The Journal of Experimental Biology 203, 1277Ð1286 (2000) 1277 Printed in Great Britain © The Company of Biologists Limited 2000 JEB2492

CONCENTRATIONS OF AND MYOGLOBIN mRNA IN HEART VENTRICLES FROM ANTARCTIC

THOMAS J. MOYLAN AND BRUCE D. SIDELL* School of Marine Sciences, University of Maine, 5741 Libby Hall, Orono, ME 04469-5741, USA *Author for correspondence (e-mail: [email protected])

Accepted 9 February; published on WWW 23 March 2000

Summary We used a combined immunochemical and molecular was used to quantify mRNA in five Mb-expressing icefishes. approach to ascertain the presence and concentrations Mb mRNA was found in low but detectable amounts in of both the intracellular -binding hemoprotein gunnari, one of the species lacking myoglobin (Mb) and its messenger RNA (mRNA) in 13 of detectable Mb. Mb mRNA concentrations in heart 15 known species of Antarctic channichthyid icefishes. from Mb-expressing species ranged from Mb protein is present in the hearts of eight species 0.78±0.02 to 16.22±2.17 pg Mb mRNA µg−1 total RNA). Mb of icefishes: rastrospinosus, Chionodraco protein and Mb mRNA are absent from the oxidative hamatus, , , skeletal muscle of all icefishes. Steady-state concentrations Pseudochaenichthys georgianus, antarcticus, of Mb protein do not parallel steady-state concentrations dewitti and . Five of Mb mRNA within and among icefishes, indicating that icefish species lack detectable Mb protein: Chaenocephalus the concentration of Mb protein is not determined by the aceratus, macropterus, , size of its mRNA pool. Champsocephalus gunnari and Dacodraco hunteri. Mb concentrations range from 0.44±0.02 to 0.71±0.08 mg Mb g−1 wet mass in heart ventricle of species Key words: myoglobin, mRNA, Antarctic fish, Channichthyidae, expressing the protein. A Mb-mRNA-specific cDNA probe , heart, , mRNA.

Introduction Low temperatures and high oxygen solubility characterize the absence of a circulating oxygen-transport protein, many the surrounding . Mean annual investigators have characterized icefishes as also lacking the temperature in McMurdo Sound is −1.86 ¡C (Littlepage, 1965), 16Ð17 kDa intracellular oxygen-binding protein myoglobin. while the temperatures of waters surrounding the Antarctic Myoglobin is a monomeric protein containing a single Peninsula range between +0.3 ¡C during the austral summer to coordinated heme group that binds oxygen reversibly with a −1.1 ¡C during the winter months (DeWitt, 1971). Despite 1:1 molecular stoichiometry. It has an important role in the these chronically cold temperatures, coastal Antarctica storage and transport of oxygen from to supports an abundant fish fauna dominated by species of the mitochondria in oxidative muscle tissues of (Covell perciform suborder , a group that has been and Jacquez, 1987; Wittenberg and Wittenberg, 1989). While evolving since the formation of the Antarctic Circumpolar the cardiovascular alterations described above may help offset Current, between 14 and 25 million years ago (Eastman and the loss of , it is difficult to envisage how these Grande, 1989). features could compensate for the reported absence of Channichthyid icefishes are thought to have diverged from myoglobin in highly aerobic heart tissues (Hamoir, 1988; other notothenioid families approximately 1Ð3 million years Eastman, 1990). ago (Bargelloni et al., 1994). The 15 species of the Although the consensus view has been that icefishes lack Channichthyidae are unique among adult vertebrates in their myoglobin, Douglas et al. (1985) reported that myoglobin complete lack of expression of hemoglobin, a characteristic was expressed in heart tissue of two icefish species, first described in the scientific literature by Ruud (1954). These Pseudochaenichthys georgianus and Chaenocephalus fishes show profound cardiovascular modifications (large aceratus. The technical basis for this report, however, was the hearts and vessels, high cardiac output, increased blood detection of the formation of pyridine hemochromagen in volume) that apparently ensure adequate delivery of oxygen to crude supernatant extracts from hearts, a method that could the tissues despite the lack of circulating hemoglobin easily lead to false positive results in tissues containing high (Hemmingsen et al., 1972; Hemmingsen, 1991). In addition to concentrations of mitochondrial cytochromes. To resolve the 1278 T. J. MOYLAN AND B. D. SIDELL question of myoglobin expression in the icefishes, our towards deciphering the molecular basis for the unusual pattern laboratory has recently used more definitive immunochemical of myoglobin expression. These measurements also permitted and molecular techniques to establish that myoglobin is us to evaluate any correlation between pools of myoglobin expressed in heart ventricles of some icefish species while mRNA and the concentration of myoglobin protein in the heart being completely absent from the same tissue in others (Sidell ventricle of notothenioid species. et al., 1997). Furthermore, the disparate positions of myoglobin non-expressers within the phylogeny of icefishes suggested that multiple independent mutational events led to the loss of Materials and methods myoglobin expression during the evolution of the family. The and tissue collection establishment of discretely different mutational mechanisms Chionodraco rastrospinosus, Pseudochaenichthys among these myoglobin non-expressers has corroborated this georgianus, Chaenodraco wilsoni, , conclusion (Small et al., 1998). The seemingly random pattern Champsocephalus gunnari, Gobionotothen gibberifrons, of loss of myoglobin among icefish species appeared to suggest Trematomus newnesi and Notothenia coriiceps were collected that the protein may not be of functional significance at the by 18 foot otter trawl net deployed from the R/V Polar Duke severely cold body temperature of Antarctic icefishes, thus while fishing off the Antarctic Peninsula in Dallman Bay near relaxing all selective pressure on the retention of its expression Astrolabe Needle (64¡10′S, 62¡35′W) in MarchÐMay 1993 and and/or structure. 1996. were transported live to the US Antarctic research The provocative suggestion that myoglobin may not station, Palmer Station, and maintained there in running function at the body temperature of Antarctic fishes led our seawater tanks (−1.5 to +1.0 ¡C). were killed by a laboratory and collaborators to examine the functional sharp blow to the head followed by severing the spinal cord characteristics of oxygen-binding by myoglobin from these immediately posterior to the head. The heart ventricle and animals and its potential physiological role in those Antarctic pectoral adductor profundus tissues were rapidly dissected on icefishes that do express the protein. Using stopped-flow a chilled stage, weighed and frozen in liquid nitrogen prior to kinetics measurements, we found that myoglobins from both storage at −80 ¡C. Antarctic and temperate-zone teleost fishes show more rapid Tissues from , Pagetopsis binding and release of oxygen at cold temperature than those maculatus, Dacodraco hunteri, Chionodraco myersi and from mammals (Cashon et al., 1997). These results indicate were generously provided by Dr A. that fish myoglobins, including those from Antarctic species, DeVries (University of Illinois) and were collected in possess alterations in protein structure/sequence that increase McMurdo Sound, Antarctica. Drs R. Acierno and G. di Prisco the speed of binding and release of oxygen at low (Italian National Antarctic Program) kindly supplied samples temperatures. Additional experiments with isolated, perfused of , collected in the vicinity of Terra hearts from two icefishes, one lacking (Chaenocephalus Nova Bay, Antarctica. Dr T. Iwami (Tokyo Kasei Gakuin aceratus) and one containing (Chionodraco rastrospinosus) University, Japan) supplied Chionobathyscus dewitti and myoglobin protein, demonstrated that hearts possessing Neopagetopsis ionah collected in the Weddell Sea. Tissues myoglobin were capable of greater mechanical performance were dissected and maintained frozen (liquid nitrogen, dry ice than those lacking it (Acierno et al., 1997). Both lines of or −80 ¡C storage) until use. evidence strongly indicate that myoglobin is functional at the normal body temperature of icefish and that it does play a Purification of myoglobin standard physiological role in the delivery of oxygen to working To determine myoglobin concentrations in icefishes, a muscle. These conclusions make the observation that myoglobin standard was purified from the related nototheniid myoglobin is absent from oxidative skeletal muscle of all species Notothenia coriiceps. Frozen Notothenia coriiceps notothenioid fishes examined to date even more perplexing heart ventricle (2.4 g) was weighed and homogenized [40 % in the light of the highly aerobic metabolism of this tissue (w/v)] in filtered 100 mmol l−1 potassium phosphate buffer, (Sidell et al., 1987). pH 7.8, with a glass tissue grinder (Tenbroeck). The The confirmation of extremely variable expression of homogenate was centrifuged at 23 000 g for 30 min at 4 ¡C. The myoglobin in the notothenioid suborder (Fig. 1) is therefore at resulting supernatant was collected and centrifuged at 23 000 g odds with strong evidence indicating a functional role for the for another 30 min. The final supernatant was applied to a protein and raises questions about the molecular mechanisms 2.5 cm×100 cm Bio-Rad P-100 gel permeation column that had responsible for myoglobin loss from aerobically poised been equilibrated with 100 mmol l−1 potassium phosphate muscles. To shed further light on this conundrum, we buffer, pH 7.8, containing 0.02 % sodium azide. The column undertook experiments (i) to estimate the intracellular was maintained at 4 ¡C. Fractions (3 ml) were collected at an concentrations of myoglobin in heart ventricle of elution rate of 13 ml h−1. Fractions containing myoglobin were channichthyid and nototheniid species expressing the protein, pooled on the basis of relative mobility compared with to help assess its potential physiological relevance, and (ii) to hemoglobin and high absorbance at 415 nm, but low quantify the concentrations of myoglobin mRNA in absorbance at 280 nm. For further purification, pooled channichthyid and nototheniid species as an essential step myoglobin-containing fractions were applied to a Whatman Myoglobin and myoglobin mRNA in fish heart 1279

DE-52 anion-exchange column (5 ml bed volume) and eluted (50Ð300 mg) by acid guanidinium thiocyanateÐphenolÐ with a 60 ml gradient of 0Ð50 mmol l−1 KCl in 10 mmol l−1 Tris chloroform extraction and reconstituted in sterile water buffer, pH 8.5. Eluted myoglobin fractions were pooled, (Chomczynski and Sacchi, 1987). The typical yield of total −1 −1 dialyzed against 10 mmol l NH4HCO3, divided into samples, RNA was 1.5 µg RNA mg tissue, with similar extraction lyophilized and stored at −80 ¡C. efficiencies among tissues. Extracted RNA was stored at −80 ¡C until use. Protein gel electrophoresis and western blotting The concentration of extracted RNA was determined in The purity of isolated Notothenia coriiceps myoglobin triplicate by spectrophotometric analysis. The integrity of the standard was established by denaturing sodium dodecyl sulfate RNA was verified by examination of 18S and 28S ribosomal polyacrylamide gel electrophoresis (SDSÐPAGE) and bands stained with ethidium bromide, following separation of immunoblot analysis as described previously (Sidell et al., the RNA by electrophoresis on 1.2 % agarose formaldehyde 1997). Confirmation that the single immunopositive band gels (Sambrook et al., 1989). The RNA was then transferred produced by purified Notothenia coriiceps myoglobin on one- to GeneScreen Plus nylon membrane (NEN) by dimensional gels represented a single protein was obtained by action and cross-linked to the membrane by ultraviolet two-dimensional electrophoresis performed according to the irradiation at a dose of 0.12 J cm−2 (Kodak IBI Ultralinker method of O’Farrell (1975) by Kendrick Labs, Inc., and by our 400). A 329 base pair (bp) myoglobin cDNA insert laboratory, using a method modified from that of Hochstrasser (corresponding to codons 5Ð114) isolated from Notothenia et al. (1988) (Theresa Grove, personal communication, data not coriiceps (kindly provided by Dr M. E. Vayda) was used to shown). construct a specific probe for myoglobin mRNA, as described previously (Sidell et al., 1997). Identification of tissues expressing myoglobin protein and estimation of intracellular concentrations of myoglobin Slot blot quantification of myoglobin mRNA The presence and intracellular concentrations of myoglobin Slot blot (Bio-Rad manifold) analysis was used to quantify in cardiac ventricular and pectoral fin adductor profundus the amount of myoglobin mRNA present in heart ventricle and muscle were determined by preparing a 10 % (w/v) pectoral adductor tissues. Prior to slot blotting, northern blots homogenate (1 vol of original wet mass of tissue plus 9 vols of (for details, see Sidell et al., 1997) were performed on all 20 mmol l−1 Hepes buffer, pH 7.8 at 4 ¡C). For each samples to verify that total RNA was not degraded and to preparation, 20Ð50 mg wet mass of frozen cardiac or pectoral verify specific hybridization of myoglobin cDNA probe to an tissue was placed in an ice-chilled glass tissue grinder with an mRNA of the proper size (0.9 kb). Denatured Notothenia appropriate volume of buffer and homogenized. The coriiceps myoglobin cDNA insert served as an internal positive homogenate was centrifuged at 10 000 g for 10 min at 4 ¡C, and control and was used to generate a standard curve within each the resulting supernatant (heart ventricle or pectoral adductor) blot. The starting concentration of cDNA was determined was collected. Protein concentrations of heart ventricle and fluorometrically, then diluted serially from 100 to 1.5 pg. Yeast pectoral muscle supernatants were determined by tRNA was used as a negative control on each blot. bicinchoninic acid (BCA) assay (Sigma), using a bovine serum Total RNA (5 µg) (serial dilution 5.0Ð1.25 µg) was loaded albumin standard, and separated electrophoretically on per sample. Myoglobin probe preparation, hybridization duplicate SDSÐPAGE gels (for electrophoresis and western conditions and blot treatment were identical to those described blotting conditions, see Sidell et al., 1997). above for the northern blot membrane. Washed blots were Myoglobin concentrations in supernatant lanes of exposed overnight to preflashed Kodak Biomax MR X-ray film Coomassie-stained gels were determined densitometrically at −70 °C. Film was preflashed with a Vivitar 283 electronic with a Sepra Scan 2001 flatbed scanner and software flash unit masked to raise the fog level of the film to between (Integrated Separation Systems). In addition to muscle 0.1 and 0.2 absorbance units above that of unexposed film. This supernatants, each gel contained a standard curve generated by procedure increased the sensitivity and linear range of the film. loading known amounts of purified Notothenia coriiceps Myoglobin mRNA was quantified by densitometry (see myoglobin. The concentration of purified Notothenia coriiceps above for scanner details) from the autoradiograph. To correct myoglobin was determined spectrophotometrically. for background, the densitometric value obtained from the Myoglobin concentrations in supernatant lanes were negative control (yeast tRNA) was subtracted from all calculated from the linear relationship between micrograms of densitometric values for unknowns. The standard curve from purified myoglobin loaded and integrated signal-area each blot was used to convert densitometric values of (r2=0.95Ð0.99 for all gels). All unknown values (myoglobin in unknowns into picograms of myoglobin mRNA detected. supernatants) consistently fell within the range of the standard These values were then expressed as picograms of myoglobin curve. mRNA detected per microgram of total RNA loaded (pg Mb mRNA µg−1 total RNA). Densitometric values of RNA gel electrophoresis and northern blotting unknowns consistently fell within the linear range of the Total RNA was isolated from finely ground, frozen heart standard curve (r2=0.96Ð0.99 for all blots). ventricle and pectoral fin adductor profundus muscle Values are presented as means ± S.E.M. 1280 T. J. MOYLAN AND B. D. SIDELL

Results were determined by SDSÐPAGE. Gels were loaded with Tissue-specific expression of myoglobin protein measured amounts of a myoglobin standard purified from Considerable variation in expression of myoglobin was Notothenia coriiceps (Fig. 2). Densitometric measurement of observed within the 13 (of the known 15) species of icefishes the curve generated by the standards was linear over the examined. Strong cross-reactivity with the polyclonal (not range of samples loaded. Myoglobin concentration estimates shown) and monoclonal anti-myoglobin antibodies (Fig. 2) for channichthyid species ranged from 0.44±0.02 to −1 was observed in heart ventricle of eight of the 13 icefish 0.71±0.08 mg Mb g wet muscle mass (N=1Ð6) and those species examined (Chionodraco rastrospinosus, Chionodraco for the two nototheniids were 0.85±0.11 and −1 hamatus, Chionodraco myersi, Pseudochaenichthys 1.12±0.07 mg Mb g wet muscle mass (N=6) (Table 1). Levels georgianus, Cryodraco antarcticus, Chaenodraco wilsoni, of myoglobin protein in icefishes are comparable with those Chionobathyscus dewitti and Neopagetopsis ionah). Five found in the red-blooded notothenioid fishes examined and are icefishes (Champsocephalus gunnari, Chaenocephalus similar to values reported by Sidell et al. (1987) for another aceratus, Dacodraco hunteri, Pagetopsis maculatus and nototheniid, Notothenia rossii. Pagetopsis macropterus) lack detectable levels of myoglobin protein in heart ventricle. Examination of pectoral adductor Tissue-specific expression of myoglobin mRNA muscles from icefishes revealed that myoglobin protein is not In the channichthyid and nototheniid species examined that expressed in this highly aerobic oxidative tissue by any species do express myoglobin protein (see above), hybridization of a of this family (see also Sidell et al., 1997). Similar tissue- myoglobin-mRNA-specific probe identified a single band of specific myoglobin protein expression was observed in the 0.9 kb in northern blots of total RNA extracted from cardiac related red-blooded nototheniid species Gobionotothen ventricle (see also Sidell et al., 1997). Four of the icefish gibberifrons and Trematomus newnesi. species that lacked detectable levels of myoglobin protein (Chaenocephalus aceratus, Dacodraco hunteri, Pagetopsis Intracellular concentration of myoglobin protein in heart maculatus and Pagetopsis macropterus) also lacked detectable ventricle myoglobin mRNA. The tissue-specific pattern observed for Intracellular concentrations of myoglobin in heart ventricle protein expression was also exhibited for myoglobin message.

Fig. 1. Hearts from three species of notothenioid fishes. The channichthyid icefish Chaenocephalus aceratus has a pale yellow ventricle (far left) and lacks myoglobin protein expression. The channichthyid icefish Chionodraco rastrospinosus expresses myoglobin protein (0.64±0.07 mg Mb g−1 wet mass) and displays a distinctly rose-colored ventricle (middle). In comparison, the related nototheniid species Notothenia coriiceps has a characteristically red ventricle (far right) associated with the presence of myoglobin protein (concentration estimate 3.04 mg Mb g−1 wet mass; T. J. Moylan and B. D. Sidell, unpublished data). Myoglobin and myoglobin mRNA in fish heart 1281 A 12345679101112 Fig. 2. Determination of myoglobin concentrations in kDa heart ventricles of notothenioid fishes. (A) 15 % tricine gel of supernatants from heart ventricle (35 µg 45.0 each) of the channichthyid icefish Cryodraco antarcticus. Lanes 3, 5, 7, 10, heart ventricle 34.7 supernatant from four individuals; lane 1, heart ventricle supernatant from Chaenocephalus aceratus (negative control). Lanes 2, 4, 6, 9, 11, standard curve of 0.4, 0.8, 1.2, 1.6, 2.0 µg of purified 24.0 Notothenia coriiceps myoglobin, respectively. Numbers alongside lane 12 correspond to sizes (in 18.4 kDa) of molecular mass standards in that lane. Quantification of myoglobin protein (Table 1) was accomplished by densitometric scan of the Mb myoglobin-associated band (Mb, arrow). 14.3 (B) Immunoblot of a duplicate gel to A (above) with monoclonal mouse anti-human myoglobin antibody. The results indicate the expression of myoglobin protein in heart ventricles of Cryodraco antarcticus and the absence of detectable myoglobin protein in Chaenocephalus aceratus. Only the molecular mass B range containing myoglobin (arrow) is shown. 18.4 (C) Plot of the standard curve from A (above). The integrated area of scanned bands from standards is Mb plotted against micrograms of protein loaded. The 14.3 area of myoglobin bands from unknowns consistently fell within this standard curve.

0.20 known amounts of unlabeled myoglobin cDNA insert. Data C obtained from identically treated samples of total RNA from 0.18 heart ventricle and pectoral adductor profundus were plotted 0.16 against the standard curve to quantify the amount of myoglobin 0.14 mRNA present in these tissues. The concentrations of mRNA were expressed as picograms of myoglobin mRNA per 0.12 − r2=0.98 microgram of total RNA (pg Mb mRNA µg 1 total RNA). 0.10 Autoradiographs of blots revealed that values for unknowns 0.08 consistently fell within the linear portion of the standard curve.

Integrated area 0.06 Myoglobin mRNA was quantified from the same animals used for the determination of myoglobin protein concentration when 0.04 sufficient tissue was available. For some species, tissue was 0.02 limiting, and sample sizes for mRNA determinations are 0 smaller than those for myoglobin protein. 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 Total myoglobin mRNA in heart ventricle for Purified Notothenia coriiceps myoglobin (µg) channichthyid species that express the protein ranged from 0.78±0.02 to 16.22±2.17 pg Mb mRNA µg−1 total RNA (N=1Ð6) (Table 1). The lowest myoglobin mRNA When detected, myoglobin message was present only in concentrations were found in Champsocephalus gunnari cardiac ventricle, while all oxidative skeletal muscle examined (0.33±0.09 pg Mb mRNA µg−1 total RNA, N=6), a species of lacked detectable signal, indicating tissue-specific expression icefish known to lack detectable myoglobin protein while still of the gene. expressing mRNA (Sidell et al., 1997). Myoglobin mRNA concentrations for the two red-blooded nototheniid species Concentration of myoglobin mRNA in heart ventricle fell between the high and low values determined for the The concentration of myoglobin mRNA was determined by Channichthyidae. We are unaware of any published values slot blot analyses using the same myoglobin-mRNA-specific for myoglobin mRNA concentrations in teleosts. However, probe utilized for northern blot analyses. Fig. 3 shows an the values reported here for Antarctic fishes are example of the standard curve constructed by hybridization of approximately five times lower than those reported by Weller a constant amount of 32P-labeled myoglobin cDNA probe with et al. (1986) for human cardiac myoglobin mRNA and fall 1282 T. J. MOYLAN AND B. D. SIDELL

Table 1. Concentrations of myoglobin (Mb) and Mb mRNA in heart ventricles of Antarctic notothenioid fishes Mb concentration Mb mRNA concentration Taxon (mg Mb g−1 wet mass) N (pg Mb mRNA µg−1 total RNA) N Channichthyidae Chionodraco rastrospinosus 0.64±0.07 6 16.22±2.17 6 Chionodraco hamatus 0.62±0.04 6 0.78±0.02 4 Chionodraco myersi 0.71±0.08 4 Pseudochaenichthys georgianus 0.46±0.04 6 1.97±0.77 4 Cryodraco antarcticus 0.44±0.02 6 2.05 1 Chaenodraco wilsoni 0.65±0.08 6 5.31 1 Chionobathyscus dewitti 0.69±0.03 2 Neopagetopsis ionah 0.70 1 Champsocephalus gunnari ND 6 0.33±0.09 6 Chaenocephalus aceratus ND 6 ND 6 Dacodraco hunteri ND 4 ND 4 Pagetopsis macropterus ND 2 ND 2 Pagetopsis maculatus ND 1 ND 1 Nototheniidae Gobionotothen gibberifrons 0.85±0.11 6 7.27±1.94 6 Trematomus newnesi 1.12±0.07 6 7.11±1.04 5

Concentration data are presented as mean ± S.E.M.; ND, not detected by western or northern blotting. within a range that can be characterized as a rare message myoglobin concentration of heart ventricle is comparable in (Alberts et al., 1994). In those species expressing the protein, both red-blooded Antarctic nototheniid fishes and those steady-state concentrations of myoglobin protein are not hemoglobinless channichthyid icefishes that express the paralleled by steady-state concentrations of myoglobin protein. Tissue myoglobin concentrations reported for mRNA within and among icefish species, indicating that the these polar fishes (Table 1) are also similar to those of synthesis and/or turnover of the protein is not directly sedentary benthic fishes from temperate-zone latitudes, such determined by the size of the myoglobin mRNA pool. as sea raven (Hemitripterus americanus) and long-horn sculpin (Myoxocephalus octodecimspinosus), 1.0 and 1.2 mg Mb g−1 wet mass, respectively (Driedzic and Stewart, Discussion 1982). However, they are lower than those of more active Expression of myoglobin protein temperate-zone species, such as striped bass Morone saxatilis The present study extends our knowledge of the pattern of (6.0 mg Mb g−1 wet mass; Sidell et al., 1987). The myoglobin myoglobin protein and mRNA expression to 13 of the 15 content of the heart muscle of fishes has been positively known species of Antarctic channichthyid icefishes. Using the correlated with the ecological physiology of a variety of same definitive immunochemical approach, we have confirmed species ranging in lifestyle from benthic sedentary to active the results of Sidell et al. (1997) and further identified three pelagic (Giovane et al., 1980). These observations make it additional species of icefish that produce myoglobin tempting to suggest that the variation in myoglobin expression (Chionobathyscus dewitti, Neopagetopsis ionah and observed among Antarctic icefishes might be similarly Chionodraco myersi) and two additional species from which attributable to life history differences. Indeed, a wide range of the protein is absent (Dacodraco hunteri and Pagetopsis relative activity levels has been reported among icefishes on maculatus) (see Table 1). Our results also document the very the basis of their feeding behaviors. Lifestyles range from high degree of tissue specificity in expression of myoglobin by sluggish demersal, such as Chaenocephalus aceratus, to semi- notothenioid fishes. In a pattern that departs from the pelagic feeders, such as Pseudochaenichthys georgianus, to the norm, myoglobin protein is detectable only in heart ventricle pelagic Champsocephalus gunnari, which feeds exclusively on and is absent from other aerobic muscles, including the in the water column (Eastman, 1993). Myoglobinless primary oxidative skeletal muscle of these labriform species occupy both ends of this range of activities within the swimmers, the pectoral adductor profundus. icefish family. Thus, no obvious causative relationship appears We included lanes containing known amounts of to exist between lifestyle and whether or not myoglobin is myoglobin purified from Notothenia coriiceps into each of expressed in hearts of these animals. This lack of correlation our SDSÐPAGE analyses of muscle extracts. In addition to a with activity level is particularly curious in the light of simple test for the presence/absence of myoglobin protein, experiments that strongly implicate a physiological role for the this approach permitted us to quantify the amount of protein in these animals (Cashon et al., 1997; Acierno et al., myoglobin in each extract. Our results reveal that the 1997). Myoglobin and myoglobin mRNA in fish heart 1283 A hearts; Vayda et al., 1997.) Consistent detection of myoglobin mRNA in Champsocephalus gunnari, despite the absence of − STD ( ) 12345 any detectable protein, has been reported previously (Sidell et al., 1997). A five-nucleotide duplication that is unique to this species causes a shift in reading frame of the message downstream from amino acid residue 91 and premature termination at residue 103 and is apparently responsible for the production of a defective translation product that is degraded immediately (Vayda et al., 1997). Finally, we have been unable to detect the presence of mRNA encoding for myoglobin in oxidative skeletal muscles from any of the channichthyid icefishes sampled and 12 other red-blooded notothenioid species, representing four different families (T. J. Moylan and 2.5 B. D. Sidell, unpublished data). This observation strongly B indicates that the event leading to loss of myoglobin expression in oxidative skeletal muscle occurred very early in the 2.0 notothenioid lineage, presumably before the divergence of the extant families. Having successfully quantified the levels of myoglobin in 1.5 those icefish species that produce the protein, we hoped to gain r2=0.98 some insight into factors that regulate its intracellular 1.0 concentration. Perhaps one of the most straightforward possibilities is that the concentration of myoglobin in the tissue Integrated area is determined directly by the pool size of mRNA encoding the 0.5 protein. To test this possibility, we sought to quantify concentrations of myoglobin-specific mRNA in the same tissues. In these experiments, one typically also probes for a 0 transcript encoding a constitutively expressed housekeeping 0 1020304050 gene to control for unequal RNA loading between samples. For Notothenia coriiceps cDNA STD (pg) this purpose, H. W. Detrich III (Northeastern University) Fig. 3. Determination of myoglobin (Mb) mRNA concentrations in generously provided a cDNA probe specific for β-tubulin heart ventricles of notothenioid fishes. Slot blot analysis performed prepared from Notothenia coriiceps, the same species from on total RNA extracted from heart ventricles. Myoglobin mRNA was which the myoglobin probe was generated. Preliminary detected by hybridization with random-primed (Boehringer- northern blot analyses with the tubulin probe indicated similar Mannheim) 32P-labeled Notothenia coriiceps myoglobin probe. levels of β-tubulin mRNA in all tissues selected. However, (A) Autoradiograph of slots used for the analysis. STD, Notothenia more sensitive slot blot analyses revealed sufficient variability coriiceps Mb cDNA insert loaded (50.0, 25.0, 12.5, 6.25, 3.10, in tubulin mRNA expression to preclude its use as a control. − 1.55 pg), ( ), yeast tRNA negative control; lanes 1, 2, 3, 4 and 5 are Because the extraction efficiencies for total RNA were representative loadings of sample total RNA from unknowns (5.0, similar among tissues, we expressed myoglobin mRNA levels 2.5 and 1.25 µg). (B) Plot of the standard curve used in determination of myoglobin mRNA concentrations. Densitometric values as a fraction of total RNA loaded in each lane. Using this associated with the integrated area of scanned slots (STD) are shown approach, total myoglobin mRNA in heart ventricles of as a function of picograms of Notothenia coriiceps Mb cDNA insert channichthyid species ranged from 0.78±0.02 to 16.22± − loaded. Absorbance values of unknowns consistently fell within the 2.17 pg Mb mRNA µg 1 total RNA in species that produce the linear range of the standard curve. protein. These are, to our knowledge, the first values reported for myoglobin mRNA content in fish tissues. The published literature contains no unifying consensus on Expression of myoglobin mRNA the relationship between concentrations of Mb protein and Using a myoglobin-specific cDNA probe, we were able to Mb mRNA in tissues. Weller et al. (1986) found myoglobin identify mRNA of the appropriate size in extracts from heart mRNA levels in human cardiac tissue of approximately ventricle of several myoglobin-protein-expressing species of 4 µg Mb mRNA µg−1 poly(A) RNA. If we assume that Antarctic fishes (see Table 1). Our northern blots did not reveal poly(A) mRNA makes up 2Ð3 % of the total cytoplasmic RNA encoding for myoglobin in hearts from four of the five RNA pool, this value corresponds to approximately species that do not express the protein. (Although not 80 pg Mb mRNA µg−1 total RNA, approximately five times detectable by northern blot analyses, we do know that very low higher than that found in icefish. The concentration of levels of transcript can be detected by polymerase chain myoglobin protein in human heart is approximately seven- to reaction amplification of cDNA from Pagetopsis macropterus 10-fold higher than that in the hearts of Antarctic icefishes, 1284 T. J. MOYLAN AND B. D. SIDELL initially suggesting that a relationship might exist between Evolutionary context of myoglobin expression in icefishes myoglobin-specific mRNA pool size and the intracellular Mapping the pattern of myoglobin protein and myoglobin concentration of the protein. However, myoglobin protein mRNA expression upon the phylogenetic tree of the icefish levels in normal human muscle are generally higher in type family permits us to address several features regarding the I fibers than in type II fibers (Jansson and Sylven, 1983), but evolution of these traits. To accomplish this, we used a no obvious difference between the amount of myoglobin consensus phylogeny based upon a combination of mRNA in the two fiber types can be detected by in situ morphological (Iwami, 1985) and more recent molecular hybridization (Mitsui et al., 1993). Weller et al. (1986) did biological mitochondrial (mt)DNA characters (Chen et al., find that levels of myoglobin protein and mRNA correlated 1998). Our first conclusion is that mutations leading to loss of in comparisons between skeletal muscle of seals and various cardiac myoglobin expression have occurred independently at muscle tissues from humans, but that mRNA levels in mouse least four times during the evolution of the family (Fig. 4). The skeletal muscle were higher than expected on the basis of five species that do not produce myoglobin protein these interspecific comparisons. (Champsocephalus gunnari, Chaenocephalus aceratus, Our results indicate that a direct relationship between Dacodraco hunteri, Pagetopsis macropterus and Pagetopsis mRNA pool size and protein concentration does not, maculatus) belong to four distinct clades, at least two of which however, exist for myoglobin in tissues from Antarctic contain members that do express the protein. Second, the icefishes. In comparing across closely related channichthyid pattern of myoglobin mRNA expression suggests that the loss icefish species, we can conclude that the steady-state of myoglobin protein has occurred by at least three concentration of myoglobin protein does not parallel the pool independent molecular mechanisms during the evolution of the size of myoglobin mRNA in the same tissue. As an icefish family. In Champsocephalus gunnari, myoglobin illustration, a difference of approximately 20-fold in mRNA is present in modest amounts, but the corresponding myoglobin mRNA level is found between the congeneric protein cannot be detected. Although not detectable by species Chionodraco rastrospinosus and Chionodraco northern blot analysis, very low concentrations of myoglobin hamatus (Table 1). However, this marked difference in mRNA are produced by Pagetopsis macropterus (Vayda et al., mRNA pool size is not reflected in the concentration of 1997). The mutations underlying the loss of myoglobin myoglobin protein, which is not significantly different expression in these two species have now been identified and between hearts from the two species. These results suggest are different (Small et al., 1998). Chaenocephalus aceratus, in that myoglobin protein content in the hearts of Antarctic contrast, apparently does not transcribe the myoglobin gene at icefishes is not determined by the size of the mRNA pool all (Small et al., 1998). All these processes are quite distinct encoding for the protein, but must be controlled at the points from the mechanism responsible for the hemoglobinless state of translation or post-translationally (e.g. degradation rate of this family, deletion of the gene encoding β-globin subunits constant for the protein). from the (Cocca et al., 1995).

Number of Myoglobin species Species Genus in genus examined Protein mRNA Champsocephalus 2 gunnari − + Pagetopsis 2 macropterus* − − maculatus − − Neopagetopsis 1 ionah ++ Pseudochaenichthys 1 georgianus ++ Dacodraco 1 hunteri − − 1 Cryodraco 1 antarcticus ++ Chionobathyscus 1 dewitti ++ Chaenocephalus 1 aceratus − − Chionodraco 3 myersi ++ rastrospinosus ++ hamatus ++ Chaenodraco 1 wilsoni ++

Fig. 4. Phylogenetic topology of myoglobin and myoglobin mRNA expression in heart ventricle among channichthyid species. The phylogenetic relationships within the Channichthyidae are based on a cladistic analysis of morphological and molecular characters (redrawn from Iwami, 1985; Chen et al., 1998). Data for Chionobathyscus dewitti, Neopagetopsis ionah, Chionodraco myersi, Dacodraco hunteri and Pagetopsis maculatus are from the present paper. Data for all other species are from Sidell et al. (1997). *Although not detectable by northern blot analyses, we do know that polymerase chain reaction amplification of cDNA from Pagetopsis macropterus hearts shows that very low levels of transcript are produced in this species (Vayda et al., 1997). Myoglobin and myoglobin mRNA in fish heart 1285

The seemingly random loss of myoglobin expression within We gratefully acknowledge the following people for the Channichthyidae at different times and by different generously providing samples; Arthur DeVries (University of mechanisms seems to be at odds with both biochemical (Cashon Illinois), Tetsuo Iwami (Tokyo Kasei Gakuin University, et al., 1997) and physiological (Acierno et al., 1997) Japan) and Raffaele Acierno and Guido di Prisco (Italian information, indicating that the protein confers functional National Antarctic Program). We are also indebted to the advantage to these animals, and molecular data (Vayda et al., masters and crew of the R/V Polar Duke and the personnel at 1997) that show very high levels of sequence conservation in the the US Antarctic Program’s Palmer Station who supported gene. Each of these latter lines of evidence suggests that our work while in Antarctica. US National Science selective pressure should mitigate towards retention of Foundation Grants OPP 92-20775 and 94-21657 to B.D.S. myoglobin expression. Perhaps the seemingly contradictory funded this work. nature of these observations is really the product of our own tendency to want to see evolutionary events through the distorted lens of ‘black-and-white fallacy’. That is, both the trait is References advantageous and its loss is lethal or, alternatively, it is at best Acierno, R., Agnisola, C., Tota, B. and Sidell, B. D. (1997). neutral in functional terms and its loss can occur randomly. The Myoglobin enhances cardiac performance in Antarctic species that pattern of myoglobin loss among Antarctic icefishes appears to express the pigment. Am. J. Physiol. 273, R100ÐR106. illustrate that evolutionary patterns can be subtler than this Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K. and Watson, J. D. (1994). Molecular Biology of the Cell, pp. 335Ð396. New binary view. Here, the truth appears to be ‘grey’ rather than York: Garland Publishing, Inc. ‘black or white’. In these animals, loss of the ability to express Bargelloni, L., Ritchie, P. A., Patarnello, T., Battaglia, B., this physiologically functional protein appears to reduce the Lambert, D. M. and Meyer, A. (1994). Molecular evolution at scope for cardiac performance, but is not lethal. subzero temperatures: mitochondrial and nuclear phylogenies of At the level of the individual organism, the non-lethality of fishes from Antarctica (Suborder Notothenioidei) and the evolution the loss of myoglobin is relatively easy to explain on the basis of antifreeze glycoproteins. Mol. Biol. Evol. 11, 854Ð863. of a combination of environmental and organismal Cashon, R. E., Vayda, M. E. and Sidell, B. D. (1997). Kinetic characteristics. Because of both very cold temperature and very characterization of myoglobins from vertebrates with vastly pronounced vertical mixing, the waters of the Southern Ocean different body temperatures. Comp. Biochem. Physiol. 117B, are characterized by exceptionally high oxygen content. 613Ð620. Controversies regarding metabolic cold-adaptation Chen, W.-J., Bonillo, C. and Lecointre, G. (1998). Phylogeny of the Channichthyidae (Notothenioidei, Teleostei) based on two notwithstanding, absolute metabolic rates of Antarctic fishes mitochondrial genes. In Fishes of Antarctica (ed. G. di Prisco, E. are relatively low because of their cold body temperature. Pisano and A. Clarke), pp. 287Ð298. Heidelberg: Springer-Verlag. Thus, the conjunction of high oxygen availability with low Chomczynski, P. and Sacchi, N. (1987). Single-step method of RNA absolute oxygen demand may explain why the loss of cardiac isolation by acid guanidinium thiocyanateÐphenolÐchloroform myoglobin is not a lethal mutation. However, these features do extraction. Analyt. Biochem. 162, 156Ð159. not address the more perplexing question that appears to fly in Cocca, E., Ratnayake-Lecamwasam, M., Parker, S. K., the face of modern population genetics theory: why is an Camardella, L., Ciaramella, M., diPrisco, G. and Detrich III, apparently disadvantageous (on the basis of cardiac H. W. (1995). Genomic remnants of α-globin genes in the performance) trait not subject to negative selection against hemoglobinless Antarctic icefishes. Proc. Natl. Acad. Sci. USA 92, those species that lack myoglobin? 1817Ð1821. The crux of the question is whether the reduction in cardiac Covell, D. G. and Jacquez, J. A. (1987). Does myoglobin contribute significantly to diffusion of oxygen in red skeletal muscle? Am. J. performance that accompanies loss of myoglobin expression is Physiol. 252, R341ÐR347. disadvantageous to icefish species. Implicit in this question is DeWitt, H. H. (1971). Folio 15. In Antarctic Map Folio Series (ed. that the disadvantage would be one that occurs through V. C. Bushnell), pp. 1Ð10. New York: American Geographical competition. Can a competitive disadvantage exist in the Society. absence of competition? We know that, some time between the Douglas, E. L., Peterson, K. S., Gyso, J. R. and Chapman, D. J. mid-Tertiary and the present, there was a massive crash of (1985). Myoglobin in the heart tissue of fishes lacking hemoglobin. species diversity in the Southern Ocean that left an ancestral Comp. Biochem. Physiol. 81A, 855Ð888. stock of demersal notothenioids to colonize this expansive Driedzic, W. R. and Stewart, J. (1982). Myoglobin content and the marine environment. Although the proximate cause of this activities of enzymes of energy metabolism in red and white fish event is not known with certainty, it is widely considered to be hearts. J. Comp. Physiol. B 149, 67Ð73. the reason for the ultimate dominance of notothenioid species Eastman, J. T. (1990). The biology and physiological ecology of notothenioid fishes. Fishes of the Southern Ocean (ed. O. Gon and in the fish fauna of Antarctica (Eastman, 1993). Perhaps the P. C. Heemstra), pp. 34Ð51. Gramstown, South Africa: J. L. B. combined evolutionary features of relatively low niche Smith Institute of Ichthyology. competition in marine habitats depauperate of fish species and Eastman, J. T. (1993). Antarctic Fish Biology: Evolution in a Unique the uniquely cold and oxygen-rich waters of the Southern Environment. New York: Academic Press. Ocean help explain why icefish species lacking functional Eastman, J. T. and Grande, L. (1989). Evolution of the Antarctic myoglobin persist and thrive. fish fauna with emphasis on the recent notothenioids. In Origins 1286 T. J. MOYLAN AND B. D. SIDELL

and Evolution of the Antarctic Biota, Geological Society Special In situ hybridization of myoglobin mRNA: results on the skeletal Publication 47 (ed. J. A. Crame), pp. 241Ð252. London: Geological muscles of normal subjects and patients with neuromuscular Society of London. diseases. Acta Neuropathol. 86, 313Ð318. Giovane, A., Greco, G., Maresca, A. and Tota, B. (1980). O’Farrell, P. H. (1975). High-resolution two-dimensional Myoglobin in the heart ventricle of tuna and other fishes. electrophoresis of proteins. J. Biol. Chem. 250, 4007Ð4021. Experientia 36, 219Ð220. Ruud, J. T. (1954). Vertebrates without erythrocytes and blood Hamoir, G. (1988). Biochemical adaptation of the muscles of the pigment. Nature 173, 848Ð850. channichthyidae to their lack in hemoglobin and myoglobin. Comp. Sambrook, J., Fritsch, E. F. and Maniatis, T. (1989). Molecular Biochem. Physiol. 90B, 557Ð559. Cloning: A Laboratory Manual. New York: Cold Spring Harbor Hemmingsen, E. A. (1991). Respiratory and cardiovascular Laboratory. adaptation in hemoglobin-free fish: resolved and unresolved Sidell, B. D., Driedzic, W. R., Stowe, D. B. and Johnston, I. A. problems. In Biology of Antarctic Fish (ed. G. di Prisco, B. Maresca (1987). Biochemical correlations of power development and and B. Tota), pp. 191Ð203. New York: Springer-Verlag. metabolic fuel preferenda in fish hearts. Physiol. Zool. 60, 221Ð232. Hemmingsen, E. A., Douglas, E. L., Johansen, K. and Millard, R. Sidell, B. D., Vayda, M. E., Small, D. J., Moylan, T. J., W. (1972). Aortic blood flow and cardiac output in the hemoglobin- Londraville, R. L., Yuan, M.-L., Rodnick, K. J., Eppley, Z. A. free fish Chaenocephalus aceratus. Comp. Biochem. Physiol. 43A, and Costello, L. (1997). Variation in the expression of myoglobin 1045Ð1051. among species of hemoglobinless Antarctic icefishes. Proc. Natl. Hochstrasser, D. F., Harrington, M. G., Hochstrasser, A. C., Acad. Sci. USA 94, 3420Ð3424. Miller, M. J. and Merril, C. R. (1988). Methods for increasing Small, D. J., Vayda, M. E. and Sidell, B. D. (1998). A novel the resolution of two-dimensional protein electrophoresis. Analyt. vertebrate myoglobin gene containing three A+T-rich introns is Biochem. 173, 424Ð435. conserved among Antarctic teleost species, which differ in Iwami, T. (1985). Osteology and relationships of the family myoglobin expression. J. Mol. Evol. 47, 156Ð166. Channichthyidae. Mem. Nat. Inst. Polar Res. Ser 36E, 1Ð69. Vayda, M. E., Small, D. J., Yuan, M.-L. and Sidell, B. D. (1997). Jansson, E. and Sylven, C. (1983). Myoglobin concentration in Conservation of the myoglobin gene among notothenioid fishes. single type I and type II muscle fibers in man. Histochem. 78, Mol. Mar. Biol. Biotech.6, 207Ð216. 121Ð124. Weller, P. A., Price, M., Isenberg, H., Edwards, Y. H. and Littlepage, J. L. (1965). Oceanographic investigation in McMurdo Jeffreys, A. J. (1986). Myoglobin expression: early induction and Sound, Antarctica. In Biology of the Antarctic Seas, vol. II (ed. G. subsequent modulation of myoglobin and myoglobin mRNA A. Llano), pp. 1Ð37. Washington, DC: American Geophysical during myogenesis. Mol. Cell. Biol. 6, 4539Ð4547. Union. Wittenberg, B. A. and Wittenberg, J. B. (1989). Transport of Mitsui, T., Kawai, H., Naruo, T., Nishino, H. and Saito, S. (1993). oxygen in muscle. Annu. Rev. Physiol. 51, 857Ð878.