10) restricted to the anterior part of the body and represent the "low" morph. A third major morph, known as the "partial" morph, typically has anterior plates near the head and posterior plates in front of the tail but has an unplated gap in between.

The first sample from Loberg Lake that we collected in 1990 was composed almost entirely of complete morphs; there were no low morphs, just a few partial morphs. This quickly changed (Fig. 2). The first low morphs appeared in 1991 and rapidly increased in frequency; by 1994 there were more lows (46.1%) than completes (39.6%) in the lake. The trend has continued, and in 2007 the frequency of lows was 89.6%, whereas only 4.9% are completes. In other words, within 25 years (about 13 generations) of establishment, the population has gone from nearly all completes, like anadromous stickleback, to numerical domination by lows, which are typically the only plate morph in lake populations of threespine stickleback in Alaska.

The last few years have brought dramatic breakthroughs concerning the genetic architecture of lateral plate phenotypes in stickleback (See Colosimo et al. 2004, 2005; Cresko et al. 2004). Although modifiers of small effect are also involved, the Ectodysplasin (Eda) gene plays a major role in regulating the expression of lateral plate morphs in threespine stickleback. The Eda alleles encoding the low morph originated from a single mutational event several million years ago, and alleles derived from it have spread around the northern hemisphere to become numerically dominant in countless freshwater populations. Low Eda alleles also occur as rare recessive variants in anadromous populations, explaining how they have spread throughout the northern hemisphere. Thus, the rapid increase of low lateral plate morph frequencies in Loberg Lake must have involved strong natural selection acting on variation that was carried by the anadromous fish that colonized the lake. Contemporary evolution of low morphs in Loberg Lake during the past 25 years give us a fascinating glimpse into the tempo and mode of lateral plate morph evolution in freshwater threespine stickleback immediately after they evolve from anadromous fish.

Fig. 2. Evolution of lateral plate morph frequencies in Loberg lake between 1990 and 2007. Besides complete and low lateral plate morphs, we recognize three intermediate phenotypes described in Bell et al. (2004). These are "intermediate partials"(IP), partial morphs, and "intermediate lows"(IL). Notice that the frequency of the three intermediate phenotypes remained relatively low and constant throughout the time series.

Evolution of Lateral Plate Number: As lateral plate morph frequencies have changed, so has lateral plate number (Fig. 3). The lateral plate number distribution in the Loberg Lake time series is bimodal with one for low morphs at 7 or 6 per side and another mode for complete morphs between 33 and 30 per side. Over time, the population went from mostly 30 or more plates to mostly between 5 and 8 plates. The low number of intermediates throughout the time series reflects the existence of a gene (Eda) with a major effect on plate number. Homozygotes with two complete morph alleles at the Eda locus or heterozygotes tend to have high numbers of plates and are complete morphs. Homozygotes with two low morph alleles have low plate counts and are low morphs. A set of modifiers exists that increases or decreases lateral plate number, producing the intermediate phenotypes.

Fig. 3. Histograms depicting evolution of lateral plate number over time. Each histogram is the lateral plate number frequency distribution of a sample of 100 fish from one year. Even years are in the left column and odd years are on the right because generation time appears to be two years in the Loberg Lake population. The Y axis is frequency and the X axis is lateral plate number.

Evolution of Low Morph Lateral Plate Number: Lateral plate number varies within morphs too. There has been a good deal of research on the functional significance of lateral plate number variation within the low morph in lakes (See papers by T. E. Reimchen). Modes of seven are considered high for low morph populations and seem to be adaptive for defense against fish predation. Lateral plates in different positions have different functions, and stickleback with seven or more plates tend to have anterior plates protecting soft tissue immediately behind the head from puncture by fish teeth. When fish predation is unimportant, lateral plate number within the low morph declines because the most anterior plates are absent. Modes of five are typical of lake populations in Cook Inlet that lack piscivorous fish. Loberg Lake originally lacked piscivorous fishes, and the original threespine stickleback population inhabiting it had a mean lateral plate count of 5.08 per side.

We counted low-morph lateral plate number from annual samples of 200 fish and found that they declined significantly over time in the direction of the extinct population originally inhabitng the lake (Fig. 4). The first low morphs in Loberg Lake had high lateral plate counts with modes of seven plates. This suggests that the high plate counts thought to be adaptive against fish predation in freshwater environments are an exaptation; when lows evolve in freshwater they probably appear with high plate counts. Over time, the mean mean low morph LP number declined to as low as 5.91 in 2006, before it shot back up to 6.65 in 2007. The means in 2007 (6.57) and 2008 (6.63) remain relatively high. The increase between 2006 and 2007 is dramatic and the reasons for it are unknown. However, this shift highlights the importance of long term monitoring programs. Dramatic fluctuations in selection pressures and phenotypic values are fairly common in populations monitored over long periods (see work by Grant and Grant on Darwin's Finches for example). Although low morph lateral plate number has declined most years, the population still had a long way to go before it unaccountably returned to nearly its original value.

Other traits are evolving in the Loberg Lake population as well. We have documented the evolution of gill-raker number (a trait related to feeding habits, see Fig. 5), body size, body shape, and operculum shape (the bone covering the gills). It's clear that this population is undergoing substantial phenotypic changes as it adapts to conditions in the lake and is telling us a lot about the evolution of postglacial threespine stickleback populations in the process.

Fig. 4. Low morph lateral plate number evolution. Each point is the mean of 200 specimens for a year, except for the early years in which 200 low morph fish were not available. The low morph lateral plate number of the extinct population is based on 100 fish collected in 1982. The decline in low morph lateral plate number in the extant Loberg Lake population is significant despite the dramatic reversal observed between 2006 and 2007.

Fig. 5. Gillraker number evolution. Each point in the Loberg lake time series is the mean of 50 male specimens for a year, except for some of the early years. The gillraker number of the anadromous sample is from 202 anadromous stickleback collected in 1992 and the estimate from the extinct population is based on 50 fish collected in 1982. The decline in gill raker number in the extant Loberg Lake population is significant and has extended beyond the value of the extinct population originally inhabiting the lake.