Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms

Volume 16, Issue 6 December 2001

EDITORIAL Pebbles of Truth 515 Martin Zatz

FEATURE Review Clockless Yeast and the Gears of the Clock: How Do They Mesh? 516 Ruben Baler

ARTICLES Resetting of the Circadian Clock by Phytochromes 523 and Cryptochromes in Arabidopsis Marcelo J. Yanovsky, M. Agustina Mazzella, Garry C. Whitelam, and Jorge J. Casal Distinct Pharmacological Mechanisms Leading to c-fos 531 Gene Expression in the Fetal Suprachiasmatic Nucleus Lauren P. Shearman and David R. Weaver Daily Novel Wheel Running Reorganizes and 541 Splits Hamster Circadian Activity Rhythms Michael R. Gorman and Theresa M. Lee Temporal Reorganization of the Suprachiasmatic Nuclei 552 in Hamsters with Split Circadian Rhythms Michael R. Gorman, Steven M. Yellon, and Theresa M. Lee Light-Induced Resetting of the Circadian Pacemaker: 564 Quantitative Analysis of Transient versus Steady-State Phase Shifts Kazuto Watanabe, Tom Deboer, and Johanna H. Meijer Temperature Cycles Induce a Bimodal Activity Pattern in Ruin Lizards: 574 Masking or Clock-Controlled Event? A Seasonal Problem Augusto Foà and Cristiano Bertolucci

LETTER Persistence of Masking Responses to Light in Mice Lacking Rods and Cones 585 N. Mrosovsky, Robert J. Lucas, and Russell G. Foster

MEETING Eighth Meeting of the Society for Research on Biological Rythms 588

Index 589 Journal of Biological Rhythms

Official Publication of the Society for Research on Biological Rhythms

EDITOR-IN-CHIEF Martin Zatz

FEATURES EDITORS ASSOCIATE EDITORS Larry Morin Michael Hastings SUNY, Stony Brook University of Cambridge Anna Wirz-Justice Ken-Ichi Honma University of Basel Hokkaido Univ School Medicine Michael Young Rockefeller University EDITORIAL BOARD

Josephine Arendt Terry Page University of Surrey Vanderbilt University Charles A. Czeisler Ueli Schibler Harvard Medical School University of Geneva Serge Daan William J. Schwartz University of Groningen U Mass Medical School

ADVISORY BOARD

Deborah Bell-Pedersen Jennifer J. Loros Texas A&M University Dartmouth Medical School Gene D. Block Robert Y. Moore University of Virginia University of Pittsburgh Vincent M. Cassone Steven M. Reppert Texas A&M University Harvard Medical School Jay C. Dunlap Till Roenneberg Dartmouth Medical School University of Munich Russell G. Foster Mark D. Rollag Imperial College of Science Uniformed Services University Albert Goldbeter Benjamin Rusak University of Brussels Dalhousie University Carla B. Green Rae Silver University of Virginia Columbia University Paul E. Hardin Kathleen King Siwicki University of Houston Swarthmore College William J. M. Hrushesky Fred W. Turek Albany VAMedical Center Northwestern University Helena Illnerova David R. Weaver Czech Academy of Sciences Harvard Medical School Carl Johnson Irving Zucker Vanderbilt University UC, Berkeley Steve Kay The Scripps Research Institute Journal of Biological Rhythms

Official Publication of the Society for Research on Biological Rhythms

The JOURNAL OF BIOLOGICAL RHYTHMS (JBR) publishes original, full-length reports in English of empirical inves- tigations into all aspects of biological rhythmicity.Of particular interest are submissions that focus on rhythms related to the major environmental cycles, including daily (circadian) rhythms, tidal rhythms, annual rhythms (including photoperiodism), and biological rhythms that interact with those rhythms influenced by the environment. Studies using genetic, biochemical, physiological, behavioral, and modeling approaches to understand the nature, mechanisms, and functions of biological rhythms in all species are welcome. A major objective of JBR is to serve as a vehicle for transmit- ting information about biological rhythms in plants and animals, as well as those studying human rhythms in both the clinical and real world setting. Preliminary or incomplete studies will not be considered. Research reported in the journal must meet the highest standards of experimental design and data analysis. Opinion papers and reviews of significant, timely issues will also be considered. JOURNAL OF BIOLOGICAL RHYTHMS (ISSN 0748-7304) is published bimonthly in February, April, June, August, October, and December by Sage Science Press, an imprint of Sage Publications, 2455 Teller Road, Thousand Oaks, CA 91320; telephone (800) 818-SAGE(7243) and (805) 499-9774; fax/order line (805) 375-1700; e-mail [email protected]; http://www.sagepub. com. Copyright © 2001 by Sage Publications. All rights reserved. No portion of the contents may be reproduced in any form without written permission of the publisher. Disclaimer: The authors, editors, and publisher will not accept any legal responsibility for any errors or omissions that may be made in this publication. The publisher makes no warranty, express or implied, with respect to the material contained herein. Subscriptions: Annual subscription rates for institutions and individuals are based on the current frequency.Prices quoted are in U.S. dollars and are subject to change without notice. Canadian subscribers add 7% GST (and HST as appropriate). Outside U.S. subscription rates include shipping via air-speeded delivery. Institutions: $550 (within the U.S.) / $574 (outside the U.S.) / single issue: $105 (worldwide). Individuals: $130 (within the U.S.) / $154 (outside the U.S.) / single issue: $33 (worldwide). Orders by MasterCard or Visa can be placed by phone at (805) 499-9774; fax/order line (805) 375-1700. Payment must be made in U.S. dollars. Indexed in Academic Search, Agricola, Biological Abstracts, Biosciences Citation Index, BIOSIS Previews, Chemical Ab- ® stracts, Comprehensive MEDLINE with FullTEXT, Corporate ResourceNET, Current Citations Express, Current Contents : Life Sciences, Elsevier BIOBASE/Current Awareness in Biological Sciences, Ergonomics Abstracts, Health Source Plus, Index Medicus, INSPEC, International Aerospace Abstracts, ISI Basic Science Index, MasterFILE FullTEXT, MEDLINE, Neurosciences Abstracts, Periodical Abstracts, Psychological Abstracts, PsycINFO, PsycLIT, Science ® Citation Index , Science Citation Index Expanded, and TOPICsearch. Back issues: Information about availability and prices of back issues may be obtained from the publisher’s order department (address below). Single-issue orders for 5 or more copies will receive a special adoption discount. Contact the order department for details. Inquiries: Address all correspondence and permissions requests to Sage Publications, 2455 TellerRoad, Thousand Oaks, California 91320; e-mail [email protected]; http://www.sagepub.com. Authorization to photocopy material for internal or personal use under circumstances not falling within the fair use pro- visions of the Copyright Act is granted by Sage Publications to libraries and other users registered with the Copyright Clearance Center Transactional Reporting Service, provided that the fee of $.50 per copy plus $.10 per copy page is paid directly to the CCC, 21 Congress Street, Salem, MA 01970. The identification code for JOURNAL OF BIOLOGICAL RHYTHMS is 0748-7304/2001/$.50 +.10. Advertising: Current rates and specifications may be obtained by writing to the Advertising Manager at the Thousand Oaks office (address above). Claims: Claims for undelivered copies must be made no later than 6 months following month of publication. The publisher will supply missing copies when losses have been sustained in transit and when the reserve stock will permit. Changes of address: Please inform the publisher at least 6 weeks prior to move. Enclose present mailing label with change of address. Effective with the 1986 volume, this journal is printed on acid-free paper. Cover photo: Plamen D. Penev, Northwestern University For Sage Science Press: Ben Sztajnkrycer, Production Editor Eric Law, Copy Editor Paul Doebler, Designer Katinka Baltazar, Designer

Pebbles of Truth

Science has gotten such a good reputation for How true is it? Is it the whole truth? Is it entirely true or answering questions that just about everybody claims just partially true? Is it strictly true, necessarily true, the adjective “scientific” for what they say. An impec- generally true, often true, true under certain circum- cable scientific approach is, however, useless for most stances? Is it conditionally true, likely true, possibly of life’s important questions like “Wherein lies the true? We thereby bypass some of the deeper, more Good?” “Why me?” “Hold ‘em or fold ‘em?” “Shall intractable, issues of truth and causality and compen- we send troops?” Scientists are no better than any- sate with the benefits of open-mindedness, disinterest body else at making most personal and political deci- (not fooling ourselves), and small hard truths. sions and can be a real pain when it comes to provid- The ability to declare a question presently unan- ing clear answers to simple questions—especially if swerable, no matter how important, and to accept their defensive scientific cloaking device is turned on. interim and partial truths without commitment, is The way we scientists deal with questions and perhaps the greatest strength of science and a hall- answers often frustrates the people who consult us mark of its different worldview. We have had the priv- and support us. ilege, so far, of choosing our questions. Although There are two reasons for this. First, we frequently everyone wants answers to big questions, we usually don’t accept the question. Many of the biggest, most prefer to settle for results that clearly answer a small urgent, or most important questions are concerned question over results that merely bear on a big ques- with what should be, and science addresses only what tion. What a peculiar way of getting answers normal is. As Richard Feynman explained, “The question: science has: nibbling at a problem, not trying to swal- ‘Should I do this?’—whether you want something to low it whole. Yet, by an invisible hand, it seems to end happen or not—must lie outside of science.” Second, up giving us a better grasp of truth and causality after our tendency, even when the questions posed are sci- all. entific, is to refuse to answer until we’re good and ready. Isaac Newton said, “I do not know what I may Unlike policy makers or executives or police officers appear to the world; but to myself I seem to have been or editors, we need not (and often refuse to) come to a only like a boy playing on the seashore, and diverting fast conclusion. We claim the privilege of uncertainty myself in now and then finding a smoother pebble or a long after others have made up their minds. Accord- prettier shell than ordinary, whilst the great ocean of ing to Feynman, a scientist is never certain. When a truth lay all undiscovered before me.” This pretty statement is made, the question is not whether it is scene embodies several of the ideals of science: mod- true or false but rather how likely it is to be true or esty, curiosity, and wonder. We have found treasures false. There is no certainty; even our best answers are on the beach: shiny shells and pebbles—what stars at least a little provisional. This chronic hedging of and people and firefly flashes are made of. The shells ours is a remarkable trait and a precious privilege— and pebbles add up and tell us about the sea. Each of rare across history and geography—the freedom to us gets to place some on the pile. doubt and to declare “I don’t know” publicly. Ever doubtful, wary of conclusions, even wary of Martin Zatz facts, we parse the truth of statements ever so fine: Editor

JOURNAL OF BIOLOGICAL RHYTHMS, Vol. 16 No. 6, December 2001 515 © 2001 Sage Publications 515 BalerJOURNAL / THE OF YEAST BIOLOGICAL HYBRID RHYTHMSSYSTEM / December 2001 REVIEW Clockless Yeast and the Gears of the Clock: How Do They Mesh?

Ruben Baler1 Unit on Temporal Gene Expression, Laboratory of Cellular and Molecular Regulation, National Institute of Mental Health, National Institutes of Health, Bethesda, MD 20892, USA

Abstract In spite of its apparent weakness as a clock model, the budding yeast has spawned a technique that has revolutionized our ability to study specific protein-protein interactions like those at the core of the molecular timekeeping mechanisms. Here, the author will summarize the evolution, power, and limitations of this technique and highlight its potential and actual contributions to the field of chronobiology.

Key words yeast two hybrid, circadian interactome, clock genes

Specific interactions between proteins form the networks (Hartwell et al., 1999), and the assignment of basis of almost every aspect in a cell’s life. These inter- likely roles to functionally uncharacterized proteins actions create a dynamic and tightly regulated com- (Brent and Finley, 1997; Uetz et al., 2000). munications network and weave a complex connec- tivity map from which cell phenotypes emerge. It is hardly surprising that the identification, characteriza- THE YTH SYSTEM: tion, modification, and exploitation of all these spe- A CATALYST OF DISCOVERY AND cific contacts would constitute such major focal points SELF-IMPROVEMENT in modern biological research. The study of protein-protein interactions was YTH relies on the modular nature of transcription largely confined, until 1989, to biochemical techniques activating factors (Sadowski et al., 1988), which tend such as crosslinking, co-precipitation, and fraction- to bind to the control region of a gene with one domain ation. However, these techniques are limited by the while activating the transcriptional machinery with fact that the interaction of interest is either initiated or another (Fig. 1A). The strategy that makes YTH possi- measured outside of a living cell. This situation was ble is based on the separation of the DNA binding and dramatically changed with the introduction of a the activation domains of a known transcription fac- genetics-based strategy (Fields and Song, 1989), tor, which are then fused to putative binding partners referred to as the yeast two-hybrid system (YTH), that (Fig. 1B). The two resulting fusion proteins are not made it possible, for the first time, to investigate pro- expected to activate transcription on their own. How- tein-protein interactions inside a living cell. ever, after their simultaneous expression into a suit- The continued application of the YTH and related able yeast reporter strain, a successful interaction will protocols led to the description of thousands of bind- reestablish the physical link between the DNA binding partners. This new information contributed greatly ing and activation domains (Fig. 1C). This event, even to the identification of novel genes (Boulton et al., if transient or of low affinity, can be recorded if the 2001), the dissection of complex signal transduction yeast strain carries an easily detectable (e.g., colori-

1. To whom all correspondence should be addressed: Building 36, Room 2A-09, National Institutes of Health, Bethesda, MD 20892, USA; e-mail: [email protected]. JOURNAL OF BIOLOGICAL RHYTHMS, Vol. 16 No. 6, December 2001 516-522 © 2001 Sage Publications 516 Baler / THE YEAST HYBRID SYSTEM 517

Reporter gene A

p b B

b p C

Figure 1. The yeast two-hybrid system. Transcription factors are often composed of separate domains (A) for DNA recognition (grabbing hand) and transcription activation (pointing finger). This characteristic allowed the physical separation of these domains onto which unknown protein moieties can be grafted. If these moieties, usually referred to as “bait” (b) and “prey” (p), are indifferent to each other, the reporter gene will remain silent (B). If, on the other hand, a specific (or nonspecific) interaction between them occurs, the event will be detected through reporter gene activation (C). This simple configuration represents the basic idea behind all the yeast hybrid protocols that allow detection of protein interactions under many different conditions. metric, such as lacZ) or selectable (e.g., prototrophic, These clear disadvantages notwithstanding, the such as his3) reporter gene. Typically, one of the bind- YTH strategy enjoys a high level of success and accep- ing partners is a known entity and referred to as the tance, possibly due to its remarkable ability to evolve “bait” (b) because it is used to “catch” an unknown rapidly in response to the frequently encountered tech- “prey” (p), via more or less stable interactions. nical obstacles (Finley and Brent, 1996; Brachmann The remarkably high sensitivity of YTH screens, and Boeke, 1997; Fashena et al., 2000). For example, however, comes at the expense of a relatively high as already mentioned, a significant fraction of YTH- number of false positives, a persistent concern for derived findings represent biologically irrelevant users of this technique. Another disturbing fact relates interactions. Commonly isolated interactors, such as to the predicted interactions that are not detected by heat shock, ribosomal, or proteasome-related pro- this technique, the so-called false negatives. Indeed, teins, can interact with approximately one-third of less than 15% of previously established binding part- randomly chosen baits. ners in yeast have been rediscovered by supposedly Hence, much of the initial efforts to improve the comprehensive screens (Hazbun and Fields, 2001). As YTH protocol focused on assessing and increasing the a consequence, researchers are constantly looking for specificity of the observed interaction. An early step the perfect balance between increased stringency and forward in this direction was the swapping of the bait maximal coverage. At the present time, YTH tech- and prey moieties between the DNA binding and niques can suggest potential binding partners, but transactivation domains of the reporting factor (Du independent confirmation by additional methods is et al., 1996). This domain swap was initially designed required to demonstrate that the interaction is specific to bypass the false positives that result from the fusion and physiologically relevant. of transcriptional activators onto the DNA binding domain, but it quickly became a standard specificity 518 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001 control. Later, the use of carefully designed double screens that target quaternary complexes is just bait systems significantly increased specificity through around the corner (Pause et al., 1999). the introduction of a second bait vector carrying a dif- Finally, YH screens have not remained confined to ferent DNA binding domain fused to a nonspecific the realm of interactions between proteins. In the interactor (Serebriiskii et al., 1999). The extent to so-called RNA three-hybrid system (Putz et al., 1996), which this alternative pathway also becomes acti- interactions between proteins and RNAmolecules can vated by the same prey provides a good indication of also be investigated. Meanwhile, a one-hybrid proto- the level of nonspecific interaction. col (Inouye et al., 1994) has been developed to identify The most significant advance in the specificity DNA binding proteins using specific cis-acting ele- front, however, has been the inclusion of multiple and ments as interaction targets. On the other hand, independent reporter systems. Thus, true binding reverse two- and one-hybrid systems (Vidal et al., partners have to activate several different genes before 1996) offer the possibility to screen for mutations or being considered for further analysis. This strategy small ligands that disrupt a particular interaction, dramatically increased the stringency of the screen adding a new dimension to the analysis of structure- and, as a result, the capacity of YTH to expose candi- function relationships and the high-throughput dates worth pursuing. search for compounds with potentially useful phar- Another limitation of the original YTH method lies macological properties. in the nature of the exclusively transcription-based readout. This feature excludes a significant number of potential interactors that are either membrane bound, BEYOND YTH unable to access the nuclear compartment in an active form, or independently active on transcription. Over Most classic biochemical techniques have been time, several variations were developed, either in used to corroborate an interaction first suggested by a mammalian or yeast cells, that address these problems yeast hybrid search. In the context of new technolo- by probing interactions at their natural site in the cell gies, however, a group of methods referred to as reso- (Johnsson and Varshavsky, 1994), in the cytosol (Rossi nance energy transfer (RET; Li et al., 2001) is emerging et al., 2000), or in association with the plasma mem- as one of the most powerful tools for confirming a sus- brane (Johnsson and Varshavsky,1994; Aronheim et al., pected interaction in vivo. RET is based on the interac- 1997; Isakoff et al., 1998; Medici et al., 1997). tion of two energetically linked luminescent (LRET), In addition, a given interaction might occur only as or fluorescent (FRET) probes (fused to interacting pro- a consequence of unique posttranslational modifica- teins). When these moieties are brought into proxim- tions. Absence of a differentially expressed modifying ity, energy resonance causes either quenching or exci- system would prevent the detection of such an inter- tation and the concomitant change in emission spectra action, a limitation of significant consequences when that can be monitored (see De Angelis, 1999, for a com- we consider a recent report regarding the dispropor- prehensive review on the major applications of FRET). tionate representation of proteins involved in signal The demonstration that the cyanobacteria circadian transduction among different species (Chervitz et al., protein KaiB can form homodimers, using biolumi- 1998). If the missing link is known, it is possible to nescence RET (BRET; Xu et al., 1999), was among the co-express it during the interaction screen (Kochan earliest successful applications of this approach. The et al., 2000); this approach appears to be particularly study of conformational changes in proteins (Wang useful when tyrosine kinases are involved (Chervitz et al., 2001) or the kinetics of proteasome targeted deg- et al., 1998). Similarly, some interactions that involve radation (Tung et al., 2000) are just two additional ex- ternary complexes are unstable if one of the compo- amples of the exciting possibilities that these techniques nents is missing. After the arrival of the three-hybrid offer for the real-time monitoring of interactive events. system (Zhang and Lautar, 1996; Kochan et al., 2000; The arrival of the proteomics age has brought along Brachmann and Boeke, 1997), formation of three-com- the intriguing possibility of a genome-wide applica- ponent complexes has been achieved repeatedly, thus tion of the YTH and derived approaches (Uetz et al., allowing us to address the issue of how larger protein 2000; Legrain and Selig, 2000; Ito et al., 2001). Several structures (or even scaffolds) are put together. Predict- consortia are combining the power of robotics, multi- ably, the development of reliable yeast hybrid (YH) cloning, and high throughput screening to lay out Baler / THE YEAST HYBRID SYSTEM 519

(more or less) comprehensive protein interaction mation in a cell is handled through “traditional” maps. It is hoped that the painstaking construction of channels, distinguishing features within individual these “interactomes” will usher in a new era of pathways are conferred by a rich tapestry of highly proteomic databases. By “mining” these resources, we specific protein-protein contacts. A real understand- might be able to generate sound hypotheses regarding ing of any particular pathway (or circuit) will depend the possible role of uncharacterized proteins, once on a comprehensive description of the sequence, loca- their likely location within such a map is identified. It tion, and timing of every specific interaction. is important to bear in mind, however, that two early The second point, of course, is that the biological attempts at defining a comprehensive catalogue of all clock (see Lowrey and Takahashi, 2000; Chang and possible interactions in the yeast system yielded dis- Reppert, 2001, for recent reviews) can be construed as appointingly few overlapping hits (Uetz et al., 2000; the signal transduction pathway that keeps track of Ito et al., 2001). This result indicates that current tech- internal time. As such, the underlying transcriptional/ niques have not yet reached the level of saturation translational mechanism that supports circadian required to detect every possible interaction (Hazbun rhythmicity can be fitted onto a similar model, once we and Fields, 2001) and that the type of data obtained recognize that it represents a special case in which, in depends strongly on the details of a particular screen addition to outside cues, key variables along the (path) design. way can feed into itself to generate a self-sustaining loop (Shearman et al., 2000). Consequently, the explosive progress in molecular DISSECTION OF GENERAL AND circadian biology in recent years can be largely accred- CIRCADIAN INTERACTION NETWORKS ited to the discovery of clock-relevant interactions (many through YTH screens) and the tacit effort to Most signal transduction pathways use a standard describe a “circadian interactome.” approach to deliver information into the nucleus. Consider the heterodimerization of PER (Period) Although different pathways display a wide range of and TIM (Timeless) in Drosophila (Sehgal et al., 1994; variations in the details, they operate in three concep- Vosshall et al., 1994, Gekakis et al., 1995); the light- tually distinct domains designed to control the all- dependent sequestration of dTIM by dCRY (Ceriani important crossing of the nuclear/cytoplasmic barrier et al., 1999); the combinatorial capacity of the three by factors that affect transcription. In the first domain mammalian PER proteins to interact among them- (domain I = input), the primary information, from selves (Zylka et al., 1998) and with the CRYs channel traffic, receptor activation, intracellular vari- (Cryptochromes) (Kume et al., 1999); the interaction, ables, and so on, is received by the proper sensor and discovered through a YTH screen, between the appar- conveyed through second messengers onto regula- ently clock gene–specific transcription factors BMAL1 tory proteins that control the fate of a nucleus-bound and CLOCK (Gekakis et al., 1998) and their trans- device (domain II = processing). The translocation, criptional repression by a PER/CRY complex (Kume final destination, activity, and partners of this nuclear et al., 1999). Each example turned out to be a pivotal device(s) determine the nature of the genetic response association for the establishment of a circadian clock. (domain III = output). Predictably, a dynamic balance More recently, we learned of the ability of Casein between the rescue and degradation of these nucleus- Kinase Iε/δ to recognize PER 1 and 2 as phosphoryl- bound proteins, and their regulators, is a recurring ation substrates, thereby regulating their susceptibil- theme in a large number of signaling pathways. This ity to proteosomal degradation (Suri et al., 2000; control mechanism is particularly common in the sec- Camacho et al., 2001; Vielhaber et al., 2000). This find- ond domain of this simplified model, where the stabil- ing retraced the previously reported light- induced ity and fate of key factors is determined, to a large tyrosine phosphorylation of dTIM by DOUBLETIME, extent, by complex phosphorylation/dephosphorylation an event that targets TIM for ubiquitination and cascades and selective interactions. proteosomal degradation (Naidoo et al., 1999). The levels of organization defined above can be Similarly, the key scaffolding role of the easily recognized in many well-characterized path- nonrhythmic protein WC-2 is of particular interest in ways, a few of which are outlined in Table 1. The first this context because it seems to control the formation point of this exercise is to emphasize that while infor- of a ternary complex as part of the timing mechanism 520 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Table 1. Many different signal transduction pathways utilize common themes of protein interactions to process information and reprogram gene expression. The mechanisms involved operate in three domains to ultimately control the activity of nucleus-bound factors. Circadian clocks are no exception.

Pathway Domain I (Input) Domain II (Processing) Domain III (Output) Reference

Heat shock (HS)/ Abnormal protein HS factor trimerization HS gene activation, protection Baler et al., 1996; Soncin stress response production and nuclear translocation. from stress et al., 2000 Modulated by calcium- activated kinases Nuclear hormone Ligand-mediated Receptor release from HSP90 Hormone-responsive gene Miyata et al., 1997 receptors activation and nuclear translocation. activation, regulation of Modulated by Tyr and reproductive (and other) CKII kinases tissues Immune, anti Extracellular stimulation Inactive NFκB subfamily Activation of anti-apoptotic Zandi et al., 1997 cell-death. (NFκB) through IL-1, TNF, members released from and immune response genes and other cytokines IκB cytoplasmic inhibitors, such as IL-2, TNFα, GM-CSF nuclear translocation. Modulated by IKK kinases Tumor suppressor Cell cycle, DNA Reduced affinity for Mdm2 Regulation of genes involved Maya et al., 2001 protein (p53) damage, apoptosis (a p53 E3 ligase), nuclear in cell cycle arrest, apoptosis, translocation. Modulated DNA damage repair by ATM kinase Circadian clock Photic input, phase Per/Tim, Per/Cry Activation/suppression of Shearman et al., 2000; of the cellular clock heterodimerization, nuclear clock and clock-controlled Camacho et al., 2001 (i.e., relative level translocation. Modulated genes of state variables) by CKIδ/ε

in Neurospora (Collett et al., 2001). This hypothesis has nents should be a carefully monitored parameter, an now been confirmed by independent biochemical assumption amply supported by the arrhythmicity of methods (Denault et al., 2001). DOUBLETIME mutant flies in which the hypo- By the same token, it can be argued that the pieces phosphorylation of TIM (Naidoo et al., 1999) and PER still missing in the circadian puzzle represent unique (Price et al., 1998) results in their increased stability. opportunities to find novel interactions that might Similarly, the tightly regulated nucleocytoplasmic affect the phase and output of the circadian clock. It shuttling of proteins like PER, TIM, and CRY is likely seems reasonable to predict that these unknown inter- to involve a number of specific (and not so specific) actions are still many and likely to play important auxiliary factors that need to be identified. A similar roles in the compartments described above. case can be made regarding the pathways and factors First, we do not yet fully understand the nature of involved in controlling the intranuclear trafficking the interactions underlying circadian photoreceptive and subnuclear localization of circadian transcription signaling. ZEITLUPE (ZTL), for example, is an factors. There is a growing number of examples in Arabidopsis protein that exerts profound effects on which these processes play an important regulatory clock-controlled processes consistent with a role in role in determining the outcome of a specific trans- light input to the clock (Somers et al., 2000), possibly duction program (Stein et al., 2000). Similar mecha- just downstream of converging photoreceptive path- nisms are likely to play a role in circadian molecular ways (Millar, 2000). The role of ZTL within or around routines. the clock is far from clear; however, the presence of Third, we are quite in the dark regarding the spe- F-box and PAS domains suggests likely classes of ZTL cific interactions that must occur around circadian binding partners belonging to the degradation and DNA consensus elements, such as E-boxes (or transcription control pathways, respectively. GATA-boxes in Neurospora), to induce selective tran- Second, we still have a far from complete descrip- scription of clock-controlled genes. As a matter of fact, tion of the components responsible for the specific we do not even know (although we might suspect this modification, degradation, and translocation events to be the case) whether clock components do contrib- that control the level and subcellular localization of all ute to other pathways, as part of different protein con- known clock proteins in different species. One would stellations, in either oscillatory or linearly responsive assume that the turnover of specific clock compo- networks. The one-hybrid system and its variations Baler / THE YEAST HYBRID SYSTEM 521

(Wilson et al., 1991) might prove useful in this context, Denault D, Loros J, and Dunlap J (2001) WC-2 mediates since they provide a means for the identification of WC-1-FRQ interaction within the PAS protein-linked circadian feedback loop of Neurospora. EMBO J 20:109-117. transacting factors and cis-acting regulatory element Du W, Vidal M, Xie J, and Dyson N (1996) RBF, a novel pairs. RB-related gene that regulates E2F activity and interacts It is interesting to consider that the apparent with cyclin E in Drosophila. Genes Dev 10:1206-1218. absence of a typical circadian clock mechanism in the Fashena S, Serebriiskii I, and Golemis E(2000) The contin - yeast might have resulted in a particularly well-suited ued evolution of two-hybrid screening approaches in environment for the search of clock-associated mole- yeast: How to outwit different preys with different baits. Gene 250:1-14. cules through the various YH systems described. We Fields S and Song O (1989) A novel genetic system to detect can expect that the wider and deeper use of these tech- protein-protein interactions. Nature 340:245-246. niques will illuminate new areas of the molecular Finley R and Brent R. (1996) Two-hybrid analysis of genetic chronobiology field, as the mapping of all possible regulatory networks. Proc Natl Acad SciUSA91:12980- interactions yields ever-denser networks of interact- 12984. ing proteins. I like to think that this illumination has in Gekakis N, Saez L, Delahaye-Brown A, Myers M, Sehgal A, Young M, and Weitz C (1995) Isolation of timeless by PER store for us many advances, few delays, and one or protein interaction: Defective interaction between time- two paradigm shifts. less protein and long-period mutant PERL. Science 270:811-815. Gekakis N, Staknis D, Nguyen H, Davis F, Wilsbacher L, REFERENCES King D, Takahashi J, and Weitz C (1998) Role of the CLOCK protein in the mammalian circadian mechanism. Aronheim A, Zandi E, Hennemann H, Elledge S, and Science 280:1564-1569. Karin M (1997) Isolation of an AP-1 repressor by a novel Hartwell L, Hopfield J, Leibler S, and Murray A (1999) From method for detecting protein-protein interactions. Mol molecular to modular cell biology. Nature 402:C47-C52. Cell Biol 17:3094-3102. Hazbun T and Fields S (2001) Networking proteins in yeast. Baler R, Zou J, and Voellmy R (1996) Evidence for a role of Proc Natl Acad Sci U S A 98:4277-4278. Hsp70 in the regulation of the heat shock response in Inouye C, Remondelli P,Karin M, and Elledge S (1994) Isola- mammalian cells. Cell Stress Chaperones 1:33-39. tion of a cDNA encoding a metal response element bind- Boulton S, Vincent S, and Vidal M (2001) Use of pro- ing protein using a novel expression cloning procedure: tein-interaction maps to formulate biological questions. The one hybrid system. DNA Cell Biol 13:731-742. Curr Opin Chem Biol 5:57-62. Isakoff S, Cardozo T, Andreev J, Li Z, Ferguson K, Abagyan Brachmann R and Boeke J (1997) Tag games in yeast: The R, Lemmon M, Aronheim A, and Skolnik E(1998) Identi- two-hybrid system and beyond. Curr Opin Biotechnol fication and analysis of PH domain-containing targets of 8:561-568. phosphatidylinositol 3-kinase using a novel in vivo assay Brent R and Finley R (1997) Understanding gene and allele in yeast. EMBO J 17:5374-5387. function with two-hybrid methods. Annu Rev Genet Ito T, Chiba T, Ozawa R, Yoshida M, Hattori M, and Sakaki Y 31:663-704. (2001) A comprehensive two-hybrid analysis to explore Camacho F, Cilio M, Guo Y, Virshup DM, Patel K, Khorkova the yeast protein interactome. Proc Natl Acad SciUSA O, Styren S, Morse B, Yao Z, and Keesler G (2001) Human 98:4569-4574. casein kinase I δ phosphorylation of human circadian Johnsson N and Varshavsky A (1994) Split ubiquitin as a clock proteins period 1 and 2. FEBS Lett 489:159-165. sensor of protein interactions in vivo. Proc Natl Acad Sci Ceriani M, Darlington T, Staknis D, Mas P, Petti A, Weitz C, U S A 91:10340-10344. and Kay S (1999) Light-dependent sequestration of Kochan J, Volpers C, and Osborne M (2000) The yeast tribid TIMELESS by CRYPTOCHROME. Science 285:553-556. system: cDNA expression cloning of protein interactions Chang D and Reppert S (2001) The circadian clocks of mice dependent on posttranslational modifications. Methods and men. Neuron 29:555-558. Enzymol 328:111-127. Chervitz S, Aravind L, Sherlock G, Ball C, Koonin E, Kume K, Zylka M, Sriram S, Shearman L, Weaver D, Jin X, Dwight S, Harris M, Dolinski K, Mohr S, Smith T, Weng S, Maywood E, Hastings M, and Reppert S (1999) mCRY1 Cherry J, and Botstein D (1998) Comparison of the com- and mCRY2 are essential components of the negative plete protein sets of worm and yeast: Orthology and limb of the circadian clock feedback loop. Cell 98:193-205. divergence. Science 282:2022-2028. Legrain P and Selig L (2000) Genome-wide protein interac- Collett M, Dunlap J, and Loros J (2001) Circadian clock-spe- tion maps using two-hybrid systems. FEBS Lett cific roles for the light response protein WHITE 480:32-36. COLLAR-2. Mol Cell Biol 21:2619-2628. Li H, Ng E, Lee S, Kotaka M, Tsui S, Lee C, Fung K, and De Angelis DA (1999) Why FRET over genomics? Physiol Waye M (2001) Protein-protein interaction of FHL3 with Genom 1:93-99. FHL2 and visualization of their interaction by green fluo- 522 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

rescent proteins (GFP) two-fusion fluorescence reso- Somers D, Schultz T, Milnamow M, and Kay S (2000) nance energy transfer (FRET). J Cell Biochem 80:293-303. ZEITLUPE, a novel clock associated PAS protein from Lowrey P and Takahashi J (2000) Genetics of the mammalian Arabidopsis. Cell 101:319-329. circadian system: Photic entrainment, circadian pace- Soncin F, Asea A, Zhang X, Stevenson M, and Calderwood S maker mechanisms, and posttranslational regulation. (2000) Role of calcium activated kinases and phos- Annu Rev Genet 34:533-562. phatases in heat shock factor-1 activation. Int J Mol Med Maya R, Balass M, Kim S, Shkedy D, Leal J, Shifman O, Moas 6:705-710. M, Buschmann T, Ronai Z, Shiloh Y, Kastan M, Katzir E, Stein G, van Wijnen A, Stein J, Lian J, Montecino M, Choi J, and Oren M (2001) ATM-dependent phosphorylation of Zaidi K, and Javed (2000) Intranuclear trafficking of tran- Mdm2 on serine 395: Role in p53 activation by DNAdam- scription factors: Implications for biological control. age. Genes Dev 15:1067-1077. J Cell Sci 113:2527-2533. Medici R, Bianchi E, Di Segni G, and Tocchini-Valentini G Suri V, Hall J, and Rosbash M (2000) Two novel doubletime (1997) Efficient signal transduction by a chimeric mutants alter circadian properties and eliminate the yeast-mammalian G protein alpha subunit Gpa1-Gs delay between RNA and protein in Drosophila. J Neurosci alpha covalently fused to the yeast receptor Ste2. EMBO J 20:7547-7555. 16:7241-7249. Tung C, Mahmood U, Bredow S, and Weissleder R (2000) In Millar A(2000) Clock proteins: Turnedover after hours. Curr vivo imaging of proteolytic enzyme activity using a Biol 10:R529-R531. novel molecular reporter. Cancer Res 60:4953-4958. Miyata Y, Chambraud B, Radanyi C, Leclerc J, Lebeau M, Uetz P,Giot L, Cagney G, Mansfield TA, Judson R, Knight JR, Renoir J, Shirai R, Catelli M, Yahara I, and Baulieu E Lockshon D, Narayan V, Srinivasan M, Pochart P, (1997) Phosphorylation of the immunosuppressant Qureshi-Emili A, and Li Y (2000) Acomprehensive analy- FK506-binding protein FKBP52 by casein kinase II: Regu- sis of protein-protein interactions in Saccharomyces lation of HSP90-binding activity of FKBP52. Proc Natl cerevisiae. Nature 403:623-627. Acad Sci U S A 94:14500-14505. Vidal M, Brachmann R, Fattaey A, Harlow E, and Boeke J Naidoo N, Song W, Hunter-Ensor M, and Sehgal A (1999) A (1996) Reverse two-hybrid and one-hybrid systems to role for the proteasome in the light response of the time- detect dissociation of protein-protein and DNA-protein less clock protein. Science 285:1737-1741. interactions. Proc Natl Acad Sci U S A 93:10315-10320. Pause A, Peterson B, Schaffar G, Stearman R, and Klausner R Vielhaber E, Eide E, Rivers A, Gao Z, and Virshup D (2000) (1999) Studying interactions of four proteins in the yeast Nuclear entry of the circadian regulator mPER1 is con- two-hybrid system: Structural resemblance of the trolled by mammalian casein kinase I epsilon. Mol Cell pVHL/elongin BC/hCUL-2 complex with the ubiquitin Biol 20:4888-4899. ligase complex SKP1/cullin/F-box protein. Proc Natl Vosshall L, Price J, Sehgal A, Saez L, and Young M (1994) Acad Sci U S A 96:9533-9538. Block in nuclear localization of period protein by a sec- Price J, Blau J, Rothenfluh A, Abodeely M, Kloss B, and ond clock mutation, timeless. Science 263:1606-1609. Young M (1998) Double-time is a novel Drosophila clock gene that regulates PERIOD protein accumulation. Cell Wang Y, Wang G, O’Kane D, and Szalay A (2001) A study of 94:83-95. protein-protein interactions in living cells using lumines- Putz U, Skehel P, and Kuhl D (1996) A tri-hybrid system for cence resonance energy transfer (LRET) from Renilla the analysis and detection of RNA–protein interactions. luciferase to Aequorea GFP.Mol Gen Genet 264:578-587. Nucleic Acids Res 24:4838-4840. Wilson T, Fahrner T, Johnston M, and Milbrandt J (1991) Rossi F, Blakely B, Charlton C, and Blau H (2000) Monitoring Identification of the DNA binding site for NGFI-B by protein-protein interactions in live mammalian cells by genetic selection in yeast. Science 252:1296-1300. beta-galactosidase complementation. Methods Enzymol Xu Y,Piston D, and Johnson C (1999) Abioluminescence res- 328:231-251. onance energy transfer (BRET) system: Application to Sadowski S, Ma J, Triezenberg S, and Ptashne M (1988) interacting circadian clock proteins. Proc Natl Acad Sci Gal4-VP16 is an unusually potent transcriptional activa- U S A 96:151-156. tor. Nature 335:245-249. Zandi E, Rothwarf D, Delhase M, Hayakawa M, and Karin Sehgal A, Price J, Man B, and Young M (1994) Loss of circa- M (1997) The I kappaβ kinase complex (IKK) contains dian behavioral rhythms and per RNA oscillations in the two kinase subunits, IKKα and IKKβ, necessary for Drosophila mutant timeless. Science 263:1603-1606. I kappaβ phosphorylation and NF-kappaβ activation. Serebriiskii I, Khazak V, and Golemis E(1999) A two-hybrid Cell 91:243-252. dual bait system to discriminate specificity of protein Zhang J and Lautar S (1996) A yeast three-hybrid method to interactions. J Biol Chem 274:17080-17087. clone ternary protein complex components. Anal Shearman L, Sriram S, Weaver D, Maywood E, Chaves I, Biochem 242:68-72. Zheng B, Kume K, Lee C, van der Horst G, Hastings M, Zylka M, Shearman L, Levine J, Jin X, Weaver D, and and Reppert S (2000) Interacting molecular loops in the Reppert S (1998) Molecular analysis of mammalian time- mammalian circadian clock. Science 288:1013-1019. less. Neuron 21:1115-1122. YanovskyJOURNAL et OF al. BIOLOGICAL / PHYTOCHROMES RHYTHMS AND / CRYPTOCHROMESDecember 2001 IN ARABIDOPSIS Resetting of the Circadian Clock by Phytochromes and Cryptochromes in Arabidopsis

Marcelo J. Yanovsky,*,1 M. Agustina Mazzella,* Garry C. Whitelam,† and Jorge J. Casal*,2 *IFEVA, Faculty of Agronomy, University of Buenos Aires, Avenida San Martín 4453, 1417-Buenos Aires, Argentina, †Department of Biology, University of Leicester, University Road, Leicester LE1 7RH, United Kingdom

Abstract The authors sought to investigate the role of phytochromes A and B (phyA and phyB) and cryptochromes 1 and 2 (cry1 and cry2) in the synchronization of the leaf position rhythm in Arabidopsis thaliana. The seedlings were transferred from white light–dark cycles to free-running conditions with or without exposure to a light treatment during the final hours of the last dark period. The phase advance caused by a far-red light treatment was absent in the phyAmutant, deficient in the fhy1 and fhy3 mutants involved in phyA signaling, and normal in the cry1 and cry1 cry2 mutants. The phase shift caused by blue light was normal in the cry2 mutant; reduced in the phyA, cry1, phyA cry1, and cry1 cry2 mutants; and abolished in the phyA cry1 cry2 triple mutant. The phase shift caused by red light was partially retained by the phyA phyB double mutant. The authors conclude that cry1 and cry2 participate as photoreceptors in the blue light input to the clock but are not required for the phyA-mediated effects on the phase of the circadian rhythm of leaf position. The signaling proteins FHY1 and FHY3 are shared by phyA-mediated photomorphogenesis and phyA input to the clock.

Key words Arabidopsis, circadian rhythms, cryptochrome, leaf movement, light input, phytochrome

Light is acknowledged as the most important envi- possess a distinctive C-terminal extension. In Droso- ronmental cue involved in resetting the circadian phila, cryptochrome appears to be the only photo- clocks. Our understanding of the light input to the receptor involved in light input to the clock, and resid- clock has improved very significantly in recent years. ual effects of light on rhythmic behavior are considered The discovery of cryptochromes first in plants (cry1, to be the indirect consequence of vision effects on cry2; Ahmad and Cashmore, 1993; Guo et al., 1999), behavioral rhythms (Emery et al., 2000). In Drosophila, then in mammals (mcry1, mcry2; Hsu et al., 1996; Todo cryptochrome physically interacts with clock compo- et al., 1996, Kobayashi et al., 1998), insects (Stanewsky nents (Ceriani et al., 1999). In mammals, the situation et al., 1998), and so on, accounts for a significant deal of has been more controversial. Cryptochromes were this advance. Cryptochromes have sequence homo- first proposed as photoreceptors involved in the light logy to distinct types of photolyases in plants and ani- input to the circadian clock (Miyamoto and Sancar, mals and, for this reason, are believed to be the prod- 1998). Although the observation that the mcry2 uct of independent evolution events (Cashmore et al., mutant of mice shows reduced sensitivity to induction 1999). Cryptochromes lack photolyase activity and of mPER1 (Thresher et al., 1998) is in favor of the latter

1. Current address: Department of Cell Biology, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037, USA. 2. To whom all correspondence should be addressed: e-mail: [email protected]. JOURNAL OF BIOLOGICAL RHYTHMS, Vol. 16 No. 6, December 2001 523-530 © 2001 Sage Publications 523 524 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001 view, the idea lost support because mcry2 displayed an increase in the length of the period not only under increased and not reduced sensitivity to light resetting blue light but also under red light, indicating that cry- of the clock (Thresher et al., 1998). The observation ptochromes are necessary for phyA signaling to the that mcry1 mcry2 double mutants were arrythmic in clock but not for phyAsignaling during photomorpho- darkness gave credit to a role of cryptochromes as cen- genesis. In the present work, we use a different proto- tral components of the mammalian clock (Van der col (Yanovsky et al., 2000a) to quantify the phase shift Horst et al., 1999). More recently, experiments with of the circadian rhythm of Arabidopsis leaf position in mutant mice lacking mcry1, mcry2 as well as rods and response to different light qualities. The results show most cones have suggested a redundant role of that phyA-mediated resetting of the circadian clock by cryptochromes and opsins in the light input to the far-red light is unaffected by the cry1 and cry2 muta- clock (Selby et al., 2000). Thus, cryptochromes in tions and impaired by the photomorphogenic mammals would have a dual role as photoreceptors mutants fhy1 and fhy3. and as central components of the clock. In plants, cry1 and cry2 are clearly important for normal control of growth and development by blue light (Lin, 2000). MATERIALS AND METHODS Mutations on these genes increase the period of rhythmic expression of a photosynthetic gene under certain Plants of Arabidopsis thaliana of the ecotype fluence rates of blue light (Devlin and Kay, 2000; Landsberg erecta or of the phyA-201 (Reed et al., 1994), Somers et al., 1998). The double cry1 cry2 mutant fhy1, fhy3 (Whitelam et al., 1993), cry1-1 (Ahmad and retains robust rhythmicity in Arabidopsis, indicating Cashmore, 1993), phyA-201 phyB-1 (Mazzella et al., that in contrast to the situation in mammals, crypto- 1997), cry1-1 cry2 (where cry2 is the fha-1 allele; Guo chromes are not essential components of the clock et al., 1999), phyA-201 cry1-1, phyA-201 cry1-1 cry2 3 (Yanovsky et al., 2000b; Devlin and Kay, 2000). In (Mazzella and Casal, 2001) were grown in 5 cm pots plants, phytochromes A, B, D, and E(phyA, phyB, containing a soil/sand medium. After sowing, the phyD, and phyE) are also involved in the input to the pots were incubated for 3 days in darkness at 7 °C, clock (Somers et al., 1998; Yanovsky et al., 2000a; exposed to a saturating pulse of red light (30 min, 30 -2 -1 Devlin and Kay, 2000). phyA is a light-modulated µmol.m .s ) to induce germination, and allowed to kinase (Fankhauser et al., 1999; Yeh and Lagarias, germinate for 24 h in darkness at 25 °C. The seedlings 1998). phyA and phyB are present in the cytoplasm in were grown at 20 °C under a 12 h white light/12 h dark dark-grown seedlings and migrate to the nucleus photoperiod provided by fluorescent lamps (80 µmol. -2 -1 upon light activation (Kircher et al., 1999; Yamaguchi m .s ). et al., 1999), where they interact with factors that bind After 2 weeks under 12-h photoperiods, during the to DNA (Martinez-García et al., 2000). A last 5 h of the dark period prior to transfer to free-run- bacteriophytochrome is involved in resetting the cir- ning conditions, the seedlings were either exposed to cadian clock in Synechococcus elongatus (Schmitz et al., the light treatment (far-red, red, or blue light) or 2000). remained as dark controls. The far-red light treatment -2 -1 The identification of cryptochromes as circadian (100 µmol.m .s ) was provided by 60-Watt incandes- photopigments requires a demonstration that the cent lamps in combination with a water filter, six blue function of cryptochromes is directly related to their acrylic filters (Paolini 2031), and two red acetate filters light-absorbing properties (Lucas and Foster, 1999). (La Casa del Acetato, Buenos Aires, Argentina). Blue -2 -1 Selective spectral effects on biological rhythms have light (20 µmol.m .s ) was provided by fluorescent not been demonstrated for cryptochromes. In plants, lamps (Philips, TLF 40W/54) in combination with a -2 -1 the effects of cry1 and cry2 on photomorphogenesis blue acetate filter. Red light (40 µmol.m .s ) was pro- are observed under UV-A/blue light but do not vided by red fluorescent tubes (Philips, 40/15). extend to the red/far-red region of the spectrum. After transfer to free running conditions (20 °C, -2 -1 Phytochromes, in contrast, operate predominantly in continuous white light 10 µmol.m .s ), the position of the red/far-red wavebands and to a lesser degree in the first pair of leaves was recorded every 2 h with a the blue region (Casal and Mazzella, 1998; Neff and digital video camera (Quick Cam, Connectix Corpora- Chory, 1998), where the pigment shows a secondary tion, San Mateo, CA, USA), and the angle between the peak of absorption. Devlin and Kay (2000) recently petioles was measured using the Scion Image analysis observed that the absence of both cry1 and cry2 causes program (Scion Corporation, Frederick, MD, USA). Yanovsky et al. / PHYTOCHROMES AND CRYPTOCHROMES IN ARABIDOPSIS 525

Data were fitted, according to Millar et al. (1995), to the π φ equation: L(t) = (co +c1t)+(a0 – a1t) sin [2 /T (t – )], where L is the leaf angle, co is the estimated value of the leaf angle at t=0, c1 is an estimate of the linear rate of change in leaf angle, a0 is the estimated value of the amplitude at t=0, a1 is an estimate of the linear change in cycling amplitude, t is time, T is the period estimate, and φ is the estimate of the phase at t=0. The first6hof measurements were not included among the data used to estimate the parameters of the above equation to avoid effects produced by shifting the plants from colored to dim white light. The phases were calculated by linear regression through the peaks in the days following the light pulse, extrapolated back to the day of the pulse (Kondo et al., 1991; Roenneberg and Deng, 1997). Phase shifts were calculated as the difference between the phases of light-treated and dark-control plants. Data are means of at least six different plants, from three independent experiments. Period length and phase shift data are presented with their standard errors.

RESULTS

Seedlings of Arabidopsis were grown under daily white light–dark cycles. Five hours before the end of the last dark period, different groups were exposed to far-red, red, or blue light while control seedlings remained in darkness. At the end of the treatments (or objective night in the controls), all the seedlings were transferred to free-running conditions, that is, constant low fluence white light and constant temperature (Fig. 1A). In control seedlings, the leaf tips moved upward, reaching the most erect position at the beginning of the subjective night. This was followed by the opposite movement leading to the least erect position at the beginning of the subjective day (Fig. 1B). This rhythm continued for several days under free-running conditions. The period of the rhythm was 25.6 h Figure 1. Far-red, red, or blue light reset the rhythm of leaf movement in Arabidopsis thaliana. The seedlings were grown under (Table 1). Exposure to5hoffar-red,red,orblue light day-night cycles, given far-red light (100 µmol.m-2.s-1), red light had no effects on the length of the period but caused a (40 µmol.m-2.s-1), blue light (20 µmol.m-2.s-1), or no light treat- persistent phase advance of the circadian rhythm, that ment during the last5hofthefinal night and transferred to free- is, each phase of the rhythm occurred at an earlier time running conditions, that is, continuous white light and constant temperature (20 °C). A. Experimental protocol. B. Leaf angle. point (Fig. 1B, Table 1). The phyA, cry1, and cry1 cry2 mutants showed normal rhythmicity when transferred to free-running response to blue light. The cry1 and cry1 cry2 mutants conditions without a selective light treatment (Fig. 2, showed a normal phase advance in response to far-red Table 1). The phyA mutant showed no phase advance light and partial phase advance by blue light (Fig. 2, in response to the far-red light treatment and only par- Table 1). This indicates that under the present experi- tial phase advance (compared to the wild type) in mental conditions, the effects of cry1 and cry2 were 526 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Table 1. Period of the leaf movement rhythm in Arabidopsis ticipate as partially redundant components of the blue thaliana under free-running conditions and phase shift caused by 5 h of far-red (100 µmol.m-2.s-1), red (40 µmol.m-2.s-1), or blue light (20 light input to the clock. µmol.m-2.s-1) compared with the controls that remained in dark- The fhy1 and fhy3 mutants show impaired seedling ness. The seedlings were grown under day-night cycles and given de-etiolation under continuous far-red light and have the light treatments during the last5hofthelast night before trans- been implicated in selective branches of phyA signal- fer to free-running conditions. ing in photomorphogenesis (Barnes et al., 1996; Light Pulse Genotype Period (h) Phase-Shift (h) Cerdán et al., 1999; Yanovsky et al., 2000c). To investi- None WT 25.6 ± 0.4 gate whether these mutants also affect phyAsignaling phyA 26.1 ± 0.4 to the clock, the seedlings were transferred to free-run- cry1 25.6 ± 0.5 ning conditions with or without exposure to far-red cry2 25.4 ± 0.5 fhy1 25.6 ± 0.4 light during the final5hofthenight. Both fhy1 and fhy3 24.7 ± 0.4 fhy3 showed normal rhythmicity but no phase phyA phyB 24.7 ± 0.3 advance of the circadian clock in response to far-red cry1cry2 26.6 ± 0.4 phyA cry2 25.7 ± 0.5 light (Fig. 4, Table 1). phyA cry1 25.8 ± 0.5 The effect of red light on the phase of the circadian phyA cry1 cry2 26.5 ± 0.6 clock was only partially reduced by the absence of Far-red WT 24.8 ± 0.4 5.7 ± 0.6 phyA 25.6 ± 0.4 0.6 ± 0.5 phyA and phyB (Fig. 5, Table 1). The analysis of seed- cry1 24.4 ± 0.5 6.1 ± 0.6 ling morphology (hypocotyl length, cotyledon un- cry1cry2 26.3 ± 0.5 5.6 ± 0.6 folding) in response to different wavebands compared fhy1 25.1 ± 0.5 0.6 ± 0.4 fhy3 24.6 ± 0.6 0.9 ± 0.5 with darkness indicates that the phyA mutant is blind Blue WT 25.8 ± 0.5 8.6 ± 0.6 to far-red light (Whitelam et al., 1993), the phyA cry1 phyA 25.6 ± 0.3 5.1 ± 0.6 cry2 mutant is blind to blue light (data not shown), and cry1 24.4 ± 0.5 4.6 ± 0.5 the phyA phyB mutant is blind to red light (Mazzella cry2 24.6 ± 0.6 8.0 ± 0.6 cry1 cry2 28.1 ± 0.6 5.1 ± 0.7 et al., 1997). Thus, additional red light photoreceptors phyA cry2 24.9 ± 0.5 5.9 ± 0.6 control the phase of the circadian clock compared with phyA cry1 25.8 ± 0.4 3.7 ± 0.5 seedling morphology during de-etiolation. phyA cry1 cry2 26.2 ± 0.6 0.4 ± 0.5 Red WT 26.2 ± 0.6 3.9 ± 0.5 phyAphyB 24.8 ± 0.4 2.2 ± 0.5 DISCUSSION wavelength selective (i.e., observed under blue light cry1 and cry2 are known to affect the period of the but not under far-red light). clock under blue light (Devlin and Kay, 2000; Somers The residual phase advance observed in the cry1 et al., 1998). Since the double cry1 cry2 mutant retains and phyA mutants in response to blue light suggested robust rhythmicity, cry1 and cry2 cannot be essential that different photoreceptors could be redundantly components of the clock (Devlin and Kay, 2000; involved in the resetting of the circadian clock under Yanovskyet al., 2000b). The results presented here go a these conditions. To investigate this issue in further step further by showing that cry1 and cry2 are specifi- detail, the phase shift was analyzed in multiple cally involved in the phase response to blue light mutants. The dark controls of the cry2, cry1 cry2, phyA (Table 1). This observation provides a link between the cry1, phyA cry2, and phyA cry1 cry2 mutants retained role of cry1 and cry2 in the control of circadian normal rhythmicity (Fig. 3, Table 1). The phase shift rhythms and their spectral properties and supports induced by blue light in the cry1 cry2 and phyA cry1 the role of cryptochromes as circadian photoreceptors double mutants was reduced compared with the wild in plants. type and similar to that observed in the cry1 single Resetting of the circadian rhythm of leaf movement mutant. Although the cry2 mutation did not reduce by far-red light is mediated by phyA (Yanovsky et al., the phase shift produced in response to blue light in 2000a). This action of phyAwas unaffected by the cry1 the wild-type background (cf. cry2 and wild type) or and cry2 mutations (Table 1, Fig. 2) and abolished by cry1 mutant background (cf. cry1 cry2 and cry1), it did the fhy1 and fhy3 mutations (Fig. 4). An effect of phyA eliminate the residual phase shift observed in the cry1 was also observed under blue light in the cry1, cry2, phyA mutant (cf. cry1 cry2 phyA and cry1 phyA) and cry1 cry2 backgrounds (Table1). Under fluences of (Table 1). This indicates that phyA, cry1, and cry2 par- red light lower than 5 µmol•m-2 s-1, the period of the Yanovsky et al. / PHYTOCHROMES AND CRYPTOCHROMES IN ARABIDOPSIS 527

Figure 2. Far-red light resets the rhythm of leaf movement in the cry1 mutant but not in the phyA mutant, whereas blue light resets the rhythm in both mutants. The seedlings were grown under day-night cycles, given far-red light (100 µmol.m-2.s-1), blue light (20 µmol.m-2.s-1), or no light treatment during the last 5 h of the final night and transferred to free-running conditions.

Figure 3. Blue light resets the rhythm of leaf movement in cry1 cry2 and phyA cry1 but not in phyA cry1 cry2. The seedlings were grown under day-night cycles, given blue light (20 µmol.m-2.s-1) or no light treatment during the last5hofthefinal night and transferred to free-running conditions. rhythmic bioluminiscence driven by the luciferace Since phyA and cry1 mutants also cause extended peri- gene under the control of a light-harvesting complex ods over the same range of fluence rates of white light gene promoter (CAB:LUC) is extended by the phyA and the phyA cry1 double mutant shows a deficiency mutation (Somers et al., 1998). A similar effect is that is not larger than that observed in any of the single caused by the cry1 mutation (Devlin and Kay, 2000). mutant parental lines, cry1 appears to be necessary for 528 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Figure 4. Far-red light fails to reset the rhythm of leaf movement in the fhy1 and fhy3 mutants involved in phyA signaling. The seedlings were grown under day-night cycles, given far-red light (100 µmol.m-2.s-1) or no light treatment during the last5hofthefinal night and transferred to free-running conditions.

genetically (Yanovsky et al., 1997; Yanovsky et al., 2000c), and molecularly (Cerdán et al., 2000). The so-called very-low-fluence response (VLFR) pathway operates under red or far-red light and saturates with infrequent (hourly) light pulses. The high-irradiance response (HIR) pathway requires sustained (continuous or very frequently pulsed) far-red light. phyA requires cry1 under red light (Devlin and Kay, 2000), that is, under conditions where phyA operates via the VLFR pathway. Hourly pulses of far-red light (instead of continuous far-red light) were fully ineffective to reset the circadian rhythm of leaf movement in Arabidopsis (data not shown), indicating that phyA operated via the HIR pathway. This is consistent with the dramatic failure to reset the clock in response to far-red light observed in fhy3, a mutant impaired in HIR but retaining VLFR (Yanovsky et al., 2000c). In conclusion, we propose that cry1 is required for phyA signaling to the clock in the VLFR but not in the HIR Figure 5. Red light resets the rhythm of leaf movement in the mode. This working hypothesis should be tested in a phyA phyB double mutant. The seedlings were grown under system where both VLFR and HIR operate. day-night cycles, given red light (40 µmol.m-2.s-1) or no light treat- Little is known about the mechanisms by which ment during the last5hofthefinal night and transferred to phytochromes and cryptochromes signal to the clock. free-running conditions. In Drosophila, for instance, cryptochrome shows direct phyA signaling to the clock (Devlin and Kay, 2000). interaction with clock components (Ceriani et al., Our data show that cry1 is not universally required for 1999). The observation that phyArequires cry1 for sig- phyA signaling to the clock. naling to the clock but not for photomorphogenesis Both sets of data can be reconciled by considering (Devlin and Kay, 2000) suggests that some elements that phyA signals via two different pathways that can could be different between both processes. The failure be dissected photobiologically (Casal et al., 2000), in photomorphogenesis and far-red-mediated phase Yanovsky et al. / PHYTOCHROMES AND CRYPTOCHROMES IN ARABIDOPSIS 529 shifting observed in fhy1 and fhy3 demonstrates that, Barnes SA, Quaggio RB, Whitelam GC, and Chua N-H at least in the HIR mode, some signaling elements are (1996) fhy1 defines a branch point in phytochrome A signal transduction pathways of gene expression. Plant J shared between the two processes. fhy1 and fhy3 are 10:1155-1161. predicted to operate downstream phyA in the light- Casal JJ and Mazzella MA (1998) Conditional synergism signaling cascade. This argues against a direct action between cryptochrome 1 and phytochrome B is shown by of phyA on clock components as observed for crypto- the analysis of phyA, phyB and hy4 simple, double and chrome in Drosophila. The effects of fhy1 and fhy3 are triple mutants in Arabidopsis. Plant Physiol 118:19-25. consistent with the earlier conclusion that phyA, Casal JJ, Yanovsky MJ, and Luppi JP (2000) Two photo- biological pathways of phytochrome A activity, only one although affecting the response to several wavebands, of which shows dominant negative suppression by is not an essential component of the clock itself. phytochrome B. Photochem Photobiol 71:481-486. The phyA phyB cry1 cry2 quadruple mutant is virtu- Cashmore AR, Jarillo JA, Wu Y-J, and Liu D (1999) Crypto- ally blind to white light for de-etiolation but retains chromes: Blue light receptors for plants and animals. Sci- robust circadian rhythmicity implicating other ence 284:760-765. photoreceptor(s) involved in the entrainment of the Cerdán PD, Staneloni RJ, Ortega J, Bunge MM, Rodriguez- Batiller MJ, Sánchez RA, and Casal JJ (2000) Sustained but circadian clock (Yanovsky et al., 2000b). The observa- not transient phytochrome Asignaling targets a region of tions that the phyA cry1 cry2 triple mutant shows no a Lhcb1*2 promoter that is not necessary for phytochrome phase shift in response to blue light whereas the phyA B action. Plant Cell 12:1203-1211. phyB double mutant retains a small but significant Cerdán PD, Yanovsky MJ, Reymundo FC, Nagatani A, phase shift under red light points toward a red-light Staneloni RJ, Whitelam GC, and Casal JJ (1999) Regula- tion of phytochrome B signaling by phytochrome A and photoreceptor as the pigment involved in changing FHY1 in Arabidopsis thaliana. Plant J 18:499-507. the phase of the rhythm in response to white light in Ceriani MF, Darlington TK, Staknis D, Mas P, Weitz CJ, and the quadruple mutant. The small effect of red light Kay SA (1999) Light-dependent sequestration of TIME- could accumulate over several days. Since the period LESS by CRYPTOCHROME. Science 285: 1599-1603. of the circadian rhythm of CAB::LUC bioluminiscence Devlin PF and Kay SA (2000) Cryptochromes are required is extended in the phyA phyB phyD and phyA phyB phyE for phytochrome signaling to the circadian clock but not for rhythmicity. Plant Cell 12:2499-2509. triple mutants compared with the phyA phyB double Devlin PF, Patel RS, and Whitelam GC (1998) Phytochrome mutant (Devlin and Kay, 2000), phyD and phyEare Einfluences internode elongation and flowering time in good candidates. phyD and phyEhave very weak Arabidopsis. Plant Cell 10:1479-1487. effects during de-etiolation (Aukerman et al., 1997; Emery P, Staneswsky R, Hall J, and Rosbash M (2000) Devlin et al., 1998), and this favors the idea that like Drosophila cryptochromes: A unique circadian-rhythm flies and mammals, plants could also have photo- photoreceptor. Nature 404:456-457. Fankhauser C, Yeh KC, Lagarias JC, Zhang H, Elich TD, and receptors predominantly or exclusively involved in Chory J (1999) PKS1, a substrate phosphorylated by phyto- circadian rhythms (Yanovsky et al., 2000a). chrome that modulates light signaling in Arabidopsis. Sci- ence 284:1539-1541. Guo H, Duong H, Ma N, and Lin C (1999) The Arabidopsis ACKNOWLEDGMENTS blue-light receptor cryptochrome 2 is a nuclear protein regulated by a blue-light dependent post-transcriptional mechanism. Plant J 19:279-289. This work was supported by grants from FONCYT Hsu DS, Zhao XD, Zhao SY, Kazantsev A, Wang RP, Todo T, (BID 1201/OC-AR - PICT 06739) and Fundación Wei YF, and Sancar A (1996) Putative human blue-light Antorchas (A-13622/1-40) to JJC. photoreceptor hCRY1 and hCRY2 are flavoproteins. Biochem 35:13871-13877. Kircher S, Kozma-Bognar L, Kim L, Adam E, Harter K, Schäfer E, and Nagy F (1999) Light-quality dependent REFERENCES nuclear import of the plant photoreceptors phytochrome A and B. Plant Cell 11:1445-1456. Ahmad M and Cashmore AR (1993) HY4 gene of Arabidopsis Kobayashi K, Kanno S, Smit B, Van der Horst GTJ, Takao M, thaliana encodes a protein with characteristics of a and Ysui, S (1998) Characterization of photolyase/ blue-light photoreceptor. Nature 366:162-166. blue-light receptor homologs in mouse and humans cells. Aukerman MJ, Hirschfeld M, Wester L, Weaver M, Clack T, Nucleic Acid Res 26:5086-5092. Amasino, RM, and Sharrock RA (1997) A deletion in the Kondo T, Johnson CH, and Hastings JW (1991) Action spec- PHYD gene of the Arabidopsis Wassilewskija ecotype trum for resetting the circadian phototaxis rhythm in the defines a role for phytochrome D in red/far-red light CW15 strain of Chlamydomonas: I. Cells in darkness. sensing. Plant Cell 9:1317-1326. Plant Physiol 95:197-205. 530 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Lin C (2000) Photoreceptors and regulation of flowering identifies cryptochrome as a circadian photoreceptor in time. Plant Physiol 123:39-50. Drosophila. Cell 95:681-692. Lucas RJ and Foster RG (1999) Something to cry about? Curr Thresher RJ, Vitaterna MH, and Miyamoto Y (1998) Role of Biol 9:214-217. mouse cryptochrome blue light photoreceptor in Droso- Martinez-García JF, Huq E, and Quail PH (2000) Direct tar- phila. Cell 95:681-692. geting of light signals to a promoter element-bound tran- Todo T, Ryo K, Yamamoto H, Toh T, Inui H, Ayaki T, and scription factor. Science 288:859-863. Nomura M (1996) Similarity among the Drosophila (6-4) Mazzella MA, Alconada Magliano TM, and Casal JJ (1997) photolyase, a human photolyase homolog, and the DNA Dual effect of phytochrome A on hypocotyl growth under photolyase-blue-light photoreceptor family. Science 272: continuous red light. Plant Cell Environ 20: 261-267. 109-112. Mazzella MA and Casal JJ (2001) Interactive signalling by Van der Horst GT, Muijtjens M, Kobayashi K, Takano R, phytochromes and cryptochromes generates de-etiolation Kanno S, Takao M, de Wit J, Verkerk A, Eker AP, van homeostasis in Arabidopsis thaliana. Plant Cell Environ Leenen D, Nuijs R, Bootsma D, Hoeijmakers JH, and 24:155-162. Yasui A (1999) Mammalian Cry1 and Cry2 are essential Millar AJ, Straume M, Chory J, Chua NH, and Kay SA(1995) for maintenance of circadian rhythms. Nature 398:627- The regulation of circadian period by phototransduction 630. pathways in Arabidopsis. Science 267:1163-1166. Whitelam GC, Johnson E, Peng J, Carol P, Anderson ML, Miyamoto Y and Sancar A (1998) Vitamin B-2-based Cowl JS, and Harberd NP (1993) Phytochrome A null blue-light photoreceptors in the retinohypothalamic mutants of Arabidopsis display a wild-type phenotype in tract as the photoactive pigments for setting the circadian white light. Plant Cell 5:757-768. clock in mammals. Proc Natl Acad SciUSA95:6097-6102. Yamaguchi R, Nakamura M, Mochizuki N, Kay SA, and Neff M and Chory J (1998) Genetic interactions between Nagatani A (1999) Light-dependent translocation of a phytochrome A, phytochrome B and cryptochrome 1 phytochrome B-GFP fusion protein to the nucleus in during Arabidopsis development. Plant Physiol 118:27-36. transgenic Arabidopsis. J Cell Biol 145:437-445 Reed JW, Nagatani A, Elich TD, Fagan M, and Chory J (1994) Yanovsky MJ, Casal JJ, and Luppi JP (1997) The VLF loci, Phytochrome A and phytochrome B have overlapping polymorphic between ecotypes Landsberg erecta and but distinct functions in Arabidopsis development. Plant Columbia dissect two branches of phytochrome Asignal- Physiol 104:1139-1149. ling pathways that correspond to the very-low fluence Roenneberg T and Deng T-S (1997) Photobiology of the and high-irradiance responses of phytochrome. Plant J Gonyaulax circadian system: I. Different phase response 12:659-667. curves for red and blue light. Planta 202:494-501. Yanovsky MJ, Izaguirre M, Wagmaister JA, Gatz C, Jackson Schmitz O, Katayama M, Willams SB, Kondo T, and Golden SD, Thomas B, and Casal JJ (2000a) Phytochrome Aresets SS (2000) CikA, a bacteriophytochrome that resets the the circadian clock and delays tuber formation under cyanobacterial circadian clock. Science 289:765-767. long days in potato. Plant J 23:223-232. Selby CP, Thompson C, Schmitz TM, Van Gelder RN, and Yanovsky MJ, Mazzella MA, and Casal JJ (2000b) A quadru- Sancar A (2000) Functional redundancy of cryptochromes ple photoreceptor mutant still keeps track of time. Cur- and classical photoreceptors for nonvisual ocular rent Biol 10:1013-1015. photoreception in mice. Proc Natl Acad SciUSA97: Yanovsky MJ, Whitelam GC, and Casal JJ (2000c) fhy3-1 14697-14702. retains inductive responses of phytochrome A. Plant Somers DE, Devlin PF,and Kay SA(1998) Phytochromes and Physiol 123:235-242. cryptochromes in the entrainment of the Arabidopsis cir- Yeh K-C and Lagarias C (1998) Eukariotic phytochromes: cadian clock. Science 282:1488-1490. Light-regulated serine/threonine protein kinases with Stanewsky R, Kaneko M, Emery P,Beretta B, Wager-Smith K, histidine kinase ancestry. Proc Natl Acad SciUSA95: Kay SA, Rosbash M, and Hall JC (1998) The cryb mutation 13976-13981. ShearmanJOURNAL and OF BIOLOGICALWeaver / DISTINCT RHYTHMS PHARMACOLOGICAL / December 2001 MECHANISMS Distinct Pharmacological Mechanisms Leading to c-fos Gene Expression in the Fetal Suprachiasmatic Nucleus

Lauren P. Shearman1 and David R. Weaver2 Laboratory of Developmental Chronobiology, MassGeneral Hospital for Children, Massachusetts General Hospital, Boston, MA 02114, USA, and Department of Pediatrics, Harvard Medical School, Boston, MA 02115, USA

Abstract Maternal treatment with cocaine or a D1-dopamine receptor agonist induces c-fos gene expression in the fetal suprachiasmatic nuclei (SCN). Other treatments that induce c-fos expression in the fetal SCN include caffeine and nicotine. In the current article, the authors assessed whether these different pharmacological treatments activate c-fos expression by a common neurochemical mechanism. The results indicate the presence of at least two distinct pharmacological pathways to c-fos expression in the fetal rat SCN. Previous studies demonstrate that prenatal activation of dopamine receptors affects the developing circadian system. The present work shows that stimulant drugs influence the fetal brain through multiple transmitter systems and further suggests that there may be multiple pathways leading to entrainment of the fetal biological clock.

Key words circadian rhythms, D1-dopamine receptor, gene expression, entrainment, SKF 38393, caffeine, nicotine

The suprachiasmatic nuclei (SCN) contain a biolog- (Bender et al., 1997; Carlson et al., 1991; Duffield et al., ical clock that regulates circadian rhythmicity in mam- 1999; Shearman et al., 1997; Strother et al., 1998a; mals (Klein et al., 1991; Weaver, 1998). The biological Weaver et al., 1992). Activation of these receptor types, clock within the SCN is oscillating prior to birth in which have opposite effects on cyclic AMP levels, sets rodents, and the timing of the fetal clock is set the phase of the fetal clock to opposite phases in ham- (entrained) prior to birth (Davis, 1997; Reppert and sters (Viswanathan and Davis, 1997). Prenatal entrain- Weaver, 1991). It appears that redundant signals from ment by these pharmacological treatments demon- the mother are normally involved in entraining the strates that functional receptors are present in the fetal fetal SCN (Reppert and Weaver, 1991; Viswanathan brain. The correlation of entrainment with c-fos gene et al., 1994; Viswanathan and Davis, 1997). induction in the fetal SCN following dopaminergic Prenatal administration of melatonin and the drug treatment suggests that induction of c-fos gene

D1-dopamine receptor agonist, SKF 38393, can set the expression is a useful marker for functionally relevant phase of the fetal biological clock (Davis and stimulation of the SCN. It is important to note, how- Mannion, 1988; Viswanathan et al., 1994; Viswanathan ever, that failure of a treatment to induce c-fos gene and Davis, 1997). Both D1 receptors and melatonin expression does not necessarily indicate insensitivity receptors are expressed in the developing rodent SCN to this input. For example, melatonin entrains the fetal

1. Current address: Department of Animal Pharmacology, Merck Research Laboratories, P.O. Box 2000, Rahway, NJ 07065. 2. To whom all correspondence should be addressed: David R. Weaver, Ph.D., Department of Neurobiology,University of Mas- sachusetts Medical School, 55 Lake Avenue North, Worcester, MA 01655; e-mail: [email protected]. JOURNAL OF BIOLOGICAL RHYTHMS, Vol. 16 No. 6, December 2001 531-540 © 2001 Sage Publications 531 532 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

SCN without induction of c-fos expression (Viswanathan receptor types leading to c-fos gene expression in the and Davis, 1997). fetal brain. The results indicate that there are multiple Nicotine also has been reported to induce c-fos gene and distinct pharmacological mechanisms leading to expression in the fetal SCN (Clegg et al., 1995; O’Hara c-fos gene expression in the fetal SCN. et al., 1998; see O’Hara et al., 1999, for review). The effects of nicotine on c-fos mRNA levels in the rat SCN are restricted to the perinatal period (O’Hara et al., MATERIALS AND METHODS 1999). Prenatal nicotine treatment leads to increased expression of other immediate early genes, including Animals and Drug Treatments junB (O’Hara et al., 1999). In these respects, prenatal treatment with nicotine is similar to prenatal treat- Timed-pregnant Sprague-Dawley rats (MBM:VAF, ment with SKF 38393, which has a transient, perinatal Zivic Miller Laboratories, Zelienople, PA, USA) were period of efficacy and leads to induction of several housed in a centralized vivarium and were exposed to immediate early genes (Weaver and Reppert, 1995; a light-dark cycle consisting of 12 h light:12 h dark, Weaver et al., 1995). The possibility that prenatal nico- with lights on at 0600 h EST (LD). During the dark tine treatment alters entrainment or other aspects of phase of the cycle, animals were exposed to red light. developing circadian function has not been examined. Gestational day (GD) 0 is defined as beginning the In adult rodents, however, nicotine treatment can morning after overnight pairing resulting in a sperm cause phase shifts (Ferguson et al., 1999; O’Hara et al., plug. Rats were studied in the afternoon, 6 to 10 h after 1999; Trachsel et al., 1995), and there are nicotinic com- lights-on, on GD 20. ponents to light-induced c-fos expression and light- Each pregnant rat received two intraperitoneal induced phase shifts (Keefe et al., 1987; O’Hara et al., injections, 30 min apart. The first injection was a 1998; Zhang et al., 1993). potential antagonist or vehicle. The antagonists used

We are interested in identifying agents that induce were the D1-dopamine receptor antagonist SCH 23390 c-fos gene expression in the fetal SCN, as these agents (0.5 mg/kg), the nicotinic cholinergic antagonist may reveal receptor-mediated mechanisms for influ- mecamylamine (Meca) (5 mg/kg), and the NMDA encing the SCN relevant to prenatal entrainment. Caf- receptor antagonist MK 801 (0.5 mg/kg). The second feine treatment induces c-fos gene expression in adult injection was an agent to induce c-fos gene expression, rodents that is limited to the striatum (Nakajima et al., SKF 38393 (10 mg/kg), caffeine (100 mg/ kg), nicotine 1989). Our interest in the functional development of (free base, 1 mg/kg), or vehicle. Limitations on the adenosine receptors (Shearman and Weaver, 1997; number of animals that could be treated on 1 day and Weaver, 1993, 1996) led us to perform preliminary the number of sections that could be processed studies on the effects of caffeine on c-fos gene expres- together for in situ hybridization dictated that three sion in fetal brain. To our surprise, maternal caffeine separate experiments be conducted, with each experi- treatment caused a widespread increase in c-fos ment consisting of all four pretreatments in combina- expression in the fetal brain, most notably in the SCN. tion with two posttreatments (one inducing drug and The anatomical pattern of c-fos expression in the fetal its vehicle). Doses were selected on the basis of prelim- brain after maternal caffeine treatment did not closely inary dose-response studies (caffeine; see also match the distribution of expression of either A1 or A2A Nakajima et al., 1989) or on the basis of responses adenosine receptor subtypes, but it did resemble the reported in the literature (Clegg et al., 1995; Weaver and pattern observed after maternal treatment with SKF Clemens, 1987; Weaver et al., 1992). In preliminary 38393 (Shearman et al., 1997; Fig. 1). The patterns of studies, we noted that the noncompetitive N-methyl- c-fos gene expression after SKF 38393 and caffeine, D-aspartate (NMDA) receptor antagonist, MK 801, while similar, are distinct from the restricted pattern of was not well tolerated by fetal rats at the doses most c-fos gene expression in the fetal brain after maternal frequently used in the literature (1.0-3.0 mg.kg; treatment with nicotine (Clegg et al., 1995). These Kennaway and Moyer, 1999; Nakazato et al., 1998; observations suggested to us that caffeine and SKF Svenningsson et al., 1996), and so we used a lower 38393 may share a convergent neurochemical mecha- dose (0.5 mg/kg). Injections were given at a volume of nism for induction of c-fos expression, while the effects 2 mL/kg. Dams were killed by decapitation 30 min of nicotine may be independent. In the present report, after the second injection. Fetuses were then removed a pharmacological approach was used to identify the from the uterus and decapitated. Fetal heads were Shearman and Weaver / DISTINCT PHARMACOLOGICAL MECHANISMS 533

Figure 1. c-fos gene expression in fetal rat brain after administration of (A) SKF 38393, (B) caffeine, or (C) vehicle. Sections shown in panels A, B, and C were hybridized with an antisense probe to detect c-fos mRNA. Panel D represents a section adjacent to the section shown in B but hybridized with a sense strand (control) probe to illustrate nonspecific hybridization. Abbreviations: CP = caudate-putamen (striatum); DE = dorsal endopiriform nucleus; SCN = suprachiasmatic nuclei. frozen in cooled 2-methybutane (–20 °C) and stored at (Weaver et al., 1992). Prehybridization, hybridization, –80 °C. Animal studies were reviewed and approved and wash conditions were as previously described by the Massachusetts General Hospital Subcommittee (Weaver, 1993). Film autoradiograms were generated on Research Animal Care. by apposition of slides to Kodak SB-5 film for 10 to 15 days. Radioactive standards (14C, 20-micron thickness; American Radiolabeled Chemical, St. Louis, MO, In Situ Hybridization USA) were included on each film. Sections (15 microns) were cut on a Bright-Hacker cryostat at –20 °C and thaw-mounted onto slides Data Analysis coated with Vectabond (Vector Labs, Burlingame, CA, USA). In situ hybridization was used to detect c-fos Optical density measurements were performed mRNA. Antisense and sense (control) cRNA probes using a computer-based image analysis system as pre- were produced from linearized plasmid DNA by in viously described (Weaver, 1993). In situ hybridiza- vitro transcription in the presence of 35S-alpha- thio- tion results are presented as SCN relative optical den- UTP (1100-1300 Ci/mmol, NEN). Probes were puri- sity (OD) values, defined as the OD of the SCN fied by extraction and ethanol precipitation prior to divided by the OD of the adjacent hypothalamus in use. The coding region of the mouse c-fos cDNA in the same section. Relative OD values from the two to pGEM3 (plasmid MUSFOS3, kindly provided by Dr. three sections having the most intense SCN hybridiza- Michael E. Greenberg) was used as the template in the tion signal were averaged to give the relative OD transcription reaction, as previously described value for each fetus. To determine whether 534 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001 pretreatments had a statistically significant impact on Caffeine-induced c-fos expression in the SCN (Fig. 2C) c-fos gene expression, data (OD values) were analyzed was not prevented by pretreatment with SCH 23390 by one-way analysis of variance and Dunnett’s test (Figs. 2F, 3B). In fact, SCN c-fos mRNA levels were using Statview version 1.03 on a Macintosh computer. actually higher following treatment with SCH 23390 Pairwise comparisons were performed using Stu- plus caffeine than after treatment with vehicle plus dent’s t test. Statistical significance was set at p < 0.05. caffeine (Fig. 3A; p < 0.05, Dunnett’s test). This increase For production of the figures, images were cap- may be due in part to an increase in c-fos expression tured using a Polaroid DMC IEdigital camera and due to treatment with SCH 23390. Remarkably, how- imported into Photoshop 5.0 for assembly of compos- ever, the induction of c-fos in the striatum was com- ite images. pletely prevented by the D1 receptor antagonist (compare Figs. 2C and 2F). Both Meca and MK 801 reduced the amplitude of the SCN response to caffeine (Fig. 2I, Chemicals and Drugs 2L; Fig. 3). MK 801 also reduced the striatal response to SKF 38393, SCH 23390, MK 801, and Meca were caffeine (Fig. 2L). obtained from Research Biochemicals, Inc., division of Experiments were also conducted to assess the Sigma-Aldrich, Inc. (Natick, MA, USA). Caffeine and pharmacological characteristics of c-fos induction in nicotine (free base) were purchased from Sigma response to nicotine treatment (Figs. 4, 5). Preliminary Chemical Co. (St. Louis, MO, USA). General lab chem- experiments demonstrated that nicotine bitartrate did icals were purchased from Sigma or Fisher Scientific not produce a substantial c-fos response in the fetal (Springfield, NJ, USA). SCN when administered at 2.9 mg/kg (equivalent to 1 mg/kg nicotine). In the free-base form, however, nicotine did cause a modest increase in c-fos expression in RESULTS the fetal SCN, consistent with previous reports (Clegg et al., 1995; O’Hara et al., 1998; O’Hara et al., 1999). Maternal treatment with caffeine (100 mg/kg) or Nicotine also induced c-fos expression in the supra- SKF 38393 (10 mg/kg) significantly increased c-fos optic and paraventricular nuclei of the hypothalamus gene expression in the fetal SCN (Fig. 1). The two drug (Fig. 4 and data not shown; see also O’Hara et al., 1999), treatments produced a grossly similar, widespread but not in the striatum, giving an anatomical pattern pattern of c-fos expression in the fetal brain, extending clearly distinct from that seen following maternal beyond the SCN and including the striatum and other treatment with caffeine or SKF 38393. Pretreatment regions (Fig. 1 and data not shown). The similarity in with each of the antagonists, SCH 23390, Meca, or MK pattern between fetuses exposed to the dopaminergic 801, significantly reduced the effect of nicotine on SCN agonist and to caffeine suggested that there might be a c-fos mRNA levels (Fig. 5 ; Dunnett’s test). SCH 23390 similar neurochemical mechanisms, for example, that and mecamylamine each blocked the effect of nico- the effects of caffeine might be exerted through indi- tine, while MK 801 reduced the amplitude of response rect effects on dopamine systems in the fetal brain but did not prevent it (Fig. 5; t tests). (Ferre et al., 1992; Fredholm, 1995; Garrett and Holtzman, 1994). To assess this possibility, pregnant dams were pretreated with antagonists prior to receiv- DISCUSSION ing a stimulant drug. The antagonists used had differential effects on The results of these experiments indicate the exis- stimulant-induced c-fos expression in the SCN (Figs. 2, tence of multiple and distinct pharmacological mech-

3). Pretreatment with the D1-dopamine receptor anisms for regulating c-fos gene expression in the fetal antagonist, SCH 23390, prevented the response to SKF SCN. In the adult SCN, major regulators of c-fos 38393 (Figs. 2E, 3A). Pretreatment with the nicotinic include glutamate and serotonergic agonists (Ebling, cholinergic receptor antagonist Meca or the NMDA 1996; Kennaway and Moyer, 1999), while in the fetal receptor antagonist MK 801 did not affect the SCN SCN, major regulators of c-fos expression include response to SKF 38393 (Figs. 2H, 2K, 3A). MK 801 pre- dopamine receptor activation and an unidentified treatment did, however, markedly reduce the induc- pathway activated by caffeine. Remarkably, none of tion of c-fos expression in the lateral portion of the fetal the treatments used to induce c-fos expression in the striatum (Fig. 2K). fetal SCN in this study causes a detectable increase in Shearman and Weaver / DISTINCT PHARMACOLOGICAL MECHANISMS 535

Figure 2. SKF 38393 and caffeine induce c-fos expression by distinct pharmacological mechanisms. Representative autoradiograms illustrate c-fos gene expression in fetal brain after various drug treatment combinations. Pregnant rats were pretreated with vehicle (A, B, C), SCH 23390 (D, E, F), Meca (G, H, I), or MK 801 (J, K, L). Thirty minutes later, they received vehicle (A, D, G, J), SKF 38393 (B, E, H, K), or caffeine (C, F, I, L). c-fos expression in the adult SCN (Clegg et al., 1995; Nakajima et al., 1989, unpublished data; Weaver and SCN. The D1-receptor antagonist, SCH 23390, blocked Reppert, 1995; Weaver et al., 1992). The mechanisms the induction of c-fos by SKF 38393. The present results regulating c-fos expression in the brain appear to vary are consistent with previous results in mice with tar- depending on the pharmacological stimulus, the geted disruption of the D1-dopamine receptor gene region, and the developmental stage examined, and and previous pharmacological data, indicating that coincident stimulation of multiple neurotransmitter dopaminergic agents induce c-fos expression in the systems may play a more important role in some SCN via activation of D1-dopamine receptors (Bender regions (e.g., striatum) than in others (e.g., SCN, see et al., 1997; Weaver et al., 1992). In contrast, induction below). of c-fos expression in the fetal SCN by caffeine was not The dopaminergic agonist, SKF 38393, acts through disrupted by SCH 23390, indicating the existence of an additional, D -receptor-independent pathway lead- D1 receptors to induce c-fos gene expression in the fetal 1 536 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

The mechanism by which caffeine induced c-fos expression in the fetal SCN is unclear. Caffeine is an adenosine receptor antagonist (Fredholm, 1995; Nehlig

et al., 1992), but A1 and A2A adenosine receptor mRNAs are not expressed at detectable levels in the fetal SCN (Weaver, 1993, 1996). Furthermore, the induction of c-fos expression in the adult striatum occurs only at pharmacological levels of caffeine and does not appear to be due to antagonism of specific adenosine receptor subtypes, suggesting a more complicated mechanism (Fredholm, 1995; Johanssen et al., 1992; Nakajima et al., 1989). Notably, the mechanism by which caffeine stimulates c-fos gene expression in the fetal striatum is pharmacologically distinct from the mechanism by which caffeine leads to c-fos in the fetal

SCN. Pretreatment with the D1-dopamine receptor antagonist, SCH 23390, prevented induction of c-fos expression in the striatum but not in the SCN.

D1-dopamine receptors are present in both the SCN and striatum, and dopaminergic innervation to both areas has been demonstrated during late fetal life (Bender et al., 1997; Duffield et al., 1999; Shearman

et al., 1997; Strother et al., 1998a). The presence of D2- dopamine receptors in the fetal striatum and their absence in the SCN (Schambra et al., 1994; Weaver et al., 1992) raises the possibility that synergistic inter-

actions between the D1 and D2 receptor subtypes occur in the striatum and that these interactions may contribute to regulation of c-fos expression in fetal striatum. In the fetal SCN, the effects of caffeine on

c-fos expression appear to be independent of D1-dopamine receptor stimulation. In the adult brain, glutamate receptor activation plays a major role in the regulation of c-fos expression (Chaudhuri, 1997). Glutamate regulates c-fos expression in the SCN and is a principal mediator conveying photic information from the retina to the SCN for entrainment (Ebling, 1996; Mintz et al., 1999). In the Figure 3. Quantitative assessment of antagonist effects on drug-induced c-fos gene expression. A. SKF 38393. B. Caffeine. adult striatum, induction of c-fos expression by Valuesrepresent the mean ± SEM of 7 to 11 brains per group. Sam- D1-dopamine-receptor activation appears to require ple sizes are shown within each bar. Values are expressed as rela- coactivation of NMDA receptors (Nakazato et al., tive optical density (OD) (OD SCN/OD adjacent hypothalamus). 1998). Similarly, in the present study, the NMDA Pretreatments were given 30 min before the treatment. Brains were collected an additional 30 min later, frozen, and processed to receptor antagonist MK801 prevented SKF 38393– detect c-fos RNA. Statistical comparisons key: *p < 0.05; NS = not induced c-fos expression in the lateral striatum. In con- statistically significant (p > 0.05) (Student’s t tests). # = signifi- trast, the induction of c-fos expression in the fetal SCN cantly different from control group that received vehicle pretreat- by SKF 38393 treatment was remarkably unaffected by ment followed by SKF 38393 (panel A) or caffeine (panel B) by Dunnett’s test (p < 0.05). pretreatment with MK 801. Thus, synergism between D1-dopamine and NMDA receptors is apparently ing to c-fos gene expression (see also Bender et al., required in the adult and fetal striatum, but not in the 1997). fetal SCN. Shearman and Weaver / DISTINCT PHARMACOLOGICAL MECHANISMS 537

Figure 4. Induction of c-fos gene expression by nicotine. Representative autoradiograms illustrate c-fos gene expression in fetal brain after various drug treatment combinations. Pregnant rats were pretreated with vehicle (A, B), SCH 23390 (C, D), Meca (E, F), or MK 801 (G, H). Thirty minutes later, they received vehicle (A, C, E, G), or nicotine (B, D, F, H; 1 mg/kg).

Dopaminergic, cholinergic, and glutaminergic Indeed, nicotinic induction of Fos in the adult brain is mechanisms each appear to contribute to the fetal disrupted by NMDA receptor antagonists and D1 response to nicotine. There may be serial activation of dopamine receptor antagonists (Kiba and Jayaraman, several receptor types in generating the response to 1994). Data from adult SCN slices indicate that nico- nicotine, making it susceptible to disruption at each of tine affects the SCN by action directly within the several points, or responsive cells may require coinci- nucleus, however (Trachsel et al., 1995). Furthermore, dent stimulation by several receptor subtypes. Nico- nicotinic receptor subunit gene expression has been tine may influence c-fos expression by acting as an documented in the fetal SCN (O’Hara et al., 1999), pro- indirect agonist, stimulating release of monoamines viding a substrate for direct action of nicotine in the and/ or glutamate through a presynaptic mechanism. fetal SCN. These lines of evidence indicate that while 538 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

induced cfos expression would vary as a function of time of day. Thus, we would not expect different results had the studies been conducted at night. The results might have been different, however, had we conducted them at a different developmental age. Rhythmicity of basal cfos levels appears in the dorsomedial SCN shortly after birth. Detection of drug-induced cfos during the daytime would have been more difficult during the postnatal period. Fur- thermore, the loss of the cfos responses to drugs during the early postnatal period limits the developmental window during which these studies could be conducted (Weaver and Reppert, 1995). Pharmacological activation of dopamine receptors in the fetal SCN can influence the development and function of the circadian timing system. Asingle injection of SKF 38393 late in gestation is sufficient to entrain fetal Syrian hamsters (Viswanathan and Davis, 1997). Sensitivity to SKF 38393 begins during Figure 5. Quantitative assessment of antagonist effects on nico- the prenatal period and continues into the neonatal tine-induced c-fos gene expression. Values represent the mean period (Grosse and Davis, 1999; Viswanathan et al., ± SEM of 5 to 10 brains per group, expressed as relative optical 1994). Entrainment to SKF 38393 may result from density (OD) (OD SCN/OD adjacent hypothalamus). Sample induction of Per1 gene expression; Per1 gene expres- sizes are shown within each bar. Pretreatment drugs, doses, and timing as described in the legends to Figures 2 and 4. Pairwise sion is increased in the fetal rat SCN by maternal treat- comparisons using Student’s t test are indicated by brackets (*p ment with SKF 38393 (Shearman and Weaver, unpub- 0.05; NS indicates not significant, p 0.05). # = significantly differ- lished data). Prenatal stimulation of dopamine ent from control group that received vehicle pretreatment fol- receptors also alters the development of SCN respon- lowed by nicotine (Dunnett’s test, p < 0.05). siveness to light at night in both rats and hamsters (Ferguson & Kennaway, 2000; Ferguson et al., 2000; the mechanism of action of nicotine is pharmacologi- Strother et al., 1998b). cally complex, this complexity may occur within the Prenatal exposure to caffeine or nicotine may also anatomical confines of the SCN. alter the circadian timing system. There is limited evi- Our studies were conducted during the subjective dence to suggest that theophylline, a methylxanthine day on GD 20. Previous studies indicate that the fetal with pharmacological effects similar to caffeine, can SCN expresses detectable but nonrhythmic (“basal”) influence the circadian timing system in adult rodents levels of c-fos mRNA(Viswanathan et al., 1994; Weaver (Ehret et al., 1975). While much of the current research et al., 1992). Nevertheless, there is abundant evidence on caffeine relates to its ability to promote arousal and that the circadian oscillator in the SCN is functioning reduce fatigue (Wright et al., 1997), the possibility that at this age (Reppert and Weaver, 1991), raising the pos- caffeine may have effects mediated by the SCN should sibility that the responses observed were influenced be considered. The absence of detectable c-fos expres- by the time of day of study. In the fetal SCN, c-fos gene sion in the adult SCN after high-dose caffeine treat- expression is induced equally well following adminis- ment (Nakajima et al., 1989; our unpublished data) tration of dopaminergic drugs during the subjective does not rule out an effect of caffeine on the circadian day and subjective night (Viswanathan et al., 1994; timing system, as c-fos gene expression is not necessar- Weaver et al., 1992). This is in contrast to the pro- ily induced by phase-shifting stimuli (Colwell et al., nounced “gating” of photic induction of cfos, junB, and 1993; Kumar et al., 1997; Rea et al., 1993; Viswanathan mPer gene expression in the adult, in which responses and Davis, 1997). are restricted to subjective night. While it is possible The present results support the concept that multi- that the basal neurochemical tone to the fetal SCN var- ple, functional neurotransmitter systems are active in ies over the course of the day,it seems unlikely that the the fetal brain, and suggest that exposure to widely pharmacological mechanisms leading to drug- used stimulant drugs during pregnancy may influ- Shearman and Weaver / DISTINCT PHARMACOLOGICAL MECHANISMS 539 ence brain development and function. Indeed, inges- Ehret CF, Potter VR, and Dobra KW (1975) Chronotypic tion of caffeine and nicotine during development have action of theophylline and of pentobarbital as circadian zeitgebers in the rat. Science 188:1212-1215. been shown to alter neurochemical development and Etzel BA and Guillet R (1994) Effects of neonatal exposure to can have long-lasting neurobehavioral effects (Aden caffeine on adenosine A1 receptor ontogeny using et al., 2000; Etzel and Guillet, 1994; Fisher and Guillet, autoradiography. Dev Brain Res 82:223-230. 1997; Grimm and Frieder, 1988; Nehlig and Debry, Ferguson, SA and Kennaway DJ (2000) Prenatal exposure 1994; Ribary and Lichtensteiger, 1989; Slotkin, 1998; to SKF-38393 alters the response to light of adult rats. Sobotka, 1989). Among the effects of prenatal expo- NeuroReport11: 1539-1541. sure to common neuromodulatory drugs may be Ferguson SA, Kennaway DJ, and Moyer RA (1999) Nicotine phase shifts the 6-sulfatoxymelatonin rhythm and altered development of the circadian timing system. induces c-Fos in the SCN of rats. Brain Res Bull 48: 527-538. Ferguson SA, Rowe SA, Krupa M, and Kennaway DJ (2000) ACKNOWLEDGMENTS Prenatal exposure to the dopamine agonist SKF 38393 disrupts the timing of the initial response of the suprachiasmatic nucleus to light. Brain Res 858:284-289. This work was supported by NIH grants HD14427 Ferre S, Fuxe K, von Euler G, Johansson B, and Fredholm BB and HD29470. LPS was supported in part by a fellow- (1992) Adenosine-dopamine interactions in the brain. ship from the Program for Training in Sleep, Circadian Neuroscience 51:501-512. and Respiratory Neurobiology at Harvard Medical Fisher S and Guillet R (1997) Neonatal caffeine alters passive School (HL07901). avoidance retention in rats in an age- and gender-related manner. Dev Brain Res 98:145-149. Fredholm BB (1995) Adenosine, adenosine receptors and the actions of caffeine. Pharmacol Toxicol 76:93-101. REFERENCES Garrett BEand Holtzman SG (1994) D1 and D2 dopamine receptor antagonists block caffeine-induced stimulation Aden U, Herlenius E, Tang L-Q, and Fredholm BB (2000) of locomotor activity in rats. Pharmacol Biochem Behav Maternal caffeine has minor effects on adenosine recep- 47:89-94. tor ontogeny in the rat brain. Pediatr Res 48:177-183. Grimm VEand Frieder B (1988) Prenatal caffeine causes Bender M, Drago J, and Rivkees SA (1997) D1 receptors long lasting behavioral and neurochemical changes. Int J mediate dopamine action in the fetal suprachiasmatic Neurosci 41:15-28. nuclei: Studies of mice with targeted deletion of the D1 Grosse J and Davis FC (1999) Transient entrainment of a cir- dopamine receptor gene. Mol Brain Res 49:271-277. cadian pacemaker during development by dopaminergic Carlson LL, Weaver DR, and Reppert SM (1991) Melatonin activation in Syrian hamsters. Brain Res Bull 48:185-194. receptors and signal transduction during development Johansson B, Svenningson P, Aden U, Lindstrom K, and in Siberian hamsters (Phodopus sungorus). Dev Brain Res Fredholm BB (1992) Evidence that the increase in striatal 59:83-88. c-fos following acute high-dose caffeine is not due to Chaudhuri A(1997) Neural activity mapping with inducible direct interaction with striatal adenosine receptors. Acta transcription factors. Neuroreport 8(13):iii-vii. Physiol Scand 146:539-541. Clegg DA, O’Hara BF, Heller HC, and Kilduff TS (1995) Nic- Keefe DL, Earnest DJ, Nelson D, Takahashi JS, and Turek FW otine administration differentially affects gene expres- (1987) A cholinergic antagonist, mecamylamine, blocks sion in the maternal and fetal circadian clock. Dev Brain the phase-shifting effects of light on the circadian rhythm Res 85:46-54. of locomotor activity in the golden hamster. Brain Res Colwell CS, Kaufman CM, and Menaker M (1993) 403:308-312. Phase-shifting mechanisms in the mammalian circadian Kennaway DJ and Moyer RW (1999) MK-801 administration

system: New light on the carbachol paradox. J Neurosci blocks the effects of a 5HT2A/2C agonist on melatonin 13:1454-1459. rhythmicity and c-fos induction in the suprachiasmatic Davis FC (1997) Melatonin: Role in development. J Biol nucleus. Brain Res 845:102-106. Rhythms 12:498-508. Kiba H and Jayaraman A (1994) Nicotine induced c-fos Davis FC and Mannion J (1988) Entrainment of hamster pup expression in the striatum is mediated mostly by dopa- circadian rhythms by prenatal melatonin injections to the mine D1 receptor and is dependent on NMDA stimula- mother. Am J Physiol 255:R439-R448. tion. Mol Brain Res 23:1-13. Duffield GE, McNulty S, and Ebling FJ (1999) Anatomical Klein DC, Moore RY, and Reppert SM (Eds) (1991) and functional characterisation of a dopaminergic sys- Suprachiasmatic Nucleus: The Mind’s Clock, Oxford Uni- tem in the suprachiasmatic nucleus of the neonatal Sibe- versity Press, New York. rian hamster. J Comp Neurol 408:73-96. Kumar V, Goguen DM, Guido ME, and Rusak B (1997) Ebling FJ (1996) The role of glutamate in the photic regula- Melatonin does not influence the expression of c-fos in the tion of the suprachiasmatic nucleus. Prog Neurobiol suprachiasmatic nucleus of rats and hamsters. Mol Brain 50:109-132. Res 52:242-248. 540 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Mintz EM, Marvel CL, Gillespie CF, Price KM, and Albers Strother WN, Norman AB, and Lehman MN (1998a) HE(1999) Activation of NMDA receptors in the supra- D1-dopamine receptor binding and tyrosine-hydroxy- chiasmatic nucleus produces light-like phase shifts of the lase immunoreactivity in the fetal and neonatal hamster circadian clock in vivo. J Neurosci 19:5124-5130. suprachiasmatic nuclei. Dev Brain Res 106:137-144. Nakajima T, Daval JL, Morgan PF, Post RM, and Marangos Strother WN, Vorhees CV, and Lehman MN (1998b) PJ (1989) Adenosinergic modulation of caffeine-induced Long-term effects of early cocaine exposure on the light c-fos mRNA expression in mouse brain. Brain Res responsiveness of the adult circadian timing system. 501:307-314. Neurotoxicol Teratol 20:555-564. Nakazato E, Ohno M, and Watanabe S (1998) MK-801 Svenningsson P, Johansson B, and Fredholm BB (1996) Caf- reverses Fos expression induced by the full dopamine D1 feine-induced expression of c-fos mRNA and NGFI-A receptor agonist SKF-82958 in the rat striatum. Eur J mRNA in caudate putamen and in nucleus accumbens Pharmacol 342:209-212. are differentially affected by the N-methyl-D-aspartate Nehlig A, Daval J-L, and Gérard D (1992) Caffeine and the receptor antagonist MK-801. Mol Brain Res 35:183-189. central nervous system: Mechanisms of action, biochemi- Trachsel L, Heller HC, and Miller JD (1995) Nicotine cal, metabolic and psychostimulant effects. Brain Res Rev phase-advances the circadian neuronal activity rhythm 17:137-170. in rat suprachiasmatic nuclei explants. Neuroscience Nehlig Aand Debry G (1994) Consequences on the newborn 65:797-803. of chronic maternal consumption of coffee during gesta- Viswanathan N and Davis FC (1997) Single prenatal injection and lactation: A review. J Am Coll Nutr 13:6-21. tions of melatonin or the D1-dopamine receptor agonist O’Hara BF, Edgar DM, Cao VH, Wiler SW, Heller HC, SKF 38393 to pregnant hamsters sets the offsprings’ circa- Kilduff TS, and Miller JD (1998) Nicotine and nicotinic dian phases 180° apart. J Comp Physiol [A] 180:339-346. receptors in the circadian system. Psychoneuroendocri- Viswanathan N, Weaver DR, Reppert SM, and Davis FC nology 23:161-173. (1994) Entrainment of the fetal hamster circadian pace- O’Hara BF, Macdonald E, Clegg D, Wiler SW, Andretic R, maker by prenatal injections of the dopamine agonist Cao VH, Miller JD, Heller HC, and Kilduff TS (1999) SKF 38393. J Neurosci 14:5393-5398.

Developmental changes in nicotinic receptor mRNAs Weaver DR (1993) A2a adenosine receptor gene expression in and responses to nicotine in the suprachiasmatic nucleus developing rat brain. Mol Brain Res 20:313-327.

and other brain regions. Mol Brain Res 66:71-82. Weaver DR (1996) A1-adenosine receptor gene expression in Rea MA, Michel AM, and Lutton LM (1993) Is Fos expres- fetal rat brain. Dev Brain Res 94:205-223. sion necessary and sufficient to mediate light-induced Weaver DR (1998) Suprachiasmatic nucleus: A25-year retro- phase advances of the suprachiasmatic circadian oscilla- spective. J Biol Rhythms 13:92-104. tor? J Biol Rhythms 8:S59-S64. Weaver DR and Clemens LG (1987) Nicotinic cholinergic Reppert SM and Weaver DR (1991) A biological clock is influences on sexual receptivity in female rats. Pharmacol oscillating in the fetal suprachiasmatic nuclei. In Biochem Behav 26:393-400. Suprachiasmatic Nucleus: The Mind’s Clock, DC Klein, RY Weaver DR and Reppert SM (1995) Definition of the devel- Moore, and SM Reppert, eds, pp 405-418, Oxford Univer- opmental transition from dopaminergic to photic regula- sity Press, New York. tion of c-fos gene expression in the rat suprachiasmatic Ribary U and Lichtensteiger W (1989) Effects of acute and nucleus. Mol Brain Res 33:136-148. chronic prenatal nicotine treatment on central catechola- Weaver DR, Rivkees SA, and Reppert SM (1992) D1-dopamine mine systems of male and female rat fetuses and off- receptors activate c-fos expression in the fetal biological spring. J Pharmacol Exp Ther 248:786-792. clock. Proc Natl Acad Sci U S A 89:9201-9204. Schambra UB, Duncan GE, Breese GR, Fornaretto MG, Weaver DR, Roca AL, and Reppert SM (1995) c-fos and jun-B Caron MG, and Fremeau RT Jr (1994) Ontogeny of D1A mRNAs are transiently expressed in fetal rodent and D2 dopamine receptor subtypes in rat brain using in suprachiasmatic nuclei following dopaminergic stimula- situ hybridization and receptor binding. Neuroscience tion. Dev Brain Res 85:293-297. 62:65-85. Wright KP Jr, Badia P, Myers BL, and Plenzler SC (1997) Shearman LP and Weaver DR (1997) [125I]4-aminobenzyl-5′- Combination of bright light and caffeine as a counter- N-methylcarboxamido-adenosine (125l)AB-MECA) measure for impaired alertness and performance during labels multiple adenosine receptor subtypes in rat brain. extended sleep deprivation. J Sleep Res 6:26-35. Brain Res 745:10-20. Zhang Y, Zee PC, Kirby JD, Takahashi JS, and Turek FW Shearman LP, Zeitzer J, and Weaver DR (1997) Widespread (1993) A cholinergic antagonist, mecamylamine, blocks expression of functional D1-dopamine receptors in fetal light-induced Fos immunoreactivity in specific regions rat brain. Dev Brain Res 102:105-115. of the hamster suprachiasmatic nucleus. Brain Res Slotkin TA (1998) Fetal nicotine or cocaine exposure: Which 615:107-112. one is worse? J Pharmacol Exp Ther 285:931-945. Sobotka TJ (1989) Neurobehavioral effects of prenatal caffeine. Ann N Y Acad Sci 562:327-339. GormanJOURNAL and OF Lee BIOLOGICAL / SPLIT ACTIVITY RHYTHMS RHYTHMS / December IN HAMSTERS 2001 Daily Novel Wheel Running Reorganizes and Splits Hamster Circadian Activity Rhythms

Michael R. Gorman1 and Theresa M. Lee Department of Psychology and Reproductive Sciences Program, University of Michigan, Ann Arbor, MI 48109-1109, USA

Abstract The phenomenon of splitting of locomotor activity rhythms in constant light has implied that the mammalian circadian pacemaker is composed of multiple interacting circadian oscillators. Exposure of male Syrian hamsters to novel running wheels also induces splitting in some reports, although novel wheel running (NWR) is better known for its effects on altering circadian phase and the length of the free-running period. In three experiments, the authors confirm and extend earlier reports of split rhythms induced by NWR. Male Syrian hamsters, entrained to LD 14:10, were transferred for 6 to 11 consecutive days to darkened novel Wahmann wheels at ZT 4 and were returned to their home cages at ZT 9. All hamsters ran robustly in the novel wheels. NWR caused a marked reorganization of home cage wheel-running behavior: Activity onsets delayed progressively with each additional day of NWR. After 11 days, activity onset in the nighttime scotophase was delayed by 7 h and disappeared completely in 2 hamsters (Experiment 1). After 6 to 7 days of NWR (Experiment 2), activity onset delayed by 5 h. Transfer of hamsters to constant darkness (DD) after 7 days of NWR revealed clearly split activity rhythms: The delayed nighttime activity bout was clearly identifiable and characterized by a short duration. A second bout associated with the former time of NWR was equally distinct and exhibited a similarly short duration. These components rejoined after 3 to 5 days in DD accomplished via delays and advances of the nighttime and afternoon components, respectively. The final experiment established that rejoining of activity components could be prevented by perpetuating the light-dark:light-dark cycle used to induce split rhythms. The data suggest that NWR causes selective phase shifting of some circadian oscillators and that component oscillators interact strongly in constant darkness.

Key words splitting, oscillator interaction, coupling, nonphotic

A multioscillator basis for mammalian circadian Illnerova, 1991; Liu et al., 1997; Pittendrigh and Daan, rhythms has been adduced through studies of 1976). Each set of studies reinforces the idea that photoperiodic control of activity duration (α), internal coherent circadian rhythms are generated from the desynchronization, splitting, and most recently, in interaction of coupled constituent oscillators with a vitro electrical recordings of single SCN cells (Aschoff, range of free-running periods, τ. Although significant 1965; Elliott and Tamarkin, 1994; Gorman et al., 1997; advances have been made in clarifying the neuroana-

1. To whom all correspondence should be addressed: Department of Psychology, University of California, San Diego, La Jolla, CA 92093-0109; e-mail: [email protected]. JOURNAL OF BIOLOGICAL RHYTHMS, Vol. 16 No. 6, December 2001 541-551 © 2001 Sage Publications 541 542 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001 tomical and physiological substrates for rhythm gen- MATERIALS AND METHODS eration and entrainment, the formal properties of oscillator interaction have received less sustained Animals and Husbandry attention. Amajor exception to this generalization is the study For all experiments, a subset of the same 24 male of split locomotor activity rhythms first reported in Syrian hamsters that were used in a separately the arctic ground squirrel, Spermophilus undulatus reported study published in this issue (HsdHan: (Pittendrigh, 1960) and elaborated further in studies AURA; Harlan, Indianapolis, IN, USA) (Gorman et al., of Syrian hamsters, Mesocricetus auratus (Pittendrigh 2001 [this issue]), 5 to 6 weeks of age at acquisition, and Daan, 1976). After prolonged (e.g., 60 days) expo- were housed with Sani-Chip bedding in polypropy- sure to constant light (LL), locomotor activity rhythms lene cages (48 × 27 × 20 cm) equipped with Nalgene of some individuals dissociate into two components (d = 34 cm) running wheels (Fisher Scientific, Pitts- that free-run initially with different frequencies. When burgh, PA, USA). Food (Purina Rodent Chow #5001, the two split activity components adopt an antiphase St. Louis, MO, USA) and water were available ad libi- relationship (180 degrees apart), they free-run with a tum. Syrian hamsters were entrained to LD 14:10 common frequency greater than that measured just (lights on 0500-1900 h; approximately 100 lux) for 3 prior to splitting. A comparable phenomenon is weeks before an initial regimen of daily NWR was obtained in a day-active species exposed to low levels initiated. of light intensity (Hoffmann, 1971). Exposure to constant lighting conditions is not the Novel Wheel Running only manipulation capable of splitting mammalian circadian rhythms. Although not discussed in the text, Following entrainment to LD 14:10 (lights off = ZT Bruce’s (1960) study of frequency demultiplication 12), hamsters were transferred within the same room includes a single actogram of a hamster maintained in to Wahmann wheels (d = 34 cm) 0 to 15 min before short cycles of 2 h light, 4.5 h dark (LD 2:4.5). In this lights were extinguished at 1100 h (ZT 4) EST. At 1600 record, two activity components 180 degrees apart h (ZT 9), the lights were turned on and hamsters were were apparent for approximately 7 days before one of returned to home cages in the light over the next 15 min. these components disappeared. Mrosovsky and Janik Thus, during NWR, animals were exposed to an (1993) reorganized the activity rhythms of hamsters LDLD 6:3:5:10 light schedule. On one day, the dark- maintained in LD 14:10 by exposing them each after- ened hamster room was entered through a light lock at noon to 3-h pulses of novel wheel running (NWR) in hourly intervals from 1200 to 1600 h to record the the dark (beginning7hbeforenormal lights-off). number of novel wheel revolutions with the aid of a When NWR was discontinued and hamsters were left dim red light. in their home cages in constant darkness (DD), locomotor activity rhythms were split into two components that rejoined after 3 to 5 days, although this pat- Analysis tern was not equally clear in all records shown (e.g., #3802 in their Fig. 2). Nighttime activity onsets in LD Wheel-running activity in the home cage was mon- 14:10 were also phase-delayed by several hours dur- itored by Dataquest III software (Mini-mitter, Sun ing NWR. Sinclair and Mistlberger (1997), using a dif- River, OR, USA) and compiled into 10-min bins. While ferent strain of hamster and a slightly modified proto- in the novel wheels, activity patterns were not moni- col, found less compelling evidence of splitting after tored, but the total number of wheel revolutions after 17 days of NWR, although nighttime activity onset the 5-h interval was recorded manually. Data analyses was delayed in some animals. Using the hamster were carried out with Excel (Microsoft, Bellevue, WA, strain employed by Mrosovsky and Janik (1993) and USA) and ClockLab software (Actimetrics, Evanston, a modification of their experimental protocol, we IL, USA). here describe marked reorganizations of locomotor Activity onset was defined as the first time point in activity rhythms induced by three regimens of daily a scotophase in which a hamster ran more than 20 rev- NWR. olutions in a 10-min interval followed immediately by Gorman and Lee / SPLIT ACTIVITY RHYTHMS IN HAMSTERS 543 an additional 10-min interval with more than 20 wheel this study and were exposed to constant darkness revolutions. Activity offset was defined as the last (DD) initiated during the subsequent 10-h scotophase time point in a scotophase that the animal ran more (i.e., the lights remained off at 0500 h). Data from the than 20 revolutions and that was preceded immedi- remaining 10 hamsters are reported here only through ately by a similar 10-min interval of activity. Activity the final day of NWR, after which they received a dif- duration (α) was calculated as the interval between ferent light treatment described in a separate study activity onset and activity offset. An interval of inac- (Gorman et al., 2001). Periods of the free-running tivity was calculated as the difference between activity rhythms of activity onset were calculated for each of offset and the subsequent activity onset. The circadian the 9 hamsters during days 1 to 4 and 8 to 11 of DD. period of activity onsets either in constant conditions (τ) or while exposed to a light-dark cycle (τ*) was esti- Experiment 3 mated with linear regression by determining slope of activity onsets over 4- to 7-day intervals. The phase Because a distinctly and evenly split home cage angle of entrainment was determined from the aver- running rhythm was obtained in Experiment 2, we age value predicted by the regression line and was asked whether these hamsters could be entrained to expressed in relation either to the time of lights-off the LDLD cycle in effect during NWR. The same ham- ψ ψ ( lights-off) or lights-on ( lights-on). When activity compo- sters (n = 20 including 1 former control hamster from nents were split, circadian parameters were calculated Experiment 2), 30 to 31 weeks of age, were re- separately for activity bouts corresponding to the entrained to LD 14:10 and treated as described in original 10-h dark period (i.e., the nighttime, n, activ- Experiment 2 except that they were not bled. After ity bout) and to the 5-h interval of NWR (i.e., the after- 6 days of NWR in LDLD 6:5:3:10, hamsters remained noon, a, activity bout). The phase angle between com- in their home cages for 11 days on the same LDLD ponents was defined as the difference between their cycle described above. Two hamsters with no prior ψ respective activity onsets ( n–a). NWR exposure (controls from Experiment 2) were exposed to identical light conditions but were not Experiment 1 transferred to novel wheels. Analyses of activity onsets were performed using Hamsters, 8 to 9 weeks of age, previously entrained data from the last 7 days of exposure to LD 14:10 prior to LD 14:10, were exposed to NWR in LDLD 6:3:5:10 to NWR and the first 7 days of continuous home cage (n = 20). After 11 days of these treatments, hamsters exposure to LDLD 6:5:3:10 after NWR. remained undisturbed in their home cages for 2 addi- Statistical tests (all two-tailed where applicable) tional days under the same light conditions. were performed with Statview 5.0 software (SAS Insti- tute, Cary, NC, USA). Experiment 2

Because 11 days of NWR in Experiment 1 phase- RESULTS delayed nighttime activity onset more than expected on the basis of published studies, we next assessed Experiment 1 whether more evenly split activity would be obtained after fewer days of NWR. Hamsters from Experiment 1, Hamsters transferred to novel wheels exhibited 12 to 13 weeks of age, were re-entrained to LD 14:10 for robust wheel-running (mean = 8186 ± 160 revolutions/ 14 days and exposed to NWR under LDLD 6:3:5:10 for 5 h, range = 6817-9210, n = 20), with no significant 7 days (n = 19). Four additional hamsters, with identi- change in amount of wheel running over the 11 days of cal light histories but no previous running-wheel the experiment (p > 0.70, repeated measures ANOVA). exposure, were equipped with Nalgene wheels. In this When measured on Day 2 of NWR, the number of experiment, these control hamsters were exposed to wheel revolutions varied over time (p < 0.001), with LDLD 6:5:3:10 without NWR. All hamsters were mini- significant monotonic increases (p < 0.05) over the first mally disturbed during days of NWR except for a sin- 4 h and a decrease from the 4th to the 5th hour (p < 0.05). gle retro-orbital bleeding conducted on the final day All 20 hamsters showed a marked reorganization of as part of another study. After the final day of NWR, 9 nighttime activity during NWR characterized by pro- of the 19 hamsters, randomly selected, remained in gressive delays in the onset of home cage activity 544 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

(Figs. 1, 2). In 2 hamsters (e.g., Fig. 1B), activity onsets tion of nighttime and afternoon activity components delayed so far as to eliminate all nighttime activity on in DD is presented in Table 1. the last 1 to 2 days of NWR. When left in the home cage Control animals exposed to identical LDLD condi- in the LDLD cycle, 19 out of 20 hamsters showed spon- tions showed no reorganization of nighttime locomo- taneous activity in the afternoon dark period, and 17 tor activity patterns or splitting in DD (data not out of 20 hamsters showed activity in both the after- shown). noon and nighttime scotophases (Fig. 1). The 1 hamster that did not run in the afternoon dark phase was Experiment 3 exceptional in having the smallest delay of nighttime activity onset. On each of the final 2 days when ham- As in Experiments 1 and 2, transfer to novel wheels sters remained in the home cage on the LDLD cycle, a elicited running in the entire sample of experimental disproportionate amount of running activity occurred hamsters (mean = 7631 ± 268 revolutions per 5 h; in the afternoon scotophase (65% ± 5%, 72% ± 5%, range = 4036-9197, n = 20), with no significant changes respectively, n = 20). in activity over the 6 days of NWR. Running induced progressive delays in nighttime activity onset. After Experiment 2 6 days of NWR, nighttime activity onset occurred approximately 6 h after lights-out. Compared to sub- As in Experiment 1, NWR was observed to be stantial further delays observed in identical light con- robust in the entire cohort of animals tested (mean = ditions in Experiment 1, discontinuation of NWR after 8153 ± 220 revolutions/5 h, range = 6,127-10,044, n = 6 days largely prevented further delays in nighttime 19). After 7 days in novel wheels, nighttime activity activity onset (Figs. 2, 4). onset was delayed approximately midway through Hamsters (18 out of 20) exposed to NWR adopted the scotophase (Figs. 2, 3). similar novel entrainment patterns in the home cage Upon release into DD, 8 out of 9 hamsters showed under LDLD: Locomotor activity was distributed into two distinct (i.e., split) activity components roughly two roughly equal components corresponding to the coincident with the prior time of nighttime running two daily scotophases (i.e., splitting occurred; Fig. 4). and novel wheel exposure, respectively (Fig. 3). In DD, Of the 2 nonsplitters, 1 showed the least amount of nighttime activity onsets occurred progressively later activity in the novel wheels (4036 rev/5 h) whereas the in the first 3 days of DD (Table 1), whereas the after- other showed typical activity levels (7820 rev/5 h). noon component was neither markedly advanced nor Neither of the 2 control hamsters exposed to this same delayed. Thus, the phase angle between components LDLD cycle, without NWR, adopted this entrainment Ψ ( n–a) and interval of inactivity initially separating the pattern (data not shown). This split pattern of locomo- nighttime and afternoon bouts in DD rapidly dimin- tor activity was sustained for a minimum of 7 days in ished. In the first 3 days of DD, the two activity bouts all 18 animals and for the duration of the experiment contained comparable amounts of activity, and bouts (11 days) for 15 of these hamsters. were characterized by short αs. A redistribution of Quality of entrainment was assessed by examining activity from the afternoon activity component to the whether the slope of the best-fitting regression line nighttime component was commonly seen on the 3rd through activity onsets differed significantly from 24 h to 5th day in DD (Fig. 3). The free-running rhythm (p < 0.05). Under baseline entrainment conditions, all derived from onsets of the nighttime component but 2 animals yielded regression lines not significantly shortened significantly after 7 days in DD (Table 1, different from 24 h, indicating that they were well Fig. 3). Acomparable analysis of the afternoon compo- entrained by the 24-h LD cycle. Slopes of these 2 ham- nent was not undertaken, because it seldom remained sters, moreover, deviated only slightly from 24 h (0.04 distinct for more than 3 days in DD. Abimodal activity and 0.05 h/day, respectively). Likewise, after NWR pattern persisted beyond the fifth day of DD, after the the afternoon activity component of split hamsters disappearance of a robust, clearly distinguishable was well entrained under LDLD with only 3 out of 18 afternoon component. However, this bimodality split hamsters producing activity onsets with slopes appeared to be indistinguishable from that character- significantly different from 24 h. Activity onsets were istic of unsplit hamsters with comparable α in DD. In less well entrained for the nighttime component of the other words, by this time the bimodal pattern does not split rhythms. The majority of hamsters exhibited τ*s suggest persistent splitting. A quantitative descrip- significantly greater than 24 h. Only 3 hamsters had Gorman and Lee / SPLIT ACTIVITY RHYTHMS IN HAMSTERS 545

Figure 1. Representative double-plotted actograms of hamsters initially entrained to LD 14:10 and later exposed to daily afternoon novel wheel running (NWR). Data are unclipped and scaled between zero and the hamster’s maximum activity count in the interval depicted. Dark rectangles above actograms represent the 10-h scotophase maintained throughout the experiment. Hatched rectangles above and on the right side of actograms represent NWRtreatment paired with darkness. This second daily scotophase was maintained for 2 final days when hamsters remained in their home cage.

tion (Table 2; p < 0.001), but the two components did not differ from one another (p > 0.05). ψ Phase angle to lights-off ( lights-off) differed significantly from the unsplit to the split state and between ψ the two split components (Table 2). lights-off was significantly less negative in the baseline condition prior to NWR than in either the afternoon (p < 0.05) or nighttime (p < 0.001) activity components in LDLD. In the ψ split condition, lights-off was far more negative for the nighttime activity component than for the afternoon ψ component. Relative to lights-on ( lights-on), the phase angle of the afternoon activity component was greater than that of the nighttime activity component (p < 0.05; Table 2). Phase angle of the two split components rela- Figure 2. Mean ± SEM activity onsets (clock hour) of hamsters ψ tive to each other ( n–a) varied from 8.40 to 12.24 h (n = 18-20) exposed to successive days of novel wheel running (NWR) in Experiments 1-3. Onsets for 2 days prior to NWR are (mean = 10.5 ± 0.22 h). Total activity was nearly designated “Pre-NWR.” Hamsters were exposed to 11, 7, and 6 equally distributed between nighttime (55% ± 2%) days of NWRin Experiments 1-3, respectively. In Experiment 3, and afternoon (45% ± 2%) scotophases. activity onsets for 5 days after NWRwas discontinued are plotted for comparison with values in Experiment 1. The arrow indicates when the Experiment 3 hamsters remained thereafter in the home cage without further NWR. DISCUSSION

τ*s not significantly different from 24 h. Finally, qual- In three separate experiments, daily NWR mark- ity of entrainment was further assessed by quantify- edly reorganized the locomotor activity rhythms of ing the sum of squared residuals of actual onsets from male Syrian hamsters maintained in an LD cycle. the best-fit regression line. Variability of both split Nighttime activity onset was progressively delayed components was greater than for the baseline condi- with subsequent days of NWR: Whereas nighttime 546 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Figure 3. Representative actograms of hamsters from Experiment 2. The asterisk on the actogram indicates beginning of exposure to constant darkness (DD). Slanted lines on right side of the actogram represent least-squares regression lines for nighttime activity onsets on Days 1-4 and 8-11 of exposure to DD. Other conventions as in Figure 1. NWR = novel wheel running.

Table 1. Circadian rhythm patterns of split hamsters (n = 8) transferred to DD in Experiment 2. Noted are significant differences over time (repeated measures ANOVA, two-tailed).

Day 1 Day 2 Day 3

Ψ Phase angle between components ( n–a)(h) 11.6 ± 0.3 10.5 ± 0.3 9.4 ± 0.5 p < 0.005 Inactive interval (h) 8.9 ± 0.4 7.8 ± 0.5 5.9 ± 0.5 p < 0.0001 Nighttime bout Activity onset (h) 1.02 ± 0.18 1.92 ± 0.20 3.21 ± 0.25 p < 0.0001 Bout duration (α)(h) 2.70 ± 0.18 2.68 ± 0.33 3.45 ± 0.82 ns Wheel revolutions 749 ± 61 913 ± 124 1190 ± 274 ns Afternoon bout Activity onset (h) 12.60 ± 0.34 12.40 ± 0.36 12.56 ± 0.52 ns Bout duration (α)(h) 2.68 ± 0.40 3.25 ± 0.42 2.55 ± 0.45 ns Wheel revolutions 778 ± 110 1031 ± 132 851 ± 135 p < 0.01

days 1 - 4 days 8 - 11 τ()h 25.. 09±± 0 15 24 .. 58 0 06 p < 0.01

activity disappeared entirely in some hamsters after period. Subsequently,when the system is released into 11 days of NWR, more modest delays were observed DD, the two bouts fuse or rejoin under the influence of after 6 to 7 days of NWR. This latter condition was strong oscillator interactions, but alternatively may be associated with distinctly split activity rhythms that effectively entrained by an LDLD cycle. NWR can rejoined after several days of DD. Perpetuation of the therefore override typical entrainment patterns estab- LDLD cycle, however, allowed the split rhythms to be lished in an LDLD cycle and reorganize activity into a sustained for at least an additional 11 days in the home second stable configuration. cage. In the absence of NWR, exposure to the LDLD It is not clear why others have failed to replicate the cycle had no marked effect on nighttime locomotor induction of splitting with afternoon NWR (Sinclair activity rhythms and yielded no evidence of splitting. and Mistlberger, 1997) and why NWR induced larger As suggested previously (Mrosovsky and Janik, 1993), phase-delays of home cage activity onset in this study afternoon NWR phase-shifts some component circa- than in others (Mrosovsky and Janik, 1993). We used dian oscillators, which thereafter give rise to the the same hamster supplier as the original report, in expression of a new activity bout in the afternoon dark contrast to the study with largely negative effects. Gorman and Lee / SPLIT ACTIVITY RHYTHMS IN HAMSTERS 547

Figure 4. Representative actograms of hamsters from Experiment 3. Conventions as in Figures 1 and 3. NWR = novel wheel running.

Table 2. Entrainment parameters of hamsters expressed in base- ened (Mrosovsky, 1993; Weisgerber et al., 1997). More- line LD 14:10 and following splitting under the experimental LDLD over, at the conclusion of NWR in the present study, 6:5:3:10 cycle of Experiment 3. the nighttime activity component free-ran in DD with Baseline Nighttime Afternoon N a long τ. Together, these results suggest an enduring τ τ* (h) 24.00 ± 0.01 24.16 ± 0.04 23.99 ± 0.02 18 effect on as opposed to a transient (e.g., phase shift) SS residuals 1.49 × 10–4 7.90 × 10–4 7.65 × 10–4 18 effect of NWR on the circadian pacemaker. Ψ lights-off (h) –0.83 ± 0.02 –6.79 ± 0.22 –1.26 ± 0.05 18 Two factors may dictate the pattern of rejoining Ψ lights-on (h) 9.17 ± 0.02 3.21 ± 0.22 3.74 ± 0.05 18 observed in DD, which in all cases was achieved via reduction of the inactive interval following nighttime activity and preceding afternoon activity. First, inde- Minor differences in intrinsic periods, propensity to pendent of any oscillator interactions, the two compo- run in novel wheels, or oscillator coupling may distin- nents may have different intrinsic free-running peri- guish splitting and nonsplitting strains. We also used ods, which would favor rejoining. The large negative longer exposures to NWR (5 h vs. 3 h) than used in pre- phase angle of the nighttime component and the rela- vious studies. Notably, running in novel wheels was tively small negative value for the afternoon compo- most intense during the 4th hour of exposure. Regard- nent suggest free-running periods, which are >24 and less of differences between studies, the significance of <24 h for the nighttime and afternoon components, this experimental paradigm is as a probe of specific respectively. Alternatively, coupling interactions aspects of oscillator function, which likely differ quan- between oscillator components may favor the titatively rather than qualitatively among different observed pattern of rejoining regardless of the periods hamster strains and experimental conditions. of the free-running rhythms. That is, the split state Various formal mechanisms may account for the may be intrinsically unstable, and oscillators may progressive delays in nighttime activity onset during interact in DD to establish a limited range of phase successive days of NWR. Amore negative phase angle angles with respect to each other. Oscillator interac- of entrainment as was obtained in each experiment tions have been invoked to understand the limited can result from a lengthening of τ. Alternatively, each decompression of α, which may be obtained in DD in day of NWR may induce a single phase-delay in activ- unsplit hamsters. The pattern exhibited in this experi- ity onset without lengthening τ. In prior studies ment is consistent with oscillators recoupling by the employing a single bout of NWR in early subjective shortest possible route (i.e., reducing the shorter of the Ψ Ψ afternoon of hamsters in DD, however, activity onset two respective phase angles, n–a versus a–n between was advanced rather than delayed, and τ was length- them), although this proposition cannot be evaluated 548 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001 against alternatives with the present data set. A role of angle as a result of the oscillator’s lengthened τ. oscillator interactions is further suggested by the Whether NWR selectively uncouples short-period marked change in τ measured from the nighttime oscillators because of a particular anatomical relation- component after several days in DD, when splitting is ship (e.g., such oscillators receive neuropeptide Y pro- presumably ended. jections) or a temporal relationship (e.g., a particular The results of Experiment 3 complement those of phase angle between NWR and short-period oscilla- Boulos and Morin (1985) who, with daily dark pulses, tors) is entirely unknown. entrained activity rhythms split by LL. In that study, Additionally, NWR apparently phase-shifts con- one component roughly coincided with a daily 2-h stituent oscillators just as it phase-shifts without dark phase, while the second activity component per- overtly splitting the pacemaker in two other para- sisted 8 to 12 h out of phase with the dark-entrained digms: After an 8-h phase-advance of the LD cycle, component. In the current study, it appears that the hamsters running in novel wheels during the new ZT split nighttime activity component, which free-runs in 13-16 re-entrained within one to two cycles, whereas DD with τ > 24 h, may be entrained solely by the nonrunning controls required several days (Mrosovsky phase-advancing effects of light onset at ZT 22. In con- and Salmon, 1987). NWR of sufficient duration begin- trast, the split afternoon activity component may be ning at ZT 5, moreover, induced rapid phase shifts in entrained by either phase-delaying effects of light excess of8hinsome hamsters transferred to DD prior to lights-off at ZT 4, by phase-advancing effects (Gannon and Rea, 1995). The present paradigm like- of lights-on at ZT 9, or by both. In hamsters split by LL, wise induces phase shifts, albeit of only a fraction of each component of the activity rhythm expresses a the oscillators formerly generating the nighttime PRC to dark pulses with defined regions of delays and activity.The present data strongly suggest that succes- advances (Boulos and Rusak, 1982). sive days of NWR recruit cohorts of oscillators to How does NWR split circadian activity rhythms? express their subjective night in the afternoon Convergent evidence from cellular, physiological, scotophase. After 6 to 7 days of NWR, activity is nearly behavioral, and mathematical paradigms (e.g., equally divided between nighttime and afternoon Illnerova, 1991; Liu et al., 1997; Pittendrigh and Daan, scotophases, whereas the afternoon scotophase con- 1976; Enright, 1980) points to the following model of tains disproportionate activity (and in some cases all) the circadian pacemaker: Overt circadian rhythms after 11 days of NWR. We suggest that a single day of reflect the output of multiple circadian oscillators that NWR produces a large phase shift of a small fraction of constitute a coupled dual oscillatory system, which component oscillators. With additional days of NWR, may be functionally described in terms of evening and a threshold fraction of oscillators may be phase- morning oscillators (Fig. 5A). A functional evening shifted to generate an activity component in the after- oscillator results from the coupling of oscillators with noon scotophase. The progressive delays in nighttime relatively short τs, and as such, its overall τ is<24h, activity onset are consistent with this model. whereas a functional morning oscillator is derived In contrast to other paradigms used (Gannon and from coupling of longer period constituent oscillators Rea, 1995; Mrosovsky and Salmon, 1987), the presence with τ > 24 h. As one effect of NWR at ZT 4 appears to of an LDLD cycle with a short (5 h) second scotophase be a marked lengthening of the period of the nighttime may prevent the entire complement of oscillators from activity component (Mrosovsky, 1993; Weisgerber being phase-shifted to express subjective night in the et al., 1997), and because NWR delayed nighttime afternoon. Moreover, in the absence of further NWR, activity onset (present studies), we hypothesize that the light pulses that bracket the afternoon scotophase early afternoon NWR preferentially lengthens the impede the recoupling of constituent oscillators back period of the oscillators underlying nighttime activity, into the unsplit state (Experiment 3), which so readily perhaps by uncoupling some of the short-period com- occurs in DD. Notably, when the intervening light ponent “evening” oscillators from the coupled oscilla- intervals were very short as in skeleton photoperiods, tor network that generates normal nighttime activity daily NWR from ZT 5 to ZT 8 induced complete inver- (Fig. 5A). In DD, the larger coupled system might sion of activity rhythms to what was previously sub- therefore free-run under the influence of its remaining jective day (Sinclair and Mistlberger, 1997). Similarly, coupled constituent oscillators, of which those with 11 days of NWR apparently also shifted the entire longer intrinsic τs predominate. Under entraining LD oscillatory system in a few hamsters of Experiment 1. conditions, one would expect a more negative phase Thus, these two factors—a titratable shifting of oscilla- Gorman and Lee / SPLIT ACTIVITY RHYTHMS IN HAMSTERS 549

Figure 5. Formal model of multioscillator basis of novel wheel running (NWR) effects on circadian rhythms. In all panels, the inverted U shape represents subjective night of hypothetical individual circadian oscillators. For clarity, the remaining phases of each oscillator are not depicted. In A, the integrated rhythm reflects a distribution through the scotophase of the subjective nights of individual oscillators. Those with short and long periods express subjective night early and late in the scotophase, respectively. The oscillator denoted with a dashed line will be phase-shifted by NWR. After a single day of NWR from ZT 4 to ZT 9, oscillators with short periods undergo large phase shifts to roughly the same circadian phase as NWR. Light pulses bracketing NWR preclude oscillator recoupling for reasons that are not yet clear. Subsequent activity onset is delayed, and τ is lengthened (τ*>τ). Additional days of NWRphase-shift additional cohorts of oscilla - tors such that approximately half are phase-shifted after 6 to 7 days but all or nearly all are shifted after 11 or more days. B. Application of the splitting model to understand effect of a single day of NWRapplied in DD. In the absence of a light pulse following NWR,the phase-shifted oscillator pulls (advances) the nighttime component that follows. Until full recoupling is achieved, the nighttime activity component may continue to express a lengthened τ. tors and countering of oscillator tendencies to rejoin— may facilitate induction and maintenance of stable, single bout of NWR is not sufficient to induce measur- split activity rhythms under an LDLD cycle and 1 or able splitting under LDLD conditions (MR Gorman, more days of NWR. unpublished observations), but if it selectively phase- Beyond the significance of NWR-induced splitting shifts a small fraction of circadian oscillators (Fig. 5B), noted previously (Mrosovsky and Janik, 1993), the overt circadian rhythms might be altered as a conse- demonstration that some oscillatory components can quence of recoupling dynamics among constituent be dissociated from others and rephased with respect oscillators. Unfortunately, little is known about these to the nighttime activity component may provide processes, except that darkness favors recoupling in insight into the mechanism of phase-shifting effects of LL-split hamsters (Earnest and Turek, 1982) and inter- NWR, or indeed, of any photic or nonphotic zeitgeber. vening light pulses appear to minimize recoupling in For example, a single bout of NWR around ZT 5 pro- NWR-induced splitting (Experiment 3). Actograms of duces large phase advances and lengthening of τ in NWR-split hamsters suggest that recoupling may be subsequent DD (Mrosovsky, 1991; Reebs and accompanied by abrupt changes in phase, with the Mrosovsky, 1989a, 1989b; Weisgerber et al., 1997). A rejoining activity bout typically expressed in an inter- 550 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001 mediate phase between the former two components Bruce VG (1960) Environmental entrainment of circadian (Mrosovsky and Janik, 1993). If no light follows a bout rhythms. Cold Spring Harb Symp Quant Biol 25:29-48. Earnest DJ and Turek FW (1982) Splitting of the circadian of NWR, strong oscillator interactions may result in rhythm of activity in hamsters: Effects of exposure to con- rapid recoupling via reduction of the shorter phase stant darkness and subsequent re-exposure to constant angle between component oscillators. This would light. J Comp Physiol [A] 145:405-411. advance rather than delay the main activity onset Elliott JA and Tamarkin L (1994) Complex circadian regula- (Fig. 5B). Consistent with this general model, light tion of pineal melatonin and wheel-running in Syrian pulses shortly after a single day of NWR, such as those hamsters. J Comp Physiol [A] 174:469-484. Enright JT (1980) Timing of Sleep and Wakefulness, Springer- that impeded oscillator recoupling in the present Verlag, New York. study, greatly attenuated the phase-advancing effects Gannon RL and Rea MA (1995) Twelve-hour phase shifts of of NWR (Mrosovsky, 1991). hamster circadian rhythms elicited by voluntary wheel An understanding of oscillator-oscillator interac- running. J Biol Rhythms 10:196-210. tions has lagged behind our knowledge of other fea- Gorman MR, Freeman DA, and Zucker I (1997) Photo- periodism in hamsters: Abrupt versus gradual changes tures of circadian rhythms, although physiological in day length differentially entrain morning and evening data suggest that such interactions must be central to circadian oscillators. J Biol Rhythms 12:122-135. an understanding of the pacemaker. Only a fraction of Gorman MR, Yellon SM, and Lee TM (2001) Temporal reor- SCN cells, for instance, receive direct retinal or IGL ganization of the suprachiasmatic nuclei in hamsters projections, and yet both of these pathways are capa- with split circadian rhythms. J Biol Rhythms 16:552-563. ble of phase-shifting, presumably,the entire SCN. Any Hoffmann K (1971) Splitting of the circadian rhythm as a function of light intensity. In Biochronometry, M Menaker, zeitgeber, therefore, first shifts a subpopulation of ed, pp 134-146, National Academy of Sciences, Washing- SCN cells that receives direct projections from the ton, DC. time-giving entrainment mechanism. These cells in Illnerova H (1991) The suprachiasmatic nucleus and rhyth- turn interact with the greater complement of SCN cells mic pineal melatonin production. In Suprachiasmatic to arrive at a steady-state phase shift. Under routine Nucleus: The Mind’s Clock, DC Klein, RY Moore, and SM conditions, this selective shifting and oscillator inter- Reppert, eds, pp 197-216, Oxford University Press, New York. action may happen in a cycle or even more rapidly. Liu C, Weaver DR, Strogatz SH, and Reppert SM (1997) Cel- The use of LDLD cycles in the present paradigm, in lular construction of a circadian clock: Period determina- contrast, facilitates a temporal dissociation of these tion in the suprachiasmatic nuclei. Cell 91:855-860. processes by impeding the recoupling process. Mrosovsky N (1991) Double-pulse experiments with nonphotic and photic phase-shifting stimuli. J Biol Rhythms 6:167-179. Mrosovsky N (1993) τ changes after single nonphotic events. ACKNOWLEDGMENTS Chronobiol Int 10:271-276. Mrosovsky N and Janik DS (1993) Behavioral decoupling of We thank Jim Donner for excellent animal care and circadian rhythms. J Biol Rhythms 8:57-65. Jeff Elliott for helpful comments on an earlier draft of Mrosovsky N and Salmon PA (1987) A behavioural method for accelerating re-entrainment of rhythms to new this manuscript. This research was supported by light-dark cycles. Nature 330:372-373. NHLBI grant HL61667 to TML and NICHD-48640 to Pittendrigh CS (1960) Circadian rhythms and the circadian MRG. organization of living systems. Cold Spring Harb Symp Quant Biol 25:159-184. Pittendrigh CS and Daan S (1976) A functional analysis of REFERENCES circadian pacemakers in nocturnal rodents: V.Pacemaker structure: A clock for all seasons. J Comp Physiol [A] 106:333-355. Aschoff J (1965) Circadian rhythms in man. Science Reebs SG and Mrosovsky N (1989a) Effects of induced wheel 148:1427-1432. running on the circadian activity rhythms of Syrian ham- Boulos Z and Morin LP (1985) Entrainment of split circadian sters: Entrainment and phase response curve. J Biol activity rhythms in hamsters. J Biol Rhythms 1:1-15. Rhythms 4:39-48. Boulos Z and Rusak B (1982) Phase-response curves and the Reebs SG and Mrosovsky N (1989b) Large phase-shifts of dual-oscillator model of circadian pacemakers. In Verte- circadian rhythms caused by induced running in a brate Circadian Systems, J Aschoff, S Daan, and G Groos, re-entrainment paradigm: The role of pulse duration and eds, pp 215-223, Springer-Verlag, Berlin. light. J Comp Physiol [A] 165:819-825. Gorman and Lee / SPLIT ACTIVITY RHYTHMS IN HAMSTERS 551

Sinclair SV and Mistlberger RE(1997) Scheduled activity Weisgerber D, Redlin U, and Mrosovsky N (1997) Lengthen- reorganizes circadian phase of Syrian hamsters under ing of circadian period in hamsters by novelty-induced full and skeleton photoperiods. Behav Brain Res wheel running. Physiol Behav 62:759-765. 87:127-137. GormanJOURNAL et al. OF / BIOLOGICAL SPLIT CIRCADIAN RHYTHMS RHYTHMS / December IN HAMSTERS 2001 Temporal Reorganization of the Suprachiasmatic Nuclei in Hamsters with Split Circadian Rhythms

Michael R. Gorman,*,1 Steven M. Yellon,† and Theresa M. Lee* *Department of Psychology and Reproductive Sciences Program, University of Michigan, Ann Arbor, MI 48109-1109, USA, †Center for Perinatal Biology, Loma Linda University School of Medicine, Loma Linda, CA 92350, USA

Abstract A dual oscillator basis for mammalian circadian rhythms is suggested by the splitting of activity rhythms into two components in constant light and by the photoperiodic control of pineal melatonin secretion and phase-resetting effects of light. Because splitting and photoperiodism depend on incompatible environmental conditions, however, these literatures have remained distinct. The refinement of a procedure for splitting hamster rhythms in a 24-h light- dark:light-dark cycle has enabled the authors to assess the ability of each of two circadian oscillators to initiate melatonin secretion and to respond to light pulses with behavioral phase shifting and induction of Fos-immunoreactivity in the suprachiasmatic nuclei (SCN). Hamsters exposed to a regimen of afternoon novel wheel running (NWR) split their circadian rhythms into two distinct components, dividing their activity between the latter half of the night and the afternoon dark period previously associated with NWR. Plasma melatonin concentrations were elevated during both activity bouts of split hamsters but were not elevated during the afternoon period in unsplit controls. Light pulses delivered during either the nighttime or afternoon activity bout caused that activity component to phase-delay on subsequent days and induced robust expression of Fos-immunoreactivity in the SCN. Light pulses during intervening periods of locomotor inactivity were ineffective. The authors propose that NWR splits the circadian pacemaker into two distinct oscillatory components separated by approximately 180 degrees, with each expressing a short subjective night.

Key words novel wheel running, melatonin, c-Fos, phase-shift, oscillator interaction

The suprachiasmatic nucleus (SCN) of the hypo- oscillator-oscillator interactions would appear to be a thalamus is the principal circadian pacemaker in central issue in understanding the mammalian pace- mammals (Weaver, 1998). Although individual cul- maker. A dual-oscillator basis for mammalian circa- tured SCN cells express circadian rhythms with vary- dian rhythms was posited following the observation ing free-running periods and phases (Herzog et al., that prolonged exposure to constant light (LL) 1997; Liu et al., 1997; Welsh et al., 1995), in vivo cellular induced locomotor activity patterns of hamsters and activity from the SCN is largely synchronized other species to diverge into two components (Bouskila and Dudek, 1993). Thus, the question of (Pittendrigh and Daan, 1976). The SCN of split ham-

552 Gorman et al. / SPLIT CIRCADIAN RHYTHMS IN HAMSTERS 553 sters express comparable split rhythms in multiunit activity rhythms could be reliably split and that com- electrical activity, suggesting that splitting is intrinsic ponents of the split rhythms could be entrained to a to the principal circadian pacemaker (Mason, 1991). light-dark:light-dark (LDLD) cycle. Studies of pineal melatonin synthesis and photo- Entrainment of NWR-induced split activity rhythms periodic regulation of activity duration (α) likewise to an LDLD cycle indirectly suggests that each of the point to a dual oscillator model of circadian rhythms two oscillators mediating the split rhythms can be sep- (Elliott and Tamarkin, 1994; Gorman et al., 1997; arately phase-shifted by light. To assess this directly, Illnerova, 1991; Pittendrigh and Daan, 1976). An eve- and to rule out the possibility that extra-SCN oscilla- ning oscillator (E), entrained by evening light, is pur- tors are recruited to mediate one or both activity com- ported to initiate nocturnal melatonin secretion and ponents (Honma et al., 1989; Stephan et al., 1979), we locomotor activity, whereas a morning oscillator (M), characterized the effects of light pulses on locomotor entrained by morning light, mediates the expression activity rhythms and on induction of Fos immuno- of these functions near the end of night. In longer day reactivity in the SCN during each of the two split activ- lengths of spring and summer, Eand M are entrained ity bouts induced by NWR. Additionally, we pre- in a phase relation that generates a short subjective sented light pulses during the inactive periods night, while short day lengths of fall and winter allow between the activity components to exclude the possi- the phase angle between Eand M to increase to gener - bility that α was simply lengthened to incorporate ate a long subjective night. Parallel photoperiodic both activity components. Finally, models of photo- modulation of light-induced phase-resetting of loco- periodic control of melatonin secretion have posited motor rhythms, as well as of expression of the imme- the existence of distinct evening and morning oscilla- diate early gene c-Fos in the SCN, suggests that com- tors timing the initiation and termination, respec- ponent oscillators are localized within the SCN (Elliott tively, of the nightly pattern of elevated pineal and Kripke, 1998; Pittendrigh et al., 1984; Sumová melatonin secretion (Illnerova, 1991). Because light et al., 1995), although identification of two distinct acutely suppresses melatonin secretion, it has not oscillators has been lacking. Unfortunately, the clear been possible to evaluate the role of component oscil- temporal resolution of distinct oscillators achieved in lators split in LL. We therefore assessed whether each LL-induced splitting is not seen in the photo- of the two split components was capable of initiating periodism literature, where, with rare exceptions melatonin secretion. (Jagota et al., 2000), evening and morning oscillator functions may greatly overlap. Conversely, notwithstanding the theoretical milestone it represents, the MATERIALS AND METHODS splitting paradigm has been of limited utility for probing the nature of underlying oscillators largely Animals and Husbandry because its requirement for LL complicates application of the bulk of analytical techniques fruitfully For all experiments except one, the same 24 male employed in photoperiodism research. Specifically, Syrian hamsters (Mesocricetus auratus, HsdHan: LL masks locomotor activity, inhibits melatonin secre- AURA, Harlan, Indianapolis, IN, USA), 5 to 6 weeks of tion, and precludes assessment of acute effects of light age at acquisition, were housed with Sani-Chip bed- pulses on pacemaker function. Thus, these two theo- ding in polypropylene cages (48 × 27 × 20 cm) retically important literatures have remained largely equipped with Nalgene (d = 34 cm) running wheels separate. (Fisher Scientific, Pittsburgh, PA, USA). Food (Purina Mrosovsky and Janik (1993) suggested that LL is Rodent Chow #5001, St. Louis, MO, USA) and water not the only stimulus capable of splitting rhythms in were available ad libitum. These hamsters are the the hamster: Repeated afternoon exposures to 3 h of same as used in another separately reported study novel wheel running (NWR) delayed the onset of published in this issue (Gorman and Lee, 2001). Eigh- nighttime wheel running and, in some cases, led to a teen additional hamsters from the same supplier and transient splitting of locomotor activity rhythms housed similarly were used only for the final study of when hamsters were later transferred to constant Fos-immunoreactivity in the SCN. Hamsters were darkness (DD). Modifying this protocol, Gorman and housed for 2 to 3 weeks in a 14 h light, 10 h dark condi- Lee (2001 [this issue]) recently demonstrated that tion (LD 14:10; lights on 0500-1900 h) before each regi- 554 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001 men of daily NWR was used to split activity rhythms. Afternoon Light Pulse (CT 13-a) Room illumination at the level of the cage lid varied from 100 to 300 lux. In the first run, lights remained off after the night- Split activity rhythms were induced by scheduled time scotophase following the final day of NWR exposures to NWR as previously described (Gorman (Fig. 1A). Activity rhythms were monitored remotely and Lee, 2001). Briefly, after entrainment to LD 14:10, for onset of wheel running the following afternoon. 20 hamsters were transferred daily for 6 to 10 days One hour after onset of this afternoon component of within the same room to Wahmann wheels 0 to 15 min the split activity rhythm, 9 randomly selected ham- before lights were extinguished at 1100 h (ZT 4) EST. sters were moved in their home cages to an adjoining At 1600 h (ZT 9), the lights were turned on and ham- room where they experienced the 15-min light pulse. sters were returned to home cages in the light over the Unpulsed split control animals (n = 10) were similarly next 15 min. With additional days of NWR, onset of jostled in their home cages but remained in the same nighttime locomotor activity in the home cage is pro- room in darkness. Following the light pulse, animals gressively delayed. The number of days of NWR were returned to the housing room and remained within each experimental run was adjusted so that undisturbed for 2 weeks. nighttime activity onset occurred approximately midway through the 10-h nighttime scotophase. After the Evening Light Pulse between final day of NWR, hamsters were left in their home Split Activity Components cages and exposed to DD by leaving lights off after the subsequent 10-h nighttime scotophase. Nonsplitting Hamsters were next re-entrained to LD 14:10 for control hamsters (n = 4) housed in the same room 2 weeks and then exposed to an additional regimen of experienced identical lighting conditions but were not NWR. Following the final day of NWR, lights were exposed to NWR and thus never split their activity turned off2hprematurely (1700 EST) to minimize rhythms. potential phase-shifting effects of the entraining Wheel-running activity in the home cage was mon- photoperiod and to unmask activity onset in unsplit itored by Dataquest III software (Mini-mitter, Sun control hamsters. Of the split hamsters, 10 were ran- River, OR, USA) and compiled into 10-min bins. domly selected and pulsed with light for 15 min begin- Activity onset (CT 12) was defined as the first bin in ning approximately 1 h after unsplit animals began each activity bout with wheel revolutions of more their normal nighttime activity (n = 4). The remaining than 20, and that was immediately followed by a sec- split hamsters (n = 9) served as unpulsed controls: they ond interval exceeding this threshold. While in the were similarly jostled at the designated time but not novel wheels, activity patterns were not monitored, exposed to light. but the total number of wheel revolutions after the 5-h interval was recorded manually. Data analyses were Nighttime Light Pulse (CT 13-n) carried out with Excel (Microsoft, Seattle, WA, USA) and ClockLab software (Actimetrics, Evanston, IL, Hamsters were again re-entrained to LD 14:10 for USA). 2 weeks and then induced to split their rhythms with NWR. Following the final day of NWR, hamsters remained in their home cages, and lights remained off Light Pulse–Induced Phase Shifts of Activity after the subsequent nighttime scotophase. Activity was monitored to verify that hamsters exhibited a Changes in the phase response to light following nighttime activity bout followed by an interval of NWR-induced splitting were assessed in three inactivity and then an afternoon activity bout. Ham- sequential experimental iterations (begun at approxi- sters were pulsed with light for 15 min (n = 9) or sham mately 12, 18, and 24 weeks of age, respectively). The pulsed (n = 8) as described above beginning 1 h after same cohort of hamsters was repeatedly first the next nighttime activity onset. entrained to LD 14:10, then exposed to a regimen of NWR to split activity, and then transferred to DD Analyses of Phase-Shift Data (Aschoff Type II methodology) where they received either a 15-min light pulse or a sham pulse (~450 lux at Because split circadian rhythms are unstable in DD the level of the cage lid). (i.e., components re-fuse rapidly and τ and phase are Gorman et al. / SPLIT CIRCADIAN RHYTHMS IN HAMSTERS 555

the experiment progressed, fragmentation of the activity rhythms occasionally precluded identification of a single clear afternoon activity component. In such cases, this component was not analyzed. Last, the free-running period of the nocturnal activity component was computed using least squares regression over the 4 days following the pulse or sham pulse. Because the afternoon component typically rejoined with the nighttime component within this interval, a comparable analysis of the free-running period of the afternoon component was not attempted.

Melatonin in Circulation

To assess whether melatonin secretion patterns were altered by NWR-induced splitting, hamsters were lightly anesthetized with methoxyflurane vapors (Metofane) and retro-orbitally bled in darkness with the aid of a dim red light (< 1 lux). Blood samples were collected into ethylene diamine tetraacetic acid–treated tubes. Plasma was harvested after centrifugation at 5000 rpm for 20 min and was stored at –70 °C until assay. Plasma samples were thawed and extracted with dichloromethane, and melatonin concentrations were determined in a single Figure 1. Schematic representation of times of experimental assay as previously described (Bae et al., 1999). The manipulations in hamsters with split activity rhythms and in intra-assay coefficient of variation was 14%, and assay unsplit controls. Each horizontal bar represents a 24-h period, with dark rectangles reflecting times of lights-off. Hatched rect- sensitivity was 17 pg/ml. Samples below the limit of angles indicate times of novel wheel running (NWR) in darkness. detectability were assigned values of 17 pg/ml for One to 2 days following the final day of NWRare plotted below. purposes of graphing and statistics. One to 3 days prior Superimposed white dumbbell shapes represent characteristic to each of the three runs of NWR described above, times of wheel-running activity in DD. Asterisks indicate time of light pulse/sham pulse for determination of behavioral blood samples were collected at ZT 16, ZT 20, and ZT phase-shifting (A), blood sampling (B), or brain collection for 21 from pseudo-randomly selected unsplit hamsters SCN Fos-ir (C). In (A), the gray bar indicates that lights were extin- (n = 3-5/time point). After the 6th day of NWR of each guished 2 h early for assessment of the earliest light pulses. experimental run, additional samples were collected at these same times or at ZT 9 from hamsters that were in flux), establishment of a prepulse baseline and cal- then split (n = 3-6/time point; Fig. 1B). Among culation of a light-induced phase shift in an individual unsplit control animals that remained at home, sam- animal are highly problematic. Therefore, phase shifts pling was performed at ZT 9 on the 6th day of the were assessed by comparing light-pulsed and afternoon dark pulse of each run (Fig. 1B). No hamster sham-pulsed animals in a between-subjects design. contributed more than one sample at any given time For each animal, nighttime and afternoon activity point, except non-NWR controls at ZT 9. As melatonin onset of the day preceding the light-pulse or concentrations were always undetectable at ZT 9 for sham-pulse were compared with the corresponding these 4 hamsters, only one determination from each activity onsets on the 2 days following the stimulus hamster was considered in statistical analyses. (see Fig. 2). Timing of the light pulse was such that we could not obtain a prepulse afternoon activity onset to SCN Fos-Immunoreactivity use as a phase reference in the second experimental run. Hence, phase shifts of this component were not Finally, we assessed whether the temporal pattern determined. Moreover, activity onsets of the night- of light-induced Fos expression in the SCN was split time component were always unambiguous, but as following exposure to daily NWR. Following the 556 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Figure 2. Representative doubled-plotted actograms (left) of hamsters split by novel wheel running (NWR), transferred to DD and pulsed with light (A) or sham-pulsed (B) 1 h after the afternoon activity onset. Data are unclipped and are scaled between zero and maximum revolutions for that hamster. The light-dark cycle prior to transfer to DD is represented above the actogram, with hatched areas representing times of NWRin darkness. Days of NWRare represented by a similar rectangle superimposed upon the right half of the actogram. After NWR, hamsters entered DD beginning during the nighttime scotophase. The first 3 days of DD exposure are enlarged and singly plotted (right) to illustrate method of phase-shift calculation. Double-headed arrows represent Day 1 and Day 2 phase shifts of each component. Onsets of nighttime and afternoon activity on the 1st day in DD served as reference points from which to measure phase shifts. The difference between initial and subsequent activity onsets was measured for each component (net phase shift). behavioral phase-shifting experiment, hamsters (now assigned to receive a 15-min light pulse or sham pulse 38 weeks of age) were re-entrained to LD 14:10 and at one of three time points (Fig. 1C; n = 3-6/group): 1 h exposed to 6 days of NWR. Areplicate experiment was after the subsequent nighttime activity onset (CT subsequently performed with 18 additional hamsters 13-n), during the morning inactive period (break) fol- at 12 weeks of age. For both cohorts, hamsters lowing nighttime activity (2-4 h after nighttime activ- remained at home and lights remained off after the ity offset), and 1 h after the subsequent afternoon night that followed the 6th day of NWR (Fig. 1C). activity onset (CT 13-a). Unsplit control hamsters were Hamsters with split activity patterns were randomly also left in DD and were exposed to light or sham- Gorman et al. / SPLIT CIRCADIAN RHYTHMS IN HAMSTERS 557 pulsed the next afternoon at the same time when split controls on Day 1 (p < 0.05; Figs. 2A, 3A), with a trend hamsters were given their light pulses. Additional in the same direction on Day 2 (p < 0.10). The nighttime unsplit control hamsters received light pulses or sham activity component, which was not itself pulsed, was pulses 1 h after the following nighttime activity onset. significantly phase-advanced by this afternoon light Hamsters were injected with a lethal dose of pulse (p < 0.05 for both days, Fig. 3A). The free-running Nembutal 55 min after the light pulse or sham pulse period of the nighttime component was unaffected by and perfused intracardially with 40 to 60 ml 0.1M the light pulse (mean τ ± SE: 25.08 ± 0.12 vs. 24.80 phosphate buffered saline (PBS; pH 7.5) followed by ± 0.18 h for unpulsed and pulsed animals, respec- 100 to 150 ml 4% paraformaldehyde in PBS. Brains tively; p > 0.20). were postfixed in paraformaldehyde at room temperature for 2 h and transferred to 20% sucrose/PBS. Evening Light Pulse between Serial coronal slices were cut at 40 µm and every fourth Split Activity Components section processed for immunocytochemistry. Free- floating sections were incubated successively in 0.1% Early evening light pulses (matched to CT 13 of hydrogen peroxide; 1:1000 anti-fos rabbit IgG (sc-52, unsplit control hamsters) had no significant effect on Santa Cruz, CA, USA) in PBS with 0.4% Triton- X the nighttime activity component of split hamsters (Fisher) and 4% normal goat serum for 24 h at 4 °C; (Fig. 3B). The free-running period of this component 1:200 biotinylated goat anti-rabbit secondary anti- also was not affected by the light pulse (25.11 ± 0.14 vs. body (Vector Labs, Burlingame, CA, USA) for1hat 25.37 ± 0.20 h for unpulsed and pulsed animals, room temperature; ABC reagent (Vector) for 1 h; and respectively; p > 0.30). 0.1% diaminobenzidine (DAB) with 0.02% peroxide for 2 min. Sections were mounted on gelatin-coated slides. The most densely stained section was selected Nighttime Light Pulse (CT 13-n) by an observer blind to experimental treatment, and the number of Fos-positive cells in the SCN was On Days 1 and 2, the nighttime activity component counted. of pulsed hamsters was significantly phase-delayed Statistical tests (all two-tailed where applicable) by the 15-min light pulse (p < 0.005, p < 0.01, respec- were performed with Statview 5.0 software (SAS Insti- tively, Fig. 3C), compared with sham light pulse con- tute, Cary, NC, USA). trols. On Days 1 and 2, afternoon activity in the split rhythm component was not significantly advanced by light pulses (p > 0.25; Fig. 3C). The free-running period RESULTS of the nighttime activity component was unaffected by the light pulse (mean τ ± SE: 24.68 ± 0.04 vs. 24.55 ± 0.11 h for unpulsed and pulsed hamsters, respectively; NWR and Splitting p > 0.28). In each induction of splitting by NWR, virtually the entire sample (>85%) of hamsters ran robustly in the Melatonin in Circulation novel wheels. A subset of the activity data during NWR has been analyzed in detail (Gorman and Lee, Prior to NWR, plasma melatonin concentrations 2001) and thus is not reported here. Figure 2 depicts were detectable in only a minority (4/12) of unsplit hamsters split by NWR, released into DD, and pulsed hamsters at ZT 16 but had increased above threshold with light (A) or sham-pulsed (B). in 9 of 10 animals by ZT 20 (Fig. 4A). One hour later (ZT 21), only 2 out of 12 remained above detectable limits. In contrast, plasma melatonin concentrations of LIGHT-PULSE INDUCED hamsters with split activity patterns in no case (0 out PHASE SHIFTS OF ACTIVITY of 10) exceeded the minimum detectable level at ZT 16, a significantly smaller proportion than among con- χ2 Afternoon Light Pulse (CT 13-a) trols at this time ( = 4.1; p < 0.05; Fig. 4B). By ZT 21, however, plasma melatonin had increased in 6 out of The afternoon activity component of pulsed ham- 10 hamsters, a proportion significantly greater than sters was phase-delayed relative to that of unpulsed among hamsters not exposed to NWR (χ2 = 4.4; p < 558 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Figure 3. Phase-shifting responses to 15-min light pulses delivered at three phases of the circadian cycle of split hamsters. In each panel, the mean (± SE) change in nighttime and afternoon activity onset relative to the first value in DD is depicted for light-pulsed (open bars) and unpulsed (sham) hamsters (filled bars). The net phase shift is indicated above each pair. Sample size is represented below or above bars (pulsed/unpulsed). Asterisks denote significant differences between pulsed and unpulsed animals (p < 0.05; Mann-Whitney U; two-tailed). (A) Day 1 and Day 2 phase shifts after light pulse or sham pulse delivered 1 h after onset of the afternoon activity component (CT 13-a). (B) Phase shifts after pulse or sham pulse during the evening inactive period (see text). Design limitations prevented collection of a prepulse reference value for the afternoon activity component. (C) Phase shifts after pulse or sham pulse delivered 1 h after the nighttime activity onset (CT 13-n). Gorman et al. / SPLIT CIRCADIAN RHYTHMS IN HAMSTERS 559

SCN Fos-Immunoreactivity

In hamsters with unsplit activity rhythms, the 15-min light pulse after nighttime activity onset induced significant Fos expression, compared with that in unpulsed controls (p < 0.05; Mann-Whitney U; Figs. 5A, 6A). A light pulse coinciding with the previous afternoon dark period had no stimulatory effect on the number of Fos+ cells in these unsplit hamsters (p > 0.50; Figs. 5B, 6A). Hamsters split by NWR, in contrast, showed two periods of increased Fos expression (Figs. 5C, 5E, 6B): Light pulses increased Fos-ir after the nighttime activity onset (p < 0.05) and after afternoon activity onset (p < 0.05). Moreover, light administered between these two time points had no effect on the number of Fos+ cells (p > 0.40; Figs. 5D, 6B). Unpulsed controls at every time point tested— whether split or unsplit—showed minimal Fos expression (Figs. 5F, 6). In groups where Fos was robustly induced by light (split hamsters in the night and the afternoon and unsplit hamsters in the night), there were no differences in the number of Fos+ cells counted (p > 0.05). Fos expression was symmetric with respect to the left and right SCN and was concentrated in the ventrolateral SCN in all groups.

DISCUSSION

As in previous experiments, daily exposure to NWR in an LDLD cycle induced split activity rhythms that free-ran and recoupled in DD (Mrosovsky and Janik, 1993). Circadian mechanisms underlying each Figure 4. Plasma melatonin concentrations (mean ± SE) of ham- bout responded similarly to timed light pulses as mea- sters with (A) unsplit or (B) split activity rhythms. Sample size and proportion of hamsters with detectable melatonin concentra- sured by behavioral phase-shifts or induction of Fos-ir tions are given above histogram. Asterisks denote significant dif- in the SCN, and both activity components were ferences (p < 0.05) in the proportion of split and unsplit hamsters accompanied by elevated plasma melatonin concen- with elevated melatonin concentrations at that sampling interval. trations. The results suggest that NWR temporally reorganizes circadian oscillators within the SCN into two distinct components. We propose that each com- 0.05), but not different from controls at ZT 20. More- ponent oscillator or group of oscillators contains a rel- over, after 6 days of NWR, nearly all NWR-split ham- atively short subjective night characterized by loco- sters (17 out of 18) had elevated concentrations of motor activity, elevated melatonin secretion, and light melatonin in circulation at ZT 9. This proportion was responsiveness. significantly greater than in non-NWR controls Conceivably,NWR might have re-entrained the cir- exposed to similar light conditions where none cadian system to generate a long subjective night that yielded detectable melatonin titers at ZT 9 (χ2 = 16.6; spanned the original night and the afternoon p < 0.001). scotophase paired with NWR. If one of the two 560 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Figure 5. Representative photomicrographs of Fos-immunostained sections from unsplit (A, B, F) and split (C-E) hamsters pulsed or sham-pulsed with light during the nighttime active phase (A, C), the morning inactive period (D), and the afternoon (B, E, F). For the latter time point, split hamsters were active, but unsplit hamsters were inactive. photophases separating dark periods masked loco- split by NWR. Among unsplit hamsters, in contrast, motor activity, the rhythm might only appear split. In plasma melatonin concentrations were elevated in a the present study, light pulses had marked effects monophasic pattern. Four hours into the night (ZT 16), when delivered during either activity component but 4 out of 12 hamsters had detectable concentrations of did not induce Fos expression in the SCN during the melatonin in circulation, consistent with onset of the morning inactive period or shift activity rhythms in rising phase in pineal melatonin production in this the early evening inactive interval. Limited resources species (Elliott and Tamarkin, 1994; Goldman et al., precluded assessment of possible light-induced phase 1981). A nighttime elevation in plasma melatonin at shifts during the morning rest period or of SCN Fos-ir ZT 20 was followed by decreases below detectable lev- expression in the early evening. In unsplit hamsters, els that were manifest by some hamsters at ZT 21 and however, Fos induction by light is restricted to subjec- all hamsters at ZT 9. If NWR delayed the circadian tive night and temporally correlates closely with peri- pacemaker as suggested by nighttime activity onsets ods of behavioral phase shifting (Kornhauser et al., (Gorman and Lee, 2001) in split hamsters, no elevation 1990; Sumová et al., 1995; Travnickova et al., 1996). of melatonin concentrations at ZT 16 would be This close relationship, moreover, is maintained in expected; importantly, none was found. Increased various photoperiods that alter α. Despite assessment melatonin secretion clearly occurred later during the by different methodologies, each inactive period thus nighttime bout of activity. However, the morning appears to represent a dead zone with respect to light decline in circadian melatonin production, evident in responsiveness. These dead zones separate intervals the majority of unsplit hamsters at ZT 21, was not of light responsiveness as measured jointly by behav- observed in split animals, providing further evidence ioral phase-shifting and Fos-ir induction. of a delayed nighttime oscillator. Most important, Elevated melatonin in circulation during both the plasma melatonin concentrations were elevated at the subjective afternoon and nighttime bouts of activity in end of the afternoon scotophase paired with NWR. hamsters with split activity rhythms supports the Among hamsters exposed to light-dark cycles, dark- hypothesis that the principal circadian pacemaker is ness in the afternoon to our knowledge has never been Gorman et al. / SPLIT CIRCADIAN RHYTHMS IN HAMSTERS 561

extra-SCN oscillators that might mediate one of the two activity components. Although their neural substrates are unknown, extra-SCN oscillators are sufficient to generate circadian rhythms in a variety of experimental paradigms (Honma et al., 1989; Stephan et al., 1979). It is not feasible to observe SCN Fos expression in a single hamster during each activity bout, but the uniformity of expression of Fos in the SCN following the afternoon or nighttime light pulse establishes that the principal circadian pacemaker itself responds to light during both activity components. In LL-induced splitting, rhythms of Fos and clock gene expression are out of phase in the left and right SCN (de la Iglesia et al., 2000), although other investigators discerned no left/right differences in electrophysiological activity in the SCN of LL- induced split hamsters (Zlomanczuk et al., 1991). The present Fos data do not suggest an anatomical or physiological distinction between the two oscillators. In the present study, all sections revealed symmetrical induction of Fos in the two SCN, indicating that both left and right are expressing a subjective night simultaneously rather than alternately. Within each SCN, the pattern of Fos induction was also similar between the two bouts, with expression concentrated in the ventrolateral SCN. How each SCN becomes Fos- inducible twice daily is not clear. At the tissue level of organization, the oscillators may be two distinct, but spatially intermixed, cell populations, each responsive to light during either the afternoon or nighttime dark period, but not both. Alternatively, a single population of SCN cells may be induced to express Fos twice daily, perhaps because individual cells in this population contain two oscillators within them, or because they are simultaneously clock-controlled by two other cell populations with rhythms out of phase. Between-group comparisons demonstrated robust behavioral phase shifts after light pulses. Phase shifts Figure 6. Number of Fos-ir cells/SCN section of (A) unsplit and calculated in this fashion might result either from dis- (B) split hamsters pulsed with light (open bars) or sham-pulsed (filled bars) at various times. Sample size is indicated in or above crete phase shifts or from changes in free-running bars. Asterisks denote significant increases (p < 0.05; Mann-Whit- period (e.g., a lengthening of τ after a light pulse ney U) in the number of Fos-ir cells relative to unpulsed controls would appear as Day 1 and Day 2 phase-delays by this at that time point. method). The latter possibility is discounted, however, by the analysis of free-running period in the days reported to induce melatonin secretion, and no eleva- following the light pulse. After all pulses, the free-run- tion was apparent in nonrunning controls. Thus, these ning period was not significantly altered; and in the data suggest that the pattern of melatonin, like that of case of the nighttime pulses, the trend was toward a activity, is split by NWR into two components. shortening of τ, which would act counter to the The ability of both nighttime and afternoon light reported phase delays. Thus, phase shifts of the activ- pulses to induce Fos expression in the SCN discounts ity components were not secondary to changes in τ but the possibility that split rhythms reflect recruitment of instead reflect discrete phase shifts. 562 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

As in LL-induced split animals probed with dark ACKNOWLEDGMENTS pulses (Boulos and Rusak, 1982; Lees et al., 1983), both NWR-split activity components responded with simi- We are grateful to Jim Donner for excellent animal lar phase shifts to light pulses 1 h after their respective care and Jeff Elliott for helpful comments on an earlier activity onsets. Afternoon light pulses additionally draft of this manuscript. This research was supported induced phase advances of the unpulsed nighttime by NHLBI grant HL61667 to TML and NICHD-48640 component, but pulses in the early nighttime to MRG. scotophase did not shift rhythms; nor did light pulses in the morning inactive period lead to induction of Fos-ir in the SCN, as is generally a prerequisite for REFERENCES light-induced phase shifts. How is this pattern of responsiveness to be understood? One possibility is Bae HH, Mangels RA, Cho BS, Dark J, YellonSM, and Zucker I (1999) Ventromedial hypothalamic mediation of photo- that the light-pulse PRC for the nighttime component periodic gonadal responses in male Syrian hamsters. J does indeed contain two light-responsive zones sepa- Biol Rhythms 14:391-401. rated by two dead zones. Alternatively, each compo- Boulos Z and Rusak B (1982) Phase-response curves and the nent may be directly light responsive only during a dual-oscillator model of circadian pacemakers. In Verte- single short subjective night associated with only one brate Circadian Systems, J Aschoff, S Daan, and G Groos, eds, pp 215-223, Springer-Verlag, Berlin. of the two scotophases. The greater fraction of each Bouskila Y and Dudek FE(1993) Neuronal synchronization oscillation may be a dead zone. The observed phase without calcium-dependent synaptic transmission in the advances of the nighttime activity component follow- hypothalamus. Proc Natl Acad Sci U S A 90:3207-3210. ing light at CT 13-a, however, could be a secondary Daan S and Berde C (1978) Two coupled oscillators: Simula- consequence of interactions with the afternoon com- tions of the circadian pacemaker in mammalian activity ponent, which is directly phase-delayed by light. rhythms. J Theor Biol 70:297-313. de la Iglesia HO, Meyer J, Carpino A Jr, and Schwartz WJ When light phase-delays the afternoon activity com- (2000) Antiphase oscillation of the left and right ponent by 90 min as in the current study, the phase suprachiasmatic nuclei. Science 290:799-801. angle between the nighttime and afternoon compo- Elliott JA and Kripke DF (1998) Photoperiodic regulation of nents is altered. If the interaction between the two light induced phase shifts and PRC amplitude in wild oscillators depends on their phase relation as com- type male and female Syrian hamsters. Soc Res Biol Rhythm Abstr 6:62. monly presumed, the phase shift of the unpulsed com- Elliott JA and Tamarkin L (1994) Complex circadian regula- ponent might reflect this altered interaction with the tion of pineal melatonin and wheel-running in Syrian pulsed component rather than a direct action of light, hamsters. J Comp Physiol [A] 174:469-484. per se. Future work can develop and test this model of Goldman BD, Hall V, Hollister C, Reppert S, Roychoud- oscillator interactions. hury P, Yellon S, and Tamarkin L (1981) Diurnal changes Multioscillatory models of circadian rhythms com- in pineal melatonin content in four rodent species: Rela- tionship to photoperiodism. Biol Reprod 24:778-783. monly distinguish between Eand M oscillators (Daan Gorman MR, Freeman DA, and Zucker I (1997) Photo- and Berde, 1978; Elliott and Tamarkin, 1994; Gorman periodism in hamsters: Abrupt versus gradual changes et al., 1997; Illnerova and Vanacek, 1985; Pittendrigh in day length differentially entrain morning and evening and Daan, 1976; Sumová et al., 1995). It is not possible circadian oscillators. J Biol Rhythms 12:122-135. to attribute the respective activity components gener- Gorman MR and Lee TM (2001) Daily novel wheel running reorganizes and splits hamster circadian activity ated here to hypothesized evening and morning oscil- rhythms. J Biol Rhythms 16:540-549. lators. Regardless of the relation to Eand M, however, Herzog ED, Geusz ME, Khalsa SBS, Straume M, and Block the circadian timekeeping system can be resolved into GD (1997) Circadian rhythms in mouse suprachiasmatic two seemingly similar oscillators, both of which can nucleus explants on multimicroelectrode plates. Brain be phase-delayed by light and both of which can drive Res 757:285-290. melatonin secretion within a single 24-h period. Fur- Honma S, Honma K, and Hiroshige T (1989) Methamphet- amine induced locomotor rhythm entrains to restricted ther probing of these two oscillators will complement daily feeding in SCN lesioned rats. Physiol Behav in vitro work to characterize the suboscillatory basis of 45:1057-1065. the mammalian pacemaker. Illnerova H (1991) The suprachiasmatic nucleus and rhythmic pineal melatonin production. In Suprachiasmatic Nucleus: The Mind’s Clock, DC Klein, RY Moore, and SM Reppert, eds, pp 197-216, Oxford University Press, New York. Gorman et al. / SPLIT CIRCADIAN RHYTHMS IN HAMSTERS 563

Illnerova H and Vanacek J (1985) Complex control of the cir- structure: A clock for all seasons. J Comp Physiol [A] 106: cadian rhythm in pineal melatonin production. In The 333-355. Pineal Gland: Current State of Pineal Research, B Mess, Pittendrigh CS, Elliott JA, and Takamura T (1984) The circa- C Ruzsas, L Tima, and P Pevet, eds, pp 137-153, New dian component in photoperiodic induction. CIBA York, Elsevier. Found Symp 104:26-47. Jagota A, de la Iglesia HO, and Schwartz WJ (2000) Morning Stephan FK, Swann JM, and Sisk CL (1979) Entrainment of and evening circadian oscillations in the suprachiasmatic circadian rhythms by feeding schedules in rats with nucleus in vitro. Nat Neurosci 3:372-376. suprachiasmatic lesions. Behav Neural Biol 25:545-554. Kornhauser JM, Nelson DE, Mayo KE, and Takahashi JS Sumová A, Trávnícková Z, Peters R, Schwartz WJ, and (1990) Photic and circadian regulation of c-fos gene Illnerová H (1995) The rat suprachiasmatic nucleus is a expression in the hamster suprachiasmatic nucleus. Neu- clock for all seasons. Proc Natl Acad SciUSA92: ron 5:127-134. 7754-7758. Lees JG, Hallonquist JD, and Mrosovsky N (1983) Differen- Travnickova Z, Sumova A, Peters R, Schwartz WJ, and tial effects of dark pulses on the two components of split Illnerova H (1996) Photoperiod-dependent correlation circadian activity rhythms in golden hamsters. J Comp between light-induced SCN c-fos expression and reset- Physiol [A] 153:123-132. ting of circadian phase. Am J Physiol 271:R825-R831. Liu C, Weaver DR, Strogatz SH, and Reppert SM (1997) Cel- Weaver DR (1998) The suprachiasmatic nucleus: A 25-year lular construction of a circadian clock: Period determina- retrospective. J Biol Rhythms 13:100-112. tion in the suprachiasmatic nuclei. Cell 91:855-860. Welsh DK, Logothetis DE, Meister M, and Reppert SM Mason R (1991) The effects of continuous light exposure on (1995) Individual neurons dissociated from rat supra- Syrian hamster suprachiasmatic (SCN) neuronal dis- chiasmatic nucleus express independently phased circa- charge activity in vitro. Neurosci Lett 123:160-163. dian firing rhythms. Neuron 14:697-706. Mrosovsky N and Janik DS (1993) Behavioral decoupling of Zlomanczuk P, Margraf RR, and Lynch GR (1991) In vitro circadian rhythms. J Biol Rhythms 8:57-65. electrical activity in the suprachiasmatic nucleus follow- Pittendrigh CS and Daan S (1976) A functional analysis of ing splitting and masking of wheel-running behavior. circadian pacemakers in nocturnal rodents: V.Pacemaker Brain Res 559:94-99. WatanabeJOURNAL et OF al. BIOLOGICAL / TRANSIENT RHYTHMS VERSUS STEADY-STATE / December 2001 PHASE SHIFTS Light-Induced Resetting of the Circadian Pacemaker: Quantitative Analysis of Transient versus Steady-State Phase Shifts

Kazuto Watanabe,* Tom Deboer,† and Johanna H. Meijer†,1 *Department of Physiology, Dokkyo University School of Medicine, 321-02 Mibu, Japan, †Department of Physiology, Leiden University Medical Center, Leiden, the Netherlands

Abstract The suprachiasmatic nuclei of the hypothalamus contain the major circadian pacemaker in mammals, driving circadian rhythms in behavioral and physiological functions. This circadian pacemaker’s responsiveness to light allows synchronization to the light-dark cycle. Phase shifting by light often involves several transient cycles in which the behavioral activity rhythm gradu- ally shifts to its steady-state position. In this article, the authors investigate in Syrian hamsters whether a phase-advancing light pulse results in immediate shifts of the PRC at the next circadian cycle. In a first series of experiments, the authors aimed a light pulse at CT 19 to induce a phase advance. It appeared that the steady-state phase advances were highly correlated with activity onset in the first and second transient cycle. This enabled them to make a reliable estimate of the steady-state phase shift induced by a phase-advancing light pulse on the basis of activity onset in the first transient cycle. In the next series of experiments, they presented a light pulse at CT 19, which was followed by a second light pulse aimed at the delay zone of the PRC on the next circadian cycle. The immediate and steady-state phase delays induced by the second light pulse were compared with data from a third experiment in which animals received a phase-delaying light pulse only.The authors observed that the waveform of the phase-delay part of the PRC (CT 12-16) obtained in Experiment 2 was virtually identical to the phase-delay part of the PRC for a single light pulse (obtained in Experiment 3). This finding allowed for a quantitative assessment of the data. The analysis indicates that the delay part of the PRC—between CT 12 and CT 16—is rapidly reset following a light pulse at CT 19. These findings complement earlier findings in the hamster showing that after a light pulse at CT 19, the phase-advancing part of the PRC is immediately shifted. Together, the data indicate that the basis for phase advancing involves rapid resetting of both advance and delay components of the PRC.

Key words circadian rhythms, suprachiasmatic nucleus, entrainment, phase shift, transients, phase response curve, evening/morning oscillators

An endogenous circadian pacemaker in the of many behavioral and physiological functions suprachiasmatic nuclei (SCN) determines the timing (Meijer and Rietveld, 1989). For proper timing, this

1. To whom all correspondence should be addressed: Department of Physiology, LUMC, P.O. Box 9604, 2300 RC Leiden, the Netherlands. JOURNAL OF BIOLOGICAL RHYTHMS, Vol. 16 No. 6, December 2001 564-573 © 2001 Sage Publications 564 Watanabe et al. / TRANSIENT VERSUS STEADY-STATE PHASE SHIFTS 565 pacemaker is responsive to the environmental light- been performed in invertebrates such as Drosophila dark cycle. Light presented during the beginning of (Pittendrigh, 1979) and Neurospora (Crosthwaite et al., the subjective dark period induces phase delays of the 1995), in mice (Sharma and Chandrashekaran, 2000), circadian activity rhythm, whereas light presented and in Syrian hamsters (Elliot and Pittendrigh 1996; toward the end of the subjective dark period produces Best et al., 1999). In the hamster, it has been established phase advances. The effects of light on the circadian that the phase-advance part of the PRC shifts within a activity rhythm can be described by a phase response few hours after a light pulse at CT 18 by applying a sec- curve (PRC). In a PRC, the magnitude and direction of ond light pulse 1 to 2 h after the first pulse (Best et al., phase shifts in activity onset are plotted as a function 1999). It was also shown that the phase-delay part of the circadian time of pulse application (Daan and shifts within a few hours after a light pulse at CT 13 by Pittendrigh, 1976). applying a second pulse 1 to 2 h after this pulse (Best After a light pulse, a phase shift is often not com- et al., 1999). However, it has not been investigated pleted within the first circadian cycle but instead whether the phase-delay part of the PRC shifts imme- grows over the course of several days. Such circadian diately after a phase-advancing light pulse. This mat- cycles are called transient cycles. They are most pro- ter is of great importance in view of recent findings, nounced following phase-advancing light pulses, but indicating that the pacemaker is composed of distinct they occur after delaying pulses as well (Pittendrigh genetic components that exhibit different responsive- et al., 1958). It is not clear whether activity onset dur- ness to phase-advancing and phase-delaying light ing the transient cycles reflects the true position of the pulses (see Daan et al., 2001). underlying pacemaker. One possibility is that the We investigated the responsiveness of the phase- pacemaker itself requires several days to complete the delay part of the PRC during a transient cycle that was phase shift. Alternatively, the pacemaker may shift induced by a phase-advancing pulse. To this purpose, immediately to its final position but the activity we applied a second light pulse shortly after activity rhythm requires several days to become fully synchro- onset at the first transient cycle. By comparing the nized with the new phase of the pacemaker. effect of a single light pulse with the effect of the dou- On the molecular level, it has been shown that the ble light pulse, we could investigate the phase-shifting expression of the mammalian homologs of the insect effect of the second light pulse. From the magnitude of period gene, mper1 and mper2, peak 1 h and 2 h after the phase shift, it can be estimated what the position of application of a light pulse, respectively (Albrecht the circadian clock was at the time point of application et al., 1997; Shearman et al., 1997; Shigeyoshi et al., of the second light pulse. To optimize this estimate, we 1997). It is clear that resetting mechanisms are induced determined the relation between activity onset during very quickly and increases in the putative protein the first transient cycles and the final steady-state products of mper may be the cause of resetting to a new phase after application of a single phase-advancing phase. The finding that inhibition of mper1 expression light pulse. This enabled us to make a reliable estimate blocks the phase shifts induced by light and glutamate of the steady-state phase shift after the first activity pulses supports the notion that mper is required for onset. phase shifting by light (Akiyama et al., 1999). It appears therefore that resetting of the clock is established within a few hours. To understand the discrep- METHODS ancy between the molecular biology and the behavior of the overt rhythm, it is important to get insight into This study was performed on 60 male Syrian ham- the position of the pacemaker during the transient sters (Mesocricetus auratus, Harlan/CPB, Zeist, the cycles in vivo. For this purpose, it is necessary to deter- Netherlands) aged 2 months at the start of the experi- mine the position of the PRC during the transient ment. The animals were individually housed in cages cycle, assuming that the position of the PRC provides (36.5 × 25.0 × 16.0 cm) with a running wheel (diameter a true reflection of the phase of the pacemaker. 26 cm) and were kept in a sound-attenuating, venti- One way to investigate the position of the PRC is by lated room at a temperature of 23 °C. Food (Hope applying a second light pulse during the first transient farms B.V. the Netherlands) and water were continu- cycle after a light pulse and to investigate the phase- ously available. Running-wheel activity was recorded shifting effect of the two light pulses on the final phase per minute to determine the animals’ circadian activ- of the activity rhythm. Two-pulse experiments have ity rhythm. In all the experiments, light pulses (15 566 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001 min) were presented to groups of animals, and the protocol was repeated three times. Positioning of the light sources on the wall behind every single cage ensured similar light levels (100-120 lux) for all animals. During presentation of the light pulses, the animals remained in their home cages. All animals received all three treatments, and treatments were presented in a randomized order. At least 1 month elapsed between application of subsequent light pulses. Data were excluded from the analysis when the actogram of running-wheel activity did not allow for unambiguous determination of activity onsets.

Experiment 1

The animals were entrained to LD 14:10 for at least 7 days before they were released into constant darkness (DD). After 7 days in DD (Day 0), the animals received a light pulse between CT 17.77 and CT 21.90. After the light pulse, the animals were kept in DD for 14 days. Lines were eye-fitted through activity onsets before and after the light pulse. The first transient cycles after the light pulse were excluded from the fit. Steady-state phase shifts were determined by measuring the difference between the fitted lines extrapolated to the first cycle after the light pulse (Day 1, Fig. 1A). In addition, the immediate phase shift on the first (Day 1) and second circadian cycle (Day 2) after the light pulse were determined. The immediate phase shift was defined as the difference between the time of observed activity onset and the time as predicted by the line through activity onsets before the light pulse. A strong correlation between the immediate phase shift on Day 1 and the steady-state phase shift was Figure 1. Experimental protocol (see Methods for explanation of found. Moreover, we found a strong correlation the protocol). A. Experiment 1: A light pulse was applied between between the immediate shift on Days 1 and 2 (for fur- CT 17.77 and CT 21.90 on Day 0. The steady-state phase shift (∆φst, ther details, see Results and Fig. 3 B,C). These correla- A1 – C1), and the immediate phase shift on Day 1 (∆φim(1), A1 – B1) ∆φ tions were used for the analysis of Experiment 2. and Day 2 ( im(2), A2 – B2) were measured. B. Experiment 2: A light pulse was applied between CT 17.75 and CT 21.71 on Day 0 and a second pulse was given 0.01 to 4.31 h after activity onset on ∆φ Experiment 2 Day 1. Steady-state phase shifts ( st, E1 – C1) and immediate phase shifts (∆φim, D2 – B2) induced by the second light pulse were Alight pulse was applied between CT 17.75 and CT measured. C. Experiment 3: A light pulse was applied between CT 10.80 and CT 15.43 on Day 1. Steady-state phase shifts (∆φ , 21.71 according to the protocol as in Experiment 1. In st C1 – A1) and immediate phase shifts (∆φim, B2 – A2) were measured. this experiment, the light pulse was followed by a second light pulse given 0.01 to 4.31 h after activity onset We considered that the steady-state phase shift on the first transient cycle (Day 1, Fig. 1B). After the induced by the second light pulse is equal to the differ- second light pulse, the animals were kept in DD for ence between the steady-state phase shift induced by 14 days. The immediate phase shift induced after the the first light pulse and the steady-state phase shift first and second light pulses and the steady-state induced by the two pulses together. To estimate the phase shift induced by the two light pulses were steady-state phase shift induced by the first light determined. Watanabe et al. / TRANSIENT VERSUS STEADY-STATE PHASE SHIFTS 567 pulse, the strong correlation between the immediate phase shift and the steady-state phase shift obtained in Experiment 1 was used. In other words, the immediate phase shift on Day 1 was determined and the steady-state phase shift (A1 – C1) for that particular animal was estimated on the basis of the regression line obtained in Experiment 1 (Fig. 3B). To determine the steady-state phase shift induced by the two light pulses, the difference between the steady-state activity onset lines before and after the two light pulses was calculated (A1 – E1). The steady-state phase shift induced by the second light pulse only is then (A1 –

E1) – (A1 – C1) = C1 – E1. Figure 2. Typical examples of activity records of all three experi- Similarly, the immediate phase shift induced by the mental conditions in the same animal. The activity record is second light pulse was estimated by measuring the double-plotted to enable visualization of the activity rhythms. difference between the time of observed activity onset Arrows and stars indicate day and time of light pulse application. Panel A shows a phase advance after application of a light pulse at on Day 2 (B2) and the time of activity onset on Day 2 CT 21. Panel B shows a phase shift after application of a light when the second light pulse would not have been pulse CT 21.5 followed by a light pulse applied 2.75 h after activity given (D2). This latter value was predicted from the onset on the next circadian cycle. Panel C shows a phase delay strong correlation between immediate shifts on Day 1 after application of a light pulse at CT 14. and Day 2 obtained in Experiment 1.

mediate phase shift were 1.42 ± 0.44 h and 0.64 ± 0.24 h, Experiment 3 respectively (mean ± SD). Although the same animals A light pulse was presented according to the proto- received light pulses several times, the difference col as in Experiment 1. The light pulse was applied between interindividual and intraindividual varia- between CT 10.80 and 15.43 on Day 1. After the light tions was not significant for the magnitudes of both pulse, the animals were kept in DD for 14 days. The steady-state and immediate phase shifts. Therefore, steady-state phase shift on the day of the light pulse phase shifts obtained from the same animal were treat- (Day 1) was determined by measuring the difference ed as independent values in all of the experiments. between the fitted lines before and after the light pulse The results of this experiment are important for the (Fig. 1C). The immediate phase shift on the first cycle protocol of the next experiment. As is evident from after the light pulse (Day 2) was measured as the dif- Figure 3A, a rather large range of phase shifts can be ference between the time of observed activity onset obtained at each circadian time. However, when plot- ∆φ and the time predicted by the line through activity ting the steady-state phase shift, st, as a function of ∆φ onsets preceding the light pulse. ANOVAand post hoc the magnitude of the immediate shift on Day 1, im(1) t tests served to compare Experiment 2 and Experi- (Fig. 3B), a strong relation is observed between imme- ment 3. diate and steady-state phase shift, which can be described as follows:

RESULTS –20 ∆φst = 1.462 ×∆φim(1) + 0.482 (r = 0.79, p < 1*10 ).

Experiment 1 This regression line was used to predict for each ani- In Experiment 1, 93 phase advances were analyzed mal the steady-state phase shift from the immediate (Fig. 2A). These light pulses were presented between shift on Day 1 in Experiment 2. CT 17.77 and CT 21.90 (mean = 19.07). The effect of the Moreover, a strong correlation between the magni- ∆φ ∆φ light pulse on the steady-state phase shift, st, and on tude of the immediate shift on Day 1, im(1), and the the immediate phase shift on the first cycle after the immediate shift on Day 2, ∆φim(2), was found (Fig. 3C). ∆φ –20 light pulse, im(1), are summarized in Figure 3A. The This correlation (r = 0.89, p < 1*10 ) can be described magnitudes of the steady-state phase shift and the im- by the following function: 568 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

∆φim(2) = 1.325 ×∆φim(1) + 0.150.

This regression line was used in Experiment 2 to predict for each animal the time of activity onset on Day 2 in the case that only the first light pulse would have been applied and to measure the difference with the activity onset on Day 2.

Experiment 2

In the double pulse experiment, 85 activity records were obtained that allowed for an unambiguous analysis (Fig. 2B). The first light pulse fell between CT 17.75 and CT 21.71 (mean = 19.48), inducing a phase advance on Day 1 in all cases (range from 0.24 to 1.26 h; mean = 0.76 h). The second pulse was given 0.01 to 4.31 h (mean = 1.61 h) after activity onset on Day 1. Those cases where the second light pulse fell before activity onset were excluded from the analysis since no estimate of activity onset on Day 1 could be made. In most cases, the second light pulse induced a phase delay but a few advances were also observed.

Experiment 3

Eighty-five clear phase-shifts were obtained from light pulses that fell between CT 10.80 and CT 15.43 (mean = 12.95) and were used to describe the phase- delay part of the PRC (Fig. 2C). The maximum delay was obtained around CT 13 (Fig. 4). Mean hourly values for the immediate and steady-state phase shift were calculated between CT 11 and CT 16 and were compared with the data from Experiment 2.

Comparison of Experiments 2 and 3

In Experiment 2, the second light pulse was given at the beginning of the subjective night. The phase shift induced by this second light pulse was compared with the shift induced by a single light pulse applied at comparable phases (CT) in Experiment 3. The circadian time of application of the second light pulse is unknown. However, there are two predictions for the circadian time of the second light pulse. (1) If the overt rhythm is the manifestation of the underlying oscilla- Figure 3. Immediate and steady-state phase shifts induced by a tor during transient cycles, the time of the activity phase-advancing light pulse. A. The phase shifts are plotted as a onset on Day 1 is equal to CT 12. (2) If the steady-state function of the time of light pulse application. Dots indicate phase shift reflects the real position of the oscillator, steady-state phase shifts, circles immediate shifts. B. The the extrapolated steady-state phase shift on Day 1 is steady-state shift is plotted as a function of the immediate shift on Day 1. C. The immediate phase shift on Day 2 is plotted as a func- equal to CT 12. tion of the immediate shift on Day 1. Watanabe et al. / TRANSIENT VERSUS STEADY-STATE PHASE SHIFTS 569

shifts immediately after application of a phase- advancing light pulse. To answer this question, we determined the position of the PRC at the time of application of the second light pulse (Experiment 2) with the delay part of the PRC induced by a single light pulse (Experiment 3). The results are in accor- dance with previous studies (Best et al., 1999) and add to this that (a) the phase-delay part of the pacemaker shifts within one circadian cycle to its new steady-state position after a phase-advancing light pulse, (b) the magnitude of a transient shift predicts the phase of the new steady-state position, and (c) the waveform and amplitude of the phase-delay part of the PRC do not change after a phase-advancing light pulse.

PRC Amplitude and Waveform Stability

Amplitude Figure 4. Comparison of phase shifts induced by the single delaying light pulse and those induced by the second of the dou- The question arises whether the properties of the ble light pulse. Hourly values (± SEM) of steady-state phase shifts pacemaker change during transient cycles. In fact, the induced by the second light pulse are plotted against CT of the data can only be interpreted on the premise that second light pulse in “steady-state phase” (triangles) or that in the pacemaker responds in its usual way to light dur- “transient phase” (circles) together with the phase shift induced by the single light pulse (squares). Asterisks indicate significant ing the transient cycles. Only then do the results from differences between the conditions (p < 0.05, two-tailed t test after Experiment 3 form a reliable prediction for the pace- significant ANOVA for factor “condition” over CT 13 to CT 15). maker’s responsiveness to the second light pulse. Previous double-pulse experiments in inverte- The two possible phase delays obtained in Experi- brates (Drosophila, Pittendrigh, 1979; Neurospora, ment 2 with the second light pulse were calculated Crosthwaite et al., 1995), and the Syrian hamster (Best and compared with the data obtained in Experiment 3 et al., 1999) were performed to investigate if there is a (Fig. 4). The analysis indicates that the steady-state refractory period in the pacemaker after application of phase shift in Experiment 2 did not differ significantly a light pulse and to determine how long this refractory from the PRC obtained in Experiment 3 when suppos- period lasts (i.e., how fast the pacemaker can react to a ing that the pacemaker shifted to its steady-state posi- second light pulse). These data show that the pace- tion within one circadian cycle (Fig. 4, Prediction 2). In maker is capable of reacting to a new light pulse contrast, significant differences were obtained at CT within a few hours. In other studies, however, it was 13 and CT 14 between the steady-state phase shifts in shown that mice and hamsters are significantly less Experiment 2 and Experiment 3 when supposing that responsive to a second light pulse when applied the first transient indicates the phase of the pacemaker within 4 h after the first light pulse (Khammanivong (Prediction 1). Also, the two predicted PRCs differed and Nelson, 2000; Nelson and Takahashi, 1999). More- significantly at CT 13 and CT 14 (Fig. 4). over, Khammanivong and Nelson (2000) indicated that responses are not even back to normal in the next circadian cycle (about 70% of expected shift). Our DISCUSSION experiments show that the amplitude of the phase delay is back to normal 17 to 21 h after a phase-advanc- The question posed in this article is whether tran- ing light pulse. This raises the possibility that respon- sient cycles accompanying a light-induced phase siveness to light is decreased 24 h after a light pulse but advance reflect the circadian time of the pacemaker or is unchanged 17 to 21 h after a light pulse. In other whether the pacemaker shifts immediately to its words, the part of the PRC that received the first light steady-state position. We specifically addressed the pulse is still affected by it after 24 h, whereas other question of whether the phase-delay part of the PRC parts of the PRC may be unaffected. 570 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Waveform maker shifts immediately, the question arises, What process induces transient behavior? Transients in The waveform of the PRC appeared unaltered on Drosophila have been attributed to the existence of two the first cycle after the light pulse between CT 12 and coupled oscillators of which one is light sensitive, CT 16. We have no knowledge of the phase response whereas the other is temperature sensitive (Pittendrigh properties at other circadian times. Our data are con- et al., 1958). In mammals, two mutually coupled circa- sistent with unpublished data of Elliot and Pittend- dian oscillators within the central pacemaker have righ (1996; Pittendrigh, 1981, Fig. 9). With respect to been proposed (Pittendrigh and Daan, 1976). One the phasedelay part of the PRC, they showed that oscillator is thought to lock onto the evening light (the between CT 12 and CT 16, the PRC had shifted to the Eoscillator) and controls activity onset in nocturnal steady-state position within one circadian cycle. How- animals, while the other locks onto the morning light ever, they showed that the onset of the phase-delay (the M oscillator) and controls activity offset. It has part of the PRC (about CT 11) had not shifted. As a con- been suggested that the differential shift in onset and sequence, the phase-delay part of their PRC was com- offset after a phase-advancing light pulse causes dif- pressed. We did not investigate responsiveness before ferent shifts of Eand M, with the M oscillator shifting CT 12 because we administered our second light pulse immediately to its new steady-state position and E after activity onset of the animal to obtain an accurate shifting more slowly (Honma et al., 1985; Meijer and prediction for the induced phase shifts. Devries, 1995; Elliot and Tamarkin, 1994; Pittendrigh and Daan, 1976). Predicting Steady-State Phase Shift Accumulating evidence at a number of research from the First Transient Cycle levels supports the proposition, initially put forward by Pittendrigh and Daan (1976), that the circadian The present analysis shows that there is a strong pacemaker is composed of different oscillators. Daan relationship between the position of the activity onset et al. (2001) summarized these results elegantly in a during the first and second transient cycles and the conceptual framework. The Eand M components are steady-state phase shift of activity onset (Fig. 3). This thought to result from a double set of circadian genes. enabled us to make a very reliable estimate of the The M oscillator consists of per1 and cry1 and is accel- steady-state phase shift induced by the first light pulse erated by light, the Eoscillator of cry2 and per2 and is in Experiment 2 on the basis of the first activity onset. decelerated by light. Evidence in favor comes, for The strong correlation indicates that the daily shifts in instance, from Albrecht et al. (2001), who demon- activity onsets during transient cycles are regulated strated that per1 knockout mice lost the capacity to with great precision and reflect the magnitude of the respond with advances to a light pulse, whereas per2 steady-state shift. Thus, the immediate shift is a fixed knockouts no longer respond with phase delays (see fraction of the steady-state shift. This indicates by itself Daan et al., 2001, for more details). An alternative that during the first transient cycle the steady-state model was proposed by Hastings (commentary to phase shift is already determined. This is consistent Daan et al., 2001). Eand M could reflect per1/per2 with our conclusion that the pacemaker has reached expression (M) versus cry1/cry2 expression (E). This its steady-state position on the first transient cycle. model is supported by differential peak times of the This result is suggestive for an interaction between pers and crys under different photoperiods. the endogenous pacemaker and a secondary down- The transients that are observed after a phase- stream oscillatory system, either inside or outside the advancing light pulse are attributed to an immediate SCN, of which the kinetics can be described with great shift of M (the oscillator that responds with advances mathematical precision. We will discuss possible sec- to light) and a delayed shift of E, as a consequence of ondary systems in the next section. coupling forces. This explanation would also be consistent with Jagota et al. (2000), who demonstrated Immediate Resetting two peaks in multiunit activity in the SCN slice preparation. One peak occurred at the onset of dawn (possi- We showed that after a phase-advancing light bly the M oscillator) and responded to glutamate with pulse, the phase-delay part of the PRC shifts within an immediate advance, while the other component one circadian cycle to the new steady-state position did not respond. The evening rise and morning while the overt rhythm displays transients. If the pace- decline of melatonin show similar differences in their Watanabe et al. / TRANSIENT VERSUS STEADY-STATE PHASE SHIFTS 571 responsiveness to light. A phase-advancing light pulse Table 1. Phase shifts induced by a phase-advancing light pulse at CT 19. The table illustrates clear asymmetry in the effect of light on results in an immediate advance of the melatonin behavioral activity and melatonin (NAT) rhythms and on multiunit decline and in a delayed advancing shift in melatonin activity (MUA). Immediate shifts were obtained in activity offset, rise (Elliot and Tamarkin, 1994; Illnerova, 1991). We melatonin offset, and morning component of MUA. No immediate have summarized the responses to phase-advancing shifts were obtained in activity onset, melatonin onset, and in the evening component of MUA. No asymmetry exists for the shifts in light pulses in Table 1. advancing and delaying parts of the PRC, as both shift immediately. Best et al. (1999) demonstrated that the phase- The shift of the rhythm in mPer is investigated in response to a 6-h advancing part of the PRC is reset within a few hours advance of the light-dark cycle (and not in response to a short light pulse at CT 19) mPer1 shifts immediately in response to this shifted by light at CT 19. Our results demonstrate that within cycle. It is unknown whether mPer2 rhythms are immediately reset one circadian cycle light presentation at CT 19 also by phase-advancing stimuli. results in advances of the delay part of the PRC, at Response to Advancing least of the delay part between CT 12 and CT 16. If the Light Pulse at CT 19 delay part of the PRC shows rapid resetting, similar to Immediate No Immediate the advance part of the PRC, transients in the onset of Shift Shift activity cannot readily be explained on the basis of dif- Behavioral activity rhythm offset ferential shifts of the Eand M component when it is Elliot and Tamarkin, 1994 assumed that Eis represented by the delay and M by Meijer and De Vries, 1995 x Behavioral activity rhythm onset the advance portion of the PRC. Elliot and Tamarkin, 1994 Although our results do not follow directly from De Vries and Meijer, 1995 x the proposed model of Daan et al. (2001), they are not Melatonin (NAT) offset necessarily in conflict with it either. An immediate Elliot and Tamarkin, 1994 Illnerova, 1991 x shift of the M component (per1/cry1) in response to Melatonin (NAT) onset light at CT 19 may result in an immediate shift of the E Elliot and Tamarkin, 1994 component (per2/cry2) as a consequence of strong Illnerova, 1991 x Multiunit activity peak dawn (M) coupling forces between the two. The same reasoning Jagota et al., 2000 x is applicable to the model of Hastings (2001), but now Multiunit activity peak dusk (E) with a different set of genes; an immediate shift of M Jagota et al., 2000 x Advance part PRC (per1/per2) may result in a shift of E( cry1/cry2) within Best et al., 1999 x one cycle as a consequence of coupling. However, irre- Delay part PRC spective of the set of genes that may underlie Eand M, This study, 2001 x our results suggest rapid resetting of both the delay mPer1 rhythm Yamazaki et al., 2000 x and advance portions of the PRC. mPer2 rhythm ? ? A different viewpoint arises when data from intact animals are compared with data obtained in slice preparations. Phase resetting of the pacemaker has secondary processes, which are downstream from the been studied in vitro by recording circadian rhythms pacemaker, outside the SCN. in sampled single unit activity from SCN brain slice The latter offers an alternative explanation for the preparations. This activity displays a circadian occurrence of transients. It has been proposed that rhythm, and the time of peak activity is used as a there might be secondary oscillatory systems (slave phase marker (Prosser and Gillette, 1989; Prosser, oscillators) regulating the coupling of the activity 1998). SCN brain slices can be kept in good condition rhythm with the circadian pacemaker (Pittendrigh and recorded from for two or three cycles. Phase shifts and Daan, 1976; Takamure et al., 1991; Yamazaki et al., are commonly induced between the first and second 2000). These secondary oscillators may require more cycle. Phase shift experiments in vitro have indicated than one circadian cycle to achieve complete re- that phase shifts in the SCN slice are immediate and entrainment. This hypothesis is supported by the find- stable (McArthur et al., 1991; Prosser, 1998; Watanabe ing that the velocity of re-entrainment of different cir- et al., 2000) also when glutamate is administered at cadian rhythms (urinary ion secretion, adrenocortical CT 19 (Ding et al., 1994). In these preparations, most levels, core body temperature) after crossing several downstream processes are eliminated because the time zones is not the same (Klein et al., 1972; slice is disconnected from most of its input and out- Moore-Ede et al., 1982; Van Cauter and Turek, 1986). put. This indicates that transients could be the result of Recently it was proposed that resetting of circadian 572 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001 time in peripheral tissue occurs via glucocorticoid sig- bling a clock for all seasons: Are there M and Eoscillators naling (Balsalobre et al., 2000), indicating that in the genes? J Biol Rhythms 16:105-116. Daan S and Pittendrigh CS (1976) A functional analysis of glucocorticoid is one of the possible candidates circadian pacemakers in nocturnal rodents: II. The vari- entraining secondary oscillators to the circadian pace- ability of phase response curves. J Comp Physiol [A] maker. Together the data indicate the existence of vari- 106:255-266. ous time lags from the circadian pacemaker to each Ding JM, Chen D, Weber ET, Faiman LE, Rea MA, and secondary oscillatory system. Gillette MU (1994) Resetting the biological clock: Media- The present experiments were not aimed to indi- tion of nocturnal circadian shifts by glutamate and NO. Science 266:1713-1717. cate where transients originate and leave open the Elliot JA and Pittendrigh CS (1996) Time course of hamster possibility that behavioral transients are attributable clock resetting following single light pulses. Fifth meet- to downstream processes, outside the SCN. The ing of the Society for Research on Biological Rhythms results have demonstrated that at least part of the [abstr] 157:105. pacemaker has shifted fully and within one circadian Elliot JA and Tamarkin L (1994) Complex circadian regula- cycle in response to light. It is possible that the first tion of pineal melatonin and wheelrunning in Syrian hamsters. J Comp Physiol [A] 174:437-447. part of the delaying area of the PRC (before CT 12) Hastings M (2001) Commentary: Modeling the molecular shows compression in response to a phase-advancing calendar. J Biol Rhythms 16:117-123. light pulse (see Pittendrigh, 1981) and that this Honma K, Honma S, and Hiroshige T (1985) Response accounts for transients in activity onset. This would curve, free-running period, and activity time in circadian lead to the conclusion that part of the delay curve locomotor rhythm of rats. Jpn J Physiol 35:643-658. shifted immediately, and part of it did not. The ques- Illnerova H (1991) The suprachiasmatic nucleus and rhyth- tion of whether the pacemaker shows immediate mic pineal melatonin production. In Suprachiasmatic Nucleus. The Mind’s Clock, DC Klein, RY Moore, and SM resetting in response to light pulses should then be Reppert, eds, pp 197-216, Oxford University Press, New reformulated into a more specified question: Which York. parts of the pacemaker show immediate resetting and Jagota A, de la Iglesia HO, and Schwartz WJ (2000) Morning which parts do not? It may appear necessary to not and evening circadian oscillations in the suprachiasmatic only distinguish between the advance and delay areas nucleus in vitro. Nat Neurosci 3:372-376. of the PRC but even within these areas. Khammanivong A and Nelson DE(2000) Light pulses sup- press responsiveness within the mouse photic entrainment pathway. J Biol Rhythms 15:393-405. Klein KE, Wegmann HM, and Hunt BI (1972) Desynchron- REFERENCES ization of body temperature and performance circadian rhythm as a result of outgoing and homegoing Akiyama M, Kouzu Y,TakahashiS, WakamatsuH, Moriya T, transmeridian flight. Aerosp Med 43: 119-132. Maetani M, Watanabe S, Tei H, Sakaki Y, and Shibata S McArthur AJ, Gillette MU, and Prosser RA(1991) Melatonin (1999) Inhibition of light- or glutamate-induced mPer1 directly resets the rat suprachiasmatic circadian clock in expression represses the phase shift into the mouse circa- vitro. Brain Res 565:158-161. dian locomotor and suprachiasmatic firing rhythms. J Meijer JH and Devries MJ (1995) Light induced phase shifts Neurosci 19:1115-1121. in onset and offset of running-wheel activity in the Syrian Albrecht U, Sun ZS, Eichele G, and Lee CC (1997) Adifferen- hamster. J Biol Rhythms 10:4-16. tial response of two putative mammalian circadian regu- Meijer JH and Rietveld WJ (1989) Neurophysiology of the lators, mper1 and mper2, to light. Cell 91:1055-1064. suprachiasmatic circadian pacemaker in rodents. Physiol Albrecht U, Zheng B, Larkin D, Sun ZS, and Lee CC (2001) Rev 69:671-707. mPer1 and mPer2 are essential for normal resetting of the Moore-Ede MC, Sultzman FM, and Fuller CA (1982) The circadian clock. J Biol Rhythms 16:100-104. Clocks That Time Us, Harvard University Press, Cam- Balsalobre A, Brown S, Marcacci L, Tronche F, Kellendonk C, bridge, MA. Reichardt HM, Schütz G, and Schibler U (2000) Resetting Nelson DEand Takahashi JS (1999) Integration and satura - of circadian time in peripheral tissues by glucocorticoid tion within the circadian photic entrainment pathway of signaling. Science 289:2344-2347. hamsters. Am J Physiol 277:R1351-R1361. Best JD, Maywood ES, Smith KL, and Hastings MH (1999) Pittendrigh CS (1979) Some functional aspects of circadian Rapid resetting of the mammalian circadian clock. J pacemakers. In Biological Rhythms and Their Central Mech- Neurosci 19:828-835. anism, M Suda, O Hayaishi, and H Nakagawa, eds, pp 3- Crosthwaite SK, Loros JJ, and Dunlap JC (1995) Light 15, Elsevier/North-Holland, Amsterdam. -induced resetting of a circadian clock is mediated by a Pittendrigh CS (1981) Circadian systems: Entrainment. In rapid increase in frequency. Cell 81:1003-1012. Handbook of Behavioural Neurobiology,Vol. 4,Biological Daan S, Albrecht U, van der Horst GTJ, Illnerova H, Rhythms, J Aschoff, ed, pp 95-1241, Plenum Press, New Roenneberg T, Wehr TA, and Schwartz WJ (2001) Assem- York. Watanabe et al. / TRANSIENT VERSUS STEADY-STATE PHASE SHIFTS 573

Pittendrigh CS, Bruce VG, and Kaus P (1958) On the signifi- Shigeyoshi Y, Taguchi K, Yamamoto S, Takekida S, Yan L, cance of transients in daily rhythms. Proc Natl Acad Sci Tei H, Moriya T, Shibata S, Loros JJ, Dunlap JC, and U S A 44:965-997. Okamura H (1997) Light-induced resetting of a mamma- Pittendrigh CS and Daan S (1976) A functional analysis of lian circadian clock is associated with rapid induction of circadian pacemakers in nocturnal rodents: V.Pacemaker the mPer1 transcript. Cell 91:1043-1053. structure: A clock for all seasons. J Comp Physiol Takamure M, Murakami N, Takahashi K, Kuroda H, and 106:333-355. Etoh T (1991) Rapid reentrainment of the circadian clock Prosser RA (1998) In vitro circadian rhythms of the mamma- itself, but not the measurable activity rhythm, to a new lian suprachiasmatic nuclei: Comparison of multi-unit light-dark cycle in the rat. Physiol Behav 50:443-449. and single-unit neuronal activity recordings. J Biol Van Cauter Eand Turek FW (1986) Depression: Adisorder of Rhythms 13:30-38. time keeping? Perspect Biol Med 29:510-519. Prosser RA and Gillette MU (1989) The mammalian circa- Watanabe K, Vanacek J, and Yamaoka S (2000) In vitro dian clock in the suprachiasmatic nuclei is reset in vitro entrainment of the circadian rhythm of vasopressin- by cAMP. J Neurosci 9:1073-1081. releasing cell in suprachiasmatic nucleus by vasoactive Sharma VK and Chandrashekaran MK (2000) Probing the intestinal polypeptide. Brain Res 877:361-366. circadian oscillator of a mammal by two-pulse perturba- Yamazaki S, Numano R, Abe M, Hida A, Takahashi R, Ueda tions. Chronobiol Int 17:129-136. M, Block GD, Sakaki Y, Menaker M and Tei H (2000) Shearman LP, Zylka MJ, Weaver DR, Kolakowski LF Jr, and Resetting central and peripheral circadian oscillators in Reppert SM (1997) Two period homologs: Circadian transgenic rats. Science 288:682-685. expression and photic regulation in the suprachiasmatic nuclei. Neuron 19:1261-1269. FoàJOURNAL and Bertolucci OF BIOLOGICAL / TEMPERATURE RHYTHMS AND / December BIMODAL 2001 PATTERN IN LIZARDS Temperature Cycles Induce a Bimodal Activity Pattern in Ruin Lizards: Masking or Clock-Controlled Event? A Seasonal Problem

Augusto Foà1 and Cristiano Bertolucci Dipartimento di Biologia, Università di Ferrara, via L. Borsari 46 I-44100, Ferrara, Italy

Abstract The daily locomotor activity pattern of Ruin lizards in the field is mainly unimodal, except for summer months when soil temperatures exceed 40 °C to 42 °C around midday. In such a situation, lizards reduce their locomotor activity around midday to avoid overheating, and thus their daily activity pattern becomes bimodal. The bimodal pattern expressed in the field is usually retained in the free-running rhythm under constant temperature and DD for a couple of weeks, after which the bimodal pattern changes into a unimodal pattern. In the present study, the authors examined whether 24-h temperature cycles (TCs) would change lizard activity from a unimodal to a bimodal pattern. Administra- tion of TCs to unimodal lizards free-running in DD is able to entrain locomotor rhythms and to induce a bimodal pattern both in summer and autumn-winter. There are, however, striking seasonal differences in the effectiveness with which TCs achieve bimodality: (a) Numbers of lizards rendered bimodal are significantly higher in summer than in autumn-winter; (b) TCs require less time to achieve bimodality in summer than autumn-winter; (c) bimodality is retained as an aftereffect in the postentrainment free-run in summer, but not in autumn- winter; (d) TCs change activity duration in summer, but not in autumn-winter. All this demonstrates the existence of seasonal changes in responsiveness of the circadian oscillators controlling activity to the external factors inducing bimodality. Oscillators’ responsiveness is high in summer, when bimodality is the survival strategy of Ruin lizards to avoid overheating around midday in open fields, and low in autumn-winter, when bimodality has no recognizable adaptive significance.

Key words circadian rhythms, locomotor activity, entrainment, temperature cycles, bimodal activity pattern, lizards

Seasonal changes in the daily pattern of locomotor unimodal in spring (with only one activity peak per activity in the field were reported in most diurnal day) to bimodal in summer (with two well-separated Lacertid lizards from southern Europe (Bowker, 1986; activity peaks per day), becoming unimodal again in Foà et al., 1992; Henle, 1988; Pough and Busack, 1978; autumn. Since lizards are ectotherms, which can be Tosini et al., 1992; Van Damme et al., 1990). Generally, active only within a limited range of body tempera- the daily pattern of locomotor activity changes from tures, seasonal changes in locomotor patterns have

1. To whom all correspondence should be addressed: Dipartimento di Biologia, Università di Ferrara, via L. Borsari 46 I-44100, Ferrara, Italy; e-mail: [email protected]. JOURNAL OF BIOLOGICAL RHYTHMS, Vol. 16 No. 6, December 2001 574-584 © 2001 Sage Publications

574 Foà and Bertolucci / TEMPERATURE AND BIMODAL PATTERN IN LIZARDS 575 been regarded more or less explicitly as a direct behav- temperature cycles in the laboratory.Hoffmann (1968) ioral response of these animals to related changes in showed that 24-h temperature cycles (TCs) of low solar radiation and ambient temperature (Avery,1980; amplitude (0.9-3 °C) are capable of entraining circa- Ouboter, 1981). Recent studies in our model animal, dian locomotor rhythms of Ruin lizards kept in con- the Ruin lizard (Podarcis sicula), showed that a great stant lighting conditions. deal of these activity changes are controlled by endog- The present study carried out completely in DD enous temporal programs (Bertolucci et al., 1999; Foà examined whether the administration of 24-h TCs is et al., 1994). Short-term experiments (20-30 days), in capable of changing a unimodal activity pattern into a which Ruin lizards collected in different months were bimodal activity pattern. When Ruin lizards were tested immediately after capture under constant tem- tested in summer, TCs were found to induce a bimodal perature (29 °C) and darkness (DD), showed that the pattern. To establish whether effectiveness of TCs in activity pattern typical of each season (bimodal/ achieving bimodality changes depending on season, unimodal) is retained in the lizard circadian locomo- we compared the locomotor behavior of lizards col- tor rhythm (Foà et al., 1994). Furthermore, while the lected and exposed to the same TCs in different sea- bimodal locomotor pattern expressed by the lizards in sons. We expected effectiveness to be maximal in the constant conditions between June and August is typi- summer, as Ruin lizards usually express a bimodal cally associated with a short free-running period (τ) activity pattern in summer but not at other times of the and a long circadian activity (α), the unimodal loco- year (Foà et al., 1992, 1994). Studies focused on the role motor pattern expressed in the remaining months is of the pineal and other clock components in establish- typically associated with long τ and short α (Foà et al., ing and/or maintaining a bimodal activity pattern in 1994). To test whether such seasonal changes in circa- Ruin lizards have been reported previously (Innocenti dian locomotor rhythms were driven by a circannual et al., 1994, 1996; Tosini et al., 2001). clock, the locomotor activity of Ruin lizards was recorded in long-term experiments (over 12-15 months) under constant temperature (29 °C) and DD. MATERIALS AND METHODS The results demonstrated the existence of circannual τ α cycles of both and and, at the same time, the Animals and absence of a circannual cycle of activity pattern from Locomotor Recording bimodal to unimodal and vice versa (Bertolucci et al., 1999). The longest τ along its circannual cycle was Ruin lizards, Podarcis sicula campestris De Betta 1857 associated with short α, and the shortest τ in the same (adult males only, 6.5-8 cm snout-vent length) from cycle with long α, so that the pattern of mutual associ- the area of Ferrara (Italy) were used. Lizards were col- ation between τ and α was found to be the same as in lected in the field in groups of 8 to 14 individuals in short-term experiments (Bertolucci et al., 1999; Foà June and November 1999 and June 2000. After cap- et al., 1994). Most lizards, however, stayed unimodal ture, each seasonal group of lizards was carried to the all the time, showing that seasonally changing envi- lab and immediately put into individual tilt-cages (30 ronmental factors (such as, for instance, photo- × 15 × 11 cm) for locomotor recording. Tilt-cages were thermoperiodic conditions) are involved in the induc- placed inside thermally programmable environmen- tion of a bimodal activity pattern (Bertolucci et al., tal chambers and connected to a computer-based data 1999). In other reptiles, such as the Namib desert dune acquisition system (DataQuest III, MiniMitter, Sun- lizard (Aporosaura anchietae) and the garter snake river, OR, USA) for monitoring locomotor activity. (Thamnophis radix), the locomotor activity pattern was Food (Tenebriomolitor larvae) and water were supplied shown to be switched between unimodal and bimodal twice a week. During experiments, lizards were kept patterns, respectively,by lowering or raising the ambi- under DD and either constant temperature (30 ± 0.2 °C) ent temperature (Heckrotte, 1962, 1975; Holm, 1973; or 14:10-h TCs. Experiments used three different TCs.

Underwood, 1992). This suggests that the expected TC1: 30:27 °C (30 °C from 0800 to 2200 h; 27 °C from 2200 seasonal changes of locomotor activity pattern (from to 0800 h); TC2: 30:28.3 °C (30 °C from 0800 to 2200 h; unimodal to bimodal and vice versa) can be achieved 28.3 °C from 2200 to 0800 h); and TC3: 30:28.3 °C (30 °C in lizards by means of appropriate manipulations of from 1400 to 0400 h; 28.3 °C from 0400 to 1400 h). ambient temperature levels and/or administration of 576 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Experimental Design mean value of τ in autumn-winter (p < 0.05), and the mean value of α in summer was significantly longer Summer 1999 (June 4 to July 27): than the mean value of α in autumn-winter (p < 0.001) Pilot Experiment (Fig. 1). Locomotor rhythms of all lizards kept in DD

Lizards (n = 8) were allowed to free-run in DD and entrained to the TC1 of 3 °C in amplitude (Figs. 2-5). constant temperature (29 °C) for 3 weeks and then Entrainment continued after reduction of the TC were subjected to the TC1 for 25 days. amplitude to 1.7 °C (TC2 and TC3; Figs. 2-3, 5) in all cases. The entrained activity phase always fell within Autumn-Winter 1999 to 2000 the 30 °C portion of the TC. When, in the final part of (November 17 to March 8) experiments, lizards were exposed again to constant temperature, activity onsets in the postentrainment Lizards (n = 14) were allowed to free-run in DD and free-running rhythm always extrapolated back to the constant temperature (29 °C) for 2 weeks and then time of activity onsets during entrainment. In winter, a were subjected to three TCs: TC for 28 days, TC for 22 1 2 6-h shift from TC2 to TC3 induced shifts of the activity days, and TC3 for 22 days. After these periods of time, onsets, which resulted, after several transient cycles, locomotor activity in constant temperature (29 °C) in entrainment to the new schedule (Fig. 5). Hence, was recorded for a further 2 to 3 weeks. 14:10-h TCs actually entrained the activity rhythm and did not merely cause masking of the underlying Summer 2000 (June 19 to September 9) oscillation. No seasonal differences in all these aspects of entrainment were found. In most lizards, activity Lizards (n = 14) were allowed to free-run in DD and onsets delayed about 1 h the daily onsets of high constant temperature (29 °C) for 2 weeks and then (30 °C) temperature phase of TC. No differences in ψ were subjected to two TCs: TC1 for 23 days and TC2 for either among TCs or among seasons were found.

17 days. After these periods of time, locomotor activ- Administration of the TC1 to lizards expressing a ity in constant temperature (29 °C) was recorded for a unimodal circadian locomotor activity pattern under further 4 weeks. constant temperature and DD was capable of inducing the appearance of a bimodal pattern in all seasons Data Evaluations (Table 1, Figs. 2-3, 5). Bimodality always persisted after reduction of amplitude of the TC from 3 °C to 1.7

For locomotor rhythms free-running in constant °C (TC2 and TC3 cycle; representative examples in temperature and DD, τ and α were estimated by the Figs. 2-3, 5). However, the percentage of lizards ren- eye-fitting method (Pittendrigh and Daan, 1976a). τ dered bimodal by TC1 is significantly higher in sum- was also measured by means of χ2 periodogram analy- mer than in autumn-winter (Fisher exact test, p < 0.04; sis (Sokolove and Bushell, 1978). Bimodality and Fig. 6A). Furthermore, the time elapsed between unimodality of the locomotor pattern were established administration of TC1 and appearance of bimodality is by visual inspection: The bimodal pattern is character- significantly shorter in summer than in autumn-win- ized by the existence of two daily peaks of locomotor ter (Kruskal-Wallis one-way ANOVA, H = 4.47, p < activity regularly repeated in subsequent days, while 0.04; Fig. 6B). Administration of TC1 also increases sig- the unimodal pattern is characterized by the existence nificantly the duration of daily activity in summer of one peak of activity each day. In presence of a (paired Student t test, p < 0.01) but leaves the duration bimodal activity pattern, either duration of circadian of daily activity unchanged in autumn-winter (paired (α) or duration of entrained activity includes duration Student t test, p > 0.10). As expected, from the very of the pause between the two daily activity peaks. beginning of locomotor recording in summer, several lizards already expressed a bimodal pattern in constant temperature. In all cases, the administration of

RESULTS TC1 lengthened the rest interval between the two daily peaks of activity that characterize the bimodal state Before administration of TCs, lizards were free- from 2 to 3.1 h, enhancing in that way bimodality running in constant temperature and DD. The mean (paired Student t test, p < 0.002; representative exam- value of τ in summer was significantly shorter than the ples in Fig. 4). Foà and Bertolucci / TEMPERATURE AND BIMODAL PATTERN IN LIZARDS 577

τ α 24.7 13

24.4 10 hours hours

24.1 7 Autumn-Winter Summer Autumn-Winter Summer

Figure 1. Estimates of τ and α of locomotor rhythms of Ruin lizards free-running in constant temperature (29 °C) and DD before starting temperature cycle administration. τ in autumn-winter was significantly longer than in summer, and α in autumn-winter was significantly shorter than in summer.

Administration of constant temperature led all liz- rhythm under constant conditions, the bimodal pat- ards to free-run (Figs. 3, 5). While in autumn-winter no tern previously expressed in the field (Foà at al., 1994). bimodal lizard maintains bimodality after release in This demonstrates that bimodality is not merely a constant temperature, in summer several lizards do so direct behavioral reaction of these ectotherm animals (Table 1 and Figs. 3, 5). The seasonal difference is sta- to the extremely high levels of soil temperatures tistically significant (p < 0.006, Fisher exact test). Fur- around midday in summer but is a season-dependent thermore, the time elapsed between administration of state of the circadian pacemaker that has evolved as an constant temperature and reappearance of unimodality adaptation to high temperatures predictably occur- is significantly shorter in autumn-winter than in sum- ring at that time of day. Long-term experiments fur- mer (Kruskal-Wallis one-way ANOVA, H = 7.17, p < ther revealed that the retained bimodal pattern disap- 0.02; Fig. 6B). Several summer lizards were still pears—that is, the pattern becomes unimodal—in the bimodal 3 to 4 weeks after release in constant tempera- course of the first 2 months of locomotor recording ture (Table 1 and Fig. 3). under constant temperature (29 °C) and DD, and bimodality never reappears (Bertolucci et al., 1999). Hence, bimodality of Ruin lizards typically shows all DISCUSSION recognized properties of aftereffects on the circadian pacemaker: (a) Its appearance in constant conditions Behavioral observations of Ruin lizards in their nat- is dictated by previous exposure to specific environ- ural environment showed that the daily locomotor mental stimuli, such as, for instance, summer pattern of focal animals in the field is mainly photo-thermoperiodic conditions; (b) once estab- unimodal, except for those summer months in which lished, it persists for several weeks in constant condi- soil temperatures exceed 40 to 42 °C around the mid- tions, after which, (c) it decays to a different state: dle of day.In such situations, lizards reduce their loco- unimodality. The present study examined whether motor activity dramatically around midday, retreat- changes in ambient temperature—which were admin- ing into shade or burrows to avoid overheating, and istered in the form of 24-h TCs—were the specific thus exhibiting the markedly bimodal activity pattern environmental stimuli capable of inducing bimodal so typical of summer (Foà et al., 1992, 1994; Tosini activity patterns at the expected time of the year. First et al., 1992). Once carried to the laboratory and placed of all, our results show that circadian locomotor in constant temperature (29 °C) and DD, summer liz- rhythms of Ruin lizards become entrained to 24-h TCs ards mainly retained their endogenous, free-running of low (3-1.7 °C) amplitude. This confirms the results 578 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Figure 2. Locomotor records of lizards free-running in constant temperature and DD, which were subjected to a 24-h temperature cycle (TC) 30:27 °C, and then to a 24-h TC 30:28.3 °C. Each horizontal line is a record of 1 day’s activity, and consecutive days are mounted one below the other. Rectangles encompass the 30 °C phase during the whole period of time of TCs administration. Arrowhead indicates the day on which the amplitude of TC was reduced from 3 °C to 1.7 °C. Records are representative examples of the fact that TC administration in summer almost immediately induces a change from unimodal activity pattern to bimodal activity pattern, while TC administration in autumn-winter does not change the pattern in 50% of the lizards. of the pioneering work done by Hoffmann in Ruin liz- ioral rhythms in ectotherm vertebrates (Hoffmann, ards, which for the first time indicated that low-ampli- 1968). More important, administration of 24-h TCs in tude TCs are capable of entraining circadian behav- the laboratory to unimodal lizards collected in the Foà and Bertolucci / TEMPERATURE AND BIMODAL PATTERN IN LIZARDS 579

Figure 3. Behavioral effects of temperature cycles (TCs) in summer. Brackets (on the right) delimit sections of records where the activity pattern was bimodal. Records are representative examples of the fact that TC administration entrained circadian locomotor rhythms and induced a fast change of locomotor activity pattern from unimodal to bimodal. Furthermore, after final release in constant temperature, lizards maintained bimodality. Further information in Figure 2. 580 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

pointed out, however, that all lizards that became bimodal in autumn-winter returned to being unimodal immediately after cessation of TCs (Fig. 5). Thus, differently from summer, in autumn-winter bimodality is not retained in the postentrainment free-running rhythm in constant conditions: It is not an aftereffect. Instead of resulting from TCs affecting the circadian oscillators that control activity, the transition from a unimodal to a bimodal pattern in autumn-winter seems to result from TCs directly affecting activity, therefore bypassing the circadian clock. In other words, bimodality achieved in autumn-winter looks like a phenomenon of masking (Underwood, 1992, p. 249). For the sake of clarity, it seems useful to underline that the case of masking proposed here exclusively concerns achievement of bimodality in autumn-winter. In fact, “true” entrainment (and not masking) of circadian locomotor rhythms of lizards to the administered 24-h TCs has occurred in all seasons, as postentrainment free-runs in constant temperature have verified. Changes in locomotor pattern in response to ambient temperature manipulations in the laboratory have Figure 4. Representative example of summer lizards exhibiting a been reported previously in other reptiles. For bimodal activity pattern already before administration of temper- instance, in the Namib desert dune lizard (Aporosaura ature cycles (TCs). Dotted lines delimit the pause between the two daily peaks of activity. Record is a representative example of the anchietae) kept under constant conditions and the gar- fact that TC administration in summer-bimodal lizards length- ter snake (Thamnophis radix) exposed to an LD cycle, ened the interruption between the two daily peaks of activity that the locomotor activity pattern was shown to be characterize the bimodal state. switched between unimodal and bimodal patterns, respectively, by lowering or raising the ambient temperature. It is unclear, however, whether the reported field during late spring to early summer was found to effects were season-dependent (Heckrotte, 1962, 1975; induce rapid appearance of a bimodal activity pattern Holm, 1973). Noteworthy, Underwood (1992) pro- in most animals (Table 1 and Figs. 2-3). posed that each activity peak of the bimodal pattern of As expected, in summer several lizards were lizards and snakes may be the overt expression of one already bimodal in constant temperature before start- of the two (or sets of) circadian oscillators described in ing with TCs administration (example in Fig. 4). TCs Pittendrigh and Daan’s (1976b) model. Underwood further enhanced bimodality by lengthening the further stated that since a rise in ambient temperature pause between the two daily peaks of the bimodal switches activity from a unimodal to a bimodal pat- activity pattern of these lizards. Most summer lizards tern and increases the duration of daily activity, tem- were still bimodal 3 to 4 weeks after cessation of the perature as well as photoperiod may be able to change TCs (after final release in constant temperature), and phase relationship between the two oscillators con- after this period of time, bimodality disappeared in trolling activity. Because the model proposes that the about 50% of the animals. These aspects of the data time interval between activity onset and activity offset confirm our assumption that bimodality is retained as each day would be a measure of the phase relationship an aftereffect in the postentrainment free-running between the two oscillators, changes in that phase rhythm in constant conditions. To our surprise, 24-h relationship would be expected to cause a change in TCs were capable of inducing bimodality even in duration of activity.According to this view, it is impor- autumn-winter, seasons in which changing from tant that in Ruin lizards the bimodal pattern achieved unimodal to bimodal activity pattern is not of appar- in response to TCs is associated with a prominent ent adaptive significance (Foà et al., 1994). It must be increase in duration of daily activity in summer—thus Foà and Bertolucci / TEMPERATURE AND BIMODAL PATTERN IN LIZARDS 581

Figure 5. Effects of temperature cycles (TCs) in autumn-winter. Records are representative examples of the fact that TC administration entrained circadian locomotor rhythms. However, it took several days for the TC to induce a bimodal pattern. Furthermore, differently from summer, autumn-winter lizards returned unimodal immediately after release in constant temperature. Other information in Figure 3. Finally, 50% of autumn-winter lizards remained unimodal after TC administration, as it is shown in Figure 2. 582 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

A overt rhythm. With regard to the present results, such 100 seasonal differences are expressed in the free-running rhythm before TC administration: in summer a short τ 75 associated with a long α, and in autumn-winter, a long τ associated with a short α (Fig. 1). Because previous

% 50 experiments in constant conditions made it clear that a short τ together with a long α typically brings about a 25 bimodal pattern, the summer effectiveness of TCs in achieving bimodality shown here is likely to depend 0 on the expression in the free-running rhythm of a short Autumn-Winter Summer τ together with a long α, facilitating a change in pattern from unimodal to bimodal as soon as Ruin lizards B from unimodal to bimodal from bimodal to unimodal perceived the TC (Foà et al., 1994). The long τ together with the short α of autumn-winter reflect a phase rela- 20 tionship between constituent oscillators that presumably hampers a change in pattern from unimodal to 15 bimodal, thus reducing autumn-winter effectiveness 10 of TCs in achieving bimodality. When, in spite of all days 5 that, TCs administered in autumn-winter were successful in changing the pattern from unimodal to 0 Autumn-Winter Summer Autumn-Winter Summer bimodal, they clearly did it by bypassing the circadian n=7 n=12 n=7 n=7 clock, as bimodality did not entail a change in activity duration and, moreover, was not retained as an after- Figure 6. Seasonal differences in effectiveness of temperature effect in the postentrainment free-run. cycles (TCs) in the induction of a bimodal activity pattern. (A) Per- Unlike bimodality, effectiveness of 24-h TCs in centage of lizards rendered bimodal by TCs is significantly higher in summer than in autumn-winter. (B) Number of days entrainment of locomotor rhythms does not change between administration of TCs and appearance of bimodality with season. This was somehow expected. In fact, (left), and between final release in constant temperature and Ruin lizards are diurnal, ecthoterm animals whose appearance of unimodality (right). Appearance of a bimodal pat- locomotor activity is necessarily limited by tempera- tern is faster in summer than in autumn-winter. Furthermore, the time elapsed between final release in constant temperature and ture. In such a situation, the circadian clock of lizards reappearance of unimodality is significantly shorter in has to entrain to the TC of the external day, to force the autumn-winter than in summer. There are also summer lizards locomotor activity it drives into the range of the ambi- that remained bimodal at the end of the experiments (Table 1). ent temperatures that are most suitable for locomotor performances, thus optimizing daily time-budget for potentially reflecting a change in phase relationship biologically significant behaviors through the entire between circadian oscillators controlling activity— year. Again, no seasonal differences in any aspect of while the bimodal pattern achieved in autumn-winter entrainment were found. Both summer and autumn- is not associated with a change in activity duration. winter entrained activity fell in the 30 °C phase and Furthermore, there are striking seasonal differ- (entrained) rest in the 27 °C (or 28.3 °C) phase of TCs. ences in the effectiveness with which TCs achieve This confirms previous work indicating that Ruin liz- bimodality: (a) The percentage of lizards that TCs ards that are held on a thermal gradient in DD sponta- have rendered bimodal is significantly higher in sum- neously select warm temperature zones of the gradi- mer than in autumn-winter; (b) the time required for ent during activity and cooler temperature zones the first administered TC to achieve a bimodal pattern during rest (Innocenti et al., 1993). On the other hand, is significantly less in summer than autumn-winter. field studies in Ruin lizards showed the existence of According to Pittendrigh and Daan’s (1976b) model, seasonal differences in body temperatures recorded such differences in effectiveness may depend on sea- during activity phase, with a mean temperature of sonal differences in the mutual phase relationship 29 °C in autumn-winter and a mean of 32.7 °C in late between the two oscillators, that is reflected in the sea- spring–summer (Tosini et al., 1992). It looks like the sonally different associated values of τ and α of the temperature of 30 °C we chose in the present experi- Foà and Bertolucci / TEMPERATURE AND BIMODAL PATTERN IN LIZARDS 583

Table 1. Summary of the results. All numbers of unimodal versus bimodal lizards in each stage of experiments in different seasons.

T Cost Temperature Cycles T Cost n Unimodal Bimodal Unimodal Bimodal Unimodal Bimodal

Summer 1999 8 6 2* 1 7 — — Autumn-Winter 1999-2000 14 14 0 7 7 14 0 Summer 2000 14 9 5* 1 13 7 7 *Summer lizards that already expressed a bimodal pattern in constant temperature.

ments is a reasonable compromise between those mea- ACKNOWLEDGMENTS sured in the field and therefore suitable for activity in each season, while 27 °C seems less suitable for activity. This work was supported by research grants of the Università di Ferrara and of the Italian Ministero dell’ Università e della Ricerca Scientifica e Tecnologica CONCLUSIONS (COFIN 1999).

Although long-term records in constant conditions did not find circannual changes between bimodal and REFERENCES unimodal patterns, the same experiments did reveal the existence of circannual changes of both τ and α of Avery RA (1980) Ecophysiology and behaviour of lacertid lizards—towards a synoptic model. Proc Euro Herp locomotor rhythms with, moreover, some phases of τ Symp CWLP, pp 71-73. the circannual cycle characterized by long associated Bertolucci C, Leorati M, Innocenti A, and Foà A (1999) with short α (potentially conducive to unimodality) Circannual variations of lizard circadian activity rhythms and other phases characterized by short τ associated in constant darkness. Behav Ecol Sociobiol 46:200-209. with long α (potentially conducive to bimodality) Bowker RG (1986) Patterns of thermoregulation in Podarcis (Bertolucci et al., 1999). The circannual data made it hispanica (Lacertilia: Lacertidae). In Studies in Herpetology, Z Rocek, ed, pp 621-626, Proc Euro Herp Meeting, clear, on one hand, that environmental factors are nec- Prague. essary to obtain a change from a unimodal to a Foà A, Minutini L, and Innocenti A (1992) Melatonin: A cou- bimodal pattern, and, on the other hand, that there are pling device between oscillators in the circadian system months in which the state of the system, expressed in of the ruin lizard Podarcis sicula. Comp Biochem Physiol the association between short τ and long α of its overt 103A:719-723. Foà A, Monteforti G, Minutini L, Innocenti A, Quaglieri C, rhythm, facilitates achievement of a bimodal pattern. and Flamini M (1994) Seasonal changes of locomotor The present results show that a change in ambient activity patterns in ruin lizards Podarcis sicula: I. Endoge- temperature—in the form of a 24-h TC—is the appro- nous control by the circadian system. Behav Ecol priate stimulus to induce bimodality, and they con- Sociobiol 34:227-274. firm the existence of pronounced changes in respon- Heckrotte C (1962) The effect of the environmental factors in siveness of the circadian oscillators controlling the locomotory activity of the plains garter snake (Thammophis radix radix). Anim Behav 10:193-207. activity to the external factors inducing bimodality, Heckrotte C (1975) Temperature and light effects on the cir- with high responsiveness in summer and low respon- cadian rhythm and locomotor activity of the plains garter siveness in autumn-winter. The data further suggest snake (Thammophis radix radix). J Interdiscipl Cycle Res 6: the possibility that a circannual rhythm in responsive- 279-290. ness of the circadian system to the bimodality-stimu- Henle K (1988) Dynamics and ecology of three Yugoslavian populations of the Italian wall lizard (Podarcis sicula lating effects of TCs parallel the previously demon- τ α campestris de Betta) (Reptilia, Lacertidae). Zool Anz 220: strated circannual rhythms of and or, alternatively, 33-48. the circannual rhythm in responsiveness may be Hoffmann K (1968) Synchronisation der circadianen merely a consequence of those of τ and α. Aktivitaetsperiodik von Eidechsen durch Temperatur- 584 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

zyclen verschiedener Amplitude. Z vergl Physiol 58: Pittendrigh CS and Daan S (1976b) A functional analysis of 225-228. circadian pacemaker in nocturnal rodents: V.Pacemakers Holm E(1973) The influence of constant temperatures upon structure: A clock for all seasons. J Comp Physiol the circadian rhythm of the Namib desert dune lizard 106:333-355. Aporosaura anchietae Bocage. Madoqua 2:33-41. Pough FH and Busack SD (1978) Metabolism and activity of Innocenti A, Bertolucci C, Minutini L, and Foà A (1996) Sea- the spanish fringe-toed lizard (Lacertidae: sonal variations of pineal involvement in the circadian Acanthodactylus erythrurus). J Thermal Biol 3:203-205. organization of ruin lizards Podarcis sicula. J Exp Biol Sokolove PG and Bushell WN (1978) The chi square 199:1189-1194. periodogram: Its utility for analysis of circadian rhythms. Innocenti A, Minutini L, and Foà A(1993) The pineal and cir- J Theor Biol 72:131-160. cadian rhythms of temperature selection and locomotion Tosini G, Bertolucci C, and Foà A (2001) The circadian sys- in lizards. Physiol Behav 53:911-915. tem of reptiles: A true multioscillatory and multiphoto- Innocenti A, Minutini L, and Foà A (1994) Seasonal changes receptive system. Behav Physiol 72(4):461-471. of locomotor activity patterns in ruin lizards Podarcis Tosini G, Foà A, and Avery RA (1992) Body temperatures sicula: II. Involvement of the pineal. Behav Ecol Sociobiol and exposure to sunshine of ruin lizards Podarcis sicula in 35:27-32. central Italy. Amphibia-Reptilia 13:169-175. Ouboter PE(1981) The ecology of the island-lizard Podarcis Underwood H (1992) Endogenous rhythms. In Biology of sicula salfii: Correlation of microdistribution with vegeta- Reptilia, C Gans, ed, Vol 18, pp 229-229, University of Chi- tion coverage, thermal environment and food-size. cago Press, Chicago, IL. Amphibia-Reptilia 2:243-257. Van Damme R, Bauwens D, Castilla AM, and Verheyen RF Pittendrigh CS and Daan S (1976a) A functional analysis of (1990) Comparative thermal ecology of the sympatric liz- circadian pacemaker in nocturnal rodents: I. The stability ards Podarcis tiliguerta and Podarcis sicula. Acta Oecol and lability of spontaneous frequency. J Comp Physiol 11:503-512. 106:223-252. MrosovskyJOURNAL OFet al. BIOLOGICAL / IRRADIANCE RHYTHMS DETECTION / December 2001 LETTER Persistence of Masking Responses to Light in Mice Lacking Rods and Cones

N. Mrosovsky,*,1 Robert J. Lucas,† and Russell G. Foster† *Departments of Zoology, Physiology and Psychology, University of Toronto, Ontario, Canada M5S 3G5, †Department of Integrative and Molecular Neuroscience, Imperial College School of Medicine, Charing Cross Hospital, London, UK W6 8RF

Key words cones, irradiance, masking, phase shifting, retinal degeneration, rods

Overt rhythms of locomotor behavior are a result of This has been ruled out for phase shifting, which has two complementary processes. The first mechanism is been shown to persist even in mice genetically modi- the synchronization of an endogenous clock, which in fied to lack both rods and cones (Freedman et al., turn directs the animal to be active in the day or the 1999). The current experiment was undertaken to night; this is called entrainment. The second mecha- determine if masking was also spared in such rodless nism involves an acute response to light, which inhib- coneless mice. its or promotes activity, depending on whether the species is nocturnal or diurnal; this is called masking (Aschoff, 1960; review in Mrosovsky, 1999). The fact METHODS that masking can occur in hamsters with lesions of the SCN (Redlin and Mrosovsky, 1999) and in mice lack- C3H/He wild-type and rodless coneless (rd/rd cl) ing cryptochromes (van der Horst et al., 1999; mice were bred in London as previously described Vitaterna et al., 1999), that is, in animals lacking any (Lucas et al., 1999) and sent to Toronto for testing. persistent circadian rhythm to be entrained, is further Assessment of inhibition of locomotion by light fol- evidence that entrainment and masking are behavior- lowed previous procedures (Mrosovsky et al., 1999). ally and physiologically different responses to light. Tests began when the mice were approximately 4 months Although masking may be of considerable impor- old. Male mice of both genotypes (wild-types n =8, tance in influencing the amplitude of overt rhythms, it rd/rd cl n = 5) were housed individually in cages 44 × 23 has received much less study than has entrainment. × 20 cm fitted with running wheels (17.5 cm d). Revo- Indeed, the main concern of rhythms researchers has lutions were monitored with Dataquest III hardware been to exclude masking as much as possible from and software (Sunriver, OR, USA). Prior to the start of their experiments. the masking tests, the animals were entrained to a 16:8 At the receptor level at least, however, masking and h light-dark (LD) cycle for 17 days; the entraining light entrainment may have considerable commonalities, provided approximately 1300 lux in the cages, as mea- in that they are both spared in retinally degenerate sured with a ISO-TECH ILM350 meter. (rd/rd) mutant mice (Foster et al., 1991; Mrosovsky, To examine the acute effects of light on wheel run- 1994). Eventually, rd/rd mice come to lose all their rods, ning, an additional set of lights were used. These were and many cones degenerate secondarily (Carter- fluorescent tubes (Sylvania Octron 32 watt 4100 K) fit- Dawson et al., 1978; Foster et al., 1991). Nevertheless, a ted with a broad-spectrum green filter (Rosco few cones remain, and it is possible that such residual Supergel filter #89, maximum transmission around cones might be enough to mediate responses to light. 520 nm, half bandwidth approximately 60 nm; Rosco,

1. To whom all correspondence should be addressed: Department of Zoology, 25 Harbord Street, University of Toronto, Ontario, Canada M5S 3G5; e-mail: [email protected]. JOURNAL OF BIOLOGICAL RHYTHMS, Vol. 16 No. 6, December 2001 585-588 © 2001 Sage Publications 585 586 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

Stamford, CT., U.S.A.). Irradiance, as measured with a Hagner E2X luxmeter, was ca 120 lux in the cages below. To reduce this illumination for subsequent tests, neutral density filters (Rosco Cinegel), calibrated in photo- graphic stops, were added as needed; illumination at three, six, and nine stops, respectively, was approximately 12, 1.5, and 0.25 lux. Tests with 1-h light pulses, starting 2 h after dark onset, were spaced not less than 3 days apart and were conducted in the following order for all animals: 0, 3, 6, 9, 7.5, 4.5, 1.5, 1, 2, 0.5 stops. The number of wheel revolutions made during a light pulse was expressed as a percentage of the number made by the same animal during the same hour on the previous day when there was no light. For additional details, see Mrosovsky et al. (1999).

RESULTS

Both wild-type and rd/rd cl mice inhibited their Figure 1. Photic inhibition of locomotor activity in wild-type and wheel running on exposure to the broad-spectrum rodless coneless (rd/rd cl) mice. Top: actograms showing wheel running over 2 days. The number of revolutions per 10 min are green light presented in the early night (Fig. 1). At the plotted in 15 quantiles on the y axis, with each quantile represent- highest irradiance studied here, wheel running was ing 26 revolutions. The open and solid bar shows the light-dark reduced to about 10% of normal levels in both geno- cycle used to entrain the mice. The arrowheads show the onset types (means ± SEMs: 12.8 ± 4.6% wild types, 10.7 ± and offset times of the 1-h light pulses (ca 120 lux in this example) used to study masking. Bottom: Means ± SEMs for the suppres- 9.3% rd/rd cl). In both genotypes, the effects were sion of locomotor activity by light pulses of decreasing irradiance. irradiance dependent but there were no significant Some symbols have been slightly shifted on the x axis for clarity. differences between the genotypes (two-way ANOVA p > 0.1). that projects to various brain areas involved in different responses. If there is a single receptor type for these different responses to light, whether or not it had been DISCUSSION called a circadian receptor in the past, it should be con- ceptualized more as a general irradiance detector. It It is evident that, as well as phase shifting, rodless has been demonstrated that besides unimpaired coneless mice retain an acute masking response to phase shifting and masking, rd/rd mice spend more light and that no impairment of this acute response is time in the dark side of a differentially illuminated box present. (Mrosovsky and Hampton, 1997). It has long been Both masking and entrainment serve the function known that they retain pupillary responses (Keeler, of confining an animal’s activity to a daytime or night- 1927) and suppression of melatonin by light pulses time niche. Both masking and entrainment require (Goto and Ebihara, 1990). Recently it has been shown only detection of overall illumination, not of that pupillary contraction and melatonin suppression spatio-temporal differences in light (Mrosovsky, are also present in rd/rd cl mice (Lucas et al., 1999, 1994). Indeed, to determine whether it is night or day, 2001). Thus, there are a variety of responses to light a system that integrates illumination over space and that do not require rods or cones. time is desirable. For it to be day, it should be light in Of course it remains possible that there are several all directions and this input should be maintained. different receptor types subserving phase shifting, Both masking and entrainment are spared in mice masking, dark preferences, melatonin suppression, lacking rods and cones and so presumably depend on and pupillary responses, perhaps with some func- some novel photoreceptor type elsewhere in the eye. tional redundancy (Selby et al., 2000). While being Given these similarities, it would be most parsimo- open-minded on this possibility, it seems more likely nious to suppose there is a single novel receptor type that photoreception for resetting circadian rhythms is Mrosovsky et al. / IRRADIANCE DETECTION 587 an aspect of a more general irradiance detection sys- Lucas RJ, Douglas RH, and Foster RG (2001) Characterisa- tem and that the hunt for the circadian receptor will tion of an ocular photopigment capable of driving pupillary constriction in mice. Nature Neurosci 4: turn up a receptor type with a much wider functional 621-626. role. Lucas RJ, Freedman MS, Muñoz M, Garcia-Fernández J-M, and Foster RG (1999) Regulation of the mammalian pineal by non-rod, non-cone, ocular photoreceptors. Sci- ACKNOWLEDGMENTS ence 284:505-507. Mrosovsky N (1994) In praise of masking: Behavioural responses of retinally degenerate mice to dim light. We thank Peggy Salmon for help. Support came Chronobiol Int 11:343-348. from the Canadian Institutes of Health Research and Mrosovsky N (1999) Masking: History,definitions, and mea- Biotechnology and Biological Sciences Research surement. Chronobiol Int 16:415-429. Council. Mrosovsky N, Foster RG, and Salmon PA (1999) Thresholds for masking responses to light in three strains of retinally degenerate mice. J Comp Physiol [A] 184:423-428. REFERENCES Mrosovsky N and Hampton RR (1997) Spatial responses to light in mice with severe retinal degeneration. Neurosci Lett 222:204-206. Aschoff J (1960) Exogenous and endogenous components in Redlin U and Mrosovsky N (1999) Masking by light in ham- circadian rhythms. In Biological Clocks, J Aschoff, ed, sters with SCN lesions. J Comp Physiol [A] 184:439-448. pp 11-27, Cold Spring Harb Symp Quant Biol, Cold Selby CP, Thompson C, Schmitz TM, Van Gelder RN, and Spring Harbor, New York. Sancar A (2000) Functional redundancy of crypto- Carter-Dawson LD, LaVail MM, and Sidman RL (1978) Dif- chromes and classical photoreceptors for nonvisual ocu- ferential effect of the rd mutation on rods and cones in the lar photoreception in mice. Proc Natl Acad SciUSA mouse retina. Invest Opthalmol Visual Sci 17:489-498. 97:14697-14702. Foster RG, Provencio I, Hudson D, Fiske S, De Grip W, and van der Horst GTJ, Muijtjens M, Kobayashi K, Takano R, Menaker M (1991) Circadian photoreception in the Kanno S, Takao M, de Wit J, Verkerk A, Eker APM, van retinally degenerate mouse (rd/rd). J Comp Physiol [A] Leenen D, Buijs R, Bootsma D, Hoeijmakers JHJ, and 169:39-50. Yasui A (1999) Mammalian Cry1 and Cry2 are essential Freedman MS, Lucas RJ, Soni B, von Schantz M, Muñoz M, for maintenance of circadian rhythms. Nature 398:627- David-Gray Z, and Foster R (1999) Regulation of mam- 630. malian circadian behavior by non-rod, non-cone, ocular Vitaterna MH, Selby CP, Todo T, Niwa H, Thompson C, photoreceptors. Science 284:502-504. Fruechte EM, Hitomi K, Thresher RJ, Ishikawa T, Goto M and Ebihara S (1990) The influence of different light Miyazaki J, Takahashi JS, and Sancar A(1999) Differential intensities on pineal melatonin content in the retinal regulation of mammalian Period genes and circadian degenerate C3H mouse and the normal CBA mouse. rhythmicity by cryptochromes 1 and 2. Proc Natl Acad Neurosci Lett 108:267-272. Sci U S A 96:12114-12119. Keeler CE(1927) Iris movements in blind mice. Am J Physiol 81:107-112. SRBR 2002

The Eighth Meeting of the Society for Research on Biological Rhythms (SRBR) will take place on May 22-26, 2002 at Amelia Island Plantation, Jacksonville, FL.

A joint one day meeting with the Sleep Research Society (SRS) will take place on May 22, 2002.

588 JOURNALINDEX OF BIOLOGICAL RHYTHMS / December 2001 INDEX To JOURNAL OF BIOLOGICAL RHYTHMS Volume 16

Number 1 (February 2001) pp. 1-96 Number 2 (April 2001) pp. 97-192 Number 3 (June 2001) pp. 193-280 Number 4 (August 2001) pp. 281-432 Number 5 (October 2001) pp. 433-512 Number 6 (December 2001) pp. 513-596

Authors: CHEMINEAU, PHILIPPE, Malpaux, B. ALBRECHT, URS, BINHAI ZHENG, DAVID LARKIN, DAAN, S., U. ALBRECHT, G.T.J. VAN DER HORST, H. ZHONG SHENG SUN, and CHENG CHI LEE, “mPer1 ILLNEROVÁ, T. ROENNEBERG, T. A. WEHR, and W. J. and mPer2 Are Essential for Normal Resetting of the Cir- SCHWARTZ, “Assembling a Clock for All Seasons: Are cadian Clock,” 100. There M and EOscillators in the Genes?” [Conjecture], ALILA-JOHANSSON, AINO, LEA ERIKSSON, TIMO 105. SOVERI, and MAIJA-LIISA LAAKSO, “Seasonal Varia- DAAN, SERGE, “A Parallactic View” [Response], 124. tion in Endogenous Serum Melatonin Profiles in Goats: A DELA IGLESIA, HORACIO, see Schwartz, W. J. Difference between Spring and Fall?” 254. DAWSON, ALISTAIR, VERDUN M. KING, GEORGE E. ARANDA, A., J. A. MADRID, and F. J. SÁNCHEZ- BENTLEY,and GREGORY F.BALL, “Photoperiodic Con- VÁZQUEZ, “Influence of Light on Feeding Anticipatory trol of Seasonality in Birds,” 365. Activity in Goldfish,” 50. DEBOER, TOM, see Watanabe, K. ARANDA, A., see Sánchez-Vázquez, F. J. DEMAS, G. E., see Bartness, T. J. ARENDT, J., see Sopowski, M. J. DIEFENBACH, KONSTANZE, see Taillard, J. BADURA, LORI L., see Imundo, J. DITTAMI, JOHN P., see Millesi, E. BALER, RUBEN, “Clockless Yeast and the Gears of the DODGE, JAMES, see Imundo, J. Clock: How Do They Mesh?” [Review], 516. DRAZEN, DEBORAH L., see Kriegsfeld, L. J. BALL, GREGORY F., Dawson, A. DUDEK, F. EDWARD, see Smith, B. N. BARTNESS, T. J., C. K. SONG, AND G. E. DEMAS, “SCN DUMONT, MARIE, DALILA BENHABEROU-BRUN, and Efferents to Peripheral Tissues: Implications for Biologi- JEAN PAQUET, “Profile of 24-h Light Exposure and Cir- cal Rhythms” [Review], 196. cadian Phase of Melatonin Secretion in Night Workers,” BENHABEROU-BRUN, DALILA, see Dumont, M. 502. BENTLEY, GEORGE E., Dawson, A. EDGAR, DALE M., see Kas, M.J.H. BERTOLUCCI, CRISTIANO, see Foà, A. EDELSTEIN, K., see Mrosovsky, N. BIELEFELD, ERIC, see Imundo, J. ERIKSSON, LEA, see Alila-Johansson, A. BIOULAC, BERNARD, see Taillard, J. FAHRENKRUG, JAN, see Hannibal, J. BLOCH, GUY, DAN P. TOMA, and GENE E. ROBINSON, FIEDER, MARTIN, see Millesi, E. “Behavioral Rhythmicity, Age, Division of Labor and pe- FOÀ, AUGUSTO, and CRISTIANO BERTOLUCCI, “Tem- riod Expression in the Honey Bee Brain,” 444. perature Cycles Induce a Bimodal Activity Pattern in BULT, ABEL, see Mahoney, M. Ruin Lizards: Masking or Clock-Controlled Event? A CAPODICE, CAMALA E., see Weaver, D. R. Seasonal Problem,” 574. CARD, J. PATRICK, see Hannibal, J. FOSTER, RUSSELL G., see Mrosovsky, N. CARRÉ, ISABELLE A., “Day-Length Perception and the FOLLETT, BRIAN, Hastings, M. Photoperiodic Regulation of Flowering in Arabidopsis,” FREE,AM. DAVID A., see Larkin, J. E.

415. GANNON, ROBERT L., “5HT7 Receptors in the Rodent CASAL, JORGE J., see Yanovsky, M. J. Suprachiasmatic Nucleus” [Review], 19. CHASTANG, JEAN-FRANÇOIS, see Taillard, J.

GARCIA, CÉLIA R. S., REGINA P. MARKUS, and LUCI- KENDALL, ADAM R., ALFRED J. LEWY, and ROBERT L. ANA MADEIRA, “Tertian and Quartan Fevers: Tempo- SACK, “Effects of Aging on the Intrinsic Circadian Pe- ral Regulation in Malarial Infection” [Review], 436. riod of Totally Blind Humans,” 87. GIEBULTOWICZ, JADWIGA M., see Ivanchenko, M. KENNAWAY, DAVID J., see Stepien, J. M. GOLDMAN, BRUCED., “Mammalian Photoperiodic Sys - KING, VERDUN M., Dawson, A. tem: Formal Properties and Neuroendocrine Mecha- KORF, HORST-WERNER, see Stehle, J. H. nisms of Photoperiodic Time Measurement,” 283. KRIEGSFELD, LANCE J., DEBORAH L. DRAZEN, and GORMAN, MICHAEL R. and THERESA M. LEE, “Daily RANDY J. NELSON, “Circadian Organization in Male Novel Wheel Running Reorganizes and Splits Hamster Mice Lacking the Gene for Endothelial Nitric Oxide Circadian Activity Rhythms,” 541. Synthase (eNOS–/–),” 142. GORMAN, MICHAEL R., STEVEN M. YELLON and KYRIACOU, C. P., see Tauber, E. THERESA M. LEE, “Temporal Reorganization of the LAAKSO, MAIJA-LIISA, see Alila-Johansson, A. Suprachiasmatic Nuclei in Hamsters with Split Circadian LARKIN, DAVID, see Albrecht, U. Rhythms,” 552. LARKIN, JENNIE E., DAVID A. FREEMAN, and IRVING GOVERNALE, MARCIAM. and THERESAM. LEE, “Olfac- ZUCKER, “Low-Ambient Temperature Accelerates tory Cues Accelerate Reentrainment Following Phase Short-Day Responses in Siberian Hamsters by Altering Shifts and Entrain Free-Running Rhythms in Female Responsiveness to Melatonin,” 76. Octodon degus (Rodentia),” 489. LEE, CHENG CHI, see Albrecht, U. GUBIK, BETTY, see Smale, L. LEE, THERESA M., see Governale, M. M. HAMPTON, S. M., see Sopowski, M. J. LEE, THERESA M., see Gorman, M. R. HANNIBAL, JENS, NIELS VRANG, J. PATRICK CARD, LEWY, ALFRED J., see Kendall, A. R. and JAN FAHRENKRUG, “Light-Dependent Induction LUCAS, ROBERT J., see Mrosovsky, N. of cfos during Subjective Day and Night in PACAP-Con- MADEIRA, LUCIANA, see Garcia, C.R.S. taining Ganglion Cells of the Retinohypothalamic Tract,” MADRID, J. A., see Aranda, A. 457. MADRID, J. A. see Sánchez-Vázquez, F. J. HARRINGTON, MARY E., “Feedback” [Poem], 277. MAHONEY, MEGAN, ABEL BULT, and LAURA SMALE, HASTINGS, J. WOODLAND, “The Colin S. Pittendrigh Lec- “Phase Response Curve and Light-Induced Fos Expres- ture: Fifty Years of Fun” [Lecture], 5. sion in the Suprachiasmatic Nucleus and Adjacent Hypo- HASTINGS, M. H., see Mrosovsky, N. thalamus of Arvicanthis niloticus,” 149. HASTINGS, MICHAEL, “Modeling the Molecular Calen- MALPAUX, BENOÎT, MARTINE MIGAUD, HÉLÈNE dar” [Commentary], 117. TRICOIRE, and PHILIPPE CHEMINEAU, “Biology of HASTINGS, MICHAEL, and BRIAN FOLLETT, “Toward a Mammalian Photoperiodism and the Critical Role of the Molecular Biological Calendar?” 424. Pineal Gland and Melatonin,” 336. HAZLERIGG, D. G., P. J. MORGAN, and S. MESSAGER, MARKUS, REGINA P., see Garcia, C.R.S. “Decoding Photoperiodic Time and Melatonin in Mam- MAYWOOD, E. S., see Mrosovsky, N. mals: What Can WeLearn from the Pars tuberalis?” 326. MAZZELLA, M. AGUSTINA, see Yanovsky, M. J. HORTON, TERESA H., and STEVEN M. YELLON, “Aging, MCELHINNY, TERESA, see Smale, L. Reproduction, and the Melatonin Rhythm in the Siberian MEIJER, JOHANNA H., see Watanabe, K. Hamster,” 243. MIGAUD, MARTINE, see Malpaux, B. ILLNEROVÁ, H., see Albrecht, U. MILLESI, EVA, HERMANN PROSSINGER, JOHN P. ILLNEROVÁ, HELENA, see Schwartz, W. J. DITTAMI, and MARTIN FIEDER, “Hibernation Effects IMUNDO, JASON, ERIC BIELEFELD, JAMES DODGE, and on Memory in European Ground Squirrels (Spermophilus LORI L. BADURA, “Relationship between Norepine- citellus),” 264. phrine Release in the Hypothalamic Paraventricular Nu- MORGAN, L., see Sopowski, M. J. cleus and Circulating Prolactin Levels in the Siberian MORGAN, P. J., see Hazlerigg, D. G. Hamster: Role of Photoperiod and the Pineal Gland,” MROSOVSKY, N., K. EDELSTEIN, M. H. HASTINGS, and 173. E. S. MAYWOOD, “Cycle of period Gene Expression in a IVANCHENKO, MARIA, RALF STANEWSKY, and Diurnal Mammal (Spermophilus tridecemlineatus): Impli- JADWIGAM. GIEBULTOWICZ, “Circadian Photorecep- cations for Nonphotic Phase Shifting,” 471. tion in Drosophila: Functions of Cryptochrome in Pe- MROSOVSKY, N., ROBERT J. LUCAS, and RUSSELL G. ripheral and Central Clocks,” 205. FOSTER “Persistence of Masking Mice Lacking Rods and KAS, MARTIEN J. H., and DALE M. EDGAR, “Scheduled Cones,” 585. VoluntaryWheel Running Activity Modulates Free-Run- NELSON, RANDY J., see Kriegsfeld, L. J. ning Circadian Body Temperature Rhythms in Octodon NELSON, RANDY J., see Prendergast, B. J. degus,” 66. NELSON, RANDY J., see Young, K. A. INDEX 591

NIXON, JOSHUA, see Smale, L. erol Responses in Simulated Night and Day Shift: Gender PAQUET, JEAN, see Dumont, M. Differences,” 272. PETRI, BERNHARD, and MONIKA STENGL, “Phase Re- SOVERI, TIMO, see Alila-Johansson, A. sponse Curves of a Molecular Model Oscillator: Implica- STANEWSKY, RALF, see Ivanchenko, M. tions for Mutual Coupling of Paired Oscillators,” 125. STEHLE, JÖRG H., CHARLOTTE VON GALL, CHRISTOF PHILIP, PIERRE, see Taillard, J. SCHOMERUS, and HORST-WERNER KORF, “Of Ro- PICKARD, GARY E. see Smith, B. N. dents and Ungulates and Melatonin: Creating a Uniform PRENDERGAST, BRIAN J., STEVEN M. YELLON, LONG T. Code for Darkness by Different Signaling Mechanisms,” TRAN, and RANDY J. NELSON, “Photoperiod Modu- 312. lates the Inhibitory Effect of in vitro Melatonin on Lym- STENGL, MONIKA, see Petri, B. phocyte Proliferation in Female Siberian Hamsters,” 224. STEPIEN, JACQUELINE M., and DAVID J. KENNAWAY, PROSSINGER, HERMANN, see Millesi, E. “Phase Response Relationships between Light Pulses PUPIQUE, M., “How to Fix the ‘Review Process’” [Letter], and the Melatonin Rhythm in Rats,” 234. 191. SUN, ZHONG SHENG, see Albrecht, U. RIBEIRO, D.C.O., see Sopowski, M. J. TAILLARD, JACQUES, PIERRE PHILIP, JEAN-FRANÇOIS ROENNEBERG, T., see Albrecht, U. CHASTANG, KONSTANZE DIEFENBACH, BERNARD ROENNEBERG, TILL, and MARTHA MERROW, “Season- BIOULAC, “Is Self-Reported Morbidity Related to the ality and Photoperiodism in Fungi,” 403. Circadian Clock?” 183. ROBINSON, GENE E., see Bloch, G. TAUBER, E., and C. P. KYRIACOU, “Insect Photoperiodism ROSE, SANDRA, see Smale, L. and Circadian Clocks: Models and Mechanisms,” 381. SACK, ROBERT L., see Kendall, A. R. TOMA, DAN P., see Bloch, G. SALDANHA, COLIN J., ANN-JUDITH SILVERMAN, and TRAN, LONG T., see Prendergast, B. J. RAESILVER,“Direct Innervation of GnRH Neurons by TRICOIRE, HÉLÈNE, Malpaux, B. Encephalic Photoreceptors in Birds,” 39. VAN DER HORST, G.T.J., see Albrecht, U. SÁNCHEZ-VÁZQUEZ, F. J., A. ARANDA, and J. A. MA- VON GALL, CHARLOTTE, see Stehle, J. H. DRID, “Differential Effects of Meal Size and Food Energy VRANG, NIELS, see Hannibal, J. Density on Feeding Entrainment in Goldfish,” 58. WAYNE,NANCY L., “Regulation of Seasonal Reproduction SÁNCHEZ-VÁZQUEZ, F. J., see Aranda, A. in Mollusks,” 391. SCHOMERUS, CHRISTOF, see Stehle, J. H. WATANABE, KAZUTO, TOM DEBOER and JOHANNA H. SCHWARTZ, W. J., see Wehr, T. A. MEIJER, “Light-Induced Resetting of the Circadian Pace- SCHWARTZ, W. J., see Albrecht, U. maker: Quantitative Analysis of Transient versus SCHWARTZ, WILLIAM J., HORACIO DELA IGLESIA, Steady-State Phase Shifts,” 564. PIOTR ZLOMANCZUK, and HELENAILLNEROVÁ, WEAVER, DAVID R., see Shearman, L. P. “Encoding Le Quattro Stagioni within the Mammalian WEAVER, DAVID R., and CAMALA E. CAPODICE, “Post- Brain: Photoperiodic Orchestration through the mortem Stability of Melatonin Receptor Binding and Suprachiasmatic Nucleus,” 302. Clock-Relevant mRNAs in Mouse Suprachiasmatic Nu- SHEARMAN, LAUREN P. and DAVID R. WEAVER, “Dis- cleus,” 216. tinct Pharmacological Mechanisms Leading to c-fos Gene WEHR, T. A., see Albrecht, U. Expression in the Fetal Suprachiasmatic Nucleus,” 531. WEHR, T. A., and W. J. SCHWARTZ, “Assembling a Clock SILVER, RAE, see Saldanha, C. J. for All Seasons: Are There M and EOscillators in the SILVERMAN, ANN-JUDITH, see Saldanha, C. J. Genes?” [Conjecture], 105. SMALE, LAURA, TERESA MCELHINNY, JOSHUA WEHR, THOMAS A., “Photoperiodism in Humans and NIXON, BETTY GUBIK, and SANDRA ROSE, “Patterns Other Primates: Evidence and Implications,” 348. of Wheel Running Are Related to Fos Expression in WHITELAM, GARRY C., see Yanovsky, M. J. Neuropeptide Y–Containing Neurons in the Intergeniculate YANOVSKY, MARCELO J., M. AGUSTINA MAZZELLA, Leaflet of Arvicanthis niloticus,” 163. GARRY C. WHITELAM, and JORGE J. CASAL, “Re- SMALE, LAURA, see Mahoney, M. setting of the Circadian Clock by Phytochromes and SMITH, BRET N., PATRICIA J. SOLLARS, F. EDWARD Cryptochromes in Arabidopsis,” 523. DUDEK, and GARY E. PICKARD, “Serotonergic Modu- YELLON, STEVEN M., see Gorman, Michael R. lation of Retinal Input to the Mouse Suprachiasmatic Nu- YELLON, STEVEN M., see Horton, T. H.

cleus Mediated by 5-HT1B and 5-HT7 Receptors,” 25. YELLON, STEVEN M., see Prendergast, B. J. SOLLARS, PATRICIA J., see Smith, B. N. YOUNG, KELLY A., BARRY R. ZIRKIN, and RANDY J. SONG, C. K, see Bartness, T. J. NELSON, “Testicular Apoptosis Is Down-Regulated SOPOWSKI, M. J., S. M. HAMPTON, D.C.O. RIBEIRO, L. during Spontaneous Recrudescence in White-Footed MORGAN, and J. ARENDT, “Postprandial Triacylglyc- Mice (Peromyscus leucopus),” 479. 592 JOURNAL OF BIOLOGICAL RHYTHMS / December 2001

ZATZ, MARTIN, “Editorial: Pebbles of Truth,” 515. “Insect Photoperiodism and Circadian Clocks: Models and ZATZ, MARTIN, “Editoral: On Telling It Like It Was,” 195. Mechanisms,” Tauber and Kyriacou, 381. ZATZ, MARTIN, “Editorial: Show Me the Data!” 99. “Is Self-Reported Morbidity Related to the Circadian ZATZ, MARTIN, “Editorial: What Do Reviewers Really Clock?” Taillard et al., 183. Want?” 3. “Light-Dependent Induction of cfos during Subjective Day ZATZ, MARTIN, “Editorial: Yes Sir, That’s My Data!” 435. and Night in PACAP-Containing Ganglion Cells of the ZHENG, BINHAI, see Albrecht, U. Retinohypothalamic Tract,” Hannibal et al., 457. ZIRKIN, BARRY R., see Young, K. A. “Light-Induced Resetting of the Circadian Pacemaker: ZLOMANCZUK, PIOTR, see Schwartz, W. J. Quantitative Analysis of Transient versus Steady-State ZUCKER, IRVING, see Larkin, J. E. Phase Shifts,” Watanabe, Deboer and Meijer, 564. “Low-Ambient Temperature Accelerates Short-Day Re- Articles: sponses in Siberian Hamsters by Altering Responsive- “Aging, Reproduction, and the Melatonin Rhythm in the Si- ness to Melatonin,” Larkin et al., 76. berian Hamster,” Horton and Yellon, 243. “Mammalian Photoperiodic System: Formal Properties and “Behavioral Rhythmicity, Age, Division of Labor and period Neuroendocrine Mechanisms of Photoperiodic Time Expression in the Honey Bee Brain,” Bloch et al., 444. Measurement,” Goldman, 283. “Biology of Mammalian Photoperiodism and the Critical “mPer1 and mPer2 Are Essential for Normal Resetting of the Role of the Pineal Gland and Melatonin,” Malpaux et al., Circadian Clock,” Albrecht et al., 100. 336. “Of Rodents and Ungulates and Melatonin: Creating a Uni- “Circadian Organization in Male Mice Lacking the Gene for form Code for Darkness by Different Signaling Mecha- Endothelial Nitric Oxide Synthase (eNOS–/–),” nisms,” Stehle et al., 312. Kriegsfeld et al., 142. “Olfactory Cues Accelerate Reentrainment Following Phase “Circadian Photoreception in Drosophila: Functions of Shifts and Entrain Free-Running Rhythms in Female Cryptochrome in Peripheral and Central Clocks,” Octodon degus (Rodentia),” Governale and Lee, 489. Ivanchencko et al., 205. “Patterns of Wheel Running Are Related to Fos Expression “Cycle of period Gene Expression in a Diurnal Mammal in Neuropeptide Y–Containing Neurons in the (Spermophilus tridecemlineatus): Implications for Intergeniculate Leaflet of Arvicanthis niloticus,” Smale Nonphotic Phase Shifting,” Mrosovsky et al., 471. et al., 163. “Daily Novel Wheel Running Reorganizes and Splits Ham- “Phase Response Curve and Light-Induced Fos Expression ster Circadian Activity Rhythms,” Gorman and Lee, 541. in the Suprachiasmatic Nucleus and Adjacent Hypothal- “Day-Length Perception and the Photoperiodic Regulation amus of Arvicanthis niloticus,” Mahoney et al., 149. of Flowering in Arabidopsis,” Carré, 415. “Phase Response Curves of a Molecular Model Oscillator: “Decoding Photoperiodic Time and Melatonin in Mammals: Implications for Mutual Coupling of Paired Oscillators,” What Can We Learn from the Pars tuberalis?” Hazlerigg Petri and Stengle, 125. et al., 326. “Phase Response Relationships between Light Pulses and “Differential Effects of Meal Size and Food Energy Density the Melatonin Rhythm in Rats,” Stepien and Kennaway, on Feeding Entrainment in Goldfish,” Sánchez-Vázquez 234. et al., 58. “Photoperiod Modulates the Inhibitory Effect of in vitro “Direct Innervation of GnRH Neurons by Encephalic Melatonin on Lymphocyte Proliferation in Female Sibe- Photoreceptors in Birds,” Saldanha et al., 39. rian Hamsters,” Prendergast et al., 224. “Distinct Pharmacological Mechanisms Leading to c-fos “Photoperiodic Control of Seasonality in Birds,” Dawson Gene Expression in the Fetal Suprachiasmatic Nucleus,” et al., 365. Shearman and Weaver, 531. “Photoperiodism in Humans and Other Primates: Evidence “Editorial: Pebbles of Truth,” Zatz, 515. and Implications,” Wehr, 348. “Editoral: On Telling It Like It Was,” Zatz, 195. “Postmortem Stability of Melatonin Receptor Binding and “Editorial: Show Me the Data!” Zatz, 99. Clock-Relevant mRNAs in Mouse Suprachiasmatic Nu- “Editorial: What Do Reviewers Really Want?” Zatz, 3. cleus,” Weaver and Capodice, 216. “Editorial: Yes Sir, That’s My Data!” Zatz, 453. “Postprandial Triacylglycerol Responses in Simulated “Effects of Aging on the Intrinsic Circadian Period of Totally Night and Day Shift: Gender Differences,” Sopowski Blind Humans,” Kendall et al., 87. et al., 272. “Encoding Le Quattro Stagioni within the Mammalian Brain: “Profile of 24-h Light Exposure and Circadian Phase of Photoperiodic Orchestration through the Suprachiasmatic Melatonin Secretion in Night Workers,” Dumont et al., Nucleus,” Schwarz et al., 302. 502. “Hibernation Effects on Memory in European Ground “Regulation of Seasonal Reproduction in Mollusks,” Squirrels (Spermophilus citellus),” Millesi et al., 264. Wayne, 391. “Influence of Light on Feeding Anticipatory Activity in “Relationship between Norepinephrine Release in the Hy- Goldfish,” Aranda et al., 50. pothalamic Paraventricular Nucleus and Circulating INDEX 593

Prolactin Levels in the Siberian Hamster: Role of Lectures: Photoperiod and the Pineal Gland,” Imundo et al., 173. “The Colin S. Pittendrigh Lecture: Fifty Years of Fun,” “Resetting of the Circadian Clock by Phytochromes and Hastings, 5. Cryptochromes in Arabidopsis,” Yanovsky, Mazzella, Whitelam, and Casal, 523. Letters: “Scheduled Voluntary Wheel Running Activity Modulates “How to Fix the ‘Review Process,’” Pupique, 191. Free-Running Circadian Body Temperature Rhythms in “Persistence of Masking Responses to Light in Mice Lakcing Octodon degus,” Kas and Edgar, 66. Rods and Cones,” Mrosovsky, Lucas and Foster, 585. “Seasonal Variation in Endogenous Serum Melatonin Pro- files in Goats: A Difference between Spring and Fall?” Poem: Alila-Johansson et al., 254. “Feedback,” Harrington, 277. “Seasonality and Photoperiodism in Fungi,” Roenneberg and Merrow, 403. Reflection: “Serotonergic Modulation of Retinal Input to the Mouse “Toward a Molecular Biological Calendar?” Hastings and Suprachiasmatic Nucleus Mediated by 5-HT1B and 5-HT7 Follett, 424. Receptors,” Smith et al., 25. “Temperature Cycles Induce a Bimodal Activity Pattern in Response: Ruin Lizards: Masking or Clock-Controlled Event? A “A Parallactic View,” Daan, 124. Seasonal Problem,” Foà and Bertolucci, 574. “Temporal Reorganization of the Suprachiasmatic Nuclei in Reviews: Hamsters with Split Circadian Rhythms,” Gorman, “Clockless Yeast and the Gears of the Clock: How Do They Yellon and Lee, 552. Mesh?” Baler, 516.

“Testicular Apoptosis Is Down-Regulated during Spontane- “5HT7 Receptors in the Rodent Suprachiasmatic Nucleus,” ous Recrudescence in White-Footed Mice (Peromyscus Gannon, 19. leucopus),” Young et al., 479. “SCN Efferents to Peripheral Tissues: Implications for Bio- logical Rhythms,” Bartness et al., 196. Commentary: “Tertian and Quartan Fevers: Temporal Regulation in Ma- “Modeling the Molecular Calendar,” Hastings, M., 117. larial Infection,” Garcia et al., 436. Conjecture: “Assembling a Clock for All Seasons: Are There M and E Oscillators in the Genes?” Daan et al., 105. Journal of Biological Rhythms Official Publication of the Society for Research on Biological Rhythms

INSTRUCTIONS TO AUTHORS

Four legible, double-spaced copies of each paper (including Examples illustrations) should be submitted to the Editor. There are no publication charges except for circumstances requiring special Aschoff J (1965) Response curves in circadian periodicity. In printing, color reproduction, or unusual length and number of Circadian Clocks, J Aschoff, ed, pp 95-111, North-Holland, Am- illustrations; in these cases, the publisher will provide cost in- sterdam. formation before the paper is accepted. Pittendrigh CS and Daan S (1976) A functional analysis of Manuscripts should be written clearly and concisely and circadian pacemakers in nocturnal rodents. I. The stability and should conform generally to the Council of Biology Editors Style lability of spontaneous frequency.J Comp Physiol [A] 106:223-252. Manual. Manuscripts should consist of the following: Richter CP (1965) Biological Clocks in Medicine and Psychiatry, 1. A title page, including the authors’ names, the title and a Charles C Thomas, Springfield, IL. short running title, and the institution(s) (with all words spelled 8. Tables and illustrations should be submitted in a form out in full) from which the paper emanates. If current addresses suitable for direct reproduction when possible. In preparation are different, then these should be indicated in a footnote. Also of figures (glossy prints, crisp laser prints, or hand-drawn graphs), authors should consider page format and letter size af- include the name, mailing address, phone and fax numbers, and ter final reproduction. Transfer (rub-on) type should be e-mail address of the person to whom correspondence and avoided. An acetate overlay containing type on a clear back- proofs should be sent. Indicate the number of pages, illustra- ground is recommended for lettering and calibration bars ap- tions, and tables. plied to halftone illustrations. Each illustration should be 2. Asuccinct Abstract, not exceeding 350 words, in a form ac- marked on the reverse side with its number, name(s) of ceptable to most abstracting services, followed by six to eight author(s), orientation, and instructions for reduction when ap- key words. propriate. Copies of the original set can be used for review pur- 3. An Introduction that provides a brief review of relevant poses but should be of suitable quality to enable the referees to background material and indicates the purpose of the study. evaluate them. Place each table on a separate sheet, following the 4. AMaterials and Methods section that provides sufficient in- figure legends. Number consecutively with Arabic numerals. formation for qualified investigators to reproduce the work in Charges for color figures are $800.00 for the first figure and similar fashion. Reference to published procedures by appro- $200.00 for each additional figure. priate succinct summary and citation is encouraged but should 9. A list of figure captions should follow the references. A brief not replace adequate methodological description. The Journal of title and description of each illustration should be included. These Biological Rhythms endorses the Declaration of Helsinki and the should be typed consecutively on the same page(s). U.S. National Institutes of Health Guide for the Care and Use of 10. Abbreviations should be introduced in parentheses after the first occurrence of the term being abbreviated. Use standard Laboratory Animals. It will be assumed that authors have consci- metric units wherever possible. entiously followed principles and practices in accord with these Submission of a manuscript implies assurance that no simi- documents in experiments involving human subjects and ex- lar paper has been submitted for publication elsewhere. Con- perimental animals. currence of all listed authors is assumed for all papers to be 5. A Results section that provides a concise description of the published. Manuscripts are evaluated on their merit by the edi- findings with appropriate reference to illustrations and tables. tors with the assistance of selected referees. Copyright is vested 6. A Discussion section that includes a summary of the main in the publisher. findings, their relation to other published work, and a statement of their significance. Mail submissions to 7. References should be listed sequentially in alphabetical Martin Zatz, Ph.D., M.D. order according to the name of the first author with (a) a full list Editor, Journal of Biological Rhythms of authors, (b) date, (c) full title of the paper, (d) journal titles ab- PMB 336 breviated as per Index Medicus, (e) volume number, and (f) first 4938 Hampden Lane and last pages. Only papers published or in press may be in- Bethesda, MD 20814 - 2914 cluded in the reference list. Papers should be cited in the text by tel: 301 - 656 - 4655 author(s) and date. e-mail: [email protected]