Ethics

Ethics approval was not required for this study as no live animals were used for this research.

Tissue Preparation

Whole eyes were removed from seven preserved specimens of Teuthowenia pellucida, which had been fixed in 5% buffered formalin and stored in 70–80% ethanol until processed. The eyes were washed several times in 0.1M Phosphate Buffer Saline (PBS) solution (pH 7.4) for 24 hours, before being submerged in 30% sucrose dissolved in PBS overnight. Eyes were either hemisected along the ventro—dorsal midline, vertically sectioning the lens in two, or were sectioned along the equator to obtain dorsal and ventral sections. Tissues were then embedded in TissueTek cryo-OCT compound (Fisher Scientifics) and frozen at -20°C before sectioning retinal samples using a LEICA cryostat (LEICA, Germany) set at a thickness of 16–20 μm. Eye sections were transferred to glycerine-coated slides and were stored at -20°C until stained.

Tissue sections were stained using a standard haematoxylin and eosin (HE) staining procedure for contrast of nuclei and cytoplasm, or were stained in fuchsin. The slides were washed in PBS, hydrated and placed into absolute ethanol for five minutes. The slides were then placed in a staining solution prepared by combining distilled water (730 ml), glacial acetic acid (20 ml), aluminum sulphate (17.6 gr), sodium iodate (0.2 gr), ethylene glycol (250 ml) and haematoxylin (2 gr) for five minutes. Slides were then washed under running distilled water and stain differentiated in 1% acid alcohol. After rinsing, the slides were dipped in 1% lithium carbonate and then washed in distilled water for five minutes. 1% eosin stain was used for contrast. Following staining, the slides were rinsed under distilled water and then rinsed consecutively in 95% ethanol, xylene, and mounted with DPX mounting medium (Sigma).

Stained samples were visualised using a LEICA DR2 bright field microscope. Images were captured using a LEICA DC 500 digital camera. A standard stage micrometer was used to calibrate the optical images.

The eye diameter and mantle length were measured using digital callipers for specimens smaller than 10 mm ML and a standard ruler for larger specimens. Values were entered into a Microsoft Excel spreadsheet and a scatter-plot graph was constructed. The Data Analysis tool was used in the construction of the best fit regression line.

Retinal sections were analysed and measured using the calibrated ruler superimposed onto the images at each magnification. The measurements were taken in central and peripheral areas of the retina (each location measured three times and averaged), from the junction of the outer segment with the posterior chamber, to the end of the support cell layer.

Relative eye size throughout ontogeny

Ontogenetic changes in eye-to-body-size ratio were investigated by measuring the eye diameter along the dorsal—ventral midline (0.5–33 mm) and mantle length (ML 1.5–165 mm) for specimens with intact eyes (Fig 1A). These two measurements maintained a linear growth ratio throughout ontogeny; Fig 1B shows the relationship plotted as the natural log of the raw values (R² value = 0.929). In specimens below ML 50 mm (n = 33), a close correlation between eye diameter and mantle length was observed, while greater variation was observed in larger specimens (n = 28). Beginning around ML 50 mm, a positive correlation was observed between body size and depth of capture (Fig 1C).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1.

(A) Schematic of T. pellucida, indicating measurements for mantle length (ML) and eye diameter (ED). (B) Log relationship between ML and ED. (C) Depth distribution of the 87 specimens examined in this study with known collection strata, plotted against ML.

https://doi.org/10.1371/journal.pone.0123453.g001

Retinal layers

The structure of the retina was examined for seven specimens (ML 9.7–102.0 mm, Table 1) representing progressive ontogenetic stages (stages A, C, E, and F, as outlined by Evans [10]). In all stages, the retina is connected to an optic ganglion through the wide optic nerve. In young specimens, retinal cross sections showed a thin anterior limiting membrane in contact with the striated outer segment of the photoreceptor (striation visible only at high magnifications); a thin, dark segment (ommin layer, the screening pigment); and the nuclei of supporting cells resting on a basal membrane that partitioned them from the inner segment of the photoreceptors. A plexiform layer was also observed between the inner segment of the photoreceptor and the axons, which form the most internal layer of the retina (Fig 2). Ontogenetic differences were noted in the thicknesses of several retinal layers (see Table 1): in younger specimens (e.g., M.070961) the outer and inner segments of each region were of nearly equal length (often differing by just 10 μm), while in older specimens, the outer segment was significantly longer than the inner and supporting cells(N = 9, p<0.0001). In general, photoreceptors in the ventral region of the retina also appeared longer than those found in the dorsal region (Fig 2B and 2D), although this was not tested statistically.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Measurements of visible retinal layers in the ventral retina of Teuthowenia pellucida.

https://doi.org/10.1371/journal.pone.0123453.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2.

(A) Cross-section of larval (ML 9.7 mm) T. pellucida eye, with insert (A’) showing retinal layers. (B–D) Changes in retinal thickness observed in serial ventral, equatorial and dorsal sections through retina of same specimen, with location of main ommin layer and migrated ommin granules shown in high magnification in B’–D’ (migrated granules indicated by white arrows). Abbreviations: basal membrane (bm), ciliary muscle (CM), ganglia (G), inner segment of photoreceptors (isp), lens (L), limiting membrane (lm), ommin layer (ol), optic nerve (O), optic fibers (of), outer segment of photoreceptors (osp), plexiform layer (pl), retina (R), supporting cells (sc). Scale bars = (A) 100 μm, all other images 10 μm.

https://doi.org/10.1371/journal.pone.0123453.g002

In paralarval and juvenile specimens (ML ≤ 50mm, n = 2), the ommin granules were distributed in two layers in the dorsal portion of the retina: a dense, dark layer between the inner and outer segments of the photoreceptors, and a sparser secondary layer at the outer end of the photoreceptor (Fig 2D’). Ventrally, ommin pigment granules were concentrated close to the main screening layer (Fig 2B’) but in the central equatorial region, some granules were observed in transit between the main ommin layer and the tip of the photoreceptors (Fig 2C’), in all three of the young specimens examined. In specimens above ML 50 mm (n = 5), the outer segment of all photoreceptors was longer than the inner segment (Fig 3A), and photoreceptors were noticeably longer in the ventral region of the retina than in the dorsal region. In mid-size specimens (ML 50–52 mm), the ommin granules appeared to have migrated toward the outer surface of the retina at the equator of the eye, and were sparsely scattered throughout the outer portion of the ventral part of the retina (Fig 3C’), while in the largest eyes analysed (from specimens of ML 102, 128 mm), all pigment granules were positioned within the main ommin layer. An oblique section through the outer segment of the photoreceptors in an eye with migrated ommin (NIWA 71673; Fig 3D) showed that the pigment granules appeared to be interspersed with, but exterior to, the photoreceptor cells themselves.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3.

(A) Cross-section of adult (ML 52 mm) T. pellucida retina showing retinal layers and progressive dorsal—ventral lengthening of outer photoreceptor segment. Migrated ommin pigment in dorsal (B) and equatorial (C) regions and representative areas are shown in high magnification (B’, C’). (D) Oblique section through outer photoreceptor segment shows that migrated ommin pigment is located between photoreceptor cells; black circles indicate areas of clustered pigment granules. Abbreviations: rudimentary cartilage (c). All other abbreviations as in Fig 2. Scale bars = (A) 100 μm, (B’–D) = 10 μm.

https://doi.org/10.1371/journal.pone.0123453.g003

Eye-size-to-body ratio

Many animals that rely on vision in the mesopelagic deep sea have proportionally large eyes [1]; the cephalopod orders Teuthida and Vampyromorpha are excellent examples. The vampire squid, Vampyroteuthis infernalis, has the largest eye-diameter-to-total-body-length ratio of any animal (about 1:6, 1:3–1:4 to mantle length) [12]. The largest known eyes in the animal kingdom belong to the colossal squid (Mesonychoteuthis hamiltoni, also in the family Cranchiidae), and attain at least 27–35 cm in diameter ([13]; pers. obs.). However, these eyes, while large in absolute terms, are about half the relative size (eye diameter about 10% ML, and 2–4% total body length, pers. obs.) of those in T. pellucida. Large, visually oriented predatory fish such as the swordfish (Xiphias gladius) have a similar eye-diameter-to-body-length ratio (about 2–3% at mature sizes of 3–5m; [13]) although their skulls could support much larger eyes, suggesting that, beyond a certain threshold, these large vertebrates are similarly subject to the law of diminishing returns in visual acuity with increased eye size [13]. Teuthowenia pellucida, however, which matures around 100 mm ML [10] likely falls within a body (and eye) size range throughout its lifespan where the extra energy expended in maximising eye size provides a worthwhile return in increased visual sensitivity; certainly the proportion of the head comprised of ocular tissue (about 88% by volume in sub-adults, pers. obs.) appears extreme.

Among cephalopods, the visual capabilities of those living in the deep sea are the least understood, in part because many of their tissues (including eye and nervous tissues) are often gelatinous and fragile, and severely damaged or destroyed entirely during capture [14]. Teuthowenia is no exception; the large, delicate eyes (observed diameter up to 33 mm, making the head width sometimes greater than the mantle width) were badly damaged on most of the material examined herein, permitting detailed histological examination of only a few specimens. When intact, these eyes appeared disproportionately large relative to the body, especially in sub-mature to mature individuals; however, observations made throughout ontogeny show a linear ratio of eye diameter to mantle length—about 1:5 (Fig 1B). Additional material, especially at the larger end of this species’ size range (reported to attain ML of 200+ mm, while the largest specimen examined herein was ML 165 mm) would be useful in examining this trend.

Voss reported that Teuthowenia, like other cranchiid squids, undergoes an ontogenetic vertical migration to the deep sea [11], an observation supported by collection data for the material examined herein (Fig 1C), and many cranchiid species are thought to migrate diurnally in the water column [15]. At the shallower depths occupied by paralarvae and juveniles, the eyes would be subjected to much higher light intensities, while larger eyes (in absolute terms) would maximise light sensitivity as the animal matures and descends into the aphotic zone. Perhaps the more surprising finding from the eye-size component of this study is therefore not that the adults have a relatively large eye-diameter-to-ML ratio compared to some other squids, but that juveniles, which inhabit brightly lit waters, have a similarly high ratio to that observed in adults. This could be, in part, due to the fact that camera-type eyes must be a certain absolute size in order to be useful/functional [16]; therefore, juveniles may require proportionally “large” eyes (relative to other invertebrates with non-camera-type eyes) simply in order for these structures to function, while adults require even larger eyes (in absolute terms) to maximise light sensitivity, both down-welling from the surface and bioluminescent signals from other deep-sea animals, within their very dark environment [17].

Muntz indicated that small animals, which move relatively slowly, may require shorter focal distances because objects of interest are often closer to them (Kirschfeld, 1974; 1976, as cited in [16]). Other authors have noted that many predators probably rely on having greater visual ranges than their prey, enabling them to identify prey before being detected themselves (e.g., Nilsson, as cited in [1]). Perhaps a squid species belonging to a genus whose main repertoire of defense behaviours relies altering the animal’s appearance (for example, ballooning the mantle and filling it with ink [18]), as an alternative to taking rapid evasive action, thus requires better visual acuity than similarly sized squid more reliant on actively fleeing—the visual deception needs to occur early enough to confuse or deter a potentially attacking predator.

Retinal structure

Although the cellular composition of the photoreceptor appears consistent throughout ontogeny, a change was observed in the proportions in the inner and outer photoreceptor lengths. In early ontogenetic stages, the lengths of the outer and inner photoreceptors were nearly equal, whereas the length of the outer photoreceptor increased dramatically in adults, comprising 89–97% of the total photoreceptor length (Table 1). Hypothetically, the onset of ontogenetic descent into deeper, darker habitats (a common life history trait in oceanic squids [19] could be partly triggered by the development of the eyes, and the lengthening of the photoreceptors—the increased sensitivity to light could drive maturing squid into lower strata, where less photic stimulation would occur. Alternatively, as small squid begin ranging into greater depths, changes in the retinal structure could be triggered by exposure to ever lower environmental light levels.

Central photoreceptors were also longer than photoreceptors located toward the retina’s periphery in the majority of examined specimens (Fig 4). This was consistent with the findings of Matsui et al. [7], who observed that photoreceptor length is greatest in areas where the most light reaches the retina, and similar to observations on the vertebrate inverted retina, where photoreceptor length has been reported to change by region (central—peripheral) and as a function of light—dark adaptation [20]. In T. pellucida, the longer outer segments found centrally suggest that the retinal area directly in the axial length is structured to gather the most incoming light.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Cross-section at the equator of adult (ML 130 mm) T. pellucida retina showing no migrated ommin in any region and peripheral-central changes in photoreceptor thickness.

Scale bar is 100μm. Arrows indicate length of outer photoreceptor segment. Abbreviations as in Figs 2 and 3.

https://doi.org/10.1371/journal.pone.0123453.g004

Ommin migration

The screening layer of ommin located between the inner and outer photoreceptors does not appear to change in absolute thickness with growth of the specimen; in most samples, the thickness of this layer remained approximately 10μm. While ommin density within this layer was not specifically examined, no obvious ontogenetic changes were observed, suggesting that the ommin protecting the shorter photoreceptors of younger individuals (which are exposed to bright lighting conditions for an extended portion of each day) may, without drastic increase, also sufficiently protect both the longer photoreceptors of older animals (which are less frequently exposed to bright light). However, the primary long-term mechanism facilitating vision in the dark environments inhabited by adults seems to be increased photoreceptor length. It remains unknown whether photoreceptor length in squid could also change (on a smaller scale) over shorter time-frames—e.g., in response to diel vertical migration—as has been observed in some vertebrates whose eyes can make short-term light/dark adaptive adjustments [21].

Daw and Pearlman [5] first reported ommin mobility in the squid Doryteuthis pealeii: when introduced to light-intense conditions, some ommin migrated from the “normal” position (within the screening layer) to the outer edge of the photoreceptor, over the course of 5–15 minutes. Since the action spectrum of the ommin migration matches that of rhodopsin (the visual pigment) [5], this adaptation presumably reduces or prevents over-stimulation by shielding the inner receptors from excess incoming light [22]. After the light source was removed, the ommin began to return to the screening layer within several minutes [5].

Similar migrations appear to happen in the eyes of Teuthowenia—all smaller specimens had a darkly pigmented “layer” at the tip of the outer receptor (Fig 3B), which became sparser or was not observed in more mature specimens. Young Teuthowenia are most commonly collected between 0 and 200 meters, while older animals are usually encountered from deeper, aphotic waters ([11]; Fig 1C); the retina thus receives far more light during early ontogenetic stages. The changes in photoreceptor length observed as the animal matures partially account for the eye’s ability to retain advanced visual capabilities across these disparate light conditions. The migratory screening pigment granules likely provide an additional means of protecting the retina during the bright light conditions experienced by young individuals (which presumably also need the short-term flexibility to see well at night). The ommin’s migratory capability appears to be retained into adulthood, but is probably less frequently utilised in the deep sea, and the thresholds and time frames of the response to light stimulus may differ from those of younger individuals.