How to go for Low- or Zero-Cost Shipping – Plasmyd

One of the consumer boons of the invention of the Internet is the abundance of online shopping opportunities, especially around the holidays. Unfortunately, many of those sale prices people see dont include shipping and handing charges. How many times have you jumped on a terrific discounted price that offers $10 off, only to shake your head when shipping costs adds $15, erasing that thrill that low prices can generate? Fortunately, there are a few ways to lower and sometimes eliminate shipping costs, though each might come with its own set of restrictions.

Recommended First Stop Online

If online shopping — and completely avoiding holiday crowds, traffic and fuel costs — is a major passion, consider your first stop This website lists stores that have either free shipping or severely reduced shipping costs or deep discount codes. This site is the impetus behind Free Shipping Day that many merchants have adopted to lure customers, guaranteeing a by-Christmas delivery schedule. This holiday season, Free Shipping Day is the 17th of December.

Regardless of the timing, this information site is available to control shipping costs all year round.

Minimum Orders

Many online retailers will offer free shipping if your order meets or exceeds a minimum purchase amount. That lowest benchmark differs from store to store, so be sure to read shipping information before you buy out the online store, then find out your shipping costs exceed the total purchase amount.

Some sites mix shipping costs per type of order. Some items may offer lower shipping costs than others. It can take a keen eye to note what costs to ship and what doesnt.

Local Store Shipping

Walmart is famous for this free shipping offer. Instead of having the items land at your door, if you or a representative are willing to travel to the store designation of your choice, the warehouses will ship to that location free of charge. Other locations have taken advantage of that consumer convenience, so if the e-store doesnt list that option, call your local store to verify.

Site Programs

Amazon has its Prime program that once joined with an annual fee, you pay nothing on designated items purchased through the website. In actuality, your shipping costs are part of your annual fee, but since you gain additional benefits with Amazon, such as free e-books and video streaming options, it doesnt feel like youre pre-paying all shipping costs for the next year.

See if your favorite online store has a similar program, especially if you make lots of online purchases there. However, if you surf their merchandise only occasionally, weigh any program fees against the actual shipping costs youd otherwise pay during the same interim. Some programs just arent cost-effective.

Rebate Sites

While the scope of rebate offers found on many of these web locations, some do offer free shipping from designated stores. If you cant find free shipping from the starter site, above, you might want to surf a few rebate sites to find that hidden shipping deal that makes a sale truly inexpensive, keeping costs low when possible.

by CreditRepairXP, a finance related website who enjoys sharing tips to keep costs low.

The cost of efficiency in energy metabolism


The cost of efficiency in energy metabolism

In a universe being dragged into disorder by the second law of thermodynamics, living cells must expend energy to maintain their complex organization. In addition to providing a carbon source for biosynthesis, the classical Embden–Meyerhof–Parnas (EMP) and Entner-Doudoroff (ED) pathways help to satisfy this energetic demand by generating ATP during glucose metabolism (1). Based on simple stoichiometry of reactants and products, the EMP pathway appears, at first blush, greatly preferable to the ED pathway, yielding twice as much ATP per glucose. If glucose breakdown and energy conservation are tightly coupled, why is the less-efficient ED pathway so prevalent? What has kept prokaryotic life in its entirety from casting off the ED pathway in favor of the more profitable EMP pathway? In PNAS, Flamholz et al. (2) address these questions by drawing on thermodynamics, enzyme kinetics, mathematical optimization, and genomics.

The first stage of glycolysis is characterized by an investment of ATP to phosphorylate glucose, which, so primed, is cleaved into two three-carbon intermediates (Fig. 1). The cell recoups its investment in the second phase of glycolysis (known as “lower glycolysis”), where oxidation of three-carbon intermediates directly generates ATP. Although the ED and EMP pathways overlap in part, they conspicuously differ in the number of three-carbon intermediates shunted down lower glycolysis. In the EMP pathway, glucose is phosphorylated twice, consuming two ATP, and both three-carbon intermediates (glyceraldehyde 3-phosphate, or G3P) enter lower glycolysis to produce two ATP each. In the ED pathway, glucose is only phosphorylated once, consuming one ATP, before being cleaved into one G3P and one pyruvate. The single G3P yields two ATP as in the EMP pathway, but pyruvate bypasses the bulk of lower glycolysis, foregoing ATP production. Thus, despite both pathways starting and ending with the same amount of glucose and lactate, the EMP pathway manages to extract two ATP per glucose, the ED pathway only one. The ED pathway is thought to predate the EMP pathway (dominant among eukaryotes) in the evolutionary timeline (3). Is the EMP pathway simply a fine-tuned adaptation of the ED pathway optimized for energy conservation, or does high ATP yield come at a cost?


Fig. 1. Structural differences between the ED and EMP pathways. This simplified diagram focuses on a few key aspects of the ED and EMP pathways (i.e., the flow of carbon and phosphate groups) to highlight the different organization of ATP-consuming and ATP-producing steps, leading to different ATP yields. Other important details (such as the stoichiometry of reducing equivalents NADH and NADPH) can be found in traditional depictions of these pathways. G6P, glucose 6-phosphate; F1,6BP, fructose 1,6-bisphosphate; G3P, glyceraldehyde 3-phosphate; Pi, inorganic phosphate.

Flamholz et al. (2) highlight a tradeoff that logically arises between a glycolytic pathway’s ATP yield and thermodynamic driving force. Free energy released during glucose breakdown can drive ATP synthesis, providing energy currency to the cell, or dissipate as heat, making the overall pathway more thermodynamically favorable (albeit less efficient) (4). By harvesting more free energy as ATP than the ED pathway, the EMP pathway operates closer to equilibrium (5). Conversely, production of pyruvate so early in the ED pathway–a highly exergonic reaction–dissipates ample free energy, leaving little for ATP generation. Flamholz et al. (2) show that, even if metabolites assume concentrations that make the least favorable reactions in each pathway as exergonic as possible, the EMP pathway faces much tighter thermodynamic bottlenecks than the ED pathway. Although a simplified representation of linear metabolic pathways (as in Fig. 1) conveys their tendency to proceed in the forward direction, it glosses over bottleneck reactions that are only weakly favorable, with a negative free-energy difference close to zero. Such highly reversible reactions can easily clog a pathway, limiting if not preventing forward flux.

Despite its thermodynamic obstacles, the EMP pathway is known to function effectively in living cells. How to reconcile these two contrasting observations? And how would we expect a microorganism to cope physiologically with a thermodynamic bottleneck? Flamholz et al. (2) address these questions by considering the enzyme production costs associated with each pathway. In its traditional form, the Michaelis–Menten equation predicts the rate of a reversible reaction as a function of kinetic parameters and the concentrations of enzyme and metabolites. Flamholz et al. cleverly use the Haldane relationship to rewrite the Michaelis–Menten rate law, expressing the reaction rate as a product of the enzyme level, a kinetic term that quantifies distance from substrate saturation, and a thermodynamic term (called the Net Flux Ratio) that quantifies distance from equilibrium. This expression shows how a high enzyme level can kinetically compensate for a reaction’s close proximity to equilibrium. One can now calculate, for each reaction, the amount of enzyme necessary to bring about a given overall flux (or steady-state reaction rate) through the pathway, in light of thermodynamic and kinetic parameters. More precisely, Flamholz et al. (2) use constraint-based optimization to find the metabolite and enzyme concentrations that minimize a pathway’s protein cost, defined as the product of enzyme abundance and molecular mass summed over all reactions. This approach falls in line with early optimization methods developed by Ebenhöh and Heinrich (6) and recent constraint-based models of metabolism, where assumed cell-level objectives partially compensate for missing knowledge (7).

Flamholz et al. (2) find that the high-yield EMP pathway incurs a greater protein cost (3.5-fold) than the ED pathway, necessitating higher enzyme levels to support the same flux. In effect, enzyme levels must strategically increase in the EMP pathway to counterbalance low thermodynamic driving force. According to experimental measurements cited by Flamholz et al., enzymes catalyzing the EMP pathway make up a sizable proportion of the proteome in both Escherichia coli and Saccharomyces cerevisiae. Altogether, Flamholz et al. depict a thermodynamically constrained EMP pathway driven forward by costly enzymatic machinery in the interest of high ATP yield. The ED pathway, by virtue of its low ATP yield, emerges as thermodynamically relaxed, requiring lower enzyme levels to operate. Interestingly, a previous computational analysis of the metabolic network in E. coli (based on elementary flux modes) also characterized the ED and EMP pathways as designed to reduce “investment cost” (high protein synthesis) and “operating cost” (poor ATP yield), respectively (8).

What circumstances in nature tip the scale between maximizing ATP yield and minimizing protein synthesis? Flamholz et al. (2) suggest that prokaryotes will adopt one strategy or the other depending on whether substrate-level phosphorylation makes up a small or large fraction of total ATP production. In completing the breakdown of lactate to CO2 downstream from glycolysis, aerobes produce 25–30 ATP through oxidative phosphorylation, rendering a single extra ATP per glucose dispensable. For anaerobes, which rely on weaker oxidants than O2 for respiration or use fermentation, substrate-level phosphorylation represents

the primary means of ATP production, justifying a greater investment in enzyme synthesis. By identifying enzymes unique to each pathway [e.g., 6-phosphogluconate dehydratase and 2-keto-3-deoxy-6-phosphogluconate aldolase in the ED pathway (9)] in over 500 annotated genomes, Flamholz et al. (2) study the distribution of glycolytic pathways among heterotrophic prokaryotes, and find that, as expected, anaerobes overwhelmingly favor the EMP pathway, while the ED pathway is statistically overrepresented among aerobes.

It is worth noting the relationship between the tradeoff studied by Flamholz et al. (2) and previous analyses of tradeoffs between metabolic rate and metabolic yield (5). In their exploration of ATP yield vs. protein cost, Flamholz et al. fix the flux through two pathways with different ATP yields and weigh the resulting protein costs. One could alternatively consider the same system under a fixed protein cost, in which case high yield comes at the expense of high flux (or reaction rate) (5). More broadly, the study of tradeoffs between the metabolic benefits of sustaining flux along a certain pathway, and the corresponding enzyme production cost, is emerging as a valuable gateway to several open research questions, from predicting evolutionary dynamics and epistasis (10, 11) to understanding the logic of metabolic regulation in the cell (12). In parallel, there is renewed interest in thermodynamic aspects of metabolism, with efforts to include nonequilibrium thermodynamics and energy balance constraints alongside flux balance constraints in stoichiometric models of metabolic networks (13⇓–15). The work by Flamholz et al. (2) offers an exciting perspective at the crossroads of these trends.

Faced with the dizzying output of almost 4 billion years of evolution, many biologists strive for a comprehensive understanding of the underlying principles that shaped cellular metabolism as we know it today. Flamholz et al. (2) dive in particular into the biological rationale for widespread adoption of the seemingly inefficient ED pathway. The unique insights they offer into this classical pathway (discovered in Pseudomonas saccharophila in 1954) (16) are enabled by recent advances in systems biology, such as the high-throughput sequencing and annotation of genomes and the rise of computational tools like constraint-based optimization. In future efforts, genome-scale metabolic models (7) might shed light on whether production of NADPH in the ED pathway, another key distinction with the EMP pathway, has a significant impact on the overall metabolic budget of the cell. Further experimental validation of the broader capacity to infer enzyme levels based on thermodynamics and kinetics would have strong implications in fields as diverse as metabolic engineering and evolutionary biology.

 Next Section Footnotes ↵1To whom correspondence should be addressed. E-mail: Author contributions: A.I.S. and D.S. wrote the paper. The authors declare no conflict of interest. See companion article on page 10039.



Rise and fall of competitiveness in individualistic and collectivistic societies


Competitiveness pervades life: plants compete for sunlight and water, animals for territory and food, and humans for mates and income. Herein we investigate human competitiveness with a natural experiment and a set of behavioral experiments. We compare competitiveness in traditional fishing societies where local natural forces determine whether fishermen work in isolation or in collectives. We find sharp evidence that fishermen from individualistic societies are far more competitive than fishermen from collectivistic societies, and that this difference emerges with work experience. These findings suggest that humans can evolve traits to specific needs, support the idea that socio-ecological factors play a decisive role for individual competitiveness, and provide evidence how individualistic and collectivistic societies shape economic behavior.

Rise and fall of competitiveness in individualistic and collectivistic societies

Individuals frequently face a decision that can affect their well-being and even survival: to compete or not to compete. Natural and social scientists argue that competitions and the right dose of competitiveness significantly determine not only the future of the individual but even the evolution of the whole species (1, 2). However, behavioral experiments with humans show that there are large differences in competitiveness between individuals that cannot be readily explained by genetic endowments, abilities, or risk attitudes (3⇓⇓⇓⇓⇓–9).

A possible explanation of the large variations in human competitiveness is based on learning theories. Observational learning describes individuals’ tendency to adapt by imitating successful behavior. Social or cultural learning models attribute an important role to individual experiences in the social and physical environments for the formation of traits and norms (10⇓⇓–13). Thus, individual variations in competitiveness may be the result of exposure to different environments and pressures.

In this study we investigate how local natural forces cause human competitiveness to change. We compare competitiveness in geographically proximate individualistic and collectivistic fishing societies with experiments. Our key exogenous variation is whether fishermen spend their lives at a lake or at the sea. The main difference between these societies is that the sea ecology favors fishermen to work in collectives, whereas the lake ecology guides them to fish in isolation. As a result, the output of the fishermen in the individualistic lake societies should depend on their willingness to compete with other fishermen for the best fishing spots, the best sales, and the most beneficial trade relations, whereas such individual competitiveness is unnecessary in the collectivistic sea societies. We hypothesize that these differences result in changes in individual competitiveness and that lake fishermen become more competitive than sea fishermen with exposure to these local pressures.

The experiments we used in the field facilitated comparisons and control of causal factors (14). Fishermen at the sea and at a nearby lake took part in experiments in which we measured their propensity to compete for high monetary stakes. We chose a task that was simple and unfamiliar to the subjects to capture competition preferences. The task was to throw a tennis ball 10 times into a bucket that was set 3 m away. Competitiveness was identified by a single choice: subjects decided, before performing the task, whether they wanted to compete. They were informed that if they decided not to compete they would earn one monetary unit per successful attempt. If they decided to compete they would earn three monetary units per successful attempt, but only if they outperformed one unknown other subject; if they scored less than this other subject they would not earn anything. In case of a tie they would earn one monetary unit per successful attempt. Subjects could earn more than an average 2-d’s salary in the competition experiment. They did not know against whom they were to compete, and to rule out fairness or other social considerations, their decision whether to enter into competition could not affect another subject’s payoff; i.e., nobody could be dragged into competition. More information on experimental procedures is reported in the SI Text.

We selected eight small traditional individualistic and three collectivistic fishing societies in Brazil (Fig. 1) that are in close geographical proximity to measure individual competitiveness. As mentioned above, the main difference between these societies was that fishermen located on the lake worked on their own in small boats, but at the sea fishermen worked on larger boats in teams (28.6% go fishing in teams of two, 35.7% in teams of three, and the remaining 35.7% in teams of four to eight individuals). Thus, as mentioned above, although fishermen at the lake spend much of their lives in isolation competing against other fishermen on the lake and fish markets (15), fishermen at the sea are together with their team members and do not compete against other individuals.


Fig. 1. Field setting. Our fishermen study was conducted in northeastern Brazil in different individualistic and collectivistic fishing societies in close proximity. The settings are connected by a river, only divided by a dam, and the collectivistic societies are at the estuary mouth of this river where fishermen fish in collectives. The individualistic societies are at the lake where fishermen go fishing alone. The societies are illustrated by pink dots.

As can be seen in Fig. 1, the lake is connected to the sea by a river, only divided by a dam. The air-line distance between the lake and sea is ∼50 km, which roughly corresponds to the distance between the west and the east side of the lake. Despite the geographical proximity, we found no evidence for migration between individualistic and collectivistic societies, and did not meet a single fisherman who moved from one setting to the other or went fishing in both settings. Immigration and emigration occur to some limited extent at the individualistic lake setting and we tested for their roles subsequently.

On average our subjects were 38.2 y (±13.3 SD, n = 289), lived for 28.3 y (±15.8 SD, n = 289) in the same fishing society, and had worked for 18.4 y (±12.4 SD, n = 289, variable = work experience) professionally as fishermen. In both settings, fishermen work for most of the year, and for 5 to 7 d a week. They are heavily dependent on the shrimp and fish resources: there are very few other types of jobs in these societies, and fishing is often the only possible profession to provide fishermen and their families with income and nutrition. Fishermen from both the individualistic and collectivistic societies are similarly educated (mean years in school = 3.45; Mann–Whitney U test, z = 0.813, P = 0.416, two-sided, n = 287) and generate equal incomes from fishing (monthly mean = 248.34 Brazilian Reais, Mann–Whitney U test, z = 0.359, P = 0.720, two-sided, n = 289).

Our first finding shows that individual competitiveness is more important in individualistic than in collectivistic societies. We observe that incomes from fishing and fishermen’s individual competitiveness measured by the competition experiment are positively correlated at the lake in the individualistic societies (Pearson’s, r = 0.227, P = 0.0016, n = 191). The lake fishermen who chose to compete in our experiment earn on average almost 50% more than those who chose not to compete (300.3 vs. 212.9 Brazilian Reais, Mann–Whitney U test, z = 3.246, P = 0.0012, two-sided, n = 191). There is no such comparable relationship in the collectivistic societies (252.2 vs. 235.1, Pearson’s, r = 0.047, P = 0.641, n = 98).

Our second finding confirms our hypothesis that fishermen in the individualistic societies are more competitive than those in the collectivistic societies: 45.6% of the lake fishermen chose to compete, compared with only 27.6% of the sea fishermen (Fisher’s exact test, P = 0.003, two-sided, n = 289). Fishermen who work in isolation were on average approximately 65% more willing to compete in the experiment than fishermen who work in collectives.

Our third finding is that the gap between individualistic and collectivistic societies in individual competitiveness emerges with exposure to the lake and sea ecology. Fig. 2 illustrates a linear estimation of the probability of competing for fishermen in the individualistic and collectivistic societies depending on work experience. First, we can see that both lines are initially very close to each other, but then significantly disperse. Second, and in line with our hypothesis, we observe that the solid line for the lake fishermen increases, whereas the dashed line for the sea fishermen decreases with work experience. The confidence intervals illustrate that the lake-sea gap in competitiveness becomes significant with ∼17 y of work experience. Thus, there are particularly large differences in competitiveness for experienced fishermen. For example, in the sample of fishermen who have worked for at least 20 y, we observed that lake fishermen were approximately 2.6-times more likely to compete than sea fishermen (54.4% vs. 21.3%, Fisher’s exact test, P < 0.0001, two-sided, n = 115). The interaction between society and work experience is significant at P = 0.019 in a Probit model (n = 289) and robust to the inclusion of control variables, as we show in the SI Text.


Fig. 2. Changes in competitiveness with work experience across individualistic and collectivistic societies. Lines show linear estimates for the probability of competition entry. The straight line is for individualistic societies, dashed line for collectivistic societies. Dotted lines indicate 95% confidence intervals for both. Fishermen took part in behavioral experiments in the field measuring their competitiveness and were asked about their work experience (years in profession).

It is hard to explain the different drifts in competitiveness between the societies by genetic endowments, but other factors could play a role, such as differential abilities (to throw the ball), risk differences across societies (16), or immigration into and emigration out of societies. To test for the effect of these additional factors, we used data from these societies on abilities, risk preferences, immigration, and emigration.

This additional data suggests that none of these alternative explanations is consistent with the primary data. First, task proficiencies in the competition experiment are unrelated to work experience in individualistic or collectivistic societies (Pearson’s, P > 0.188), and controlling for successful attempts does not affect the impact of society on competitiveness. Second, risk preferences identified in a lottery experiment are also unrelated to work experience in individualistic and collectivistic societies (Pearson’s, P > 0.32) and controlling for lottery investments does not affect the impact of society on competitiveness. Third, fishermen who did or did not immigrate into the lake society were not differently competitive (43.9% vs. 46%, Fisher’s exact test, P = 0.861, two-sided, n = 191) and fishermen who did or did not emigrate out of the lake society or stopped fishing were also not differently competitive (39.5% vs. 46.2%, Fisher exact test, P = 0.579, two-sided, n = 170). Thus, immigration and emigration cannot drive the changes in competitiveness in the individualistic and collectivistic societies.

Another potential driver is the differences in individualistic and collectivistic societies other than local natural forces affecting the manner in which members generate their living. To test for such other potentially unobservable differences, we conducted two additional competition experiments. First, we conducted the same competition experiments with women living in the individualistic and collectivistic societies who do not fish and are thus not differently affected by local natural forces. Second, we conducted group competition experiments with fishermen at the lake and sea to test whether there are differences in group competitiveness (17, 18). Because group—in contrast to individual—competitiveness is not crucial at the lake, we hypothesized that we should observe that group competitiveness is not more pronounced at the lake than at the sea. The task, choice, and parameters were identical to the individual group competition experiment. The only difference was that participants were told that they could either be paid depending on their own and an unknown partner’s performance if they decided to not compete or by their pair performance relative to another pair if they decided to compete.

The additional competition experiments suggest that society differences other than differential local natural forces are not responsible for the findings, as we found no differences at the lake and sea in women’s competitiveness and fishermen’s group competitiveness. Women in the individualistic societies who do not fish were as competitive as women in the collectivistic societies who do not fish (15% vs. 14.7%, Fisher’s exact test, P = 1, n = 66), suggesting that traits that evolve at work do not easily spread to other society members. Furthermore, fishermen in the individualistic societies were similarly likely to enter into group competitions as fishermen in the collectivistic societies (36% vs. 35.8%, Fisher’s exact test, P = 1, n = 103).

By combining a unique spatial feature affecting living patterns with experiments in the field, we are able to gain insights into the underpinnings of human competitiveness. Our results show that local work experience resulting from different technologies and socio-ecological factors can have an important impact on the shaping of competitiveness. We find that competitiveness changes with exposure to local forces: in the individualistic society where nature constrains humans to work in isolation, individuals become considerably more competitive, whereas the opposite holds in the collectivistic society where there is teamwork.

Our findings may also provide evidence in favor of endogenous preference formation (19, 20) and highlight that natural pressures can have a large impact on norms of competition. Finally, our study informs the literature that has investigated the relationships between individualism, collectivism, and economic outcomes (21⇓⇓⇓⇓⇓⇓–28) and the role of the social environment for human traits (29⇓⇓⇓–33).

Previous SectionNext Section Footnotes ↵1To whom correspondence should be addressed. E-mail: Author contributions: A.L., U.G., and J.A.L. designed research; A.L. performed research; A.L. and J.A.L. analyzed data; and A.L., U.G., and J.A.L. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. This article contains supporting information online at


Competitiveness pervades life: plants compete for sunlight and water, animals for territory and food, and humans for mates and income. Herein we investigate human competitiveness with a natural experiment and a set of behavioral experiments. We compare competitiveness in traditional fishing societies where local natural forces determine whether fishermen work in isolation or in collectives. We find sharp evidence that fishermen from individualistic societies are far more competitive than fishermen from collectivistic societies, and that this difference emerges with work experience. These findings suggest that humans can evolve traits to specific needs, support the idea that socio-ecological factors play a decisive role for individual competitiveness, and provide evidence how individualistic and collectivistic societies shape economic behavior.




Locomotion dynamics of hunting in wild cheetahs



Measurements of instantaneous speed, acceleration and manoeuvring during athletic competition or hunting are rare1, 2, 3, 4, even for humans, horses and dogs, the most studied species. The cheetah (Acinonyx jubatus) is acknowledged as the ultimate cursorial predator, and its published5 top speed of 29 m s−1 is considerably faster than racing speeds for greyhounds2 (18 m s−1), horses1 (19 m s−1) or humans (12 m s−1; see ‘Analysis of Bolt’s 100m’ at Quantitative measurements of cheetah locomotion mechanics have only been made on captive animals chasing a lure in a straight line, with few studies eliciting speeds faster than racing greyhounds6, 7. For wild cheetahs, estimates of speed and track have been made from direct observation or film only, and are limited to open habitat8, 9 and daylight hours.

Tracking collar design

To collect free-ranging locomotion data on wild cheetahs during hunting in their normal environment, we designed and built a tracking collar similar in size and weight to a conventional wildlife collar10, 11 (Fig. 1a; mass of 340 g), equipped with a GPS module capable of delivering processed position and velocity data, and raw pseudo-range, phase and Doppler data for individual satellite signals at 5 Hz, and an inertial measurement unit (IMU) consisting of triaxial microelectro-mechanical systems (MEMS) accelerometers, gyroscopes and magnetometers (Methods). The collar was powered by a rechargeable battery charged from solar cells, plus a non-rechargeable auxiliary battery. Data download and configuration upload was via radio. Collar software monitored the accelerometers to create activity summaries and detect the brief hunting events, buffered accelerometer data to capture the start of hunts, and adapted collar operation to battery voltages, time of day and activity. We increased the effective sample rate of the positioning system to 300 Hz, and reduced noise in the kinematic parameters, by fusing data from GPS and the IMU with a loosely coupled extended Kalman smoother (Methods). This was especially important during hunting because GPS accuracy was degraded both during initialization, and under conditions of high acceleration and high jerk12.

Figure 1: Cheetah with collar and anatomical features contributing to performance. a, Cheetah with a mark 2 collar is shown. b, Gravitational and centripetal accelerations acting on a turning cheetah; g denotes acceleration due to gravity, v2 r−1 denotes centripetal acceleration, and a is the resultant acceleration (effective gravity). c, Non-retractable cheetah claws that enhance grip. d, Low posture used in deceleration, which prevents pitching and engages hind limb musculature to absorb kinetic energy.

Collection of hunting data

We recorded GPS–IMU data from 367 runs by three female and two male adult cheetahs (100, 66, 61 and 84, 56 runs respectively) over 17 months. A further 530 runs were identified in the activity data because the collar did not trigger on every run owing to the time of day and conservative trigger thresholds. An episode of feeding after a run indicated hunting success, and was identified in the activity data by consistent, low-magnitude acceleration on all three axes13 and was confirmed on a subset of hunts with field observations (Methods). Run routes were overlaid on Google Earth to identify terrain. The total number of GPS fixes recorded depended on activity, with an average of 180 ± 171 (mean ± s.d.) per cheetah per day, and a range of 7 to 1,571.

Runs started with a period of acceleration, either from stationary or slow movement (presumably stalking) up to high speed (Fig. 2). The cheetahs then decelerated and manoeuvred before prey capture. About one-third of runs involved more than one period of sustained acceleration (all 369 runs are presented in Supplementary Video 2). In successful hunts, there was often a burst of accelerometer data after the speed returned to zero, interpreted as the cheetah subduing the prey.

Figure 2: An example day and hunt. Track of cheetah over 11 h (GPS data are available as a Google Earth file in Supplementary File 1). Each circular mark represents a GPS-derived position. Cheetah track and marks are colour-coded to collar state (detailed in Supplementary Fig. 1) as follows: alert, blue; mooch, green; ready, yellow; chase, red. b, Hunt track magnified from bottom right of a, hunt track is anticlockwise and marked with an arrow. Warmer (bright red) colours on track represent higher speed. c, Activity summary calculated in the collar from the accelerometer (Methods) for the 11-h period shown in a; shaded regions of the graph represent collar states as labelled. Line colours: peak accelerometer signal amplitude recorded in each 30-s period X, blue; Y, green; Z, red; mean of peak amplitude values extracted for each 2-s in each 30-s (that is, 15 bins) period X, cyan; Y, magenta; Z, black. The relative values for each axis differentiate between a single high-acceleration cycle and consistent movement in the 30-s window. Coordinate system: X lateral, positive left, Y fore–aft, positive forwards, Z vertical, positive upwards. Time is local (coordinated universal time (utc) + 2 h). ‘Hunt’ time is labelled. d, Doppler-derived velocity profile for hunt determined by the GPS receiver at five updates per second. e, GPS–IMU-derived velocity profile for the chase; in b, d and e warmer (bright red) colours represent faster speeds. f, Accelerometer data recorded at 300 Hz for chase; X, blue; Y, green; Z, black. Red circles indicate forward acceleration peak used as event marker for stride cutting at, approximately, hindlimb foot contact. The high accelerations at zero velocity at t = 12–13 s suggest subduing prey and a successful hunt. An animation of a hunt is in Supplementary Video 1, plots of further runs are available in Supplementary Fig. 5, and all runs are in Supplementary Video 2.

As well as hunting runs, cheetahs play and run from larger predators, but we had insufficient data validated by direct observations to provide secure separation of these activities, although only a few runs did not involve the tight turns and rapid speed changes characteristic of hunting (for example, runs 5, 32 and 49 in Supplementary Video 2). We therefore compared successful hunts to all other runs recorded by the collar. In total, 94 of the 367 runs (26%) were successful hunts. Including the 530 additional runs detected solely from IMU data did not change the success rate (223 out of 897; 25% success), which is lower than previously reported for individual cheetah9, 14, 15, perhaps due, in part, to the inclusion of non-hunting runs. Cheetah are reported to move in predominantly open habitats using vegetation-edge to stalk their prey, often at dawn and dusk8, 9, 14, 15, 16. Although almost half of the runs here occurred at/after dawn, runs occurred throughout the day and night (Fig. 3e). The individual cheetahs varied in their predilection for running in open grassland or dense shrub (Supplementary Fig. 6). On average, the cheetahs ran most often in open habitat (48%, 176 of 367 runs); 28% of runs occurred in open shrub/around large trees, and 24% occurred within dense vegetation. Only 20% of runs occurring in the open grasslands were identified as successful hunts, compared with 31% of runs in dense cover. This difference in outcome was not significant (P = 0.054, chi-squared test) and is confounded by individual variation and habitat, but it does demonstrate that cheetahs do hunt successfully in all terrains8, 15. Vegetation may confer an advantage by permitting stalking and limiting prey options for escape by manoeuvring; however, there was little difference in the distance or speed between terrains (Supplementary Table 1).

Figure 3: Descriptive hunt statistics. Top speed, averaged over a stride, reached in each run colour-coded for outcome. b, Distance covered in each run. c, Top speed in each run coded for terrain type. d, Peak acceleration and deceleration recorded in each run. e, Plot of time of day of runs recorded in period when collar was set to trigger at any time of day, time local. f, Example hunt file colour-coded for speed (bright red denotes fastest), and with horizontal acceleration vectors drawn, to scale, for each stride. n = 367 (a–d) and n = 254 (e).

Description of hunts

The average run distance was 173 m (±116 m) (Fig. 3b) though recorded run distance will be shorter than the true value in the runs where the start of the run was missed (Methods, Supplementary Video 2). The longest runs recorded by each cheetah ranged from 407 to 559 m; the mean run frequency (including information from activity data) was 1.3 times per day, so, even if some hunts were missed, high speed locomotion only accounted for a small fraction of the 6,040-m average daily total distance covered by the cheetahs. The mean top speed was 14.9 ± 3.4 m s−1 and was usually only sustained for 1–2 s. The highest speed we recorded was a stride-averaged 25.9 m s−1 in run 250 (Fig. 3a, c and Supplementary Video 2). The top speeds attained by the other cheetahs were 25.4, 22.0, 21.1 and 20.1 m s−1. The cheetahs studied here mostly hunted impala (Aepyceros melampus)17, which made up 75% of their diet, although one male cheetah (Qamar), which frequently hunted in thicker vegetation (Supplementary Fig. 6), never exceeded 20.1 m s−1 and was often observed on warthog (Phacochoerus africanus) kills. Cheetah hunting the (anecdotally) faster Thompson’s gazelle (Eudorcas thomsonii) on open East African savannah may use higher speeds.

Successful hunts involved greater deceleration on average (−7.5 m s−2 versus −5.5 m s−2; P < 0.05; Fig. 3d), but there was no significant difference in peak acceleration (Fig. 3d), distance travelled (Fig. 3b) number of turns (6.7 versus 6.5) or total turn angle (347° versus 260°) (generalized linear mixed model (GLMM); Methods). This indicates that outcome was determined in the final stages of a hunt rather than hunts being abandoned early to save energy or reduce risk of injury, and the higher deceleration values may reflect actual prey capture. Equivalent locomotion and outcome data for coalition-hunting cheetah might clarify the importance of the final manoeuvring phase in hunt outcome.

Comparison with other athletic animals

The greatest acceleration and deceleration values were almost double values published for polo horses18 and exceeded the accelerations reported for greyhounds at the start of a race18. The cheetahs sped up by up to 3 m s−1 and slowed by up to 4 m s−1 in a single stride (Supplementary Fig. 5d). Mass-specific change in kinetic energy over a stride (Fig. 4c and Supplementary Fig. 7) exceeded 30 J kg−1 stride−1 across the broad speed range of 10 to 18 m s−1. On the basis of forward acceleration, the greatest stride-averaged whole animal powers often exceeded 100 W kg−1 (body mass) (Fig. 4d), and also occurred between 10 and 18 m s−1. For comparison, we calculated a stride-averaged power of 25 W kg−1 for Usain Bolt’s 9.58-s 100-m world record (Methods and, consistent with other measurements on human sprinters19; polo horses achieve 30 W kg−1 (ref. 18) and racing greyhounds 60 W kg−1 (ref. 18).

Figure 4: Performance summary. a, Stride frequency plotted against speed; each point is colour-coded for tangential (forwards) acceleration, bright red points represent the greatest forward acceleration, and are plotted last (on top). Lines are linear regression of stride frequency against speed for each individual cheetah. b, Tangential (forwards, positive) acceleration and deceleration (y axis) against speed (x axis). Horizontal lines represent acceleration and deceleration of 13 m s−2, equating to the proposed grip limit of 1.3 (see text). Curved lines represent stride-averaged whole-body powers of ±30, 60, 90 and 120 W kg−1; points outside the outer dashed line equate to a mean stride power in excess of ±120 W kg−1. c, Body, mass specific, horizontal kinetic energy change performed in each stride (work per stride). d, Stride-averaged whole-body acceleration power plotted against speed, with horizontal lines showing powers of ±30, 60, 90 and 120 W kg−1. e, Horizontal speed against turn radius, region around origin magnified in inset. Slanting straight lines show different rates of heading change in degrees per second, with values (2, 6, 10, 16, 25, 43 and 112) at the top of the line. The solid curved line (μ = 1.3) represents a grip limit/coefficient of friction of 1.3; the curved shorter-dashed line (μ = 0.6) denotes the 0.6 grip limit reported for polo horses3; points above each line require a higher grip level. The curved longer-dashed line (LFL) represents a limit to turning defined by the maximum force the legs can withstand. f, Plot of tangential acceleration against lateral acceleration. Total horizontal acceleration is the distance from the origin, circles represent mean total horizontal acceleration of 6 and 13 m s−2 (equating to average grip limits of 0.6 and 1.3). Each point on each plot represents data centred on a single stride, with data smoothed over three strides. Points are colour-coded by individual, except in plot a. The number of strides from each cheetah were 5,031, 4,022, 3,211, 2,657 and 1,895 giving a total n of 16,816 for plots b, c, d and f. The total n is given in each plot and was slightly different for plots a and e owing to the mathematics of generating those plots but the individual contributions were in proportion.

The locomotor (limb and back) muscle accounts for 45 ± 4% of body mass20, 21 in captive cheetah. The wild cheetahs had similar limb and back lengths to those captive cheetahs, but were heavier at 53 kg versus 33 kg (means, n = 5, 5), and visibly more muscled (mean mid-thigh girth 540 mm versus 450 mm, n = 5, 5), so much of their body mass is locomotor muscle. Major propulsive muscles such as the hamstrings (biceps femoris, semimembranosus and semitendinosus) at the hip and gastrocnemius at the tarsus have 64% and 60% longer moment arms, respectively, than in the greyhound and similar muscle fibre lengths21. Stride frequency and posture are similar at the same speed in the two species7 so the muscle sarcomeres (and fibres) will be shortening considerably faster in the cheetah than in the greyhound at the same speed (like the engine of a car in a lower gear). This fast muscle contraction velocity will enable large muscle powers and hence deliver the very large acceleration powers observed22. The high muscle speed and power are consistent with our measurements on contracting skinned fibres from cheetahs23. The cheetah deceleration magnitudes (Figs 3d and 4b), cycle works (Fig. 4c) and powers (Fig. 4d) were greater than during acceleration and up to three times higher than polo horses18; however, comparative figures are sparse. Cheetah can crouch to engage locomotor muscle to enable these deceleration magnitudes (Fig. 1d), and sliding or colliding with the prey may dissipate some energy.

Grip and manoeuvrability key to hunting success

Hunts involved considerable manoeuvring, with maximum lateral (centripetal) accelerations often exceeding 13 m s−2 at speeds less than 17 m s−1 (Fig. 4e, f; polo horses achieve 6 m s−2; ref. 3). A lateral acceleration of 13 m s−2 (Fig. 1b) requires a coefficient of friction with the ground of at least 1.3. Ridged footpads and substantial claws24 (Fig. 1c) act as cleats to augment friction and deliver this level of grip. The maximum centripetal acceleration observed was smaller at speeds greater than 17 m s−1 (Fig. 4e), which may be behavioural in origin; that is, cheetahs do not perform tight turns at their highest speeds. Studies on other animals show that, although grip limits turning performance at low and moderate speed, a model based on the capacity of the limbs to withstand the combination of centripetal acceleration and gravity (Fig. 1b) is appropriate to account for reduced speed on bends in humans, mice and racehorses3, 25, 26, 27 but not greyhounds2. The dashed line labelled LFL (leg force limit) in Fig. 4e is calculated using published models3, 25, 27, published stride data7 and the maximum speed recorded here. The equations and assumptions are presented in the Supplementary Information. The LFL line seems to follow the upper bound of the data points at higher speeds, but confident verification would, however, require stance times or limb forces during manoeuvring26. When combined with gravity, a lateral acceleration of 13 m s−2 equates to a 66% increase in the cheetah’s effective weight and hence average limb force (Fig. 1b). Cheetahs have relatively large limb bone cross-sectional areas (compared with greyhounds20, 21), which may be an adaptation to resist the large peak limb forces that occur during high speed manoeuvring.

The cheetah should run little faster than its prey in the manoeuvring phase of the hunt28, 29 if it is to capture an agile and quick-turning prey. A cheetah running at 25.9 m s−1 with the maximal observed lateral acceleration of 13 m s−2 would have a turn radius of 52 m and would take 6 s to perform a 180° turn (πr v−1)—peak running speed is therefore unlikely to be, and was not found to be, a feature of the final stage of successful hunts. A cheetah can slow by 4 m s−1 in a stride (Supplementary Fig. 5d), and the cheetahs often decelerated sharply before turning, which would enable much tighter turns. Slowing from 16 m s−1 to 4 m s−1 (three strides, 1 s) would drop the turn radius with v2 r−1 = 13 (lateral acceleration of 13 m s−2) from 19.7 m to 1.2 m, and heading velocity (v r−1) would rise from 46 to 190° s−1. This demonstrates the value of slowing down before manoeuvering. The cheetahs did not use highest tangential and centripetal accelerations simultaneously, consistent with grip limiting maximal horizontal acceleration (there are few data points in the corners of the square in Fig. 4f). Rapid deceleration would unload the hindquarters, which could result in yaw instability when manoeuvring because the centre of mass (COM) is behind the forelimbs (like a ground loop in a tail wheel aircraft). The pitch limit proposed in ref. 18 may apply at low speed, but insufficient low-speed data exist to consider this further, and it can be circumvented by posture due to the cheetah’s flexible spine (Fig. 1d). The active movements of the high-inertia tail that are observed in wildlife documentaries will help in positioning and banking the body (and limbs) to apply appropriate forces to prevent this and for turn initiation and manoeuvring.


Equivalent data for other wild cursorial species would enhance what we know about natural speed, agility, endurance and locomotor physiology, and provide detailed information on ranging behaviour in the wild. For example, such fine-scale data on habitat selection by endangered species detailing where animals are commuting, hunting and resting will be informative when attempting to evaluate landscape scale connectivity, corridors and wildlife-protected areas. Tightly coupled GPS–IMU processing can deliver 0.2-m position accuracy (the level of individual shrubs and footfalls) during hunts, enabling detailed analysis of context variables (such as habitat characteristics and prey visibility), modes of hunting success and failure, and the effect of slope, camber and foot-surface interaction on stride-by-stride performance. These data on hunt environment would inform about the determinants of preferred hunting habitats, risk of injury (of paramount importance for solitary predators), risk of detection by kleptoparasites (open versus closed habitat), available palatable grazing and habitat-dependent risk of predation (detection).



The cheetahs used in this study were part of a continuing study by Botswana Predator Conservation Trust ( in the Okavango Delta region of Northern Botswana. Initially, three ‘mark 1’ prototype collars were fitted to three cheetahs in July 2011. All collars successfully collected data as intended, two collars for 7–9 months whereas the third suffered a memory card failure after 6 months. Three collars of a new ‘mark 2’ design were used in April 2012, and two more collars in July 2012 (fitted to the original three cheetahs plus two new individuals). Data were again successfully collected from these collars, and they continue in operation.

The cheetahs were immobilized by free darting from a vehicle by A.M.W. using medetomidine (2 mg) and ketamine (80–120 mg) and reversed after 60 min with 10 mg atipamezole. While sedated, dimensions including limb lengths, thigh girths and back lengths and body mass were recorded. Collar data were downloaded by radio link every few weeks to a ground vehicle or a light aircraft.

Collar design and fabrication

The major design challenges included the measurement and logging of data at a sufficiently high rate and accuracy, timely remote retrieval of substantial volumes of data from the collar and maintaining the very low average power consumption required in a wildlife collar. To conserve power, careful management of the internal readiness of the GPS subsystem allowed this and other sensor systems to be started quickly enough to capture data at maximum rate only during these events.

All collars were constructed in-house. In the original collars (mark 1, used in 2011), a commercial radio-tracking collar (Sirtrack, New Zealand) was used as a base, our custom electronics package being mounted on the top of the collar in a clear cast resin case and wired to the collar’s original battery box at the bottom of the collar. The revised mark 2 collars (Fig. 1a) were entirely constructed in-house, with a revised lower-profile electronics enclosure (cast from polyurethane resin using a silicon mould and a rapid prototyped former; Aprocas GmbH) and a vacuum-formed polycarbonate battery box holding larger rechargeable and back-up battery in potting compound. The actual electronics package was similar on both versions, with an identical chip set as described below, and with almost identical software functionality. Collar mass was approximately 340 g.

Collar design: electronics payload

The collar circuit was based around a low-power MSP430 16-bit microcontroller (Texas Instruments), running custom software written in the ‘C’ programming language developed using an integrated development system from IAR Systems. The microcontroller contains several internal peripheral blocks, including an 8-channel 12-bit analogue-to-digital converter (ADC), four serial communications modules, plus various timers, general-purpose digital input and output lines, and other support modules. A connected 2-GB micro-SD flash memory card (Sandisk) provided data storage.

GPS position was obtained from an LEA-6T GPS module (u-Blox AG). In addition to internally computed position and velocity, the module is able to generate raw pseudo-range, phase and Doppler data for the signal from each satellite enabling detailed GPS performance evaluation, and use of customized differential techniques for increased accuracy. The data rate was five position, velocity and raw data points per second during continuous operation (for example, during a chase).

The collar circuit also included an inertial measurement suite, based on MEMS devices. Acceleration was measured using an MMA7331 three-axis accelerometer module (Freescale Semiconductors), providing acceleration with a ±12 g range. The roll and pitch rotation rate was measured by a dual-axis gyroscope (ST Microelectronics), and yaw rotation rate by a single-axis gyroscope (ST Microelectronics), both set to the 2,000° s−1 range. Sensor outputs were filtered by simple single-pole analogue filters (100 Hz knee), and then sampled by the microcontroller ADC at 300 or 100 samples per second (Accelerometers or Gyroscopes, respectively). Three-hundred hertz was chosen as giving an overhead to a frequency of 30 Hz; that is, 1/minimum published stance time7. A three-axis magnetometer (Honeywell), connected via I2C, provided magnetic compass functionality at 12 measurements per second.

Primary communication with the collar, for tasks such as data file download and configuration file upload, was via a 2.4-GHz chirp-spread-spectrum communication module (Nanotron Technologies Gmbh), communicating at 1 Mbit per second using a custom communications protocol. A 173-MHz VHF radio transmitter (Radiometrix) provided longer-range transmission of current GPS-derived position, for tracking purposes. An original equipment manufacturer (OEM) conventional wildlife tracking transmitter in the 149-MHz band (Sirtrack) facilitated long-range animal location using conventional direction-finding techniques.

Collar design: power

Primary power supply for the collar was a 900 mAh lithium-polymer rechargeable battery (Active Robots), charged by a solar cell array consisting of 10 monocrystalline silicon solar cells (Ixys Koria). On the mark 2 collars, a 13 Ah lithium thionyl chloride primary battery (Saft) provided a back-up power source (on the original collars, a 7.7 Ah lithium thionyl chloride primary battery was used). Both battery voltages, together with the charge current from the solar cell array, were measured by the microcontroller, which switched the collar electrical load from one battery to the other depending on battery state.

Collar design: software states and movement detection

In operation, the collar software moved between several different operating ‘states’, the particular state at any moment being dependent on a combination of animal activity level (measured using the accelerometers) and time of day (from a GPS-synchronised software clock). Each state required a different mix of hardware sub-systems to be powered on or off, and different intervals between GPS module operation, and thus the power consumption of the collar varied depending on the operating state. Thus, the inevitable compromise between average power consumption on the one hand, and quantity and resolution of data gathered on the other, could be optimized by setting the parameters for the state transitions. The different operating states and associated average power consumption for the collar are summarized in Supplementary Fig. 1.

To keep the average power consumption as low as possible, the collar would generally default to operating in state 1 (‘alert’ state). In this state, to detect when the cheetah was moving, the accelerometer was sampled at 30 Hz for a period of 10 s in every minute. Within each 10-s sampling period, the peak-to-peak acceleration was computed for each axis every 2 s, and an accumulator incremented by a specified value for each 2-s window in which the peak-to-peak acceleration exceeded a pre-set threshold; For each 2-s window in which the peak-to-peak acceleration did not exceed the threshold the accumulator was decremented by a (different) specified value. Thus, periods of movement could be given higher ‘weight’ than periods of no movement or vice versa to identify stalking. If the accumulator total exceeded a specified value, the cheetah was deemed to be consistently moving and the collar switched to a higher operating state, the exact state depending on time of day. A similar algorithm with different weights and thresholds was then used to determine when the animal had settled back to rest, at which time a switch back to the lower state was executed.

When consistently moving between local times of 06:00 and 09:00, and 17:00 and 19:00 (times when hunting was most likely from previous work), the operating state would transition to state 3 (‘ready’ state). The GPS was refreshed every 30 s and position recorded every 60 s. Accelerometer data were recorded into a circular buffer at 100 Hz, the buffer storing the latest 3 s of data. If the fore–aft accelerometer data then exceeded a threshold equivalent to galloping, state 4 (‘chase’ state) would be entered. The buffered data were stored and 5 Hz GPS data, 300 Hz accelerometer, 100 Hz gyroscope and 12 Hz magnetometer data recorded. A record was defined as valid if five further peaks (strides) were detected, and then recording would continue until there were no peaks above the threshold for 5 s. When moving consistently but outside of the peak hunting times, the lower-powered state 2 (‘mooch’ state) would be invoked, with GPS positions being taken every 5 min and simple activity measurements being taken as described below. The GPS delivered a first fix in 1.30 s after triggering (median), accurate position data (<10 m s.d.) after 1.58 s, and full rate data (5 Hz) after 5.4 s (Supplementary Fig. 3). The unexpectedly long delay in the GPS module delivering 5 Hz data prevented open-loop GPS–IMU integration back to the beginning of the run in some cases. This is why many runs in Supplementary Video 2

Collar power handling and power consumption

Average collar power consumption varied between individual animals (owing to differing patterns of activity and hence a different distribution of collar operating states), but was typically around 4 mA when averaged over 24 h. The main contributor to this average was the time spent in the ready state when the animal was active during hunting times of day (Supplementary Fig. 1), in which average consumption was around 16 mA with a 30-s GPS refresh time. By comparison, the time spent in the mooch state (animal active but outside hunting times) had a lower consumption of about 5 mA, whereas ‘sleep’ or alert states (animal inactive) contributed only about 0.6 mA. The ‘chase’ state, used only when the animal is running, required some 90 mA, but time spent in this state was very small. Solar charge currents ranged from 35 mA with the animal in full sunlight, to typically 10 mA in dappled shade and almost zero in deeper shade. Average charge current over a 24-h period was typically 2 mA, with some variation between animals due to terrain preferences, indicating little time spent in full sunlight even in the winter study period. The solar cells, via the rechargeable battery, contributed roughly 75% of the collar power, the remainder being supplied by the non-rechargeable battery. Collar battery life was predicted at approximately one year with these settings, but was very dependent on collar settings and animal behaviour.

On cheetahs four and five, the ready state GPS refresh interval was changed from 30 s to 300 s—this resulted in a typical power saving of around 30% over a 24-h period, with unexpectedly little effect on GPS start-up time (Supplementary Fig. 3f). We reduced power consumption on mark 2 collars (254 runs) by not pre-buffering data, and moving directly from mooch to chase state (and allowing this to happen at any time of day, enabling Fig. 3e to be generated), so that IMU data logging began on the first accelerating stride when the cheetah was already in motion. The time that could be recovered through backwards integration was therefore reduced, and the first 1–2 acceleration strides lost.

Collar design: generation of activity summaries

Throughout all states, a background measurement of animal activity was also recorded. For every 2-s ‘window’, the maximum peak-to-peak acceleration range is recorded separately for all three accelerometer axes. After 15 ‘windows’ have passed, an activity record is generated, containing GPS time, the largest X, Y and Z peak-to-peak acceleration amplitudes seen in any of the 15 windows, and the average of the 15 2-s peak-to-peak X, Y and Z accelerations amplitudes. This enabled differentiation of transient high acceleration events and consistent activity. This record is generated continuously in the mooch and ready state, every 3 min in the alert state, and every 30 min in the sleep state. Amplitudes are higher than body acceleration, because the collar can move relative to the centre of mass.

All settings that affected the state transitions (times, acceleration thresholds, and so on), and many other settings besides, could be modified by uploading a new configuration file over the 2.4-GHz communications link. In addition, a complete new version of the collar firmware could be uploaded over this link, allowing for in-field program updates while the collar is on the animal.

Sensor fusion and signal processing to capture hunting dynamics

In the collar data collected here, the power management features used gave different sampling rates for accelerometer (300 Hz) and gyro (100 Hz) in the chase state. To capture the full acceleration profile within the microcontroller, 3 s of accelerometer measurements were continually buffered in ready state at a reduced sampling frequency (100 Hz) and recorded when entering the chase state (gyro-power consumption was too high to permit continuous pre-buffering). GPS position and velocity measurements were usually (but not always) available within 1 s after entering the chase state (Supplementary Fig. 3).

The unique characteristics of these data required a custom-designed GPS–INS (inertial navigation system) integration method written in Visual C++ and MATLAB. Calibrated IMU measurements were first linearly interpolated to 300 Hz. Orientation changes were assumed to be minimal during the buffer period, and hence the unmeasured gyro angular rates assumed to be zero. GPS and IMU measurements were fused using a 12-state extended Kalman filter30 in loosely coupled architecture. The total state formulation used propagates position, velocity and orientation states with time using the IMU measurements in a simplified form of the strap-down inertial navigation equations31. The associated process noise was estimated from the known error characteristics of the inertial sensors used. GPS position and velocity updates were used as measurement updates, and receiver accuracy data for each fix used to estimate measurement noise to appropriately weight the GPS to the inertial solution.

The filter was run in reverse time from the last GPS observation of each run to the beginning of the buffered inertial data. During the short time period in which only inertial data was present (throughout buffer and between GPS measurements), the filter propagation was equivalent to open-loop inertial navigation. The filter was initialised using last GPS position and velocity data, and Euler angles assumed zero with covariances appropriate for the uncertainty in that assumption. A Rauch-Tung-Striebel (RTS) smoother32 was then applied in forward time on the Kalman-filtered data. This is equivalent to combining backward and forward solutions, effectively halving the open-loop INS integration period between GPS observations. It was not always possible to reconstruct the period before the first GPS observation, as this period was often too long or the accuracy of the initial GPS observations insufficient (Supplementary Fig. 3c–f). This will result in a somewhat short measurement of hunt distance in those cases (apparent qualitatively in Supplementary Video 2).

GPS–INS processing was used to reduce noise and improve precision in the position and velocity solution (Supplementary Fig. 3), as well as increasing the temporal resolution of the data. It also allowed determination of orientation, which is otherwise not directly measured. Because the GPS receiver also records raw pseudorange, Doppler and carrier phase measurements for each satellite, future data processing may use a stationary reference station to calculate a more accurate differential GPS solution. Use of a tightly coupled GPS–INS solution may also provide increased accuracy and robustness, especially during periods when a reduced number of satellites are tracked (for example, turns).

Extraction of parameters for analysis: speed, distance and stride timing

Stride timings for data cutting and stride frequency were determined from the axis of accelerometer aligned approximately in the cranio-caudal direction. These accelerations were first low-pass filtered at twice anticipated stride frequency (8 Hz), and a peak detection algorithm was used to detect forward acceleration peaks at minimum duration of 0.2-s apart (equal to a maximum stride frequency of 5 Hz).

Horizontal speed was calculated from filtered velocity and averaged over the calculated strides ( ) to remove the effects of speed fluctuation through the stride and collar oscillation relative to the centre of mass. These data were then smoothed with a rolling average (see below). Run distance was calculated by zero-order hold integration of the stride averaged horizontal speeds over the duration of the run. Maximum speed during each run was determined from these values. Stride frequency was calculated from the duration between stride timing peaks. For consistency in comparison, other parameters were then determined using the same method as in ref. 3, using only two-dimensional position and speed measurements. Position data were first down-sampled to the calculated stride times. The displacement vectors between consecutive positions were then calculated:

and in which is the two-dimensional position at sample/stride i.

Extraction of parameters for analysis: acceleration and power

A signed change of heading ( ), and hence heading angular velocity ( ), were then calculated from the angle between the two vectors:


in which is the sampling interval.

The tangential or forward acceleration ( ) and centripetal acceleration ( ), as well as instantaneous turn radius ( ) were then calculated:

Finally mass-specific COM power was calculated as the dot product of stride averaged acceleration and stride averaged velocity (that is, multiply forward acceleration by forward speed):

Mass-specific COM stride work (net COM kinetic energy change in a stride) was calculated as change in speed over a stride multiplied by stride average speed.

Extraction of parameters for analysis: improving accuracy through averaging

One important consideration when calculating heading, change of heading, and heading angular velocity from position measurements is that accuracy will decrease as speed decreases. Although averaging over a stride and across strides markedly improves the accuracy, lower average speed values will still be less accurate. The noise present is of a level that does not unduly influence extreme values even at very low speeds.

Although validations carried out on the stride timing show that it is generally accurate (Supplementary Fig. 2f), detection of an incorrect or spurious peak for end of stride would result in one stride duration being under or overestimated, and the adjacent stride duration being affected in the opposite manner. This would introduce error in parameters that do not change smoothly through a stride, such as acceleration and kinetic energy. We therefore applied a weighted average in which the stride period was averaged, with the mean of the duration of the preceeding and following stride. The weighted average was of the form:

in which S represents the parameter being weighted, and i is the stride number.

This approach was used as follows: tangential acceleration and hence acceleration power were calculated based on a weighted average stride speed. Centripetal acceleration was based on weighted stride speed and weighted heading rate. Stride duration was also weighted. Where these parameters have been plotted against horizontal speed, the weighted stride speed was also used. Applying more averaging than this did not change the distribution of outliers to a discernible extent (Supplementary Fig. 4), but applying no averaging did result in more outliers giving us confidence in our extreme values with this treatment.

Extraction of parameters for analysis: grip and manoeuvring

Maximum traction has been proposed as a potential constraint to turning performance3. Coefficient of friction, μ, is the maximum achievable ratio of horizontal force (acceleration) with respect to vertical force (acceleration). Average vertical force is equal to acceleration due to gravity and assuming that vertical and horizontal forces are always in proportion:

So that maximum horizontal force and horizontal acceleration (a) are:

in which g is acceleration due to gravity, and m is mass. Substituting for horizontal acceleration in terms of tangential ( ) and centripetal components ( ):

This demonstrates the potential trade-off between tangential and centripetal accelerations. Given that maximum centripetal acceleration will occur at constant speed ( ), and likewise that maximum tangential acceleration will occur in a straight line ( ):

Remembering that centripetal acceleration:

in which v is horizontal speed, and r is radius of turn. We form an equation for maximum speed ( ) in terms of turn radius (r):

A maximum limit for tangential acceleration based on maximum available muscle power (K) is derived as follows. When force and velocity are in the same direction:

Where F is force magnitude, v is horizontal speed, is tangential acceleration and m is body mass. Given specific power by body mass (k):

Substituting gives:

Geometric limit to acceleration

A pitch limit for acceleration was previously proposed20 that assumes that propulsion is derived purely from hip extension. This gives an acceleration limit for greyhounds of 10 m s−2 at all speeds derived from back length and leg length, and the limit for cheetahs would be similar as body height and length are similar7. Such a limit is not exceeded in our data (Fig. 4b), but there are few low speed acceleration strides.

Collar validation

A lurcher (greyhound/whippet/terrier cross in this case) dog was fitted with a mark 2 collar and encouraged to undertake maximal accelerations and sharp running turns on a beach in England, UK (the dog was accustomed to collar-testing experiments). The position of each footfall was determined using Survey grade GPS (OEM4, Novatel). Dual frequency Doppler and pseudorange and phase GPS data were post-processed relative to a local base station data using Waypoint GrafNav 8.10 (Novatel) with a horizontal accuracy of 20 mm. The timing of each footfall was determined from simultaneous high-speed video at 500 frames per second (f.p.s.) (X-Pri 1280 × 1024 AOS Gmbh). The camera trigger event was captured via an interrupt channel on an RVC GPS logger module with sub-millisecond accuracy, and used to express footfall events in GPS time for comparison to collar data (Supplementary Fig. 2e). The four footfalls per stride were easily identified in the position data (Supplementary Fig. 2a, b), and the distance between subsequent non-lead forefootfalls was defined as stride length, and the time between those foot falls as stride duration. Stride duration by video and by processing of collar data was compared by subtracting stride time from foot falls on high-speed video from stride duration from collar data and plotting the difference as a histogram (Supplementary Fig. 2f). Speed was calculated by dividing stride length by stride duration, and data were smoothed with a three-stride centre weighted rolling average as described for the collar data and the results plotted (Supplementary Fig. 2d). These data show that qualitatively the collar reproduces the track of the footfalls and that the speed time (and hence acceleration) data are indistinguishable between the two approaches. Further trials and analysis are required for a full assessment of the two methods.


To establish which aspects of a run correlate with success, GLMMs were performed in R statistical software (R, version 2.14.1, 2011. R Development Core Team 2011, Foundation for Statistical Computing, Vienna, Austria). In the model, all the descriptive parameters of each hunt (terrain, distance, top speed, peak acceleration and deceleration number of turns and total turn angle) were included as fixed effects. To control for individual variation, a subject was included as a random effect. If an effect was not significant, and removing it from the model improved the Akaike information criterion (AIC), then it was removed. A chi-squared test was used to evaluate the effect of terrain on outcome.

Human acceleration power

Ten-metre split times for the 9.58 s world 100-m record run by Usain Bolt in 2009 were retrieved from the IAAF website ( A fifth order polynomial was fitted through the distance–time data. This polynomial visually fitted the data points and was differentiated to give formulae for speed and acceleration through the race and a function for instantaneous power through the race calculated as the product of the functions for speed and acceleration. This gave a peak centre of mass power of 25 W kg−1 body mass at 7 m s−1, which is similar to previously published values for human sprinters19.

Hunting, terrain, and outcome (success)

Runs were identified in activity summaries by very high-peak acceleration amplitudes in all three axes, but particularly high accelerations in the cranio-caudal direction were the best indicator, confirmed from GPS speed where present. If two run events were within 10 min of one another, they were considered to be the same event for outcome measures. Terrain was determined from Google Earth; georeferencing of known landmarks and road junctions was confirmed to be accurate to within 5 m in the study area.

We identified feeding as a consistent signal on all three accelerometer axes (mean amplitude similar to mean of mean amplitudes), with particularly low cranio-caudal accelerations (compared with walking) and no change in location. See ref. 13 for more discussion. We classified a run as a successful hunt if 6 min of this feeding behaviour occurred in the 30 min after a run was identified. These methods correctly identified nine out of the ten known successful hunts using only the activity data (that is, without using GPS data), and correctly identified all nine as successful hunts. When applied to the main data set, the classification outcome correlated to other markers of success in 97% of known hunts. The other markers were: prey struggling captured in the accelerometer signal; cheetah remaining at hunt location for over two hours after the run; observing the cheetah on a kill.

List of symbols

i, stride number; , two-dimensional position; Δθi, signed change of heading; ωi, heading angular velocity; ΔT, sampling interval; ai, horizontal acceleration; at,i, tangential or forward acceleration; ac,i, centripetal acceleration; ri, instantaneous turn radius; vi, stride averaged horizontal speed; K, whole-body power; ki, mass-specific whole-body power; Si, parameter to be weighted; Si,w, parameter after weighting; μ, coefficient of friction; m, body mass; g, acceleration due to gravity.


Download references


We thank S. Amos and M. Dickson for fabricating collars, and F. Broekhuis, R. Furrer and N. Jordan for working with us in the study area; P. Apps for many discussions; T. Hubel and A. Wills for helping to collect and analyse validation data; P. Apps, J. Usherwood and A. Wilson for comments on the manuscript; and the EPSRC (EP/H013016/1), BBSRC (BB/J018007/1) and DARPA M3 Program (W91CRB-11-C-0048, with Boston Dynamics) for funding. This work was approved by RVC Ethics & Welfare Committee and was carried out under a Botswana Government Research Permit held by J.W.M. and Botswana Veterinary Registration held by A.M.W.

Author information

Supplementary information

In Vivo–Directed Evolution of a New Adeno-Associated Virus for Therapeutic Outer Retinal Gene Delivery from the Vitreous


Inherited retinal degenerative diseases are a clinically promising focus of adeno-associated virus (AAV)–mediated gene therapy. These diseases arise from pathogenic mutations in mRNA transcripts expressed in the eye’s photoreceptor cells or retinal pigment epithelium (RPE), leading to cell death and structural deterioration. Because current gene delivery methods require an injurious subretinal injection to reach the photoreceptors or RPE and transduce just a fraction of the retina, they are suitable only for the treatment of rare degenerative diseases in which retinal structures remain intact. To address the need for broadly applicable gene delivery approaches, we implemented in vivo–directed evolution to engineer AAV variants that deliver the gene cargo to the outer retina after injection into the eye’s easily accessible vitreous humor. This approach has general implications for situations in which dense tissue penetration poses a barrier for gene delivery. A resulting AAV variant mediated widespread delivery to the outer retina and rescued the disease phenotypes of X-linked retinoschisis and Leber’s congenital amaurosis in corresponding mouse models. Furthermore, it enabled transduction of primate photoreceptors from the vitreous, expanding its therapeutic promise.

In Vivo–Directed Evolution of a New Adeno-Associated Virus for Therapeutic Outer Retinal Gene Delivery from the Vitreous

Supplementary MaterialsBack to Top Fig. S1. Glycan dependencies of 7m8 compared to AAV2. Fig. S2. GFP expression in photoreceptors after transduction with 7m8. Fig. S3. GFP expression in the optic tract and brain histology after transduction with 7m8. Fig. S4. Comparison of photoreceptor transduction by 7m8-rho-GFP and AAV2-4YF-rho-GFP. Fig. S5. Immunolabeling for activated microglia and infiltrating macrophages after 7m8 readministration. Fig. S6. Structural damage caused by subretinal injections in Rs1h−/− mice. Fig. S7. Expression of control vectors and structural improvements at the photoreceptor–bipolar cell junction in Rs1h−/− mice. Fig. S8. Expression of 7m8-CMV-GFP in nonhuman primate photoreceptor flat mounts. Fig. S9. Expression of 7m8-cx36-GFP in nonhuman primate. 1LP3-mutant-swiss.pdb: PDB format data file for the modeling of the 7m8 capsid.

AbstractBack to Top

Inherited retinal degenerative diseases are a clinically promising focus of adeno-associated virus (AAV)–mediated gene therapy. These diseases arise from pathogenic mutations in mRNA transcripts expressed in the eye’s photoreceptor cells or retinal pigment epithelium (RPE), leading to cell death and structural deterioration. Because current gene delivery methods require an injurious subretinal injection to reach the photoreceptors or RPE and transduce just a fraction of the retina, they are suitable only for the treatment of rare degenerative diseases in which retinal structures remain intact. To address the need for broadly applicable gene delivery approaches, we implemented in vivo–directed evolution to engineer AAV variants that deliver the gene cargo to the outer retina after injection into the eye’s easily accessible vitreous humor. This approach has general implications for situations in which dense tissue penetration poses a barrier for gene delivery. A resulting AAV variant mediated widespread delivery to the outer retina and rescued the disease phenotypes of X-linked retinoschisis and Leber’s congenital amaurosis in corresponding mouse models. Furthermore, it enabled transduction of primate photoreceptors from the vitreous, expanding its therapeutic promise.

IntroductionBack to Top

Inherited forms of retinal degeneration, which afflict 1 in 3000 people worldwide, arise primarily from mutations in cells of the eye’s outermost retinal layer (Fig. 1). These include photoreceptor cells—light-detecting neurons in the retina of vertebrate eyes—or cells of the retinal pigment epithelium (RPE)—a layer of pigmented cells that lies just outside of and supports the photoreceptors. The outer retina is, therefore, the primary target for ocular gene therapies (1), which deliver a wild-type copy of the mutated gene to the appropriate cells (transduction) typically by using an adeno-associated virus (AAV) vector.


Fig. 1 Three libraries were created: an error-prone AAV2 Y444F library, an AAV shuffled library, and a random 7mer insertion library. Each library had a diversity of 107, and they were mixed in equal parts and injected intravitreally into adult transgenic rho-GFP mice eyes. One week after injection, eyes were enucleated and retinas were dissociated using a mild papain protease treatment, followed by FACS isolation of photoreceptors (representative FACS plot). Viral cap genes from the isolated cells (representing the AAV variants from the library that successfully transduced photoreceptors) were then PCR-amplified from genomic extractions for cloning and repackaging. One round of evolution consisted of initial library diversification followed by three selection steps. RGCs, retinal ganglion cells; RPE, retinal pigment epithelium.

Current gene delivery vehicles require administration into the subretinal space between the RPE and the photoreceptors to transduce these neurons. To achieve delivery of the vector, a needle must penetrate the retina and, in doing so, detaches the photoreceptor cell layer from the RPE (2, 3). Three recent clinical trials for the retinal disease type 2 Leber’s congenital amaurosis (LCA2) used subretinal injections to deliver AAV that carried the retinal isomerase–encoding gene RPE65 to the RPE; the trial protocol benefited from the atypical pathology of LCA2, which exhibits a loss of photosensitive function without significant structural disruption of retinal layers for many years (4–8). In contrast, most retinal degenerative diseases (including retinitis pigmentosa and macular degeneration, which account for half of all retinal degeneration cases) are characterized by the progressive loss of photoreceptor cells and increasingly fragile retinal architecture across the entire retina (9, 10). In such disease states, subretinal surgery can induce mechanical damage, reactive gliosis, and loss of function (8, 11). These procedural effects have even been documented in one LCA2 trial, as patients receiving a subretinal injection under the foveal region lost retinal thickness and visual acuity; these results led investigators to conclude that LCA gene therapy is efficacious in the extrafoveal retina but offers no benefit and some risk in treating the fovea (12).

X-linked retinoschisis (XLRS) is a retinal degenerative disease caused by mutations in transcripts expressed in the outer retina and is amenable to clinical gene therapy. XLRS mutations reside in the gene encoding the retinoschisin protein and lead to particularly severe structural abnormalities and compromised vision. This protein, ordinarily secreted from photoreceptors, is required for maintenance of normal retinal organization including the photoreceptor–bipolar cell synapse (13). Gene therapy studies in the XLRS knockout mice, null for retinoschisin (Rs1h), illustrate the challenges for XLRS gene replacement therapy in patients. Intravitreal and subretinal approaches have been used to deliver AAV carrying a functional copy of the Rs1h gene in this model (14). However, photoreceptors were efficiently transduced only with the subretinal injections, which both limited the region of therapeutic effect and caused surgical disruption at the injection site (15, 16).

How well retinas of patients with retinal degenerative disease can tolerate subretinal surgery will depend on the nature of the mutation and the stage of the retinal degeneration at which the surgery is performed. As an additional concern, because outer retinal defects are expressed across the tissue, an effective treatment should be pan-retinal or reach cells along the entire width of the retina; subretinal injections transduce only cells that contact the local “bleb” of injected fluid. Therefore, a new delivery technology is needed for gene transfer to deeper layers of the retina. Such a technique could have broader implications for treating human diseases that involve cells within structurally complex tissues that are inaccessible to AAV as a result of physical (for example, diffusion and membranes) and cellular barriers (for example, endothelial cells).

To endow AAV with the ability to overcome such complex tissue barriers, we developed an in vivo–directed evolution strategy that enabled us to iteratively enrich for AAV variants capable of reaching the outer retina from the vitreous. Here, we describe an evolved AAV variant (7m8) that mediated highly efficient delivery to all retinal layers in mice and nonhuman primates. 7m8 also mediated therapeutic gene delivery to photoreceptors in two mouse models of retinal disease, enabling noninvasive, long-term histological and functional rescue of disease phenotypes across the entire retina. These findings have important and immediately clinically relevant implications for the development of gene therapies for LCA2, retinoschisis, and other retinal diseases requiring robust, pan-retinal gene expression without retinal detachment.

ResultsBack to Top

Naturally occurring AAV serotypes cannot transduce photoreceptors from the vitreous side of the retina (17, 18) because the inner limiting membrane (ILM) and intervening neuronal and glial cells and processes of the inner retina (Fig. 1) form a formidable diffusive barrier with abundant binding sites for AAV particles that is several hundred micrometers thick in rodents and primates (19, 20). Rational mutagenesis of surface-exposed tyrosine residues, which allows AAV particles to avoid intracellular ubiquitination and degradation, increases intravitreal photoreceptor gene delivery through improved intracellular nuclear trafficking and thereby indicates that AAV2 has the partial ability to infect photoreceptors from the vitreous (21, 22). However, decreasing intracellular ubiquitination is secondary to the substantial physical barriers a vector encounters during infection, which in this case may include viral sequestration, poor diffusion in the ILM, extracellular spaces, and intervening cell layers, or both.

Library screening converges on one dominant variant

To develop an AAV capable of outer retinal gene delivery upon intravitreal injection, we designed an in vivo–directed evolution approach (Fig. 1). Three AAV libraries, each with a diversity of 107, were used: an AAV2 capsid protein (cap)–encoding library with a random seven–amino acid sequence inserted into loop 4 (within the heparin-binding domain) of the capsid (23, 24); a library that encoded a tyrosine mutant version of the AAV2 genome (25) subjected to random mutagenesis (AAV2 Y444F EP); and a chimeric capsid–encoding library generated by shuffling AAV1, 2, 4, 5, 6, 8, and 9 (26). The libraries were combined and injected intravitreally into the eyes of transgenic mice that expressed a rhodopsin–green fluorescent protein (GFP) fusion specifically in their rod photoreceptors (27). Fluorescence-activated cell sorting (FACS) was optimized to isolate pure, GFP-positive photoreceptors from harvested retinas 1 week after library injection. Successful AAV cap variants were then recovered from these neurons by polymerase chain reaction (PCR), and virus was again packaged. Two more such selection steps were conducted, followed by error-prone PCR to introduce further diversity into the library, and three additional in vivo selection steps were carried out.

After this extended evolution process, 48 variant cap genes were sequenced (Fig. 2A). Notably, 46 of these clones originated from the AAV2 seven–amino acid peptide (7mer) insertion library, with 31 containing the same seven–amino acid motif (LGETTRP). The next most prominent variant (5 of 46) contained a similar motif (NETITRP) with a positively charged arginine residue in between a polar threonine and a nonpolar proline residue (TRP). This represents a convergence from ~107 input variants down to a single dominant consensus sequence. In addition, 34 of the 46 clones coming from the 7mer library harbored a V708I mutation. We focused on one of these prominent clones, 7m8, for in-depth characterization.


Fig. 2 (A) Sequencing of evolved variants revealed a high degree of convergence in the selected viral pools. All but two sampled variants originated from the AAV2 seven–amino acid library, and 67% of all clones (32 of 48) contained the same 7mer motif (LGETTRP). The remainder, not included in the table, were represented only once within the population. (B to D) Molecular model of AAV2 containing the insertion LALGETTRP (shown in orange) after amino acid 587. The interactions between the inserted loop and the other surface loops of the capsid likely play a role in the novel properties of the virus.

We modeled the major capsid protein of 7m8, that is, AAV2~588LALGETTRP, superimposed on its parent AAV2 (Fig. 2B). The 588LALGETTRP insertion disrupted basic arginine residues in loop 4 implicated in AAV2 binding to its primary receptor, heparan sulfate proteoglycan (HSPG) (28) (Fig. 2, C and D). In addition, the peptide’s location in loop 4 and proximity to loop 3 (also involved in receptor binding) place it in a position to further alter viral tropism. 7m8 was used to package recombinant virus carrying the gene encoding GFP (7m8-GFP), and its glycan dependencies (fig. S1A) were analyzed in vitro. 7m8 infection was still HSPG-dependent (28), although its heparin affinity was lower compared to AAV2 (fig. S1B). In addition, like AAV2, 7m8 showed no sialic acid dependence; however, it was >10- to 100-fold more infectious than AAV2 in Chinese hamster ovary (CHO) cell lines.

In vivo characterization shows pan-retinal reporter gene expression with 7m8

To assess the capacity of 7m8 to mediate gene delivery to the outer retina, we injected 7m8-GFP intravitreally into adult wild-type mice, which resulted in strong pan-retinal GFP expression that was readily visualized via fundus imaging (Fig. 3A). In addition, 6 weeks after injection, retinal flat mounts, imaged with the outer nuclear layer facing upwards, showed marked GFP expression in photoreceptors across the retina (Fig. 3B and fig. S2), and a montage of transverse cryosections further demonstrated pan-retinal expression (Fig. 3C). Higher-magnification imaging revealed that although AAV2-mediated expression was limited to RGCs and some Müller glia (Fig. 3D), 7m8 led to marked expression, not only within RGCs and Müller cells but also in amacrine cells, bipolar cells, rods, cones, and RPE (Fig. 3, E and F). These results establish that positive selection for localization to photoreceptors resulted in substantially enhanced transduction of this important cell type, although not surprisingly, it did not select against transduction of other neurons or RPE. Because 7m8 is capable of transducing cells far away from the injection site, we assessed whether this variant remained restricted to the retina by examining brain sections of mice after intravitreal 7m8-GFP delivery. The optic nerve from the injected eyes showed high levels of GFP expression detectable through the optic tract, with the GFP-positive axons of RGCs traveling to the suprachiasmatic nuclei and lateral geniculate nuclei; however, there were no GFP-positive cell bodies (fig. S3) in either brain region. Furthermore, 7m8 carrying a rod-specific rhodopsin promoter (7m8-rho-GFP), rather than the ubiquitous CAG promoter, successfully restricted expression to photoreceptors (Fig. 3, H and I). These results indicate that we created an AAV variant able to transduce all retinal cell types after intravitreal administration and that coupling this capsid with a promoter of interest can target expression to a specific cell type.


Fig. 3 (A) Fundus image of a wild-type (WT) mouse retina 3 weeks after injection with 7m8-scCAG-GFP, showing pan-retinal gene expression. (B) A retinal flat mount oriented with the outer nuclear layer facing upwards shows GFP expression in photoreceptor cell bodies and outer segments across the retina. (C) Montage of confocal images showing unamplified pan-retinal GFP expression in a WT retina after injection with 7m8-scCAG-GFP. (D and E) Confocal microscopy of transverse retinal sections 6 weeks after intravitreal injection of 1 μl of AAV2 [1 × 1012 viral genomes (vg)/ml] (D) or 7m8 (E) in adult WT mice shows greater transduction of the outer retina in eyes injected with 7m8 compared to the parental serotype. (F) GFP expression in flat mounted RPE after injection of 7m8-scCAG-GFP. (G to I) GFP expression restricted to photoreceptors using a rhodopsin promoter. (G) Flat mount retina, (H) cross section, and (I) fundus images showing in vivo expression of 7m8-rho-GFP.

We next compared the transduction properties of 7m8 to the AAV vector that reportedly provides the best photoreceptor transduction profile when injected into the vitreous, a quadruple tyrosine mutant AAV2 (AAV2-4YF) (21, 22). Particles (1011) of both 7m8-rho-GFP and AAV2-4YF-rho-GFP were injected into the vitreous cavity of four wild-type mice eyes. Reverse transcription PCR (RT-PCR) was conducted to quantify GFP mRNA levels, and 7m8 showed a fivefold increase over AAV2-4YF in the murine retina (fig. S4).

Last, we investigated immune reactions to the 7m8 capsid upon vector readministration. Intravitreal administration of either AAV2 or AAV5 vectors has been reported to generate a humoral immune response against the viral capsid that blocks vector expression upon subsequent readministration into the partner eye (29, 30). We investigated whether such immune responses to 7m8 were similar to these previously reported findings and did indeed find that after injection of 7m8-GFP in one mouse eye, later injection of 7m8-GFP into the contralateral eye did not lead to GFP expression. Cryosections of these eyes showed the presence of a few activated microglial and macrophage cells in the second eye but no structural damage in either retina (fig. S5).

7m8-RS1 improves structure and rescues function in the Rs1h−/− mouse

To evaluate the potential of the 7m8 vector for photoreceptor gene therapy, we used the retinoschisis knockout (Rs1h−/−) mouse (13), which has difficulty tolerating subretinal injections (fig. S6). After injection at P15, early in the development of the Rs1h−/− pathology when the retina is largely intact, 7m8-rho-GFP strongly and pan-retinally transduced photoreceptors (fig. S7J). In contrast, AAV2-rho-GFP and AAV8-rho-GFP—which was previously reported to improve the RS1 phenotype when injected 6 to 9 weeks after birth, closer to the peak of cavity formation (31)—led to little GFP expression in photoreceptors when administered at P15 (fig. S7).

Intravitreal administration of 7m8-rho-RS1 led to high-level, pan-retinal RS1 expression in photoreceptors and throughout all other retinal layers (Fig. 4, A and B), comparable to wild-type levels of this secreted protein (Fig. 4C). Furthermore, imaging of the photoreceptor–bipolar cell synapse showed that 7m8-rho-RS1 treatment improved synaptic organization (fig. S7, K and L).


Fig. 4 (A) RS1 immunostaining in control Rs1h−/− eyes shows absence of the protein. (B) Staining of 7m8-rho-RS1–injected eyes shows strong RS1 expression in photoreceptor inner segments, as well as in the outer plexiform layer, inner plexiform layer, and inner nuclear layer. (C) Expression of RS1 in a WT eye. (D to F) Representative high-resolution SD-OCT images of retinas injected with 7m8-rho-GFP (D), 7m8-rho-RS1 (E), or uninjected WT animals (F). Fundus images were taken through the inner nuclear layer of the superior retina and exclude other layers. (G to I) Transverse images of the superior (upper image) and inferior (lower image) retina were collected using the optic nerve head as a landmark. (J) Quantification of the mean full-field scotopic ERG b-wave amplitude resulting from a high-intensity (1 log cd × s/m2) stimulus recorded on a monthly basis beginning 1 month after injection at P15. The increase in b-wave amplitude in 7m8-rho-RS1–treated eyes was significant at every time point (comparison to 1 month: P = 0.006, 2 months: P = 0.019, 3 months: P < 0.0001, and 4 months: P < 0.0001), whereas the response recorded from AAV2-rho-RS1–, AAV8-rho-RS1–, and 7m8-rho-RS1–injected eyes was the same as that from untreated Rs1h−/− eyes. n = 7 for all groups. Statistical analysis was a one-way analysis of variance (ANOVA) with post-hoc Tukey’s multiple comparison performed with GraphPad Prism software. (K) Analysis of ERG responses under scotopic (upper traces, stimulus range from −3 to 1 log cd × s/m2) and photopic (lower traces, range from −0.9 to 1.4 log cd × s/m2) conditions. (L) Representative ERG traces from 7m8-rho-RS1–injected eyes show improved amplitude of the a- and b-wave and a waveform closer to WT eyes, compared to 7m8-rho-GFP–injected eyes.

The structural improvement observed by immunohistochemistry was corroborated by high-resolution spectral domain optical coherence tomography (SD-OCT) imaging (Fig. 4, D to I). Four months after intravitreal injection of 7m8-rho-GFP into the Rs1h−/− mice, retinas were marked by large and pervasive cavities (Fig. 4, D and G). In stark contrast, 7m8-rho-RS1–treated retinas had few cavities, which were barely visible in fundus images (Fig. 4E) and cross sections (Fig. 4H). Also, treated retinas were only slightly thinner than wild-type retinas of the same age (Fig. 4, F and I). By comparison, no structural improvements were observed in retinas treated with AAV2-rho-RS1 or AAV8-rho-RS1 4 months after treatment (fig. S7).

Electroretinography (ERG), which reports retinal function in response to a flash of light, was conducted to analyze rescue of the hallmark functional deficits of retinoschisis. Specifically, the Rs1h−/− mouse progressively loses amplitude in the ERG b-wave, which arises from the inner retina, and displays relative preservation of the a-wave, which originates from photoreceptors (32).

A time course analysis over 4 months showed that ERG b-waves in 7m8-rho-GFP–injected eyes steadily decreased by 27% to 165.9 ± 17.5 μV, whereas the amplitude in 7m8-rho-RS1–treated eyes was preserved at a significantly higher level of 333.4 ± 15.7 μV (Fig. 4J; P < 0.0005, 7m8-rho-GFP versus 7m8-rho-RS1 4 months after injection, two-tailed paired Student’s t test). In contrast, ERG b-wave amplitudes recorded from eyes injected with AAV2-rho-RS1, AAV8-rho-RS1, or AAV2-rho-GFP were undistinguishable from untreated eyes. In addition, the amplitude of ERGs recorded from 7m8-rho-RS1–treated mice 4 months after injection revealed improvement of the b-wave over a range of stimulus levels under both scotopic (dark-adapted, upper traces) and photopic (light-adapted, lower traces) conditions (Fig. 4K). Furthermore, representative scotopic ERG recordings from 7m8-rho-RS1– and 7m8-rho-GFP–treated eyes at 4 months with the highest stimulus intensity illustrated restoration of ERG amplitude and waveform (Fig. 4L). These results indicate that intravitreal administration of 7m8-rho-RS1 led to substantial and stable improvements of rod and cone photoreceptor-mediated visual function, as well as synaptic transmission over a wide range of lighting intensities.

7m8-RPE65 expression rescues the rd12 phenotype

To generalize the potential of the 7m8 vector for outer retinal gene therapy to another important disease model, we investigated gene replacement therapy in the rd12 mouse model of LCA, which differs from the retinoschisin model in several key ways. First, in contrast to Rs1h−/− mice, rd12 retinas remain structurally intact until 3 months after birth. In addition, because the rd12 phenotype results from mutations in the gene encoding RPE-specific protein RPE65, rescue in this model requires transduction of the RPE, which lies beyond the photoreceptor layer and is thus an even more distant and challenging target to infect from the vitreous.

We analyzed the ability of 7m8 to deliver a wild-type copy of the RPE65 gene in this animal model by administering 7m8-RPE65 into one eye and injecting 7m8-GFP into the contralateral, control eye. Labeling of RPE65 protein in RPE flat mounts (Fig. 5A) revealed expression of the protein in 7m8-RPE65–injected eyes, whereas 7m8-GFP– and AAV2-RPE65–injected eyes lacked any labeling, similar to untreated eyes. RT-PCR (Fig. 5B) revealed that eyes injected with 7m8-RPE65 had increased amounts of RPE65 mRNA, whereas 7m8-GFP–injected eyes expressed amounts of RPE65-encoding mRNA similar to those of untreated eyes.


Fig. 5 (A) Anti-RPE65 labeling. WT C57BL/6 mice expressed RPE65 across the RPE monolayer, whereas rd12 mice lacked the RPE65 protein. Mice injected with 7m8-GFP and AAV2-RPE65 had no labeling of RPE65 protein. Mice injected with AAV2-RPE65 displayed low amounts of RPE65 in scattered RPE cells. In contrast, mice injected with 7m8-RPE65 displayed high levels of RPE65 expression in RPE cells. (B) RT-PCR from RNA extracted from RPE cells (which were dissected away from other retinal cells) revealed that levels of RPE65 mRNA were elevated in mice injected with 7m8-RPE65 but not in mice injected with 7m8-GFP or AAV2-RPE65. (C) ERG recordings revealed that the amplitude of the ERG b-wave was significantly increased (P < 0.0001) in mice injected with 7m8-RPE65 relative to other cohorts, whereas no significant differences were observed among ERGs recorded from mice injected with 7m8-GFP, AAV2-GFP, or the AAV2 vector expressing RPE65. (D) Representative ERG traces illustrate improved amplitude of the ERG in 7m8-scCAG-RPE65–injected (solid trace) compared to 7m8-GFP–injected eyes (dashed trace). n = 7 for all groups. Statistical analysis was a one-way ANOVA with post-hoc Tukey’s multiple comparison performed with GraphPad Prism software.

We assessed functional recovery in rd12 retinas by electroretinographic analysis of the full-field scotopic b-wave measured 35 days after vector injection (n = 7). 7m8-CAG-RPE65 treatment led to a significant restoration of the scotopic b-wave amplitude compared to the 7m8-CAG-GFP–treated contralateral control eye, or to AAV2-CAG-RPE65 (Fig. 5C; P < 0.0001, one-way ANOVA with post-hoc Tukey’s multiple comparison). Representative ERG traces from 7m8-CAG-RPE65–and 7m8-CAG-GFP–injected eyes further illustrate this rescue of a- and b-wave amplitudes (Fig. 5D).

These results demonstrate that intravitreal 7m8-CAG-RPE65 administration yields substantial amounts of RPE65 protein expression in the RPE, resulting in considerable functional improvements in vision. Therefore, 7m8-mediated gene therapy can reach the deepest layers of a structurally intact retina and deliver therapeutic amounts of an important gene.

Intravitreal delivery of 7m8 transduces photoreceptors in the nonhuman primate retina

We further evaluated the clinical potential of our vector in nonhuman primates (male cynomolgus monkeys) in which the ILM is a significantly thicker physical barrier to the retina than in rodents (19, 20, 33) and thus poses greater challenges for intravitreal gene delivery. To date, only AAV2 has been intravitreally administered in macaques, leading to restricted transduction in a limited region around the macular RGCs (34–37). To assess the ability of 7m8 to extend gene delivery beyond the foveal region or deeper than the RGC layer, we intravitreally administered 5 × 1012 viral particles of 7m8 with the cytomegalovirus (CMV) promoter driving expression of GFP (7m8-CMV-GFP) or the AAV2 quadruple tyrosine mutant (AAV2-4YF-CMV-GFP), again used as a comparison, in an adult macaque. GFP expression was subsequently assessed in vivo by funduscopic imaging.

In both cases, there was an initial moderate but self-resolving vitreal inflammation. At 3 weeks after injection, fluorescence imaging showed GFP expression for both vectors (Fig. 6, A and B). By 8 weeks, fluorescence levels increased for both vectors (Fig. 6, C and D); however, 7m8 mediated substantially higher gene expression both across the retina (Fig. 6, C and D) and inside the fovea (Fig. 6, G and H). At 12 weeks after injection, fundus imaging revealed signs of an immune response, likely a result of high levels of GFP, known to be an immunogenic protein (38, 39). The study was halted, and histological analysis showed GFP expression in the extrafoveal photoreceptors in the 7m8-treated eye but not in the AAV2-4YF–treated eye (Fig. 6, E, F, and I to K, and fig. S8). We did not observe GFP expression in the RPE, in contrast to the results observed in mice, which showed high levels of transgene expression in this outermost layer. This may result from the fact that fewer viral particles were able to overcome the more substantial barriers in the primate retina, and, thus, fewer particles were able to reach the RPE adjacent to the transduced photoreceptors of the primate retina. The late-developing inflammation caused retinal thinning and damage in the highest GFP-expressing regions within the fovea, consistent with previous studies using high doses of AAV vectors that express GFP under the control of the CMV promoter in the nonhuman primate retina (40). Another macaque that received an injection of 5 × 1012 viral particles of 7m8 expressing GFP under the control of the neuronal promoter cx36 yielded a similar GFP expression pattern starting at 3 weeks after injection (fig. S9) but with no observable inflammation even after 16 weeks, further indicating that CMV promoter usage can yield a transgene-specific immune response against GFP. Regardless, these results indicate that 7m8 mediates strong amounts of transgene expression in the outer retina upon intravitreal administration to both rodents and nonhuman primates.


Fig. 6 (A to D) Fundus images of an adult macaque injected bilaterally with 5 × 1012 viral particles of either (A) AAV2-4YF-CMV-GFP or (B) 7m8-CMV-GFP 3 weeks after injection, and (C) AAV2-4YF-CMV-GFP and (D) 7m8-CMV-GFP at 8 weeks after injection. (E and F) Agarose sections through a brightly expressing fluorescent spot from peripheral retina in (E) AAV2-4YF-CMV-GFP– or (F) 7m8-CMV-GFP–injected eyes imaged at equal confocal acquisition settings. (G and H) Equal setting fundus imaging of the fovea in (G) AAV2-4YF-CMV-GFP– or (H) 7m8-CMV-GFP–injected retinas. (I and J) Confocal imaging through agarose sections of retinal sections injected with (I) AAV2-4YF-CMV-GFP or (J) 7m8-CMV-GFP imaged at adjusted settings for optimal detection of signal, costained with anti–m-opsin labeling in red. (K) High-magnification imaging of native GFP in macaque rods and cones resulting from injection of 7m8-CMV-GFP into the vitreous.

DiscussionBack to Top

Our results illustrate the substantial potential of 7m8 for retinal gene therapy to treat diverse forms of inherited retinal degenerations affecting the inner and outer retina. More broadly, this work demonstrates that engineering AAV vector variants with properties designed to overcome critical biological transport and transduction barriers promises to make a wide range of diseases affecting complex tissues amenable to clinical gene therapy.

The landmark clinical trials for LCA2 gene therapy support the safety and efficacy of retinal gene transfer via AAV vectors (5–8) when the retinal structure is intact; however, the deteriorating retina in most degenerations make it challenging to build on the success of these clinical trials. We have demonstrated that AAV can be engineered to traverse several hundred micrometers of dense tissue—filled with extracellular matrix network, cell bodies, and processes—and dramatically increase gene delivery to important target cells far from a noninvasive injection site. This capability may substantially broaden the therapeutic potential for AAV to treat retinal neurodegeneration. Moreover, gene therapy approaches that require delivery across other complex tissue structures may also benefit from this engineering strategy, including endothelial barriers (for example, blood-brain, -retina, -muscle, or -tumor) upon systemic administration, or intraorgan or tumoral barriers upon local delivery. This approach, which selected viruses for the ability to penetrate the retina and infect a therapeutically relevant cell, advances on previous work that selected AAV libraries in vitro (41, 42) or selected the virus in vivo for general localization to an organ after injection into the bloodstream (43, 44).

The new AAV variant provided therapeutic amounts of transduction of photoreceptor cells and RPE after noninvasive administration to the vitreous humor of the eye in both normal murine retina and in models of human retinal disease. Furthermore, transduction was strong in the nonhuman primate fovea—a region essential for high-acuity vision that may be damaged by detachment from the RPE during a subretinal injection (12)—as well as numerous regions beyond the fovea. Future engineering within nonhuman primates may enable even broader pan-retinal expression.

The creation of a new AAV variant that successfully transduces all retinal layers from the vitreous offers therapeutic potential for a broad array of additional inherited retinal degenerative diseases. The pan-retinal infective properties of 7m8, coupled with promoters or microRNA target regions to mediate cell-restrictive expression, make it a valuable vector for targeting other subsets of neurons in the retina. For example, 7m8 gene delivery of engineered light-sensitive channels, such as LiGluR (45) or other optogenetic tools including channelrhodopsin variants, to ON bipolar cells (46) or cone photoreceptor inner segments (47) may allow for the artificial restoration of light sensitivity in late stages of retinal degeneration.

The increased retinal transduction efficiency of 7m8 compared to its parental serotype AAV2 may arise from this variant’s reduced heparin affinity, which may both decrease capsid sequestration in the ILM and enable enhanced penetration through retinal layers. Also, the peptide sequence could confer binding to a novel cell surface receptor or enhance intracellular viral trafficking. In any case, the high infectivity of the 7m8 capsid compared to previous vectors may enable the use of relatively low dosages, thereby reducing the chance of immune response to vector capsid protein, an important consideration for intravitreally delivered AAV vectors (30).

Materials and MethodsBack to Top

Library generation and viral production

To generate the Y444F mutation on the AAV2 cap gene, as well as all subsequent tyrosine-to-phenylalanine mutations, we used a site-directed mutagenesis kit (QuikChange Lightning, Agilent Technologies). As we have previously reported (48, 49), random mutagenesis libraries were generated by subjecting the resulting AAV2 Y444F cap to error-prone PCR, as described. A 7mer peptide display library, created essentially as previously described (24), and an AAV library constructed by DNA shuffling of cap genes from AAV1, 2, 4, 5, 6, 8, and 9, as we have reported, (26) were also used. The rcAAV libraries and rAAV vectors expressing GFP under a CAG or Rho promoter were packaged as previously described (26, 48), and deoxyribonuclease-resistant genomic titers were obtained through quantitative PCR.

Library selection and evolution

The libraries were pooled, and two rounds of evolution were performed, each consisting of initial library diversification followed by three in vivo selection steps. In each such step, P30 rho-GFP mice (27) were intravitreally injected with 2 ml of iodixanol-purified, phosphate-buffered saline (PBS)–dialyzed library with a genomic titer of about 1 × 1012 vg/ml. One week after injection, harvested retinas were dissociated with a light papain protease treatment, followed by FACS (Cytopeia Influx Cell Sorter, BD Biosciences) isolation of photoreceptor cells. DNA was extracted, and successful virions were PCR-amplified, recloned into the AAV genomic plasmid, and repackaged for the next injection. After three selection steps, the recovered cap genes were subjected to error-prone PCR, and then three additional selection steps. After this process, the cap genes of 48 variants were sequenced.

Molecular modeling

The amino acid residues were inserted into the Protein Data Bank (PDB) file 1LP3 (AAV2) at position 588 with Maestro software. This resultant data file was used in the Swiss Model homology mode with default, automated settings to build the VP3 monomer (Fig. 2B). The monomer was superposed onto the wild-type AAV2 VP3 monomer, and Viper ( was used to reconstruct the capsid.

Histological characterization of AAV transduction

P30 wild-type mice were used to analyze evolved variants. One to 3 months after intravitreal injection, retinas were extracted, and 10-μm-thick transverse cryosections were cut as described previously (17). Sections were analyzed by confocal microscopy (LSM5; Carl Zeiss). For immunohistochemistry on agarose embedded retinas, eyecups were embedded in 5% agarose and cut on a vibratome in a PBS bath. Sections were briefly postfixed, rinsed in PBS, and blocked before antibody labeling overnight at 4°C. Antibodies were the following: peanut agglutinin (PNA) (Molecular Probes, 1:200), anti–M-opsin (Chemicon International, 1:500), anti-RS1 (3R10 mouse monoclonal antibody, gift from R. Molday, 1:5), anti-synaptophysin (Abcam, 1:1000), anti-vimentin (Dako, 1:1000), anti-mGluR6 (Abcam, 1:1000), anti-IbaI (Abcam, 1:500), or anti-CD68 (Abcam, 1:500).


Electroretinograms were recorded (Espion E2 ERG System; Diagnosys LLC) in response to six light flash intensities ranging from −3 to 1 log cd × s/m2 on a dark background as described previously (17). Each stimulus was presented in series of three. For photopic ERGs, the animal was first exposed to a rod-saturating background for 5 min. Stimuli ranging from −0.9 to 1.4 log cd × s/m2 were presented 20 times on a lighted background. Data were analyzed with MATLAB (v7.7; MathWorks).

High-resolution SD-OCT

Images of retinal cross sections were averaged from eight contiguous slices. Histological imaging was performed with an 840-nm SDOIS (Spectral Domain Ophthalmic Imaging Systems) OCT system (Bioptigen). Retinal thickness, outer nuclear layer thickness, and photoreceptor inner and outer segment length measurements were gathered with InVivoVue software.

Primate retina

Intravitreal injections were made, with methods described previously (34), in one 11.2-kg male cynomolgus monkey that had tested negative for serum antibodies against AAV2. The right and left eyes were injected with 5 × 1012 viral particles of 7m8-CMV-GFP and AAV2- 4YF-CMV-GFP, respectively. Confocal scanning laser ophthalmoscopic images (Spectralis HRA, Heidelberg Engineering) were obtained from the two retinas at 3 and 8 weeks after injection, with autofluorescence settings, which leads to effective GFP visualization. At 12 weeks after injection, the monkey was euthanized, both retinas were lightly fixed in 4% paraformaldehyde, and tissue was examined via confocal microscopy. Pieces of primate retina were then embedded in 5% agarose and sectioned at 200 μm for immunocytochemistry.

Statistical analysis

Data were analyzed in GraphPad Prism version 5.00 (GraphPad Software) and are presented as means ± SD. As indicated in the results, data were compared with either a paired or an unpaired Student’s t test or one-way ANOVA with post-hoc Tukey’s multiple comparison. The difference was considered statistically significant if the P value was less than 0.05. In the figures, P values <0.05 are indicated by a single asterisk. P values <0.01 are indicated by a double asterisk. P values <0.001 are indicated by a triple asterisk. All analyses were two-tailed.

Supplementary MaterialsBack to Top

Fig. S1. Glycan dependencies of 7m8 compared to AAV2.

Fig. S2. GFP expression in photoreceptors after transduction with 7m8.

Fig. S3. GFP expression in the optic tract and brain histology after transduction with 7m8.

Fig. S4. Comparison of photoreceptor transduction by 7m8-rho-GFP and AAV2-4YF-rho-GFP.

Fig. S5. Immunolabeling for activated microglia and infiltrating macrophages after 7m8 readministration.

Fig. S6. Structural damage caused by subretinal injections in Rs1h−/− mice.

Fig. S7. Expression of control vectors and structural improvements at the photoreceptor–bipolar cell junction in Rs1h−/− mice.

Fig. S8. Expression of 7m8-CMV-GFP in nonhuman primate photoreceptor flat mounts.

Fig. S9. Expression of 7m8-cx36-GFP in nonhuman primate.

1LP3-mutant-swiss.pdb: PDB format data file for the modeling of the 7m8 capsid.

References and NotesBack to Top