Control of directional change after mechanical stimulation in Drosophila
© Zhou et al.; licensee BioMed Central Ltd. 2012
Received: 30 August 2012
Accepted: 23 October 2012
Published: 29 October 2012
Skip to main content
© Zhou et al.; licensee BioMed Central Ltd. 2012
Received: 30 August 2012
Accepted: 23 October 2012
Published: 29 October 2012
Proper adjustment of moving direction after external mechanical stimulation is essential for animals to avoid danger (e.g. predators), and thus is vital for survival. This process involves sensory inputs, central processing and motor outputs. Recent studies have made considerable progress in identifying mechanosensitive neurons and mechanosensation receptor proteins. Our understandings of molecular and cellular mechanisms that link mechanosensation with the changes in moving direction, however, remain limited.
In this study, we investigate the control of movement adjustment in Drosophila. In response to gentle touch at the anterior segments, Drosophila larvae reorient and select a new direction for forward movement. The extent of change in moving direction is correlated with the intensity of tactile stimuli. Sensation of gentle touch requires chordotonal organs and class IV da neurons. Genetic analysis indicates an important role for the evolutionarily conserved immunoglobulin (Ig) superfamily protein Turtle (Tutl) to regulate touch-initiated directional change. Tutl is required specifically in post-mitotic neurons at larval stage after the completion of embryonic development. Circuit breaking analysis identified a small subset of Tutl-positive neurons that are involved in the adjustment of moving direction.
We identify Tutl and a small subset of CNS neurons in modulating directional change in response to gentle touch. This study presents an excellent starting point for further dissection of molecular and cellular mechanisms controlling directional adjustment after mechanical stimulation.
Proper adjustment of moving direction is essential for animals to forage and to escape from predation. Animals use cues such as light, odor, temperature and mechanical stimuli to make their movement decisions. The focus of this study is to understand the mechanisms that regulate the adjustment of moving direction after gentle touch.
Reorientation of movement after mechanical stimulation requires activation of mechanosensitive neurons, the integration and processing of information in the central nervous system (CNS), and motor outputs (as reviewed by[2, 3]). Recent studies in genetic model systems such as Drosophila and C. elegans have shed light on molecular mechanisms underlying the activation of mechanosensitive neurons[4, 5]. For instance, genetic screen in C. elegans led to the identification of mec-4 and mec-10, which encode mechanotransducers (i.e. DEG/ENaC channels). Genetic dissection of mechanosensation in Drosophila also identified NompC, a member of the TRP channel family, as a mechanotransducer[7, 8]. However, less is known about how the information from mechanosensory neurons is processed in the CNS for animals to adjust their moving direction.
Drosophila is an excellent model system for understanding molecular and cellular mechanisms underlying directional change after mechanical stimulation. The anatomy and development of mechanosensory organs in Drosophila have been well studied[4, 9]. Molecules important for mechanotransduction have been identified in Drosophila, such as mechanotransducers Pickpocket, Piezo and NompC[7, 8], as well as other proteins that are required for maintaining the structural integrity of mechanosensitive neurons (e.g. NompA). Recent development of sophisticated techniques that allow spatial and temporal manipulation of circuit activity in living flies (e.g.[13–15]), greatly facilitates the study of neuronal circuitry underlying specific behaviors.
In this study, we investigate the mechanisms that regulate the adjustment of moving direction by Drosophila larva in response to gentle touch. We examined the modulation of directional change by gender difference, the intensity of tactile stimuli, and the nociceptive pathway. We also performed genetic analyses to gain insights into underlying molecular and cellular mechanisms. We show that the adjustment of moving direction after gentle touch requires the turtle (tutl) gene, which encodes an evolutionarily conserved Ig-superfamily transmembrane protein. Our results also implicate a role for a small subset of Tutl-positive neurons in modulating the pattern of directional change.
To quantify the data, we measured the angle (“θ” in Figure 1A”’) between the directions of original and reoriented forward movement. Similar navigational pattern was observed in Canton-S (CS), Oregon-R (OR), and w1118 larvae (Figure 1B). We also found that male and female larvae showed similar navigational pattern in response to gentle touch (data not shown). No significant difference in withdrawal response (data not shown), responding time (data not shown), or selection of new moving direction (data not shown), was observed between male and female larvae.
To determine if the level of sensory inputs affects navigational pattern, we applied different intensities of tactile stimuli (i.e. 1 mN, 3 mN, 7 mN and 10 mN) with calibrated filaments to the anterior segments (see Methods). Interestingly, we found that the extent of directional change after tactile stimuli was correlated with the intensity of stimuli (Figure 1C). In response to an increase in intensity from 1 mN to 10 mN, the average change in forward movement direction was increased from 69.4° to 93.8° (Figure 1C). The data fit a linear regression model, indicating a significant correlation between the intensity of stimulus and directional change (Figure 1C).
Previous studies in Drosophila suggest that the mechanisms of sensing gentle touch are different from that of nociception[7, 10, 11, 16]. If so, one would predict that directional change after gentle touch should not require the activation of nociceptive pathway. To test this, we examined the response of painless (pain) mutants to gentle touch. pain encodes a member of TRPN channels. pain is expressed in multidendritic neurons (md) and chordotonal organs, and is required for both mechanical and thermal nociception.
We then examined navigational pattern of pain 1 and pain 3 mutant larvae in response to gentle touch. Compared to wild type, no significant difference in navigational behaviors was observed in pain 1 and pain 3 mutant larvae (Figure 2B). This result suggests strongly that directional adjustment after gentle touch involves a Pain-independent pathway.
Previous studies suggest that chordotonal organs are involved in sensing gentle touch in larvae. To determine the potential role of chordotonal organs in navigational pattern after gentle touch, we examined the effect of blocking synaptic transmission from chordotonal organs by expressing a temperature-sensitive form of shibire (shits) that encodes the fly homolog of dynamin. The expression of shits was under control of the chordotonal-specific driver iav-GAL4. This allows the blockage of synaptic transmission in targeted neurons at restrictive temperature.
Previous studies report that class IV dendritic arborization (da) sensory neurons mediate mechanotransduction in response to noxious mechanical (>30 mN) stimuli[10, 20]. To determine if class IV da neurons also play a role in sensing gentle touch, we examined navigational pattern in larvae in which sensory inputs from class IV da neurons were blocked by expressing shits under control of class-IV-da-specific driver pickpocket 1.9-GAL4 (ppk-GAL4) at restrictive temperature. We found that blocking class IV da neurons also significantly affected withdrawal response and subsequent directional change after 1 mN stimulus (Figure 3A and B), while no effect was observed after 7 mN stimulus (Figure 3C and D). Together, these results suggest strongly that class IV da neurons and chordotonal organs are involved in sensing gentle touch.
To understand molecular and cellular mechanisms that modulate directional change after gentle touch, it is necessary to elucidate molecular networks that regulate the formation and function of neuronal circuitry involved. In a search for genes controlling larval navigational pattern, we found that mutations in the turtle (tutl) gene caused a severe defect in adjusting moving direction after gentle touch. tutl encodes an evolutionarily conserved Ig-superfamily transmembrane protein. It is highly homologous to Dasm1 in mice and IgSF9 in humans[22–24], whose function in mammals remains unknown.
To determine if the above defects were due to a reduction in sensation of gentle touch, we examined withdrawal response, which occurs before selection of new moving direction after gentle touch. Surprisingly, we found that tutl mutant larvae, like wild type, displayed normal withdrawal response after gentle touch (Figure 4E). This result indicates that tutl mutant larvae could still sense gentle touch.
Above results indicate a specific role for tutl in the control of navigational pattern after gentle touch, which presents an excellent starting point for genetic dissection of molecular networks and neuronal circuitry involved. Previous studies show that tutl is exclusively expressed in the nervous system[22, 28, 29]. To identify neurons in which tutl functions to regulate directional change, we performed rescue experiments.
Transgene rescue of the navigational phenotype by expressing a UAS- tutl transgene under control of cell-type-specific GAL4 drivers
Change in moving direction (°) (Mean±SEM)
All post-mitotic neurons
Many PNS and CNS neurons
Cholinergic neurons in PNS and CNS
Glutamatergic neurons (motor neurons and neuronal clusters in the brain)
Dopaminergic and serotonergic neurons
RP2, aCC and pCC
Subsets of neurons
Motor neurons and PNS neurons
CNS neurons expressing TrpA1 gene
Neurons expressing serotonin receptor 1B
CNS neurons with diffuse expression throughout brain lobes
All sensory neurons
Class I , bd neurons and chordotonal organs
md neurons, chordotonal organs and some CNS neurons
Class IV da neurons
Neurons co-expressing Appl-GAL4 and Cha-GAL4 are broadly distributed in the peripheral (PNS) and CNS (data not shown), suggesting that proper navigation decision after gentle touch requires the function of tutl in both sensory and central compartments. Consistently, we found that expression of tutl under control of the SN (5–40)-GAL4 driver, which drives gene expression in all PNS sensory neurons but not in CNS neurons, was not sufficient to rescue the phenotype (Table 1). That Cha-GAL4 is not expressed in motor neurons (data not shown), together with a failure of rescue with the drivers (e.g. ftz.ng-GAL4 and OK371-GAL4) for motor-neuron expression (Table 1), argue against a requirement of tutl in motor neurons.
There are a large number of tutl-positive neurons co-expressing Appl-GAL4 and Cha-GAL4, which are widely distributed in the nervous system (data not shown). Such a large number of tutl-positive neurons are likely involved in regulating many different behaviors. To gain insights into neuronal circuitry underlying the control of directional change, it is necessary to identify tutl-positive neurons that are specifically involved in regulating navigational behaviors.
GMR91F06-GAL4 was generated by placing GAL4 under control of an enhancer element in the tutl gene, and is expressed in a small subset of tutl-positive neurons exclusively in the CNS (Figure 8A-C). tutl- GAL4 was generated by inserting GAL4 into the tutl gene. tutl- GAL4 is expressed in a subset of tutl-positive neurons including class III da neurons in the PNS (data not shown) and a subset of neurons in the CNS (Figure 8D-F). Blocking synaptic transmission in GMR91F06-GAL4-positive neurons or tutl- GAL4-positive neurons by shifting from permissive temperature to restrictive temperature, caused a significant decrease in directional change after tactile stimuli (Figure 8G). Whereas expression of shits under control of GMR60G12-GAL4, a driver in which GAL4 is driven by an enhancer element in the Appl gene, did not affect navigational pattern (Figure 8G). Interestingly, while larvae in which GMR91F06-GAL4-positive neurons or tutl- GAL4-positive neurons were silenced, displayed significant changes in navigational pattern, they were still able to withdraw from the stimuli (data not shown). Since withdrawal response is the first response after gentle touch before larvae reorient, this result is consistent with a role for these tutl-positive neurons in central information processing, but not in sensation of gentle touch.
We then examined the effects of blocking synaptic transmission simultaneously in both tutl- GAL4-positive neurons and GMR91F06-GAL4-positive neurons. We found that silencing both types of neurons simultaneously generated an even greater effect (Figure 8G). This suggests that tutl- GAL4-positive neurons and GMR91F06-GAL4-positive neurons function together to modulate navigational pattern in response to tactile stimuli.
We also took an alternative approach to block synaptic transmission in tutl-positive neurons by expressing tetanus toxin light chain (TeTxLC), which blocks evoked synaptic transmission by cleaving synaptic vesicle protein synaptobrevin. UAS-TeTxLC was expressed under control of GMR91F06-GAL4 or tutl- GAL4. Consistent with the results from circuit breaking analysis with shits (Figure 8G), we found that blockage of synaptic transmission in GMR91F06-GAL4-positive neurons or tutl- GAL4-positive neurons with TeTxLC, also significantly affected navigational pattern after tactile stimuli (Figure 8H).
Together, above results suggest strongly that small subset of tutl-postive neurons defined by tutl- GAL4 and GMR91F06-GAL4 are required specifically in neuronal circuitry that modulate navigational pattern in response to tactile stimuli.
In this study, we investigated the control of directional change in response to gentle touch in Drosophila. We showed that navigational pattern was affected by the intensity of stimuli, but not by gender difference. Consistently, reducing sensory inputs by blocking inputs from chordotonal organs or class IV da neurons significantly affected navigational pattern in response to light touch. Our genetic analysis revealed a role for the tutl gene in the control of navigational behaviors. Circuit analysis identified a small subset of tutl-positive neurons that are specifically required for modulating directional change in response to gentle touch.
Consistent with the correlation between stimulus intensity and the extent of directional change, our results showed that reducing sensory inputs by blocking synaptic transmission in chordotonal organs or class IV da neurons, led to a significant decrease in directional change in response to light touch (i.e. 1 mN). The role of chordotonal organs in larval mechanosensation has been reported by several previous studies. For instance, several genes whose mutations caused defects in response to tactile stimuli, were shown to be expressed and functionally required in chordotonal neurons[32, 33]. Moreover, disrupting the structural integrity of chordotonal organs, or disrupting the connection of chordotonal neurons with their post-synaptic targets in the CNS, caused a decrease in sensitivity to touch and vibration, respectively.
Our results indicate that in addition to a role in mechanical and thermal nociception[16, 20], class IV da neurons also mediate mechanosensation in response to light touch. Previous studies show that larvae in which class IV neurons carry mutations in genes encoding mechanotransducers such as pain, pickpocket and piezo, displayed defects in mechanical nociception, but showed normal sensitivity to gentle touch[11, 16, 20]. Together, these studies suggest that class IV da neurons mediate mechanotransduction in response to gentle touch by employing a mechanism different from that in mechanical nociception. Further studies are needed to elucidate the exact mechanism by which class IV da neurons mediate mechanotransduction in response to gentle touch.
Interestingly, we found that when the intensity of tactile stimuli was increased from 1 mN to 7 mN, blockage of sensory inputs from chordotonal organs or class IV da neurons did not affect withdrawal response nor the pattern of directional change. One possible explanation is that stronger stimulus intensity may significantly increase mechanoreceptor currents in other types of mechanosensitive neurons, for instance, external mechanoreceptive sense organs inserted in the cuticle, which may compensate for loss of inputs from chordotonal organs or class IV da neurons leading to normal navigational behaviors.
Behavioral analysis of tutl mutant larvae reveals an interesting phenotype in the adjustment of moving direction after gentle touch. While tutl mutant larvae were able to withdraw from tactile stimuli similarly as wild-type larvae, they displayed severe defects in adjusting moving direction after gentle touch. That tutl mutant larvae were capable of making large-angle turns during the course of free movements, argues against a general defect in the sensorimotor system. Consistent with this notion, we found that tutl mutant larvae displayed normal phototaxis behaviors. These results suggest strongly that mutations in the tutl gene specifically affect the circuits that modulate the changes in moving direction in response to gentle touch.
Our results from transgene rescue indicate that Tutl is required exclusively in post-mitotic neurons at larval stage after the completion of embryonic development, which is consistent with neuronal-specific expression pattern of endogenous Tutl. Restoration of tutl expression in Appl-positive neurons or cholinergic neurons also substantially rescued the navigational phenotype. Consistently, triple labeling highlighted a large population of cholinergic neurons positive for both Tutl and Appl in the nervous systems (data not shown). Appl-positive neurons are distributed broadly in the larval nervous system, including most of sensory neurons in the PNS and interneurons in the CNS. Mutations in the Appl gene caused mild defects in locomotor reactivity, suggesting a role for Appl-positive neurons in the control of fly locomotion. Similarly, the larval cholinergic system includes many sensory neurons (e.g. chordotonal and da neurons) and a large group of interneurons in the CNS[37, 38]. Blockage of synaptic transmission in all cholinergic neurons caused paralysis, while silencing communication between random cholinergic neurons caused several types of locomotor defects such as sluggish movement, failure in initiation or maintenance of locomotion, uncoordinated movement, and arrest of locomotion. Taken together, those studies suggest that Appl-positive cholinergic neurons may form a functional circuit consisting of sensory neurons in the PNS and interneurons in the CNS, which controls larval sensorimotor decision making.
Tutl may function in Appl-positive cholinergic neurons in both PNS and CNS for proper navigational pattern in response to gentle touch. Consistent with a role for Tutl in sensory neurons, previous studies showed that mutations in the tutl gene caused defects in dendritic patterning of class I, II, III and IV da neurons in the PNS[29, 40]. Two lines of evidence support that Tutl also plays a role in the CNS for adjusting moving direction after gentle touch. First, expression of tutl transgene in all peripheral sensory neurons was not sufficient for rescuing the navigational phenotype. And second, blockage of synaptic transmission in a small subset of tutl-positive neurons in the CNS significantly affected navigational pattern in response to gentle touch. These tutl-positive CNS neurons may function in the circuits that integrate and process information from tactile stimuli, thus allow animals to adjust their moving direction properly.
Tutl may play a role during the development of larval nervous system for hardwiring of neuronal circuits that are specifically involved in directional adjustment in response to gentle touch. Such a role for Tutl in circuit development is supported by several recent studies. For instance, our recent studies show that Tutl is involved in regulating axonal tiling and dendrite self-avoidance[28, 29], two important cellular mechanisms that pattern neuronal circuitry during development. It is also suggested that Tutl play a role in regulating axonal pathfinding at embryonic stage.
Alternatively or additionally, Tutl may also play a role in modulating the activity of the circuits for adjusting moving direction in response to gentle touch. In vitro analysis shows that Tutl can function as a homophilic cell adhesion molecule. Many homophilic cell adhesion molecules have been shown to mediate synaptic function[43, 44]. For instance, the well-known homophilic cell adhesion molecule Fasciclin II (FasII), and its mammalian homolog NCAM, have been implicated in regulating synaptic plasticity[45–47]. In this context, it is also worth noting that interfering with the function of Dasm1, the mouse homolog of Tutl, prevents synapse maturation in cultured hippocampal neurons. Further studies are needed to elucidate the exact action of Tutl in the development and/or function of the circuits that control navigational pattern in response to gentle touch.
Our study identifies Tutl and a small subset of CNS neurons in modulating directional change in response to gentle touch. The function of mammalian homologs of Tutl (i.e. Dasm1 in mice and IgSF9 in humans) is still unknown. Given high homology between Tutl and its mammalian homologs[22–24], it is possible that Dasm1/IgSF9 play a similar role in directional change after mechanical stimulation in mammals. The implication of Tutl and a small subset of CNS neurons in the control of directional change after gentle touch, presents an excellent starting point for further dissection of underlying molecular networks and neuronal circuitry.
Flies were reared in plastic vials containing standard fly food or in grape juice plates at 25°C with ~50% humidity. Grape juice plates were prepared by mixing 30 g agar, 30 g sugar and 354 ml grape juice in 1.2L ddH2O. Flies for behavioral tests were kept in incubators with 12h light/dark cycle.
The following fly stocks were obtained from the Bloomington Stock Center: Appl-GAL4 (BL#30546), Cha-GAL4 (BL#6798), OK371-GAL4(BL#26160), Ddc-GAL4(BL#7009), RN2-GAL4(BL#7472), G11-1-GAL4 (BL#7030), ftz.ng-GAL4(BL#8767), D42-GAL4(BL#8816), TrpA1-GAL4(BL#27593), 5-HTR1B-GAL4(BL#27637), C81GAL4(BL#3738), Pain-GAL4(BL#27894), GMR91F06(BL#47170), GMR60G12 (BL#45360), tubP-GAL80ts(BL#7017), UAS-mCD8-GFP (BL#5136), UAS-CD4-tdGFP(BL#35838), UAS-TeTxLC(BL#28838), pain 1 (BL#27895), pain 3 (BL#31432), tutl 01085 (BL#10979). tutl 23 , tutl-GAL4, and UAS-tutl, were generated in our previous studies[28, 29]. pBac[WH] [f03313] and pBac[WH]CG16857[f02225] were used to generate tutl Df , which removes tutl and CG16857 by using the FLP/FRT-based strategy.
For cell-type-specific transgene rescue, genetic crosses were performed to generate tutl 23 homozygous mutant larvae carrying UAS-tutl and GAL4 driver. Their navigational pattern was then compared to that in tutl 23 homozygous mutant larvae carrying only GAL4 driver. For temporal control of UAS-tutl expression in tutl mutants using the TARGET system, larvae were raised with 12 hr light/dark cycle and moved between 18°C and 29°C incubators to turn on or turn off tutl transgene expression in tutl 23 mutants. For circuit breaking analysis, flies carrying GAL4 drivers were crossed with UAS-shits flies, and were raised at 22°C. Larval behaviors at permissive temperature (i.e. 22°C) or restrictive temperature (i.e. 32°C) were examined in a transparent box with precise temperature control (Kooland incubator).
3rd-instar larvae were collected and gently washed in ddH2O before transferred to 60 mm petri dish containing 2.5% agar substrate. Larvae were allowed for 3-min free locomotion prior to tactile stimuli. Gentle touch was applied to anterior segments of a larva at 25°C (22°C or 29°C for circuit breaking analysis). Filaments used for applying different stimulus intensities (i.e. 1 mN, 3 mN, 7 mN, 10 mN) were calibrated similarly as described previously. Navigational pattern of each larva in response to tactile stimuli was tested four times during the course of forward movements. Larval navigational behaviors were recorded with a digital monochrome camera (LTC 0335, BOSCH), and analyzed using the MB-ruler software (MB-Software solutions).
Mechanical nociception assay was performed similarly as described previously[10, 16]. Briefly, 3rd-instar larvae were stimulated with a 50 mN calibrated Von Frey filament. Noxious mechanical stimuli were delivered by rapidly touching the larva with the fiber at abdominal segments (i.e. four to six). A positive escape response was scored if at least one 360° revolution around the anterior/ posterior axis occurred in response to the stimuli. Each larva was tested only once. For each genotype, three trials (20–30 larvae per trial) were performed.
A slightly modified version of the Darth Vader assay was used. Larvae were raised on grape juice plates with 1.25g/L β-carotene (Jamieson.). A 100 mm petri dish containing 2.5% agarose was divided into four quadrants, and two of which were covered by black paper (as shown in Figure 6A). The dish was illuminated from above with incandescent light (40W). All experiments were done at night in a dark room. After the release of larvae at the center of the plate, the number of larvae in each sector were counted at every 1-min interval for 10 minutes. A preference index (PI) was calculated as: PI = (number of larvae in two dark quadrants - number of larvae in two bright quadrants) / (number of larvae in two dark quadrants + number of larvae in two bright quadrants).
After 1-min adaptation time, free movements of 3rd-instar larvae on a 100 mm plate containing 2.5% agarose were recorded with a digital monochrome camera (LTC 0335, BOSCH) for 3 min at 25 images/sec, and analyzed with the Videotrack 3.1.1 software (ViewPoint, Life Sciences Inc.). Turnings are defined as >30° in directional change, followed by linear locomotion.
Larval CNS and/or body wall were dissected in phosphate buffer (pH 7.2), fixed in 3.2% paraformaldehyde for 50 min, washed three times with PB-TX (0.5% Triton-X 100 in 1x PBS), and incubated with primary antibody in 10% normal goat serum at 4°C for three hours. Primary antibodies used were: mouse monoclonal anti-GFP (1:500 dilution) (Invitrogen/Molecular Probes), chick anti-GFP (1:500 dilution) (Abcam), and rabbit anti-Tutl polyclonal antibody (1:60,000 dilution). Following secondary antibodies were used: Alexa-488 dye-conjugated anti-mouse antibody (1:500 dilution), Alexa-568 dye-conjugated anti-rabbit antibody (1:500 dilution), or Alexa-647 dye-conjugated anti-mouse antibody (1:500 dilution) (Invitrogen/Molecular Probes). Images were captured using an Olympus FV1000 Confocal LSM microscope.
For generating anti-Tutl antibody, PCR fragments encoding the extracellular region of Tutl was subcloned into the pIB/Fc expression vector for producing Tutl-Fc fusion protein in S2 cells. Tutl-Fc fusion protein was purified using Protein A-conjugated Sepharose column, and used to raise antibodies in rabbits by using standard methods. Specificity of anti-Tutl antibody was confirmed by immunostaining showing absence of tutl staining in tutl mutant larvae (data not shown).
Student’s t-test and/or ANOVA test were used for statistical analysis. A best-fit linear-regression analysis was used to determine the correlation between navigation decision and the intensity of stimuli. Statistical analysis was performed with Excel 2007 (Microsoft Corp) or GraphPad Prism 5.0 (GraphPad software).
We thank members of the Rao laboratory and Donald van Meyel laboratory for comments and discussions; the Bloomington Stock Center and the Exelixis collection at Harvard Medical School for piggyback insertion lines; Dr. Craig Montell for iav-GAL4; Dr. Yuh Nung Jan for SN(5–40)-GAL4, NompC-GAL4 and ppk1.9-GAL4; Dr. Greg Suh for UAS-shits. This work was supported by an operating grant (MOP-14688) from Canadian Institutes of Health Research, and an operating grant (Rgpin 341431–10) from Natural Science and Engineer Research Council of Canada awarded to Yong Rao.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.