Abstract Mitochondrial DNA from the livers of adult rockfishes were examined for differences in restriction fragment lengths. Äfter the mitochondrial DNA from kelp rockfish, blue rockfish, black rockfish, and other rockfishes was isolated, it was digested with seven dif¬ ferent six-base restriction endonucleases, run on agarose gels, transferred to nylon membranes, probed with radioactive mitochon¬ drial DNA, and the resulting autoradiographs examined for polymor¬ phisms. We found that all the enzymes showed differences among the three principal species, and that only the kelp rockfish displayed polymorphisms within the species. This method might be used to identify larval rockfish. The results also have interesting implica tions for the population genetics of the kelp rockfish, S. atrovirens. Introduction Purpose of the Study There is a good deal of interest in managing rockfish populations because rockfish are commercially important. Part of successful management of a population is an understanding of its recruitment patterns. Recruitment is the state of bringing new individuals from one life stage to another (Powers, et. al., unpubl.). In this study we are interested in ways to study larval recruitment of rockfish. In order to study recruitment, it is important to be able to iden¬ tify larval fish, and rockfish larvae are difficult to identify. Ways to identify larvae accurately and consistently need to be developed. Studies of larvae are quite difficult because a fertile mother needs to be kept until she bears her live young, and to investigate variation within a species requires raising several mother fish until they release their larval offspring. A graduate student from Moss Landing who defended his master's thesis in May had some success in looking at pigment differences of different species' larvae in a study he recently finished, but there are still many technical problems with this approach. Another way to tackle the problem is through the use of isozyme studies. However, in rockfish there is little variation within spe¬ cies and not enough between species so that recruitment studies using isozymes are impractical (Seeb, 1986). Besides, it is not clear if the isozymes that a larva uses are the same ones he uses as an adult. DNA studies of rockfish have not been investigated until recently. Seeb has begun to investigate mitochondrial DNA (mtDNA) patterns in a group of rockfish to better understand their phylo¬ genetic relationships. MtDNA is particularly useful because of its pattern of inheritance. It is possible that a species has one or a few distinct types of mtDNA that no other species shares while it is harder to find a unique version of an enzyme that sets that species apart in isozyme studies. We chose to study mitochondrial DNA over nuclear DNA for a few reasons. First, there are a lot more copies of an individual fish's mtDNA than its nuclear DNA because there are more mitochondria per cell than nuclei, and each mitochondrion has several copies of its genome (Alberts, et. al., 1985). Second, mtDNA has a pattern of inheritance which helps in examining branching patterns on phylo¬ genetic trees. Third, its possibly higher mutation rate might show more differences than would nuclear DNA. Ihe Genus Sebastes The genus Sebastes is made up of nearly 100 species of rock- fishes worldwide, and more than sixty of these are found in Cali¬ fornia waters (Burgess and Axelrod, 1984; Miller, 1972). It is the most speciose fish in the eastern North Pacific. As a group, rockfish make up the biggest catch for the fishing industry in California (Amidei, 1986). Both commercial fisheries and sportfishers fish for rockfish. Some rockfish live near the shore on rocky bottoms or in kelp beds while other species live offshore as far out as 300 miles and as deep as 450 m (Phillips, 1957). Sebastes is in the only subfam¬ ily of Scorpaenidae that gives birth to live young (Seeb, 1986). That is, rockfish have internal fertilization and viviparity. Adults range in length from 15 cm to 90 cm and longevity varies from less than ten years to as long as 100 years (Amidei, 1986). They are charac¬ terized by spines found on their dorsal and anal fins and also on their heads. It can be very difficult to identify certain species if not because there are just so many species to tell apart but because many of the species look very similar. Color and pigmentation are important in identifying species as are morphometric and meristic characters (Phillips, 1957). Some researchers are turning to other means of identification like gel electrophoresis. The three species that were examined in this study were S. atrovirens, S. melanops, and S. mystinus which are commonly referred to as the kelp rockfish, black rockfish, and blue rockfish respectively. All three are nearshore varieties with maximum lengths ranging from 40 to 60 cm. All three have mottled coloring with a predominant color that corresponds with that species name (the kelp being a brownish-green color) (Phillips, 1957). The kelp belongs to the subgenus, Pteropodus (seeb, unpubl.). The black and blue varieties are more closely related to each other than they are to the Pteropodus subgenus Seeb, 1986). Mitochondria and Their Special Genome Mitochondria are found in almost every cell of every eukaryotic organism, so it is not surprising that they provide a unique and essential function. It is generally believed that many millions of years ago, they were independent energy-producing bacteria, and they evolved into endosymbiotic organelles important for cell res piration. The mechanism for the electron transport chain resides in the mitochondria. Mitochondria divide independently of the cell and are passed on to daughter cells through the cytoplasm. A cell has from a few to a few thousand mitochondria depending on its energy requirements (Alberts, et. al., 1985). Although not always, mitochondria are usually inherited from the mother (Park, pers. comm.). This is because the mitochondria in the sperm (from the father) either do not survive or never enter the egg. and the mother contributes thousands of mitochondria from her egg As they used to be independent organisms, mitochondria use a genetic code which is different from the nuclear genomes of their "hosts“ and even from bacteria, which they are related to (Brown, 1985). Mitochondrial DNA is a circular duplex with a range of 16.3 to 19.2 kilobases (kb) in vertebrates (Brown, 1985). Beckwitt and Petruska (1985) found the mtDNA of rockfishes to vary between 16.9 and 17.4 kb. The genome is haploid, but there are multiple copies of it in each mitochondrion. mtDNA encodes 37 genes in animals: 22 are for tRNA's, two for TRNA's and 13 encode proteins. There is also a control region, which is important in the regulation of mtDNA rep¬ lication and gene expression. Much of the sequence and length varia¬ tion occurs in this control region (Brown, 1985). There is some debate concerning the evolutionary rate of mtDNA compared to that of nuclear DNA. In mammals for instance, mtDNA seems to have five to ten times more variation. The data for fish are not available, but even the data that do exist for other organisms are difficult to interpret (Park, pers. comm.). What makes mtDNA particularly useful in population genetics is its means of inheritance alluded to earlier. It is maternally inher¬ ited and haploid so non-deleterious mutations in the mtDNA of one mother are inherited in all of her descendants without mixing of genes. That means it is possible to look at phylogenies among groups of organisms. New branches on a phylogenetic tree are crea- ted with new mutations, but branches on the tree die when the last of a female individual's female descendants fail to reproduce. Sophisticated math proves that with an initial population size of n, all the descendants of that population will be able to trace their mtDNA to one female ancestor in 4n generations (Avise, 1986). BELP Analysis To analyze DNA, it is either cut into pieces using sequence-spe¬ cific restriction endonucleases or the nucleotide bases are se¬ quenced. Restriction endonucleases are enzymes which recognize a certain sequence of nucleotide bases and cut the DNA molecule at that sequence. When these fragments of DNA are exposed to an elec tric potential in a gel, they travel at different rates according to their length. Differences in these fragment lengths between indi¬ viduals are called polymorphisms. A restriction fragment length polymorphism (RFLP) analysis is useful for many things from explaining phylogenies to doing recruitment studies. It is possible to get mtDNA from a rockfish larva but only a lim¬ ited amount (Park, pers. comm.). How to use this DNA with economy presents a problem. Different restriction endonucleases recognize different sequences, so choosing which ones to use in a study is essential for economy of samples. Mitochondrial DNA happens to be particularly rich in adenine and thymine (Brown, 1985), so maybe the investigator wants an enzyme that recognizes a sequence with A's and T's so the mtDNA gets cut more. What he really wants is an en¬ zyme that can point out a lot of differences in RFLP patterns be¬ tween individuals. The best way to find out which enzymes are best for this is empirical testing of several different enzymes. I tested how good certain enzymes were at producing polymorphisms, so that these enzymes could later be used as diagnostic tools when dealing with the limited amount of mtDNA in a larval rockfish. Methods Rockfish were collected by the California Department of Fish and Game, and the livers from the fish were dissected. Liver was the tissue of choice because it is so rich in mitochondria. As will be described later, mitochondrial DNA was extracted according to the Chapman and Powers (1984) method. The DNA samples were diges¬ ted with an assortment of restriction endonucleases, and the products of these digestions were run on an agarose gel. A Southern blot transferred the DNA to a nylon membrane which was probed with mitochondrial DNA from Fundulus heteroclitus in a lambda phage. The radioactive membrane was exposed to X-ray film, and the film was developed and interpreted. Collection of Samples The fish were collected by Dave VanTresca and others at the Monterey office for the California Department of Fish and Game. Using SCUBA, the collectors speared fish and tried to keep them cool until they could be dissected. We had asked them to look for spe¬ cific species on their dives, and they identified the species. Either they would dissect the livers for us or we dissected the livers and the hearts also in some cases. The tissues were placed labelled bags on ice. The people from Fish and Game kept track of measurements of the fish, their sex, and other data for their own records. (These data may prove to be useful for further population studies of these samples.) There was no more than a two hour delay between collection and dissection. The first collection was off of Del Monte Beach in Mon¬ terery. In addition to several kelp rockfish, blues, and blacks, a copper (S. caurinus) and a gopher rockfish (S. carnatus) were col¬ lected at this site. The second also came from Del Monte Beach. A spearfishing contest at Carmel River State Beach was the third col¬ lection site. These fish were not actually caught by the Department of Fish and Game, but they were the sponsors. We took kelps, blacks, and blues from the event. Linda and 1 accompanied Dave VanTresca and Peter Nelson, a graduate student from UC Santa Cruz, on the final collection trip just north of the boardwalk in Santa Cruz. They speared kelps, blacks, and blues and an olive rockfish (S. serranoi¬ des). All of this is summarized in Table 1. Isolation of mtDNA Although these methods were not all carried out throughout the study, the following is the most refined version of the methods that were eventually performed. On a piece of glass on ice, a small (- 500 ul) slice of liver was cut, and saved at -70°C. About three grams of the remaining liver were minced with a razor to a liquid consistency. If the liver was big enough and parts of it could be spared, then pieces of liver not containing the tough connective tissues were chosen. The minced liver tissue was placed in a glass tube with 15 ml of TEK (.O5M Tris, pH 7.5, .OIM EDTA, and 1% KCl.) The tube was placed in an ice bucket and the sample was ground with a teflon pestle. Grinding continued until it was possible to make two complete passes with the pestle. The liver homogenate was then placed in a centrifuge tube and layered with a 15% sucrose/TEK solution that was passed through pasteur pipet under the homogenate. This was spun at 1000 G for ten minutes. The supernatant was drawn off from the lower layer of debris, but the thin layer of fat above the supernatant was not saved. This step of layering with sucrose/TEK, spinning, and drawing off the supernatant was repeated once. The supernatant from the second spin was spun at 18,000 G for thirty minutes, which leaves a pellet of mitochondria in glycogen. The supernatant was discarded, and the mitochondrial portion of the pellet was resuspended in one milliliter of TEK. At this point, each sample was divided equally into two eppendorf tubes. The mitochondria were lysed by adding ten microliters of 20% SDS. To increase the effectiveness of the SDS, the samples were placed in a 37°C water bath for about five minutes. To purify the DNA, 500 ul of phenol was added next, the mixture shaken vigor- ously, and then spun at full speed in the microfuge for ten minutes. The top layer was drawn off from the lower phenol layer and the unwanted white material at the interface. The phenol extraction was repeated once. To clean the phenol out of the sample, the same extraction procedure was performed with chloroform. However, the spin for the chloroform extraction was only five minutes. To cause the DNA to precipitate out of the TEK solution, 25 ul of 5M Nacl and one milliliter of -20°C ethanol were added to the solu¬ tion. After several hours in the freezer, the tubes were spun at full speed in the microfuge for fifteen minutes. The supernatant was discarded, and the pellet of DNA resuspended in 100 ul of 1X TE. Finally, the two tubes for each sample were combined in one tube. Enzyme Digestions of DNA Samples Seven different restriction endonuclease enzymes were used to examine differences in cutting patterns between individual samples. Six-base cutters were chosen because they generally have from zero to five or more restriction sites on mitochondrial DNA. The enzymes were BamHl, Clal, ECORI, ECORV, Hindill, Pstl, and Xhol. Enzymes were placed in distilled water and the appropriate salt for that enzyme so that the digestion would be in a 10% salt buffer solution (i.e. 2.5 ul salt for a 25.0 ul reaction.) For each reaction, 0.3 ul of enzyme in 15.0 ul was used to digest 10.0 ul of DNA sus¬ pended in TE. The labelled reaction tubes were finger-vortexed and placed in the 37°C water bath for several hours as the enzymes acted. To ensure that the digestion was complete, another 0.3 ul of the same enzyme in an extra 5.0 ul of buffered solution was added to each reaction tube. Äfter several hours in the warm water bath. some of the liquid in the tubes had condensed on the side of the tubes, so the tubes were microfuged for about fifteen seconds so that its contents would all be interacting together at the bottom of the tube. Once again, the tubes were finger-vortexed and placed in the 37°C water bath for several hours. If a double digestion was being performed, then 0.3 ul of the second enzyme would be added instead of the extra 0.3 ul of the same enzyme. Otherwise double digestions were done the same way. Bunning Gels The agarose gels were made with 0.8% agar in 1X TAE. The agar solution was heated in the microwave for several minutes until it boiled. While the agar boiled, an 11 X 14 cm gel tray was prepared by taping its ends tightly and placing a fourteen-toothed (thin tooth variety) comb at one end. The very hot agar was poured, making sure there were no bubbles and no leaks from around the tape. This cooled for about thirty minutes, at which time the tape and comb were removed. The fourteen lanes were each filled with the full contents (30u)) of a reaction tube using a predetermined order. One lane in each ge¬ was reserved for the standard, lambda phage cut with Hindlll in a solution containing a blue dye. The dye has two components, one which travels at the same pace as a 100 base pair fragment and one at 5000 base pairs. This standard dye is used to estimate how far the DNA has migrated in a gel. Ten microliters (.25 to .50 ug) of the standard were loaded in the standard lane. The gel trays were placed in the gel rigs and the rigs filled with about one liter of 1X TAE buffer. The rig was filled with TAE to just under the top of the gel. The power was turned on to 100 volts for about five minutes, and the DNA entered the gel. Then, the gel was covered with more TAE and run at a lower voltage, to achieve good separation of the fragment bands. Voltages used ranged from twenty-five to seventy-five volts, and the gels ran for several hours and often overnight. When the leading dye front had gone about three-fourths of the way across the gel, the power was turned off, and the gel was prepared for visualizing the DNA. The gel was placed in 500 ml of the buffer it just ran in with 100 microliters of ethidium bromide solution (10 mg/ml). The gel soaked in the dilute ethidium solution for at least ten minutes. Then, the gel was rinsed twice with tap water. The gel was placed on the ultraviolet light source in the dark room, and a transparent ruler was placed over the standard lane to serve as a reference. At an f-stop of 11, the shutter was released for three seconds, and a polaroid picture was developed immediately. Southern Blots The gel was covered with denaturing solution and gently rocked on a shaker for ten minutes. The denaturing solution was replaced with fresh denaturing solution, and the gel was once again shaken for ten minutes. Äfter a rinse with neutralizing solution, the gel was covered with neutralizing solution and gently rocked for fifteen minutes. The neutralizing solution was replaced, and the gel was shaken for another fifteen minutes. A blot was made by layering several materials around a gel. Two pieces of Whatman's paper cut to the size of the gel had soaked in 20X SSC and were then placed on a sheet of plastic wrap on a flat surface. Air bubbles between the paper and plastic were removed by stroking them out on the sides. The gel was placed on the paper and air bubbles removed again. A sheet of nylon membrane which was cut to size and wetted in water was soaked in 20X SSC, and it was placed on the gel. Air bubbles were carefully removed before ano¬ ther piece of 20X SSC-soaked Whatman's paper was layered next. A stack of paper towels was placed on the gel layer, and a glass plate placed on the whole stack. Finally, an object with some degree of mass such as a full flask, was placed on the plate. The blot sat on the bench for at least eight hours before the nylon membrane was removed and saved in a sealable plastic bag. This bag was placed on the ultraviolet light source with the DNA side of the membranes facing the light and then exposed to UV light for two minutes. This covalently bonds the DNA to the nylon. Hybridization The filters were prepared for hybridization with a pre-hybridiza¬ tion solution. This solution is three parts 20X SSC and one part each of SDS and nonfat milk in fifteen parts water. It was poured in the bag and the air bubbles removed. Then the bag was sealed and placed in a 65°C water bath for two hours, at which time it was ready for hybridization. did not grow the probe, so I cannot precisely explain the me¬ thods of growing it. However, I do know that the probe we used is a lambda phage with a Fundulus mtDNA insert, and it was grown in E. coli. The probe was labelled with cytosine nucleotides containing the phosphorus radioisotope 32P. To label the probe, 12 ul (.25 ug) of the probe DNA and 1.0 ul of random DNA primers were boiled in a screw-top tube for five minutes. While the DNA was boiling, a mix¬ ture of 2.0 ul of Klenow, 2.0 ul of adenine, thymine, and guanine, and 2.5 ul of Klenow buffer was made. The hot DNA was cooled quickly on ice before the Klenow enzyme mixture was added. This safeguard prevents denaturing of the enzyme when the two solutions are mixed. Äfter the Klenow was added to the DNA, 5.0 ul of the radio¬ active dCTP was added, and the reaction occurred for the next two to four hours in a lead container. Eighty microliters of 1X TE were added to the radioactive probe to stop the reaction. To remove the unincorporated dCTP's, the probe mixture was run through a sephadex column and microfuged for about five minutes. The filtered solution was boiled in a screw-top tube for ten minutes to denature the DNA strands. Then, the bag with the nylon membranes was cut open, the radioactive probe injected into the bag, the bag resealed, the contents mixed around inside the bag, and the bag placed into a 65°C water bath for twelve hours. Next, the membranes had to be washed. The liquid contents were carefully disposed of into the radioactive waste, and the membranes were rinsed in a wash solution of 0.2 % SDS and 2X SSC. This rinse requires only enough wash solution to rinse off the membranes and still be a reasonable amount to throw away in the radioactive waste, For the next wash, the membranes soaked in more of the same wash solution for ten minutes. For the last wash, the wash solution was replaced with fresh wash solution which had been diluted by one¬ half, and the membranes soaked for fifteen minutes. The membranes were rinsed with distilled water, and they were ready for exposure to film. Developina The membranes were patted dry with paper towels and wrapped in one layer of plastic wrap. The wrapped membranes were taped to a developing screen. Then, X-ray film was placed between the mem¬ branes and another developing screen. This was closed tightly into a developing cartridge and placed in a -70°C freezer. The film was exposed to the radioactive bands in the membranes for three days at which time the film was run through the developing machine, produ¬ cing the final product, an autoradiograph. Results The first result to report is the success of the probe we used. It appears that a mtDNA probe from Fundulus is sufficient for hybri¬ dizing with rockfish mtDNA. At least earlier in the study, the auto¬ radiographs were clear, and the probe seemed to be working fine. Later, unexpected problems arose with the autoradiographs. It was difficult to get the probe radioactively labelled, and this may have been the problem. However, the problem could have been in any of a number of steps from denaturing the gels to UV-fixing the mem¬ branes to hybridizing and washing the membranes. Nevertheless, was able to get enough data from the autoradiographs that did deve¬ lop and from the polaroid pictures of the ethidium-stained gels, that could quantify the data in some ways. was able to see fragment patterns of blacks, blues, and kelps for all the enzymes considered. A maximum of around ten to fifteen individuals in each species were informative for any given enzyme, although sometimes fewer than five individuals in one species could be clearly analyzed for an enzyme. The four individuals from other species did not always show up clearly for every enzyme they were digested with. No polymorphisms appeared in black rockfish with any of the enzymes tested. Only one polymorphism was detected in the blue rockfish, and that was with Xhol; two of the blue rockfish had a restriction site that was not shared by the other blues. Kelp rockfish usually had two polymorphisms with each enzyme tested: only with BamHI and ECORV were no polymorphisms detected. Hindlll had two sites that were variable within the species, and the other four enzymes each had one variable site. In eight kelp rockfish individuals that were collected together, two groups of four segre¬ gated together in four of the six variable sites. Only one individual kelp differed from the other kelps at the fifth site, and the sixth site that is variable within kelps is the same one that seemed to be variable in the blue rockfish (a cut that creates a 3 kb fragment in Xhol.) Double digests were used to more closely examine the poly¬ morphisms within the kelp species. Clal and BamHl were used to¬ gether (see figure 1), and Pstl and Clal were tested together also. The polymorphisms we had suspected were confirmed by the double digests. In addition to restriction fragment length polymorphisms in the kelps, length variation was also observed. One of the individuals is consistently smaller than the others, and there are possibly other length differences that are harder to detect. All of the enzymes showed some differences between the species (see figures 2-8). It is not possible to make accurate maps of the restriction sites because there are no sequence data available for these samples and because not enough double digests were per¬ formed. Therefore, the following figures are only proposed restric¬ tion maps for each of the enzymes. They are schematic representa¬ tions of my best interpretations of the bands I was able to detect on either the autoradiographs or the polaroid pictures. For example, in figure , we see that BamHI, had two polymorph¬ isms. The blue, kelp, gopher, and olive rockfishes had a pattern con¬ sisting of two bands, a pattern which would be produced by two re¬ striction sites. The black rockfish's DNA digest shows only one band which means there was one restriction site. These sites are labelled in the diagrams at the bottom of the figure. According to the most parsimonious interpretation of the data, site a is shared by all the species, and site b has been lost in black rockfish. The top line in the schematic of the bands represents the well where the lane started. The numbers next to the other lines give the approx¬ imate lengths of those fragments in kilobases. (O.C. is an abbrevia¬ 12 tion for open circle, a distinctive band which travels about half the distance from the well to the band where linear fragments greater than 20 kb cluster. think the figures are fairly clear upon exam¬ ination, and I will discuss them in greater detail in the discussion section. From the variability of the sites, genetic distances were calcu¬ lated and a cladogram was also constructed. Only the data from the kelp, blue, and black rockfish were used because no other species had complete data for all the enzymes. As it is, these results are merely speculative. Not nearly enough individuals were examined to give significant results; however, it is still interesting to see what phylogenetic information these data give. Genetic distance, d, can be calculated according to Nei's genetic distance d= Ein(S))/r where r is the number of nucleotide bases per restriction site and (S) is the probability that a pair of sequences X and Y share a given restriction site according to (S) = 2nxy/(nx + ny). In the last equation, the ny and ny terms refer to the number of DNA fragments in all the digests for species X and Y while nyy is the number of shared fragments (Hartl and Clark, 1990). Table 2 gives calculated values for all the groups of individuals that had the same polymorphisms for every enzyme, or in other words, all the groups of individuals that segregated together. The cladogram was constructed with McClade on a Maclntosh SE (see figure 9). Cladistics attempts to build phylogenetic trees based on informative restriction sites. For all the sites which show dif¬ ferences between groups, presence or absence of the site is recor¬ ded. It two or more groups share the presence or absence of a site, then that site is phylogenetically informative. Based on the number of steps (site changes) it would take for the groups to arrange themselves in a particular order, a best tree is selected which requires the least number of steps. This tree is presented in fig. 9. 13 Discussion We looked at mitochondrial DNA in rockfishes in the first place because we wanted to see if it would be a useful tool for identifying larvae in recruitment studies. With the limited scope of these findings, the answer appears to be yes because mitochondrial DNA is cut into different patterns of fragment sizes by certain enzymes according to species. All the enzymes we used showed differences between species. It is important to note that much of the raw data from the auto radiographs and polaroids was quite difficult to interpret due to experimental problems and the poor resolution we could see. was able to see the band patterns for about ten individuals for each spe¬ cies with each enzyme; however, I did not always see that many individuals. Nevertheless, the results I reported seem consistent for all the enzymes with the exception of Xhol. In the case of Xhol. could detect two patterns: one which appeared to be a single frag¬ ment which had one restriction site resulting in a 17 kb fragment and one which resulted in two bands of lengths 14 and 3 kb, presu¬ mably due to one additional site. It is often difficult to see the bands of small DNA fragments because they bind less ethidium, and the radioactive probe has less to bond to on a small piece of DNA. Larger fragment sizes are difficult to estimate because they tend to cluster together due to the logarithmic nature of their migration rate through the gel. Thus, it is easy to see how I could not have seen a small band in some lanes and how 1 could have mistaken a 14 kb fragment for a 17 kb fragment in the Xhol digests. For this and other reasons which will be brought up in this section of the paper, would propose that the data 1 reported from the Xhol digests be con¬ sidered invalid. The data from the other enzyme digests show that RFLP analysis using mtDNA is useful for differentiating between at least kelp, blue, and black rockfishes. EcoRl has different banding patterns for each of these species, so it could conceivably be used as the diag¬ nostic enzyme for them. However, the enzymes we used did not show differences between the individuals from some of the other species (copper, gopher, and black and yellow rockfishes) and some of the kelp rockfish. However, this is a very interesting result which I can discuss only after describing the variation within the kelp rockfish species. 14 The kelp species divides into two main groups of RFLP patterns, which I call kelp A and kelp B. Eight kelp individuals that were col¬ lected together could always be analyzed in each of the digests, and four were A type while four were B type. In the six reliable enzymes we tested, there are five restriction sites which show polymorph¬ isms in S. atrovirens, the kelp rockfish. One of these polymorph¬ isms was a single site difference in Pstl in one individual (of the A type) which set it apart from the other kelp individuals. With the other four site differences, the individuals of the A type always segregate together as do all the individuals of the B type. The black and yellow rockfish (S. chrysomelas) segregates with the kelp B type in digests with Clal, EcoRI, and Hindlll, and it segregates with both kelp types together in the ECORV digest. The copper rockfish (S. caurinus) segregates with the kelp B type in the Hindlll digest at both sites, and with both kelp types together in the ECORV digest. The gopher rockfish (S. carnatus) segregates with kelp B type in the Clal digest, and with both kelp types together in the BamHI, ECORV and Pstl digests. It is not surprising that these three individuals from other species segregate closely with the kelp rockfish when one considers that all four belong to the same subgenus, Pteropo¬ dus. If the five groups of fish (including the two groups within S. atrovirens) are all in the same subgenus, then why is it that kelp A is so different from all the other groups of fish? The variation within the kelp species is greater than the overall variation of the subgenus, an obviously significant result. There are at least four explanations for the result that in terms of mtDNA, the kelp A group is so different from the other fish in the subgenus. The first is simply that the kelp rockfish individuals were not identified correctly in the first place. The species most easily confused with the kelp rockfish is the grass rockfish (S. rastrelliger) which is also in the same subgenus with the kelp rockfish. One would expect that the grass rockfish's mtDNA would look similar to the mtDNA of the other species in the subgenus, which all look very similar in RFLP analysis. If it is not actually kelp rockfish and not grass rockfish either, it might be a more dis¬ tantly related fish in the genus. The other three explanations of the strange result try to explain the result assuming the species were identified correctly. They also try to account for the fact that it is unusual for a species to support two different strains of mtDNA because branches seen with mtDNA on the tree of an organism are expected to die out with time as dis¬ cussed earlier. The second explanation is that there is actually a cryptic species that has never been recognized before. This seems unlikely, but some support comes from the kelp rockfish's depth dis¬ tribution. Often, two species of rockfish which occupy a similar niche will live in mutually exclusive depth ranges. For instance, Hallacher (1977) found that the black and yellow rockfish occupies the shallower water, and the gopher rockfish lives in the deeper water in the Monterey Bay area. In the same study, he found the kelp rockfish distributed evenly through the different depths. Perhaps, two species could be recognized, one would find that the spatial distribution would get divided as it did in the above example. Using allozyme studies, it should be simple to test the hypothesis that there is a cryptic species. Comparing the kelp A mtDNA type to the kelp B type, if one were to find that an allozyme is fixed in one type, and a different allozyme is fixed in the other type, then that would be proof of the cryptic species. The third and fourth explanations are that the present kelp species comes from two different popula¬ tions that were not mixing genes but are now, and their mtDNA types both persist. The two populations could have been different species which hybridized, resulting in the present species of kelp rockfish. Hybridizations are not at all uncommon in Sebastes (Seeb, 1986), and the possibility exists that a hybrid was once fertile. The fourth explanation is that a population of kelp rockfish was once isolated from another population of kelp rockfish long enough for a different mtDNA type to develop, and now those two populations are mixing. Support for this would come from allozyme studies which have un¬ fortunately not been conducted in depth for the kelp species. A stu¬ dent from last year's spring class found little allozymic variation within the kelp species (Bachmann, 1989), a finding which supports the last hypothesis. Another one of her findings was a significant difference in allozymes between the kelp rockfish and the black and yellow rockfish, which contradicts the earlier hypothesis that there exists a cryptic species. If there were a cryptic species, one would expect that there would be more allozymic variation within the kelp species than Bachmann found. Öther evidence which suggests that isolation may have once occurred is the current evidence for some geographical separation. It is known that there is a geographic boundary between central and southern California populations of S. 16 atrovirens which is manifested in differing numbers of tympanic spines (Love and Larson, 1978). Whatever the case may be, this sub¬ ject warrants further investigation. As far as the other species which has not been discussed vet. there is not enough information to make a conclusion concerning the olive rockfish (S. serranoides) and its mtDNA. With allozyme studies, it appears to be a sister species with S. mystinus, the blue rockfish (Seeb, 1986) . In the only digests for which I have included olive rockfish data, it seems to segregate with the blue rockfish. However, inconclusive data from other digests seemed to show differences between the two species. If this is true, then it would demonstrate how much more effective mtDNA studies are in showing differences between species as compared to isozyme studies. More samples need to be analyzed, and more enzymes need to be tested to get a more accurate assessment of mtDNA variation in the genus Sebastes. To further resolve the question of the kelp rock¬ fish species, allozyme studies should be conducted. In addition. direct sequencing may give the higher degree of resolution that is needed to make the best conclusions concerning the interesting results we found here. Acknowledgments I will never forget my experience at the Hopkins Marine Station, and I have several people to thank for it. Without Linda Park, my project would not have existed, and I appreciate the great deal of time, work, patience, and care she gave to me this quarter. I would also like to thank my advisor, Dennis Powers for his help and instruction and for letting me work in his lab. Rob Rowan, Lynna Hereford, Lani West, Doug Crawford, and all the others in the lab really helped me get through my project. Dave VanTresca and the other collectors did excellent work. Sam Wang and Jenifer Levitt put great effort into making the class a success. I thank all the other faculty and staff at the Marine Station for keeping things going and making the class feel so welcome. Without the encouragement of my peers who took the class in previous years, my advisor Bob Simoni, and other professors back at Stanford 1 may not have come, so 1 thank them. Finally, I would like to thank my father and my parents for their support and encouragement during the quarter. Table 1 Rockfish Collections Site Date Species 4-25-90 Monterey Kelp Del Monte Beach Black Blue Copper Gopher 5-1-90 Black and Yellow Black Blue Carmel River State 5-17-90 Kelp Beach Black and Yellow Black Blue 5-22-90 Santa Cruz Kelp Boardwalk Blue Olive Black and Yellow Black Number Table 2 Nei's Genetic Distances Kelp B Kelp A 1 ind. Kelp A 3 ind. Black Blue Kelp B 18 0274 0209 0293 .0460 Kelp A1 14 .0055 .0402 0278 15 Kelp A3 15 16 .0341 .0460 Black 13 13 .0304 Blue 11 Values on the diagonal are how many fragments that group has in total for all six reliable enzymes. Values below the diagonal are shared fragment numbers. Values above the diagonal are Nei's genetic distance (see text.) Literature Cited Alberts, Bruce, et. al. Molecular Biology of the Cell. Chapter 9: Energy Conversion: Mitochondria and Chloroplasts. Garland Publishing: New York. pp 483-548, 1983. Amidei, Rosemary. Editor. Rockfish: A Focus for Research? Proceedings of a California Sea Grant Workshop April 4, 1986. Avise, J.C. Mitochondrial DNA and the evolutionary genetics of higher animals. Phil. Trans. R. Soc. Lond. B 312, 325-342, 1986. Bachmann, Anja. Electrophoretic Analysis of Sebastes atrovirens and Sebastes chrysomelas in the Monterey Bay Area. Unpublished MS on file at Hopkins Marine Station. Beckwitt, Richard and John Petruska. Variation in Mitochondrial DNA Size Among Fishes of the Family Scorpaenidae. Copeia. Vol. 4, Pp 1056-1058, 1985. Brown, Wesley M. The Mitochondrial Genome of Animals. In Molecular Evolutionary Genetics. (R.J. Maclntyre, ed.) Plenum; New York. pp 95-130, 1985. Burgess, Warren E. and Herbert R. Axelrod. Fishes of California and Western Mexico. Pacific Marine Fishes Book 8. T.F.H.: Neptune City, N J. pp 2118-2140, 1984. Chapman, Robert W. and Dennis A. Powers. A Method for the Rapid Isolation of Mitochondrial DNA from Fishes. Maryland Sea Grant Program Technical Report, 1984. Hallacher, Leon Ernest. Patterns of Space and Food Use by Inshore Rockfishes (Scorpaenidae: Sebastes) of Carmel Bay, California. Doctoral Dissertation from UC Berkeley, 1977. Hartl, Daniel L. and Andrew G. Clark. Principles of Population Genetics., Second Edition. Sinauer Associates: Sunderland, MA. Pp 331-335, 1989. Love, Milton S. and Ralph J. Larson. Geographic Variation in the Occurrence of Tympanic Spines and Possible Genetic Differentiation in the Kelp Rockfish (Sebastes atrovirens). Copeia. Vol. 1, pp 53-59, 1978. Miller, Daniel J. and Robert N. Lea. Guide to the Coastal Marine Fishes of California. California Fish Bulletin No. 157. pp 1-249. 1972. Phillips, Julius B. A Review of the Rockfishes of California. California Fish Bulletin No. 104, pp 1-158, 1957. Powers, Dennis A. et. al. Application of Molecular Techniques to the Study of Marine Recruitment Problems. Unpublished, Seeb, Lisa Wishard. Biochemical Systematics and Evolution of the Scorpaenid Genus Sebastes. Doctoral Dissertation from University of Washington, 1986. Seeb, Lisa W. Untitled Project Description for Analyzing Mitochondrial DNA of the Sebastes subgenus Pteropodus. figure 1. Two autoradiographs. Above is the double digest of three kelp individuals with BamHl and Clal. The upper of the three bands for the individual is the BamHl digest, the middle lane is the double digest, and the lower lane is the Clal digest The middle individual (of the Kelp A varirty) seems to have a restriction site in Clal, but the 17 kb fragment is DNA that has been physically nicked open. This is true because the double digest shows the BamHl fragments not getting cut again. In the lower gel, several individuals are digested by Hindll. From the top, a copper, a black and yellow, a black, and seven blues. The black and blue rockfishes have the same RFLP, and the other two share an RFLP. The middle band of the blues and black has been cut once more in the other two lanes. figures 2 - 8. Schematics of the enzyme digestions. Proposed band lengths are labelled. O.C. stands for open circle configuration. Below the bands are the proposed maps. They are not necessarily accurate, just a visual aid to understanding the fragment patterns. figure 9. Cladogram of the kelp groups, black, and blue rockfishes from the McClade program. Figure 1 Ban 8/ . 2a Nif 8/ 1B 84 2 5 4 Cano 6 0 1 15. 5 Ms-12 N Blue Kelp Olive Gopher 15 2 — 15 BamHI RFLP'S Figure 2 Black 17 — Blue Kelp A Olive — O.C. — Clal RFLPS Figure 3 Kelp B Black Gopher Black and Yellow 17 — o s - + + — S Blue Black 14— 3 — 14 2 Kelp Gopher Black and Yellow Copper — — ECORV RFLP'S Figure 5 HindlII RFLP'S Kelp B Blue Black & Yellow Black Copper — O — 10 - 5 — 4 — 2 — 2 — — 10 Figure 6 6 Kelp A 6 — 5 4 2 - a 4 Blue Black Kelp (1 indiv.) 10 — 4— 2.5 — 5 — 5.5 4.5 ) 25c PStI RFLP'S Figure 7 Kelp (all others) Gopher 5.5 — 4.5 — 4 2.5— 5 — 10 1.0 25/c Blue (2 indiv.) Black Kelp (4 indiv.) Gopher 14— 3 — Xhol RFLP'S 17 Kelp (4 indiv.) Blue (most indiv.) Black and Yellow Figure 8 5 e . ha Ctar step- reversible Figure 9